There is a wide range of transliterations and transcriptions of Japanese, but there is little variation in the standard pronunciation of Japanese sounds. So, how should the words be pronounced?
After the table, several most common phenomena are described.
| Romanization | Cyrillization | IPA | Comments |
|---|---|---|---|
| a | а | ɑ | Latvian <a>. The non-prolonged version of the <a> in English word <father>. |
| i | и | i | Spanish, Latin etc. <i>. The non-prolonged version of the <ee> sound in English words <deep>, <see>. |
| u | у | ü̜ spelled as ɯ | Between Spanish, Latin etc. <u> or the <u> in English words <put>, <look> and Russian and Polish <ы> (Polish <y>). Like the Spanish/Latin sound, but with the lips compressed towards each other rather than rounded (but not spread to the sides). |
| e | э | e | Spanish, Latin etc. <e>. The first sound in the diphthong in English words <lay>, <play>. The vowel sound produced by <е> in Russian words <снег>, <цвет>. |
| o | о | o | Spanish, Latin etc. <o>. The first sound in the General American version of the diphthong in English words <hole>, <stone>. |
| k | к | k | |
| ky, k [before i] | к [before и, я, ю, ё] | kʲ | Palatalized version of [k], as in Russian (<кино>, <звонкий>) or Polish. In particular, the sequence /kʲi/ is approximated by the English sequence /kiː/ (except for the wovel’s length), as in <keen>. |
| g | г | ɡ, ŋ | All speakers can be classified into three groups: those who use both phones interchangeably (A-speakers), those who use [ŋ] consistently (B-speakers) and those who use [ɡ] consistently (C-speakers). The majority are A-speakers, the minority C-speakers. |
| gy, g [before i] | г [before и, я, ю, ё] | ɡʲ, ŋʲ | Palatalized version of [ɡ], as in Russian (<гений>, <биология>) or Polish. In particular, the sequence /gʲi/ is approximated by the English sequence /giː/ (except for the wovel’s length), as in <geek>. |
| s | с | s | |
| sh, sy, s [before i] | с [before и, я, ю, ё] | ɕ | Polish palatalized version of [s], as in <siedem> and <miesiąc>. This phone lies between [sʲ] (Russian palatalized version of [s], as in <семь> and <месяц>) or, to less extent, [s], and [ʃʲ] or [ʃ] (the <sh> in English words <shall> and <shield>). See Ishwar’s pictures of places of articulation, alveolo-palatal fricatives and palato-alveolar fricatives for a visual reference (this phone is alveolo-palatal). |
| z, dz, d [before u] | дз | ʣ, z | [ʣ] is the traditional version. |
| j, jy, zy, dy, z [before i], d [before i] | д(з) [before и, я, ю, ё] | ʥ, ʑ | Polish palatalized versions of [ʣ] and [z] respectively. [ʥ] is much more popular than [ʑ], it is also the traditional version. The former lies between [ʣʲ] (Russian palatalized version of [ʣ] and [d], as in <день> and <диво>) or, to less extent, [ʣ], and [ʤʲ] or [ʤ] (the <j>, <g> and <dg> in English words <jump>, <age> and <edge>), the latter between [zʲ] (Russian palatalized version of [z], as in <земля> and <зима>) or, to less extent, [z] and [ʒʲ] or [ʒ] (the <s> and <z> in English words <pleasure> and <azure>). See Ishwar’s pictures of places of articulation, alveolo-palatal fricatives and palato-alveolar fricatives for a visual reference (this phone is alveolo-palatal). |
| t | т | t | |
| ch, ty, t [before i] | т [before и, я, ю, ё] | ʨ | Polish palatalized version of [ʦ], as in <kwiecień> and <stulecie>. This phone lies between [ʦʲ] (Russian palatalized version of [ʦ] and [t], as in <тень> and <тьма>) or, to less extent, [ʦ], and [ʧʲ] or [ʧ] (the <ch> and <tch> in English words <reach> and <witch>). See Ishwar’s pictures of places of articulation, alveolo-palatal fricatives and palato-alveolar fricatives for a visual reference (this phone is alveolo-palatal). |
| ts, t [before u] | ц, т [before у] | ʦ | |
| d | д | d | |
| n | н | n | |
| ny, n [before i] | н [before и, я, ю, ё] | ȵ spelled as ɲ | Polish palatalized version of [n]. This phone lies between [nʲ] (the Russian palatalized version of [n], as in <небо>) and [ɲ] (the <gn> in French word <champignon>) and shares its articulation point with [ɕ], [ʑ], [ʨ] and [ʥ]. See Ishwar’s pictures of places of articulation, alveolo-palatal fricatives and palato-alveolar fricatives for a visual reference (this phone is alveolo-palatal). |
| h | х | h | The <h> in English words <halt>, <happy>. This phone is mere aspiration. |
| hy, h [before i] | х [before и, я, ю, ё] | ç | Palatalized version of [x], as in Russian (<махина>, <хилый>) or Polish. The <ch> and <g> in the ich-Laut (i. e. standard) pronunciation of German words <ich>, <Mädchen>, <niedrig>, more specifically its variety that sounds more like [h] than like [ʃ]. |
