Nasalization

The articulation

The roof of the mouth is split into two zones: a hard palate at the front and a soft palate (velum) at the back. The velum is a movable flap. When raised against the back wall of the pharynx, it seals off the nasal cavity, and all air leaving the lungs flows out through the mouth. When lowered, it opens a passage so that air can also flow out through the nose.

For most oral sounds — /a/, /i/, /s/, /p/, /k/ — the velum is up. For nasal consonants /m/, /n/, /ŋ/, the velum is down and the mouth is closed somewhere, forcing all air through the nose. For nasal vowels and nasalized consonants, the velum is down but the mouth stays open: air flows through both passages at once.

The acoustic consequence is a coupling effect. The nasal cavity adds an extra resonator to the vocal tract, introducing nasal formants (around 250 Hz, 1000 Hz, 2200 Hz) and damping the oral formants. The result is a vowel with a more "spread out" spectrum — perceptually, a "muffled" or "stuffy" quality.

You can produce nasalization deliberately. Pinch your nose shut while saying man: the /m/ is impossible because the only escape route is closed. Now hold your nose during cat with no nasal consonants: it sounds normal. The pinch test is the field linguist's quick check for nasal airflow.

Nasal vowel vs nasal consonant

Nasalization in vowels and nasalization in consonants are produced the same way (velum down) but distinguished by the oral configuration.

Sound type	Velum	Oral cavity	Airflow	Example
Oral vowel	Raised (closed)	Open	Mouth only	English /a/ in cat
Nasalized vowel	Lowered (open)	Open	Mouth + nose	French /ã/ in chant
Oral consonant	Raised (closed)	Closed at some point	Mouth only, after release	English /b/ in bat
Nasal consonant	Lowered (open)	Closed at some point	Nose only	English /m/ in mat
Nasalized consonant	Lowered (open)	Partially constricted	Mouth + nose	Portuguese /w̃/ in mãe
Pre-nasalized stop	Lowered then raised	Closed throughout	Nose then mouth	Swahili /ᵐb/ in mbwa

The pre-nasalized stop is a single consonant, not a sequence of two — it patterns as one segment in the language's phonology. Several Bantu and Austronesian languages have prenasalized stops as their basic stop inventory.

Cross-linguistic nasalization

The same articulation — velum lowered during a vowel — has very different status from language to language.

Language	Status	Inventory	Example contrast / pattern
French	Phonemic	3-4 nasal vowels (/ɑ̃, ɛ̃, ɔ̃/, sometimes /œ̃/)	beau /bo/ vs bon /bõ/; pas /pɑ/ vs pan /pɑ̃/
Portuguese	Phonemic	5 nasal vowels + nasal diphthongs	vi /vi/ "I saw" vs vim /vĩ/ "I came"; mãe /mɐ̃j̃/
Polish	Phonemic	2 nasal vowels (/ɛ̃, ɔ̃/, written ę, ą)	kąt /kɔ̃t/ "corner" vs kat /kat/ "executioner"
Hindi	Phonemic	Most vowels have nasal counterparts	hai /hɛ/ "is" vs haiñ /hɛ̃/ "are"
Guarani	Phonemic, with nasal harmony	Every vowel has a nasal counterpart	káva "wasp" (oral) vs kãva (nasal harmony spreads through the word)
Yoruba	Phonemic	5 nasal vowels (no /e, o/ nasal counterparts)	ṣe "to do" (oral) vs ṣe with nasal vowel — different word
English	Allophonic	None phonemic; vowels nasalize before /m, n, ŋ/	bean [bĩn] vs bee [biː] (nasalization predictable)
German	Allophonic + loanwords	Native: allophonic. Loans: /ɛ̃, ɑ̃/ from French	Native kann [kãn]; loan Pension [pɑ̃ˈzioːn]
Spanish	Mostly allophonic	Vowels lightly nasalized adjacent to nasal consonants	mano [ˈmãno]; no minimal pairs
Mandarin	None on vowels	Nasal coda /-n, -ŋ/ counts; no phonemic nasal vowels	Distinction is in the coda, not the vowel

The pattern is clear: nasalization is universal as a phonetic phenomenon (velum lowering happens in every speech tradition near nasal consonants), but languages differ on whether they let it become contrastive. Where a former /-n/ has weakened or deleted, the nasal vowel that replaced it tends to become phonemic.

English nasalization in detail

English speakers nasalize vowels constantly without noticing. The rule is simple: a vowel adjacent to a nasal consonant inherits nasalization. The effect is anticipatory (the velum starts dropping during the vowel before the nasal) and carry-over (the velum stays partly down into the vowel after a nasal).

