How New Loanwords Are Coined in Japanese (and Why Some Fail)
To understand how new loanwords are coined in Japanese, trace one journey: a foreign concept arrives, gets spelled in katakana, has its sounds approximated, drifts in meaning, and then either settles, gets clipped, or is quietly swapped for a native paraphrase.1 The same pipeline that turned "personal computer" into the settled パソコン also produced the contested オーバーシュート, which never stuck.23
Overview
A brand-new borrowing does not enter Japanese by official decree. It is introduced from the bottom up, passes through several predictable stages, and only later meets the language-policy bodies and broadcasters that try to steer it.45
This article models that path as a five-stage pipeline, maps the institutions that sit alongside it, and explains why some candidate words never settle.42 The 2020 COVID-19 katakana wave serves as a dated case study of success and failure side by side.36
Where New Loanwords Come From
Borrowing in Japanese is constant and mostly comes from the bottom up. A new word spreads through use first. Only afterward does any institution standardize its spelling or propose a clearer alternative.45
The trigger: a concept with no ready Japanese word
A borrowing usually begins because a concept arrives with no ready Japanese word for it: a new technology, food, or social idea, or a register and nuance the native term lacks.5 This gap-filling motive is the standard explanation for why the gairaigo (loanword) layer keeps growing, and it is the foundation laid out in "What Is Gairaigo? A Guide to Loanwords in Japanese."
The 1991 cabinet directive on loanword notation frames its scope as a reference point for writing loanwords "in everyday social life." It explicitly excludes specialized fields and proper nouns such as personal, company, and brand names.5 In other words, loanwords are treated as an open everyday layer of the lexicon, not a closed official list.
Who actually introduces them
New loanwords usually enter from the bottom up: through media, advertising, industry, and online or youth speech, not committees.52 None of the institutional bodies covered below coins or licenses a new everyday loanword. They issue notation conventions and advisory paraphrases after the fact.
This ordering is visible in the record. NINJAL's paraphrase project was created to address loanwords that were already circulating in public and administrative contexts yet remained "not well-established among the general population and difficult to understand."47 Institutions react to words that have already entered use.
The Adoption Pipeline, Stage by Stage
A foreign word that settles in Japanese typically passes through five stages: phonological approximation, orthographic settling, semantic shift, grammatical naturalization, and a final fork between settling, clipping, and replacement.
The diagram below shows that path and the three-way fork at its end.
Stage 1: Phonological approximation (English sound to katakana)
First, the word's sounds are remapped onto Japanese phonology. Japanese has a small set of consonants and a (C)V mora structure. It has no native /f, v, θ, ð/, and one liquid sound covers both English /l/ and /r/.1 A mora is the basic rhythmic unit of Japanese; most kana represent one mora.
Consonant clusters and word-final consonants are broken up by an inserted (epenthetic) vowel. The default is /u/, the short weak vowel that produces the smallest acoustic change while satisfying Japanese phonotactics; /i/ appears after sounds like [ʃ] and [tʃ].1 This is why "smartphone" begins スマ: the /sm/ cluster is split by an epenthetic /u/.
Experimental work shows that loanword adaptation begins in perception: Japanese listeners hear an illusory native vowel where the source has none, mapping foreign sounds onto the closest native sounds. French word-final [n] with a strong vocalic release is perceived with an epenthetic [ɯ], whereas English word-final [n] is heard as the moraic nasal ン.8
Modern extended katakana (ファ, ヴィ, ティ, トゥ) gives a closer option for foreign sounds. A long vowel is written with the chōonpu ー; both are addressed in the 1991 notation directive's tables of kana and symbols for loanwords.5 The mechanics of those extra kana are covered in "Extended Katakana for Loanwords (ファ, ヴィ, ティ, トゥ, and the Full Small-Vowel System),". The same sound-approximation logic governs "How to Write Your Name in Katakana: Foreign-Name Transcription Rules with Examples."
Two everyday words show the result of this approximation:
パソコンは使える？9
"Can you use a computer?"
新しいスマホが欲しい。9
"I want a new smartphone."
Both パソコン and スマホ are register-neutral, high-frequency words, and both are clipped (see Stage 5). The unclipped パーソナルコンピューター and スマートフォン survive in formal and technical contexts.2
Stage 2: Orthographic settling (which katakana spelling wins)
Before one spelling stabilizes, competing katakana spellings may coexist. They vary in long-vowel marks and small-kana choices. The 1991 内閣告示第2号「外来語の表記」 (Cabinet Directive No. 2, "Notation of Loanwords") sets the default conventions used in official documents, laws, newspapers, magazines, and broadcasting, while explicitly not binding specialized fields, proper nouns, or individual preference.5
The directive was based on the 国語審議会's recommendation of 7 February 1991. It consists of a preamble, a main text with tables of kana and symbols plus usage notes, and an appendix of example words.510 A documented variation case is word-final English -er and -or words, which vary between a chōonpu ending and none: コンピューター versus コンピュータ.5
The guideline's usage notes (留意事項) address this long-vowel-symbol variation, allowing both forms with a stated default.5 The chōonpu-final コンピューター is the everyday and broadcast default; the chōonpu-less コンピュータ is common in older technical and engineering writing.5
Stage 3: Semantic shift (the meaning drifts from the source)
Borrowed words rarely keep the full source meaning. They may narrow, broaden, or get repurposed.11 The NINJAL paraphrase entries document this directly. The committee splits アクセス into three distinct Japanese senses: 接続 "to connect to and use information" (connect or log on), 交通手段 "convenience of transport or contact" (transport access), and 参入 "entering a market" (market entry).11
The single English word "access" has spread into three context-bound Japanese uses. Two corpus sentences show that split in everyday text:
そのホテルは駅からのアクセスが便利だよ。9
"The hotel is within easy access of the station."
