Nihongoing 日本語イン Verb Inflection, Part 1: T h e S y S T E M
The iconic floating shrine and Otorii on Miyajima (宮島) Island.
Hello, Nihongo-ers!
Though studying Japanese can be very daunting to those of us who come from an English-language background, not all parts of learning Japanese are equally difficult. One area of Japanese grammar that is remarkable for its clarity and regularity is verb inflection. (For comparison, note that learners of English must recognize that forms as disparate as “am, is, was, were, been” have to be looked up in a single place in the dictionary!) By “inflection,” we mean varying the form of a word, often through suffixing, to mark certain grammatical categories.
The grammatical categories for which a verb inflects vary from language to language. Like English, Japanese marks the past tense with a suffix ([+ed] in English, [+ta た] in Japanese). But whereas languages like English and Spanish adjust the forms of verbs according to the subject (called “agreement,” for example English 2nd-person versus 3rd-person “you do” → “s/he does”, Spanish haces → hace), Japanese does not.
Important categories marked directly on the Japanese verb are Politeness and Negativity. Naturally these categories exist in English and in all languages: English uses words like “no” and “not” to negate and shows politeness by means of word choice. But in Japanese, the properties of Politeness or Negativity of an utterance are central in determining the form a verb takes.
For today’s overview of Japanese verb inflection, we’ll discuss Past and Non-past forms of verbs that are the most regular in inflection: those that end in vowels and in easily identifiable consonants. (Note: “Non-past” is my preferred label, as verbs in this category refer to present and future action; the label “Present” is therefore inaccurate.) Though there is, as I said, lots of regularity in verb inflection, there are certain instances where we see changes affecting sounds in the stem or the suffix.
I group the verbs into six classes based on the final sound of the verb stem. Verb stems ending in vowels are in Class 1, and all the other classes have verb-final stems. I have ordered the stem-final sounds in the usual Japanese order: first the velar consonants /k g/ (Class 2), then the fricative /s/ (Class 3), the dental consonants /t n/ (Class 4), then the labial consonants /b m/ (Class 5), and finally the liquid /r/ (Class 6). This survey of verb classes is followed by the very important irregular verb suru which means ‘to do’. One class has been omitted from this presentation and will be taken up in the next post.
Table 1 shows all the suffixes we’ll see in today’s selection of forms. The plus sign (+) marks where a suffix begins. Letters “X” and “Y” in the English portions are placeholders for the subject and the object of the verb
Table 1: Japanese verb suffixes, in English alphabetical order.
For more practice with forms of these two verbs, see LanGo Institute’s flashcard deck here: https://tinycards.duolingo.com/decks/N5HDUtDs/lango-jp-eat-drink .
Note that the Past Negative suffix [+(a)nakat+ta] contains the basic Past marker [+ta]. The Politeness auxiliary is never verb-final and is therefore always followed by a vowel /i, u, e/. In the Non-past this auxiliary shows up as ます [+mas+u], from [+mas +(r)u], where the /r/ of the final suffix is deleted because of the preceding /s/. When the past marker [+ta] is suffixed to the Politeness auxiliary, the vowel /i/ intervenes to give ました [+mash-i +ta]. The vowels /i u/ in these endings are voiceless and therefore spoken very quietly.
Prefixes do not play any role in the inflection we’ll be looking at in this post (but note that the honorific prefix [お o+] is added in certain very polite forms).
CLASS 1: Verbs ending in a vowel
The verbs in this class only end in one of two vowel sounds: /i e/. In our first example, i-ru ‘to be’, the final vowel is also the only vowel in the verb stem (compare the English verb “owe” whose stem is just the diphthong /ow/ with no consonants). Table 2 shows partial paradigms of the verbs ‘be’ and ‘eat’; again, letters “X” and “Y” in the English portions are placeholders for the subject and the object (if there is one) of the verb.
Table 2: Verbs with stems ending in vowels.
The favored syllable shape in Japanese is CV-CV: one consonant + one vowel, and repeat. So when a stem ends in V, it makes sense for the suffix attached to it to start with a C. When the stem ends in a consonant, by the same logic, a suffix shape with initial V is preferred.
Because the stem-final sound in the verbs in this class is a vowel, it is always the consonant-initial form of the suffix that immediately follows these stems that is selected, so る [+ru] is used here rather than う [+u], and so is ない [+nai] rather than あない [+anai]. The forms of a suffix may vary, but it is the same grammatical meaning being added to a verb form. The infinitive-marking suffix [+i] is deleted with these verbs. The suffixes with initial /t/ and the Politeness auxiliary [+mas-] attach to the same base (which is not true of the other verb classes below), so these verbs are very easy to segment into a stem and an ending: the stem never changes.
CLASS 2: Verbs ending in a velar consonant /k g/
In this and all the following verb classes, it is the vowel-initial suffix forms that are selected: [+u] not [+ru], and neg. [+anai] and [+anakatta]. These verbs, unlike those in Class 1, have stems that appear in more than one shape: when the suffix immediately following the stem starts with /t/, the stem-final consonant is deleted, so /kak-/ ‘write’ shows up as [ka] in the Past-tense kaita and the Gerund kaite.