| f, h [before u] | ф, х [before у] | ɸ | Similar to [f] (the <f> in English words <fight> and <friend>), but the lip does not touch the teeth. |
| p | п | p | |
| py, p [before i] | п [before и, я, ю, ё] | pʲ | Palatalized version of [p], as in Russian (<писать>, <песня>). In particular, the sequence /pʲi/ is approximated by the English sequence /piː/ (except for the wovel’s length), as in <peek>. |
| b | б | b | |
| by, b [before i] | б [before и, я, ю, ё] | bʲ | Palatalized version of [b], as in Russian (<бить>, <обида>). In particular, the sequence /bʲi/ is approximated by the English sequence /biː/ (except for the wovel’s length), as in <bee>. |
| m | м | m | |
| my, m [before i] | м [before и, я, ю, ё] | mʲ | Palatalized version of [m], as in Russian (<мелкий>, <милый>). In particular, the sequence /mʲi/ is approximated by the English sequence /miː/ (except for the wovel’s length), as in <mean>. |
| y | [the first sound in stand-alone я, ю, ё] | j | |
| r, l | р, л | ɺ | To an English speaker’s ears, the pronunciation lies somewhere between a flapped <r> [ɾ] (the <tt> and <dd> in <better> and <ladder>), an [l] and a [d], sounding most like [l]. Wikipedia says that Spanish speakers often associate it with a soft <r>, as in <pared>. First, position the tip of your tongue to where you pronounce /ʤ/ (the <j> and <g> in English words <jump> and <age>); it should touch the upper alveolar ridge a little above the teeth. If you are familiar with the General American sound [ɾ] described above, pronounce it. Keep pronouncing it and try to make it sound more like [l], while at the same time retaining all properties of [ɾ]. The only thing you have to do is to allow the air flow over the sides of the tongue rather than its middle. If you end up having a [l] or a [t] or [d], you have gone too far, just try the same technique again. If you are not familiar with the [ɾ] sound, try to pronounce a [d] or [r] with a single contraction of the muscles, retracting the tongue tip behind the alveolar ridge; then follow the same instructions. Note that a [ɾ] is also okay (as for understanding each other, a [r] or a [l] will also be fine), but are we not learning the precise sounds? ;‑) |
| ry, ly, r [before i], l [before i] | р [before и, я, ю, ё], л [before и, я, ю, ё] | ɺʲ | Palatalized version of [ɺ]. The palatalization technique follows the same scheme as for [kʲ] or [pʲ], with the sound being pronounced at the same articulation point as for the non-palatalized version (i. e. at the same position as for /ʤ/), however the palatalized phone sounds more like [ɾ] and [d] than like [l]. |
| w | в | ɰ | This sound is somewhat the non-syllabic version of the vowel /ü̜/. It relates to /ü̜/ as /w/ relates to /u/. |
| N, n’, n̄, n [before a consonant other than y], m [before a consonant other than y] | нъ, н [before a consonant], м [before a consonant] | n̩ before n, ȵ, t, ʨ, ʦ, d, ʥ, ʣ; ŋ̩ before k(ʲ), g(ʲ), ŋ; m̩ before m(ʲ), p(ʲ), b(ʲ); ɴ̩ word-finally before a pause; ã, ĩ, ɯ̃, ẽ, õ (the preceding vowel nasalized) in all other occasions and also word-finally before a pause | The actual sound is homorganic (i. e. shares the same point of articulation) with the following stop, affricate or nasal consonant; if it is not followed by a consonant or if the following consonant is neither of these, the phoneme is realized as a nasal vowel (remember [ɺ] is a flap and not a stop). If followed by nothing at all, [ɴ̩] is used, but a nasal vowel is also possible (sometimes a [m̩] is also found in final position). In isolation, only [ɴ̩] or [m̩] is possible. Note that the sentence-final [m̩] is rare; do not use it. This phoneme occupies a whole mora, i. e. it lasts same long as a syllable with a non-prolonged vowel. The phone [ɴ] is similar to [ŋ], but the back of the tongue touches the uvula rather than the soft palate (this is about the deepest that the back of the tongue can actually touch). |
Things described here are phenomena, which means they are possible but not required or even neccessarily often heard, except for the dropping of [ɯ]!
Some phenomena do not affect the perception of words:
Here is a brief list of phenomena when a phone is replaced with a phone usually belonging to another phoneme or disappears altogether that are presumed the most common by Chortos‑2 (these occur mostly in informal speech):
