The Japanese ら-Row (ら り る れ ろ): The Alveolar Tap That Is Neither R Nor L

The Japanese ら-row (ら り る れ ろ) is built on a single consonant phoneme, conventionally written /r/. Its citation-form realization is the voiced alveolar tap [ɾ]: a single, brief contact of the tongue tip against the ridge behind the upper teeth.¹² English ears tend to file each token under either /r/ or /l/ because English contrasts those sounds and Japanese does not. That is why "neither R nor L" is the most honest one-line description of the sound.¹³

Overview

The five kana ら り る れ ろ in hiragana and ラ リ ル レ ロ in katakana all share the same onset consonant, followed by /a, i, u, e, o/.¹⁴ That onset is one phoneme, not two. The perceptual scatter that English speakers hear comes from the listener, not the sound.³⁵

Why this sound trips English speakers

English has two contrastive liquid phonemes (R-like and L-like sounds): /r/, a postalveolar approximant (typically [ɹ̠] in General American, often with rhotic constriction), and /l/, an alveolar lateral approximant.⁶³ Swapping one for the other changes the word, as in right vs. light or row vs. low.⁶

Japanese has exactly one liquid phoneme.¹⁷⁸ Its phonetic realization lands acoustically between the two English categories. English ears therefore tend to assign each token to whichever English bin it most resembles, producing the impression that the same kana is "sometimes R, sometimes L."

Best and Strange (1992) document the mismatch in both directions. Japanese listeners hear English /r/, /l/, and /w/ as variants of a single Japanese category, while English listeners hear Japanese /r/ tokens scatter between English /r/, /l/, and /d/.³

The diagram below shows the asymmetry. English maintains two boxes where Japanese maintains one; both languages cover roughly the same articulatory region, but they slice it differently.

One phoneme, two English bins

The mismatch is not that Japanese has "an R that is also an L." Japanese has one liquid phoneme that does not participate in the R/L contrast at all; English splits a region that Japanese leaves whole.¹³

Where you meet it

The same /r/ phoneme appears in the particle より (yori, "from / than"), in the conditional and provisional endings ～たら and ～れば, in Sino-Japanese morphemes such as 六 (roku, "six") and 料理 (ryōri, "cooking"), and in nearly every English loanword written with the ら-row kana.⁹⁴ Standard Modified Hepburn romanization writes the row as ra ri ru re ro, and Kunrei-shiki uses the same letters.¹

The kana row also carries every English /r/ and every English /l/ that gets borrowed into Japanese. The two source phonemes collapse onto one Japanese phoneme in loanword adaptation.⁹ This is why ライト covers both English light and English right with no spelling distinction.

The phoneme: one Japanese /r/, not two

IPA: [ɾ] as the citation form

The standard citation form of Japanese /r/ is the voiced alveolar tap [ɾ]. This means a single, brief contact of the tongue tip against the alveolar ridge, with voicing throughout.¹²⁶ The IPA Handbook's Japanese illustration (Okada 1999) transcribes the phoneme as [ɾ] in citation. It notes that the prototype is "an apical tap, either alveolar [ɾ] or postalveolar [ɾ̠]."²

Labrune summarizes the same point in one sentence: Japanese /r/ is "consistently voiced, apical, alveolar or post-alveolar, with a short or weak closure."⁷ Ladefoged and Maddieson classify the sound as a tap (a single quick gesture) rather than a flap (a glancing movement from a different starting position). The two IPA terms are often used interchangeably for [ɾ].⁶

The Japanese phonetic-terminology label for this category is 弾き音 (hajikion, "flicked sound"), which is the term used in the NHK pronunciation reference and in mainstream Japanese phonetics teaching.¹⁰

Why "neither R nor L" is a phonemic statement, not just acoustic

The R/L contrast in English is phonemic: swapping the two changes the word, as in right vs. light, row vs. low, and correct vs. collect.⁶³ In Japanese, no minimal pair distinguishes a hypothetical /r/ from a hypothetical /l/, because there is only one phoneme. The contrast is unaskable, not just absent.¹⁷⁸

Loanword adaptation makes the collapse visible. English light and English right both borrow as ライト (raito). English glass and English grass both borrow as グラス (gurasu). English lock and English rock both borrow as ロック (rokku).⁹ No mechanism inside Japanese spelling can keep the source distinction.

