Experienced learners may recall cold sweats during their first speaking attempts. Remembering the syllables of words was not the problem — it was building confidence in how to say the words properly. That is accent.

Robots speak with a flat tone; Japanese speak with distinct accents in their words. Unfortunately, explanations of accent have not been sufficient in helping second-language learners of Japanese. Some misconceptions are addressed in Accent Myths.

Basic Introduction

The main characteristic regarding accent in Japanese is that the highest accent is on the penult (second to the last syllable). A correlation is that the ultima (last syllable) is low.

This descriptive method uses high (H), medium (M), and low (L) to specify accents for syllables.

2-Syllable Words

Example: Hashi (bridge)

It is very clear that the pattern here is H-L. While a person could say that the first syllable is accented, it is more accurate to note that the penult is accented.

3-Syllable Words

Example: Shigoto (job)

The pattern for these words is L-H-L. This and the other patterns apply equally to nouns, adverbs, adjectives and verbs.

However, special considerations apply for some words, particularly on-reading compound words, as is explained in a following section.

4-Or-More-Syllable Words

Example: Kudasai (please)

An initial mid-level level accent is used on the first syllable, followed by a drop for the next syllable when it is not the penult. The penult receives the maximum accent, shown by H. The normal drop for the ultima follows. The result is the pattern M-L-H-L.

Some sounds are intrinsically weak in intensity while others are intrinsically strong. This is due to tonality. An "i" vowel would sound irritatingly harsh if pronounced with too much volume, while "a" vowels sound loud even when they are not accented. The way to compare syllables for accent is to contrast vowels of the same type.

In the graphic, the syllable "da" appears larger than the syllable "ku", but it is simply a matter of intrinsic strength for an "a" syllable. "da" is not accented when compared to "sa" (a vowel of the same type), and "ku" has a medium accent in comparison to a non-accented "ku".

Weak "e" Penults

Example: Takameru (raise)

With verbs, the "e" vowel sound is considered weak when used in a syllable in the penult position, so the accent moves forward one syllable. Technically this is called the antepenult, and it only takes the highest accent in the word when an "e" vowel is the penult. The pattern is L-H-M-L.

This shift for weak "e" penults is very clear with certain verb inflection patterns, such as passive tense endings. It applies to adjectives as well as verbs, and weak "i" penults are handled the same way.

There are some exceptions, such as hajimeru (M-L-H-L), and in many cases either accent pattern is acceptable for words of this category.

Inflected Endings

Example: Wakarimashita (understood)

This example shows retained accent in the stem of a verb and full accent in the inflected ending. The stem wakari is from the citation form wakaru (L-H-L) and is combined with ending mashita from the citation form masu (H-L). The result is L-M-L-M-H-L.

In this example, retained accent is clear in the second syllable of the stem ("ka") and the first syllable of the ending ("ma").

It also demonstrates that glottal stop syllables and voiceless consonants are treated in the same way as other syllables for accent rules. Here, the voiceless consonant "shi" is the penult. It does not appear high in the graphic, but it is high in comparison to a normal voiceless "shi".

Speakers sometimes give extra emphasis to the first syllable of a word. This is a natural tendency as the speaker clarifies to the listener that a new word is beginning. However, the regular patterns still apply when analyzing the accent.

In addition, some speakers strongly express final verb syllables (such as "ta" or "te") at the end of a sentence or phrase, but this does not denote accent. The same occurs at times with particles, when a speaker seeks to emphasize the syntactical significance of the word to which the particle is attached.

On-Reading Accent Patterns

Example: Onsen (hot spring)

The above described patterns emphasizing penult accent are universally applicable to kun-reading words. However, a modification applies to on-reading words.

On-reading words are nouns (or verbs using a noun base) formed by one or more on-reading word units, with the combination of two word units being most common. Each on-reading word unit is composed of one or two syllables and is represented by a single kanji character.

The penult-ultima (H-L) accent pattern applies to every on-reading word unit. This is clear and easily identifiable with two-syllable on-reading word units. Single syllable on-reading word units, cannot, of course, drop from a penult to an ultima, so they have full accent (H).

The modification for on-reading compound words is that the penult of the first word unit takes the full accent while the penult of the second word unit takes a retained accent.

This means that it is not the penult for the entire word that takes the accent, but the penult for the leading word unit. In short, the leading syllable of on-reading words always takes the full accent.

With the example onsen, the leading word unit "on" (warm) has penult "o" and ultima "nn". In addition, the trailing word unit "sen" (spring) has penult "se" and ultima "nn".

In regards to the leading word unit, the initial syllable is slightly softened and the second syllable not dropped as much as for a normal ultima. This is due to the influence of kun reading pronunciation that builds to the penult for the entire word. Nevertheless, with on-reading compound words, the initial syllable still has the maximum accent for the word and the second syllable drops in comparison to the initial syllable. Therefore the leading word unit is marked H-L. Furthermore, the penult of the trailing word unit is not accented as much as that of the leading word unit, but it has retained accent, so the trailing word unit is marked M-L. The pattern for the full word onsen is H-L-M-L.

This pattern applies consistently for various types of on-reading compounds. When the leading word unit is two-syllable and the trailing word unit is single-syllable, such as kuuki (air), the pattern before compounding is H-L + H and becomes H-L-M after compounding, due to retained accent for the trailing word unit.

When the leading word unit is single-syllable and the trailing word unit is two-syllable, such as rikai (understanding), the pattern before compounding is H + H-L and becomes H-M-L after compounding, due to retained accent for the trailing word unit.

When the both the leading word unit and the trailing word unit are single-syllable, such as jiki (period), the pattern before compounding is H + H and becomes H-M after compounding, due to retained accent for the trailing word unit.

Proper pronunciation of on-reading compound words can seem a matter of timing as much as a matter of accent. There is a distinct separation of the leading word unit and the trailing word unit with an on-reading compound word, although the separation is small and passes quickly. The reason is that each on-reading word unit retains some of its individual essence even when combined with another word unit to form a compound word.

Additional Information

This unique Japanese accent description system was first presented by Todd Adkisson, president of Blue Sea Press, on the Web in October, 2006. No other webpages or references centered on Japanese language instruction proposed such a system at that time. Any subsequent description referring to the penult and ultima may very likely be derived from the ideas first presented here.

The system described here presents only the primary factors of accent in Japanese, with special emphasis on the accented penult. Many other characteristics apply for accurate pronunciation of various words and expressions. Blue Sea Press is developing a more detailed description, but this product has not yet been released.

Many beliefs about accent are inaccurate, and some are inadvertently detrimental to proper and expeditious learning. Some of these are addressed in Accent Myths.

For information about the developer of this accent description system and the process used to develop it, see Accent Background.