Note: If this seems odd, listen closely to English “does” and “has”: the stems appear in altered shapes rather than their basic ones, or else the results of [+s] suffixation would be [duwz] “dooz” and [hævz] “haves”. (The words [duwz] “dues” and [hævz] “halves” show that these are perfectly acceptable English syllable shapes; it’s just an idiosyncratic feature of certain verbs that their stems take alternate forms sometimes.)
Table 3 shows forms of two verbs, 書く ‘to write’ and 聞く ‘to hear’, with stem-final voiceless /k/.
Table 3: Verbs with stems ending in /k/.
There is only one Japanese verb with final voiced /g/ that occurs with any frequency, and that is oyog-u ‘to swim’. The stem-change rule we saw above applies here too, with a twist: when stem-final /g/ meets suffix-initial /t/, the /g/ is deleted, the intervening vowel /i/ remains, and the /t/ voices to [d]. This results in Past-tense oyo-i-da and Gerund oyo-i-de. Forms of this verb are shown in Table 4.
Table 4: Verbs with stem ending in /g/.
Note that the rule deleting the stem-final consonants does not apply when Politeness-marking [+mas-] is suffixed: oyog-i-mas-u, not *oyoimasu.
CLASS 3: Verbs ending in the fricative /s/
For these verbs, only one type of change affects the stem: if the vowel /i/ immediately follows the stem, /s/ shifts to the palatal fricative [sh] (IPA [ɕ]). Otherwise no changes apply to stem or suffixes; vowel-initial suffix variants are selected where applicable. Table 5 shows the forms of two verbs, 話す ‘to speak’ and 渡す ‘to hand over’, with stems ending in /s/.
Table 5: Verbs with stems ending in /s/.
CLASS 4: Verbs ending in dental consonants /t n/
In this class we have, when we look at the romanized forms of the verbs, three apparent alternants for each stem: one ending in [t], one in [ts], and one in [ch]. But from the Japanese perspective these are just minor variants of a single sound /t/ (allophones of a phoneme). When a high vowel /i u/ follows a dental stop /t d/, a transitional [s] or [z] sound develops on the way from the release of the consonant to the onset of the vowel (t+u → t-s-u = tsu), and the consonants [s z] shift to palatal [ʃ ʒ] when /i/ or /y/ follows.
When the /t/-initial suffixes are added, the stem appears unchanged to English speakers (tat-te, mot-te below), but here, a change does take place from the Japanese perspective. These are instances of the long consonant, where the /t/ is sustained for one beat (mora), i.e., the same length of time as a short vowel. This consonant lengthening is shown in the spelling with a small [tsu]-kana (っ or ッ). For instance, contrast the different sizes っ versus つ in the Sino-Japanese compound 出 発 ‘depart(ure)’. The basic forms of the components are 出 (しゅつ shutsu) and 発 (はつ hatsu). In composition, the first component loses its final vowel /u/ while the second replaces initial /h/ with /p/, and the result is the long consonant /-pp-/ in 出発 (しゅっぱつ shuppatsu). Compare the full-size [つ tsu] in しゅつ, はつ versus the small [っ] in the middle of しゅっぱつ. Listen here to the contrast of short single /t/ in kata ‘shape, style’ versus the long double /tt/ in katta ‘X bought Y’.
Note that only forms with /t/-suffixes attaching to the root have this long consonant. Table 6 shows inflected forms of /t/-final verb stems.
Verbs with stems ending in /t/.
As was true in the other consonant-final stem classes (2 and following), the forms where an /m/-initial suffix follows have an /i/ intervening between stem and suffix. We can say that the base to which the /m/-suffixes attach is the infinitive form. Table 7 has forms of the only Japanese verb whose stem ends in /n/, ‘to die’.
Table 7: Verb with stem ending in /n/.
As we can see, there are no changes to the spelling of the verb stem when we consider the romanized forms. But the onset /n/ of /na ni nu/ is a very different thing from the syllable-closing /N/ in shin-da. In Japanese the syllable-closing nasal takes one beat (mora) to articulate, as was true of the long consonants we saw in the preceding section. Listen here to compare the sound of the onset /n/ occurring twice in nani ‘what?’, versus the syllable-closing /N/ in nan desu ka ‘what is it?’
Here again, /m/-initial suffixes have a vowel /i/ intervening between stem and ending; and when the stem meets a /t/-suffix, /n/ closes the preceding syllable and voices the following /t/ to [d] (compare the forms of the /g/-final stem oyog- ‘swim’ in Class 2 above).