Vance frames the conclusion sharply: the Japanese phoneme is not "in between" English /r/ and /l/ in any phonological sense; it simply does not participate in a contrast that English maintains.¹

Articulation walkthrough

Step 1: find the alveolar ridge

The alveolar ridge is the bony shelf immediately behind the upper front teeth. In IPA terminology, "alveolar" means an articulation that places the tongue tip or blade at that ridge.⁶¹¹ For Japanese /r/, the active articulator is the tongue tip (apex), and the passive articulator is the alveolar ridge. This is what "apical alveolar" means in Labrune's description.⁷¹²

English /r/, by contrast, is articulated with a bunched tongue body raised toward the postalveolar region, with no contact at the alveolar ridge.⁶³ English /l/ holds the tongue tip in sustained contact at the alveolar ridge with airflow around the sides of the tongue.⁶ The Japanese target sits at the same place as English /l/ but with the contact released almost immediately.

Step 2: a single voiced tap, not a sustained contact

A tap is one quick gesture: the tongue tip moves to the alveolar ridge, makes momentary contact, and releases.⁶¹¹ The contact is too brief to build up oral pressure for a stop release, which is what keeps [ɾ] acoustically distinct from [d].²⁶

Voicing is continuous through the gesture, which is what distinguishes [ɾ] from a voiceless tap.⁷² Spanish single-tap /ɾ/ (as in pero, "but") is articulatorily very close to Japanese /r/. Spanish trilled /r/ (as in perro, "dog") involves multiple taps and is not the Japanese target.¹³

Step 3: the "butter / water" English-tap anchor and its limits

In American English, /t/ and /d/ between vowels often surface as an alveolar tap [ɾ]. This is the intervocalic flapping rule: butter [ˈbʌɾɚ], water [ˈwɑɾɚ], city [ˈsɪɾi], ladder [ˈlæɾɚ].⁶ This intervocalic [ɾ] is phonetically the same gesture as the Japanese /r/ tap and is widely used as a pedagogical anchor for English-speaking learners.⁸

The limit of the anchor: American English never starts a word with [ɾ]; intervocalic flapping is a context-sensitive allophone of /t/ and /d/, not a word-initial option.⁶ Japanese uses /r/ in word-initial position freely in Sino-Japanese vocabulary and loanwords.⁴ Pure Yamato (native Japanese) vocabulary historically restricted morpheme-initial /r/. This is one reason native-stem ら-row words are sparse compared with Sino-Japanese and loanword ら-row words.¹⁴

Borrow the butter, then move it

Drill the tap on the second syllable of "butter," "water," "city," and "ladder" until the gesture is automatic. Then drill the same tongue motion at the start of a syllable. The gesture is the same; only the position in the word is new.⁸⁶

Step 4: word-initial vs. intervocalic positions

Intervocalic /r/ (between two vowels) is the closest to the textbook tap, as in あれ (are), それ (sore), 桜 (sakura), and 心 (kokoro).¹²

桜さくらの花はながきれいです。¹⁵
"The cherry blossoms are pretty."

Word-initial /r/ and /r/ after the moraic nasal /N/ are often realized with momentary alveolar-ridge contact before release. They are "described variably as a tap, a 'variant of [ɾ]', 'a kind of weak plosive', and 'an affricate with short friction, [d̠ɹ̝̆]'."²⁷ Arai (2013) documents that Japanese-acquiring children commonly substitute a stop for word-initial /r/, producing りんご "apple" as [dingo] or [gingo]. This is consistent with word-initial position being articulatorily closer to a stop than the intervocalic tap is.¹⁶

来年らいねん、東京とうきょうに行いきます。¹⁵
"I'm going to Tokyo next year."

For learners, the safe target in all positions is the textbook tap [ɾ]; the [d]-like onset is a native-speaker variant, not a pedagogical goal.⁴⁸

Allophones and contexts

The single Japanese /r/ phoneme surfaces in a range of phonetic variants depending on position, speaker, and register. The map below collects the variants discussed in the descriptive literature. The citation form sits at the centre, and the others are positional or stylistic deviations from it.

Word-initial [d]-like onset

Okada describes the word-initial realization as one in which "the tip of the tongue is at first momentarily in light contact with the alveolar ridge before being released rapidly by airflow." He gives the IPA transcription [d̠ɹ̝̆] for this affricate-like variant.² Labrune treats this as the same phoneme: the increased closure duration in initial position is a positional allophone, not a separate sound.⁷

The variant is most audible in careful or emphatic word-initial speech. In fluent connected speech, the difference between initial and medial tokens compresses.⁷²