彼はその保存情報にアクセスできる。9
"He has access to the stored information."
The first uses the 交通手段 "transport access" sense; the second uses the 接続 "connect and retrieve" sense.9 Both are the same katakana word, doing semantic work that the committee later untangled into separate native glosses.11
The committee glosses バリアフリー (from "barrier-free") as 障壁なし, meaning "a living environment in which physically disabled people can act without obstruction." This is a narrowed, accessibility-specific sense rather than a general "without barriers."12
Stage 4: Grammatical naturalization
A naturalized loanword stops behaving like a quoted foreign word and takes on native grammatical machinery. It can add する to form a verb (アクセスする), add な to form an adjectival noun, and compound with native and kango (Sino-Japanese) elements.9 The sentence above, アクセスできる, already shows アクセス taking the potential form of a する-verb.9
Once naturalized, the word also heads predicates and takes particles like any native noun:
スタジアムへはバスでのアクセスが可能です。9
"The stadium is accessible by bus."
Here アクセス takes が, heads a 可能です predicate, and compounds freely with スタジアム and バス. Grammatically, it is indistinguishable from a native noun.9 A naturalized loanword also receives a Japanese pitch-accent pattern, a further sign it is inside the grammar rather than quoted from outside it.
Stage 5: Settling, clipping, or replacement
At the final stage, the word reaches one of three outcomes. It can settle in its long form, without a paraphrase, as パンデミック did (see the case study). It can be clipped toward roughly four morae, the most common fate of long borrowings. Or it can be swapped for a native or kango paraphrase.13911
Clipping is corpus-attested for the everyday clips パソコン (from パーソナルコンピューター) and スマホ (from スマートフォン).9
スマホ持ってないの。9
"I don't have a smartphone."
That shortening, and the pull toward four morae, is the subject of "Shortened Loanwords in Japanese: Why パソコン, リモコン, and アポ Get Clipped." The third outcome, replacement by paraphrase, is exactly the NINJAL 言い換え lever covered next: アクセス → 接続／交通手段／参入, バリアフリー → 障壁なし.1112
The Gatekeepers: Who Tries to Steer Loanwords
Three kinds of institutions sit alongside the pipeline: national language-policy bodies, NINJAL's loanword-replacement committee, and broadcasters. All of them react to words already in circulation; none coins everyday loanwords.45
The language-policy bodies (国語審議会 and its successors)
The 国語審議会 (National Language Council) was established in 1934, reorganized in 1949, and abolished in 2001 as part of the central-government reorganization (中央省庁再編, central ministry reorganization).14 It is a historical body, not a current one.
The 国語審議会 was abolished in 2001. Its national-language functions passed to the 文化審議会国語分科会 (Cultural Council, Japanese Language Subcommittee) under the 文化庁 (Agency for Cultural Affairs), whose first session was held on 1 May 2002.141516 Any source presenting 国語審議会 as a body that still meets is out of date.
The 文化審議会 (Cultural Council) itself was created on 6 January 2001 by merging the former 国語審議会 with three other councils.1415 The 国語審議会's last major outputs, in 2000, included 「現代社会における敬意表現」 and 「表外漢字字体表」. This was the year before its abolition.14
These bodies set notation conventions and give advice. For loanwords, the relevant convention is the 外来語の表記 directive.515 They do not police everyday speech.
NINJAL's loanword-replacement proposals (言い換え提案)
The loanword-paraphrase work belongs to NINJAL (国立国語研究所, National Institute for Japanese Language and Linguistics). In August 2001, it set up an 「外来語」委員会 (Loanword Committee) to propose clearer Japanese paraphrases for hard-to-understand loanwords used in public and administrative contexts.7
The committee published its proposals in rounds. Its first final proposal was issued on 25 April 2003 and covered 62 words. The work continued through March 2006, with a general-readership handbook following in June 2006.427 The proposals pair a loanword with one or more native or kango glosses, such as アクセス → 接続／交通手段／参入 and バリアフリー → 障壁なし.1112
NHK and broadcast style decisions
Broadcasters make practical calls about whether to air a katakana word or gloss it. NHK maintains a 放送用語委員会 (Broadcasting Terminology Committee) under its 放送文化研究所 (Broadcasting Culture Research Institute). The committee decides on broadcast wording, including how to handle loanwords and foreign words prone to spelling variation.17
The committee's remit is to standardize broadcast vocabulary. Its practice has historically included substituting or glossing foreign terms for clarity.17 This is where a loanword can be endorsed in practice for a mass audience, or sidelined in favor of a plainer word.