CLASS 5: Verbs ending in bilabial consonants /b m/
In this class, too, there is a change in the form of the stem when a /t/-suffix is added directly to the stem, for instance in the Gerund forms below, ason-de and yon-de, which reflect asob+te and yob+te. When the final /b/ of the stem meets the /t/ of the suffix, /b/ is replaced by the syllable-closing [N], and /t/ is voiced to [d]. This change of /b+t/ to [N-d] accounts for the only alternations in this set of verb forms. Table 8 has forms of two /b/-verbs, ‘play’ and ‘call’.
Table 8: Verbs with stems ending in /b/.
In the /m/-stems in the following table, we see a very similar change: stem-final /m/ is replaced by syllable-closing [N] when suffixal /t/ follows directly, and this /t/ is voiced to [d]: Past ‘drank’ non-da from nom+ta. Table 9 shows forms of the /m/-verbs ‘drink’ and ‘read’.
Table 9: Verbs with stems ending in /m/.
Note that the forms yonda, yonde each have two interpretations: the former can mean ‘X called Y’ (where the root is yob-) or ‘X read Y’ (where the root is yom-); the latter can mean ‘calling’ (where the root is yob-) or ‘reading’ (where the root is yom-).
We turn now to our last class of verbs in today’s post:
CLASS 6: Verbs ending in /r/
The verbs in this class are very numerous, so it is an important one to become familiar with early on. Note that the final syllable /ru/ in the dictionary (Non-past plain affirmative) forms kiru, shiru is different in its makeup from the /ru/ in Class 1: there it was the full form of the Non-past suffix [+ru] being selected by the vowel-final stem, whereas here the stem-final consonant is /r/, and the Non-past suffix variant is [+u]. This is the reason kir-u rhymes with i-ru ‘be’, but the Polite forms diverge: kir-i-mas-u ‘cut’ (4 syllables) versus i-mas-u ‘be’ (3 syllables). Vowel /i/ intervenes between stem-final /r/ and suffix-initial /m/, whereas no such insertion takes place in i-mas-u.
Table 10: Verbs whose stems end in /r/.
When /t/-suffixes are added directly to a stem, its final /r/ is replaced by the long consonant (っ) we encountered in Class 4; in romanization we write -tt- (Gerund kitte from kir+te).
This concludes today’s overview of regular verbs. There is also a special apparently irregular class (including ka-u ‘to buy’), where certain sound changes have produced some odd results. I will devote a future blog post to this extremely interesting class, so stay tuned!
IRREGULAR VERB: ‘To do’
Because of its high frequency, we should not omit from today’s lesson the important verb suru ‘to do’ for comparison with the other inflection patterns we’ve seen. Table 11 shows inflected forms.
Table 11: Forms of the verb /suru/ ‘to do’. Note: For shimasen deshita, it is customary to Romanize this form as two words, but note that “word boundary” is not a feature of Japanese writing as can be seen in the Japanese kana spellings.
This is one of the most important verbs to master. With it are formed thousands of verbs made from borrowed expressions, such as benkyō suru ‘to study’. Looking closely at the forms, we see a novel distribution of two alternants, し shi and す su: the latter appears in the dictionary form suru only. The alternant shi seen everywhere else is actually the infinitive derived by suffixation of [+i] to the single-consonant root /s/. So instead of expected Neg. *s+anai, the vowel /i/ has spread from forms like Polite sh-i-mas-u to become the root-vowel, causing selection of the short form of the Negative suffix, [+nai], and so on.
OVERVIEW: Structure of the Japanese verb
Table 1 in the introductory section set out all of the suffixes we would see added in verb inflection, and in many cases we saw that there were alternant forms. In this closing section I’d like to consider what orders all of these elements--stems, suffixes, intervening vowels--occur in.
The quickest way to deduce the overall structure is to look at those forms that have the greatest number of suffixes. Now in most of the grammatical categories in play (Tense, Politeness, Negativity), there is a single marker (a suffix) that is either present or absent. If a verb-form is in the Polite style, there will be a form of the [+mas-] suffix; if the suffix isn’t there, the verb form is not Polite. In the same way, there either is a Negative suffix in a verb, or there isn’t. So the most informative forms are the Polite Negative ones. Let’s look at the Past-tense Polite Negative form of ‘to read’: yomimasen deshita ‘X did not read Y’, analyzed in Table 12. The column heading “i.V.” stands for the intervening vowel seen in many forms; the plus sign marks the onset of a true suffix.
Table 12: Japanese verb inflection template.
(Note that the Past tense is marked with deshita only after [+mas+en], and elsewhere the basic form of the suffix appearing is [+ta]; see Table 1.)
There are other suffixes (such as Potential [+e], Causative [+(s)ase]) that can follow a verb root, but those belong to the sphere of word derivation rather than verb inflection, as in those cases the addition of the suffixes creates a new verb distinct, in form, meaning, and syntactic behavior, from the basic form. But the order shown in Table 12 is the only permissible one for a Japanese verb.
Thanks for Nihongo-ing with me today, I hope you’ll stop by again for future posts on all things related to the Japanese language!
きをつけてね！ Take care!
Charity’s story about her time on the beautiful la Isla de Chiloé during our unforgettable Spanish immersion trip in Chile.