Retroflex flap [ɽ] in some speakers

Some speakers, particularly in informal or emphatic registers, retract the tongue tip (curl it slightly back) and produce a retroflex flap [ɽ] instead of the alveolar tap.¹⁷ An even more extreme variant is the alveolar trill [r] (multiple rapid taps), which Japanese phoneticians call 巻き舌 (makijita, "rolled tongue"); it is marked as rough or aggressive and is associated with stereotyped "tough-guy" speech.¹⁷

These variants appear in Vance and Labrune as part of the phoneme's allophonic range. They are not targets for learners and are not the citation form for broadcast Japanese.¹⁷¹⁰

Trill carries social weight

A consistent Spanish-style multi-tap trill on every ら-row token reads as the makijita register. Japanese listeners hear it as rough or theatrical. The unmarked target is one tap.⁷¹³

Lateral-like realizations

Vance notes that "the apical alveolar or postalveolar lateral approximant [l] is a common variant in all conditions" of Japanese /r/.¹ Akamatsu argues this lateral variant is better described as a lateral flap [ɺ] than as a sustained lateral approximant [l]. The reason is that the contact duration is too brief to match the English /l/ profile.¹² A retroflex lateral [ɭ] also appears, particularly before /i/ and /j/, in some speakers.⁷

The existence of these lateral variants is why an English /l/ substitution by a learner sometimes sounds "acceptable" to a native ear. Still, the substitution is acceptable as a phonetically nearby variant, not as the citation form.¹¹²

Why the L1 substitutions fail

English /r/ substitution: the bunched-tongue trap

General American /r/ uses a bunched or retroflex tongue body, with no tongue-tip contact at the alveolar ridge and frequent pharyngeal constriction.⁶³ Substituting this articulation into Japanese yields a long, rhotic vowel-like onset that bears no resemblance to a single tap. The tongue is in the wrong place and stays there too long.⁶

The corrective drill is mechanical: move the active articulator from the tongue body (English /r/) to the tongue tip (Japanese /r/). Then time the contact as a single touch, not a sustained gesture.⁸⁴

Bunched-tongue /r/ on every ら-row kana

Pronouncing 寿司を食べる as [sɯʃi o tabeɹɯ] (English bunched [ɹ] for る) is the clearest English-L1 error. The target is [sɯɕi o tabeɾɯ], with one tongue-tip touch at the alveolar ridge for る.⁶³

English /l/ substitution: sustained contact

English /l/ holds the tongue-tip contact at the alveolar ridge for the duration of the consonant, with lateral airflow.⁶ The Japanese tap releases the contact almost immediately. Sustaining contact produces a mushy, English-accented impression.¹⁴

The corrective drill is to shorten the contact rather than change the place of articulation. The alveolar ridge is correct; the duration is not.⁸

Spanish or Italian trill: too many taps

Spanish has a contrast between /ɾ/ (single tap, as in pero) and /r/ (trill, multiple taps, as in perro).¹³ Japanese has only the single-tap variant. Substituting a Spanish-style trill is hyper-articulated and lands in the marked makijita register rather than neutral speech.¹⁷

Spanish-L1 learners typically have the right articulator and place but need to suppress the multi-tap habit on every ら-row kana.¹³

Korean ㄹ and Chinese r: close but not the same

Korean ㄹ is a single phoneme with two allophones (context-based pronunciations): an alveolar tap [ɾ] between vowels (e.g. 나라 nara, "country") and an alveolar lateral approximant [l] syllable-finally and before consonants (e.g. 말 mal, "word").¹⁷ The Korean intervocalic tap is articulatorily very close to Japanese /r/. Korean learners' main adjustment is to use the tap variant everywhere, including syllable-final position, rather than letting the lateral allophone surface.¹⁷

Mandarin /r/ (pinyin r, as in 日 rì) is a voiced retroflex approximant [ɻ] or fricative [ʐ], articulated with the tongue tip curled back without contact. It is acoustically and articulatorily distant from a tap.¹⁸ Mandarin-L1 learners need to flatten the retroflex curl and add a brief tongue-tip contact, rather than fine-tune the existing gesture.¹⁸

L1 background	Default reflex	What to change
English	Bunched [ɹ] (postalveolar approximant)	Move from tongue body to tongue tip; touch ridge once and release.
English	Sustained [l] (lateral approximant)	Keep the ridge contact; release immediately.
Spanish, Italian	Trilled [r] (multiple taps)	Cap at exactly one tap per kana.
Korean	[ɾ] intervocalic, [l] elsewhere	Use the tap variant in every position.
Mandarin	Retroflex approximant [ɻ] or fricative [ʐ]	Flatten the curl; add a brief alveolar contact.