Why Some Loanwords Fail
Not every candidate settles. A loanword may fail when speakers cannot decode it, when a native or kango synonym already fills the slot, or when a paraphrase is actively pushed in its place.43
The comprehension barrier
When too few speakers can decode a loanword, it stalls. Opaque katakana in public messaging triggers pushback and 言い換え (paraphrase).47 NINJAL's paraphrase project rests on this premise. It targeted loanwords that were "not well-established among the general population and difficult to understand" in public and administrative use.4
The native or kango paraphrase tends to win where clarity matters most: in medical, administrative, and emergency messaging.4 That is precisely the domain NINJAL prioritized.
Competition from a native or kango synonym
A loanword is sidelined when an existing wago or kango word already covers the same meaning, or when a paraphrase is pushed. The clearest case is クラスター, which the source itself glosses through the native 集団感染 ("group infection"). クラスター is defined as a 集団感染 and the cluster of infected people it produces.3
The native 集団感染 remained available as the clearer competitor for the same referent. The inverse case, where the loanword wins and pushes the native word out, is the subject of "Loanwords That Replaced Native Words: When Gairaigo Pushes Out Wago."
Case study: the 2020 COVID katakana wave (a dated example)
The early 2020 COVID-19 pandemic produced a burst of public-health katakana in Japanese: クラスター, オーバーシュート, ソーシャルディスタンス, パンデミック, and ロックダウン.36 This 2020 wave works like a controlled experiment, because some terms settled and others did not.
クラスター settled into wide use but drew pushback in 2020, competing with the clearer native 集団感染 that the term itself was glossed against.3 パンデミック also settled: it rode the World Health Organization (WHO) Director-General's declaration of 11 March 2020 (reported the next day) into mainstream Japanese media and stuck.13
Other terms moved in the opposite direction. オーバーシュート was intended to mean "an explosive surge in cases." It was criticized as opaque, did not settle as everyday vocabulary, and faded after 2020.36 ソーシャルディスタンス was flagged as confusing and as quasi-wasei-eigo. English public-health usage is "social distancing" (a behavior), while "social distance" is a sociology term. WHO itself shifted in 2020 from "social" to "physical" distancing.618
At a press conference on 24 March 2020, 河野太郎 (Kōno Tarō) said, "Is there any need to deliberately say in katakana what can be said in Japanese? People are saying it's hard to understand." He also signaled he would ask the health ministry to use plainer wording.3 This is the 外来語疲れ (loanword fatigue) reaction in public discourse, tied to the 2020 pandemic.
These five terms are presented as word-level glosses with their 2020 dates, rather than as sentence examples. The reason is that the wave is a dated case study, and clean corpus sentences for the COVID katakana were not available.3136 The point is the contrast: in 2020, the same pipeline ran on five words at once. Some settled, while others were paraphrased or faded.
Good to know
Coining is bottom-up; regulation is top-down and late
Committees and broadcasters react to words that have already spread. NINJAL's paraphrase project explicitly targeted loanwords already circulating in public and administrative use. The notation directive also standardizes spellings after words are in use.45 If you wait for an "official" katakana word, you misread the order of events: coining comes first, regulation second.
A "failed" loanword often half-survives
A sidelined loanword rarely disappears completely. オーバーシュート faded from everyday use after 2020 but persists in finance and epidemiology jargon. The mainstream slot went to plainer wording.36 Failure usually means demotion to a niche register, not disappearance.
When the pipeline produces something that isn't English at all
Semantic shift plus Japanese re-compounding can yield wasei-eigo: an "English-looking" coinage with no native-English equivalent. ソーシャルディスタンス is a borderline case. English uses "social distancing," and "social distance" carries a different sociological meaning, so the Japanese form behaves like a re-coined term.618 Coining sits upstream of this phenomenon, which is the subject of "Wasei-Eigo: The English-Looking Japanese Words That Aren't English."
The 4-mora gravity well
Long borrowings tend to get clipped toward roughly four morae as they settle. パーソナルコンピューター became パソコン, and スマートフォン became スマホ.92 When you meet a long katakana word, expect a shortened form to emerge. This mechanism is detailed in "Shortened Loanwords in Japanese: Why パソコン, リモコン, and アポ Get Clipped."
See also
- What Is Gairaigo? A Guide to Loanwords in Japanese
- Pre-English Loanwords in Japanese: Portuguese, Dutch, German, and More
- Wasei-Eigo: The English-Looking Japanese Words That Aren't English
- Shortened Loanwords in Japanese: Why パソコン, リモコン, and アポ Get Clipped
- Loanwords That Replaced Native Words: When Gairaigo Pushes Out Wago
- Japanese Internet, Tech, and Smartphone Vocabulary: ネット, アプリ, SNS, and the Gairaigo-Density Words