Nasal g (が行鼻濁音)

If you listen closely to an older Tokyo speaker or an NHK newscaster, the g in the middle of a word often sounds gentler than the g at the start of one — softer, slightly hummed, almost swallowed. That is 鼻濁音 (bidakuon), the nasalized pronunciation of the が-row (が・ぎ・ぐ・げ・ご). It is one of the quiet markers of traditional standard Japanese, and understanding it explains why certain speakers sound the way they do. Crucially, this is a comprehension and register-awareness topic, not a production target: no one will misunderstand you if you never use it, and most younger Tokyo speakers no longer do.

What the sound actually is

The nasal g is the velar nasal [ŋ] — the ng in the English word singer (not the ng-g of finger). Your tongue makes the same back-of-the-mouth closure as for a hard [ɡ], but instead of releasing a little burst of air through the mouth, you let the air flow out through the nose. The result is a soft, humming g that blends into the surrounding vowels rather than popping out of them.

Compare the hard plosive and the nasal within the が-row:

Because it is a nasal, [ŋ] is a close relative of the moraic ん, which is itself [ŋ] before a k or g (as in 音楽, ongaku). See The Moraic N (ん) for that neighboring sound, and Consonants for the が-row's plosive baseline.

The rule: initial hard, medial nasal

Traditional standard Japanese follows a clean positional split:

• Word-initial が-row → hard plosive [ɡ].
• Word-medial が-row → nasal [ŋ].
• The subject/object particle が → nasal [ŋ], because inside a phrase it is never word-initial.

This sentence contains two nasal g's for a bidakuon speaker: the が inside 鏡 (kagami → ka-*ŋa-mi) and the particle が right after it (*ga → ŋa). Contrast that with a word-initial が, which stays a firm [ɡ]:

学校 (gakkō) begins with が, so the g is a hard [ɡ] even in the most traditional speech. Two more everyday medial cases:

In りんご (ringo, "apple") the ご is medial → [ŋo]; in 誰かが the particle → [ŋa].

The exceptions

The medial → nasal rule has a few well-established holdouts where the g stays plosive even mid-word:

• Reduplicated onomatopoeia keep a hard [ɡ] in both halves: ガラガラ (garagara, "empty/rattling"), ゲラゲラ (geragera, "guffawing").
• Transparent compounds whose second element is still felt as its own word often keep the plosive at the seam: 小学校 (shōgakkō, "elementary school") is frequently said with a hard [ɡ] because 学校 is heard as a standalone word.
• Some speakers keep the number 五 (go) plosive inside compounds, though usage varies.

Both が's in ガラガラ stay hard reduplicated onomatopoeic shape blocks nasalization.

Status: prestige, receding, regional

Three things make bidakuon a register feature rather than a pronunciation rule you must obey:

1. It is prestige/traditional. Bidakuon was the norm of refined Tokyo (山の手) speech and is still taught to broadcasters: NHK announcers are trained to nasalize, and the NHK pronunciation-and-accent dictionary historically marked the nasal が-row with a small ring (a handakuten-like °) over the kana. It also survives in classical singing, enka, and traditional stage speech. For the wider prestige "standard language" question, see 標準語 vs 方言.

2. It is receding. Among younger Tokyo speakers, the hard plosive [ɡ] is spreading into medial position, and full command of bidakuon is increasingly a mark of older or professionally-trained speakers. A learner who never nasalizes sounds simply modern, not wrong.

3. It is regional. Bidakuon was strongest in eastern and northern Japan; large parts of western Japan (including the Kansai region) traditionally used the plosive everywhere. So even among native speakers, its presence or absence is partly a map of where someone is from (regional).

Why textbooks and dictionaries mark it

If bidakuon is optional, why do some materials flag it at all? Because dictionaries built on the broadcast standard — chiefly NHK's — codified the traditional distribution, and older textbooks inherited it. When you see a が-row kana printed with a little ring, or a note that "medial が is nasal," it is documenting the prestige broadcasting norm, not a rule that ordinary conversation enforces. Knowing this saves you from two anxieties: thinking your ear is broken when speakers differ, and thinking your Japanese is deficient because you use the plosive.

Common misconceptions

Because production is not the goal here, the "mistakes" are misunderstandings rather than pronunciation errors.

Misconception 1: The nasal g is a different word or a mistake. It is neither.

Misconception 2: You must learn to produce it to sound native. You do not. Native Tokyo speech is drifting away from it. A consistent plosive [ɡ] sounds contemporary and standard.

Misconception 3: The subject particle が has a special "harder" sound. The opposite is true in traditional speech: because が is never phrase-initial, the particle が is exactly where the nasal [ŋ] shows up most reliably.

For a bidakuon speaker, that particle が is [ŋa]; for a younger speaker it may be [ɡa]. Either is correct today.

Misconception 4: A soft, hummed g means the speaker mumbled. It means they are using bidakuon — often a sign of careful, trained, or older speech, the opposite of sloppiness.

Key takeaways

• 鼻濁音 is the nasal [ŋ] realization of the が-row — the g of English singer — used word-medially (and on the particle が) in traditional standard Japanese, while word-initial が-row stays a hard [ɡ].
• Exceptions keep the plosive: reduplicated onomatopoeia (ガラガラ) and transparent compound seams (小学校).
• It is allophonic (never changes meaning), prestige/broadcast-associated, receding in modern Tokyo, and regional — so treat it as something to recognize, not something you must produce.