Drills and minimal contrasts

The ら-row tongue twister

Sequential single-mora practice is the standard Japanese elementary-school drill for the ら-row. It is presented as a 早口言葉 (hayakuchi-kotoba, "fast-mouth words") ladder.⁴ The drill target is one clean tap per kana with no carry-over of contact between adjacent moras.

ら り る れ ろ。⁴
"ra ri ru re ro (recited as a row, like an English A B C D E)."

The classic ら-row twister stacks /r/ tokens in close succession. It is a standard test of whether the single-tap gesture survives at speed:

すもももももももものうち。¹⁹¹⁵
"Plums and peaches are both kinds of peaches."

Words that flush out the English-R reflex

Drill these in different positions so the tap appears word-initially, between vowels, and after vowels. Recite slowly enough to feel one touch per kana, then build up speed.

りんごを食たべます。¹⁵
"I eat an apple."

これは桜さくらの花はなです。¹⁵
"This is a cherry blossom."

よろしくお願ねがいします。¹⁵
"Pleased to meet you."

Borrowed words that look like English

Loanwords are the highest-risk environment for this sound. The source word still echoes in the learner's ear and pulls the gesture back toward the English /r/ or /l/. The corrective tactic is to commit to the Japanese tap, not to bend back toward the source word.

部屋へやのライトをつけてください。⁹¹⁵
"Please turn on the room light."

レストランで晩ばんご飯はんを食たべました。¹⁵
"I ate dinner at a restaurant."

ラジオを聞ききます。¹⁵
"I listen to the radio."

Good to know

The butter / water / city anchor and its word-initial limit

The intervocalic flap in American English (the [ɾ] in butter, water, city, ladder) is articulatorily identical to Japanese intervocalic /r/. That is why it is the standard pedagogical anchor in introductory Japanese-phonology texts.⁶⁸ The anchor has a limit. American English never starts a word with [ɾ], so drilling らりるれろ in word-initial position with the butter anchor still requires moving the gesture from "between vowels" to "at the start of a syllable."⁶

弾き音 hajikion as the Japanese-phonetics name

The Japanese phonetic-terminology label for the tap category is 弾き音 (hajikion, literally "flicked sound"). It comes from the verb 弾く (hajiku, "to flick, to strike off").¹⁰ The Japanese label captures the motor instruction ("flick the tongue") more directly than the English "flap" or "tap" does. Keeping the Japanese term in a study notebook keeps the articulation in mind.¹⁰

Sustaining the contact (English /l/ leakage)

A common error is to pronounce ありがとう with the tongue holding contact at the alveolar ridge through り, as in an English /l/. The correct form releases the contact immediately:

ありがとう。¹⁴
"Thank you." (with a single, brief tap on り, not a sustained lateral contact)

The rhythm of Japanese is mora-timed. If the consonant of one mora bleeds into the next, the word's timing collapses along with its sound.¹⁴

Bunching the tongue (English /r/ leakage)

A second common error is to pronounce 寿司を食べる with English-style bunched [ɹ] on the る. This gives [sɯʃi o tabeɹɯ] instead of the target [sɯɕi o tabeɾɯ]. The English approximant /r/ bunches the tongue body and never makes alveolar-ridge contact. Substituting it produces a rhotic colouring across the vowel that natives hear as a foreign accent.⁶³ The corrective is mechanical: move the active articulator from the tongue body to the tongue tip. Time the contact as a single touch.

The trilled R (巻き舌) is socially marked

The multi-tap alveolar trill [r] is a real allophone of Japanese /r/, but it is socially marked. It is associated with rough, aggressive, or stereotyped "tough-guy" speech (the so-called "yakuza R").⁷ Learners producing a Spanish-style trill on every ら-row token sound either comical or threatening depending on context. The unmarked target is the single tap.⁷¹³

Loanwords collapse English R and L

English light and right both become ライト. Grass and glass both become グラス. Lock and rock both become ロック.⁹ The collapse is direct evidence that the Japanese phoneme covers both English categories at once, not that it sits "between" them. Memorising one collision pair, such as ライト for light and right, anchors the abstract phonemic claim to something concrete.⁹

Don't trust romaji for this sound

Writing the kana as r in Hepburn is a one-way mapping. It tells a Japanese reader which kana is meant, but it tells an English reader to produce an English /r/, which is the wrong articulation.¹ Treat the romaji r as an arbitrary label for the alveolar tap, not as a phonetic instruction.