Word	Phonemic	Phonetic (narrow)	Notes
bee	/biː/	[biː]	No nasal nearby; pure oral vowel
bean	/biːn/	[bĩːn]	Anticipatory nasalization on /iː/
numb	/nʌm/	[nʌ̃m]	Carry-over from /n/ + anticipation toward /m/
mat	/mæt/	[mæ̃t]	Carry-over from initial /m/
cat	/kæt/	[kæt]	No nasal context; vowel is fully oral
sing	/sɪŋ/	[sĩŋ]	Anticipatory toward velar nasal

The nasalization is so automatic that English speakers struggle to suppress it when learning a language that contrasts oral and nasal vowels. Conversely, French speakers learning English often produce nasal vowels in words that should be oral, because their velum drops earlier than English requires.

Worked example: French beau vs bon

The classic French minimal pair shows nasalization at work. Beau "beautiful" is /bo/; bon "good" is /bõ/. The two words differ in exactly one feature: the vowel's [±nasal] specification.

1. Articulation comparison. In beau, the lips part for /b/, the tongue is back, the lips round, and the velum stays raised throughout the /o/. In bon, everything is identical except that the velum lowers for the duration of the vowel.
2. Acoustics. A nasal formant appears around 1 kHz; the F1 of the oral vowel is damped. Spectrograms show the difference plainly.
3. Historical origin. Latin bonum /'bonum/ → Old French /bon/ → Middle French /bõn/ → Modern French /bõ/. The /n/ left phonetic residue (nasalization on the vowel) before deleting; once /n/ was gone, the nasal vowel was the only meaning-distinguishing cue.
4. Phonemic status. Because bo and bõ form a minimal pair, /o/ and /õ/ are separate phonemes of French. In English, no such minimal pair exists — there is no English word that contrasts a nasal /õ/ with an oral /o/.

Theoretical perspectives

Framework	How it represents nasalization	Predictions	Notes
SPE features	Binary [±nasal]	Nasalization is a single feature; spreads via rule	Cannot distinguish degrees of nasality
Feature geometry	[nasal] under Soft Palate node	Nasal feature spreads as a unit; explains harmony	Predicts nasal harmony patterns; matches Guarani
Articulatory phonology (Browman & Goldstein)	Velum gesture overlapping with vowel gesture	Degree and timing of nasalization vary continuously	Captures English carry-over and anticipation
Aerodynamic models (Ohala)	Nasalization predicted from coupling and pressure	Some nasal sounds are unstable; explains historical paths	Connects phonetics to sound change
Optimality Theory	Constraints like *NASALV banned in some grammars	Languages without nasal vowels rank *NASALV high	Encodes typological distribution as constraint ranking

Variants

• Anticipatory (regressive) nasalization. Vowel before a nasal consonant: English [bĩn]. The velum lowers during the vowel in preparation.
• Carry-over (progressive) nasalization. Vowel after a nasal consonant: English [mæ̃t]. The velum hasn't fully raised yet.
• Nasal harmony. Nasalization spreads through a whole word or domain. Guarani: once a vowel is nasal, all vowels in the morphological domain become nasal. Tucano languages have similar patterns.
• Phonemic split (historical). Once-allophonic nasal vowels become phonemic when the conditioning consonant deletes — the path to French nasal vowels.
• Pathological nasalization. Cleft palate or velopharyngeal insufficiency causes hyper-nasalization on vowels that should be oral; speech-language pathology treats this clinically.
• Stylistic / sociolinguistic. Some American dialects (Southern, Appalachian) have heavier nasalization than General American; some West African Englishes lighter.

Common pitfalls

• Confusing "nasal vowel" and "vowel + /n/". A French nasal vowel has no consonant; the /n/ is gone. Beginners write bon as /bon/ when it is /bõ/.
• Missing English allophonic nasalization. Native speakers don't notice that bean has a nasal vowel, but acoustic measurement and the pinch test confirm it. Phonetic transcription must mark it.
• Treating velum lowering as binary. Velum height is continuous; nasalization can be partial. The IPA diacritic is binary, but real speech is gradient.
• Forgetting carry-over. Anticipatory nasalization (V before N) is well-known; carry-over (V after N) is real but often ignored in textbook descriptions.
• Assuming nasal vowels need a former /n/. Most arose that way historically, but contact-induced nasalization and emergent harmony systems can produce nasal vowels by other paths.
• Confusing the IPA tilde with Spanish ñ. Spanish ⟨ñ⟩ is the palatal nasal /ɲ/; the IPA tilde over a vowel marks vowel nasalization. They are unrelated despite the visual similarity.