See also

• How to Pronounce つ (tsu) in Japanese: The Voiceless Alveolar Affricate English Lacks
• How to Pronounce ふ (fu) in Japanese: The Voiceless Bilabial Fricative English Lacks
• Long vs. Short Vowels in Japanese: The Distinction Beginners Miss
• Why "Tokyo" Is Two Syllables in English and Four Morae in Japanese: Loanwords as a Timing Drill
• The Japanese Vowel Inventory: Five Vowels, Done Right
• Geminate Consonants (Sokuon っ): The Silent Pause

References

Footnotes

1. Vance, Timothy J. The Sounds of Japanese. Cambridge University Press, 2008. ISBN 978-0-521-61754-3. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰

2. Okada, Hideo. "Japanese." In Handbook of the International Phonetic Association. Cambridge University Press, 1999, pp. 117–119. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

3. Best, Catherine T., and Winifred Strange. "Effects of phonological and phonetic factors on cross-language perception of approximants." Journal of Phonetics, vol. 20, no. 3, 1992, pp. 305–330. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰

4. Banno, Eri, et al. Genki I: An Integrated Course in Elementary Japanese. 3rd ed., The Japan Times, 2020. ISBN 978-4-7890-1730-5. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰

5. Goto, Hiromu. "Auditory perception by normal Japanese adults of the sounds 'L' and 'R'." Neuropsychologia, vol. 9, no. 3, 1971, pp. 317–323. ↩

6. Ladefoged, Peter, and Ian Maddieson. The Sounds of the World's Languages. Blackwell, 1996. ISBN 978-0-631-19815-4. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰

7. Labrune, Laurence. The Phonology of Japanese. Oxford University Press, 2012. ISBN 978-0-19-954583-4. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶

8. Tsujimura, Natsuko. An Introduction to Japanese Linguistics. 3rd ed., Wiley-Blackwell, 2014. ISBN 978-1-4443-3773-1. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

9. Kang, Yoonjung. "Loanword phonology." In The Blackwell Companion to Phonology, edited by Marc van Oostendorp et al., Wiley-Blackwell, 2011, pp. 2258–2282. (Treatment of English /r/ and /l/ both mapping to a single Japanese liquid in loanword adaptation.) ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

10. NHK放送文化研究所, ed. 『NHK日本語発音アクセント新辞典』. 日本放送出版協会, 2016. ISBN 978-4-14-011345-5. ↩ ↩² ↩³ ↩⁴

11. International Phonetic Association. Handbook of the International Phonetic Association: A Guide to the Use of the International Phonetic Alphabet. Cambridge University Press, 1999. (IPA chart and conventions for tap/flap [ɾ], lateral flap [ɺ], retroflex flap [ɽ], alveolar trill [r].) ↩ ↩²

12. Akamatsu, Tsutomu. Japanese Phonetics: Theory and Practice. LINCOM Europa, 1997. ISBN 978-3-89586-095-9. ↩ ↩² ↩³

13. Hualde, José Ignacio. The Sounds of Spanish. Cambridge University Press, 2005. ISBN 978-0-521-54538-9. (Spanish /ɾ/ tap vs. /r/ trill contrast.) ↩ ↩² ↩³ ↩⁴ ↩⁵

14. Frellesvig, Bjarke. A History of the Japanese Language. Cambridge University Press, 2010. ISBN 978-0-521-65320-6. (Historical phonology; native vocabulary morpheme-initial /r/ restriction in Yamato-kotoba.) ↩

15. Tanaka, Yasuo, and Jim Breen. Tanaka Corpus, integrated into the Tatoeba Project (https://tatoeba.org). CC BY 2.0 FR; example sentences cross-verified against Genki I⁴ and elementary-vocabulary entries in Tsujimura⁸. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

16. Arai, Takayuki. "On Why Japanese /r/ Sounds Are Difficult for Children to Acquire." Proceedings of Interspeech 2013, ISCA, 2013. https://www.isca-archive.org/interspeech_2013/arai13b_interspeech.pdf ↩

17. Lee, Hyun Bok. "Korean." In Handbook of the International Phonetic Association. Cambridge University Press, 1999, pp. 120–123. (Korean ㄹ as tap intervocalically, lateral elsewhere.) ↩ ↩²

18. Duanmu, San. The Phonology of Standard Chinese. 2nd ed., Oxford University Press, 2007. ISBN 978-0-19-921579-9. (Mandarin /r/ as retroflex approximant [ɻ] or fricative [ʐ].) ↩ ↩²

19. 国立国語研究所 (NINJAL). 『現代日本語書き言葉均衡コーパス』(BCCWJ). https://clrd.ninjal.ac.jp/bccwj/ (2011, balanced corpus of contemporary written Japanese; cited here for high-frequency lexical attestations.) ↩