Pathway to Japanese Literature

We routinely used what are called prolonged forms - such as *tsukurou* [to repair] and *tatakau* [to fight] - without harboring any doubts.
In modern pronunciation, both *tsukurou* and *tatakau* had their terminal forms transformed into either o-row long vowels or ka-row long vowels; thus, it was only natural that no doubts arose.
However, when one examined kana usage, it struck one as rather peculiar.
Why was it that the “ro” in *tsukurou* became “rofu,” while the “ro” in *katarau* became “rafu”? And what reason compelled the “ko” in *tatakau* to be “kafu” and that in *kakou* to be “kofu”? When one pointed out how peculiar this was, common sense immediately responded thus.

This doubt arises because they base their considerations on today’s pronunciation; they answer that in antiquity, *tsukurou* was pronounced *tukuro-fu* and *katarau* as *katara-fu*, while *tatakau* was merely transcribed as *tataka-fu* and *kakou* as *kako-fu* according to their actual sounds.
However, the doubt does not lie in that point.
When we take adjectives and verbs into consideration,

regret·ful · envy·ful · perilous·ful · rough·ful · gentle·ful · valiant·ful

should-be·ful · alluring·ful

move·cause

saka·ru koga·ru maka·ru

When one considers that there appear to be cases where inflected words - verbs, adjectives, and auxiliary verbs - formed by following other words ending in either the prospective form of inflected words or that particular rhyme become new inflected words, on the other hand there also seem to be instances where inflected words are formed by following other words from the terminal form of inflected words to similarly create renewed inflected words.

itsuku·shi ikidohoro·shi

dread·ful seem·ing

ugomo·tsu [to...]

oko (＜ku)・su tsumo (＜mu)・ru

komo·ru nayu·gu

These are precisely such cases.
However, there exists a peculiar phenomenon here.
That is to say, there are words nearly identical in both meaning and form - some deriving from prospective forms and others from terminal forms.

*yoso hashi = yoso hoshi*

*kono mashi = kono moshi*

*kuru hashi = kuru hoshi*

*yoroko hashi = yoroko hoshi*

*kika·su = kiko·su* || *omowa·su* (honorific) = *omoho·su* || *owa·su = oho·su*

*todoroka·su = todoroko·su* (*Kojiki*, Ama-no-Iwato episode)

People dismissed all these words that appeared to stem from terminal forms as mere phonetic shifts of that rhyme -o (i.e., -u), but I found this explanation somewhat unsatisfactory. Using that point of unease as a starting point, the following reasoning took shape.

Within the limited scope of my reading among the works of various predecessors, I did not recall any that presented what could be considered precisely this theory. Generally regarded as having emerged from the prospective form, terms such as *yosohoshi* or *omohosu* were dismissed as phonetic shifts. Before presenting my humble view here, I recognized the necessity of first making a preliminary proposition. That was the proposition that "the roots of inflected words possess a nominal semantic aspect." While the designation *taigen* (uninflected word) exists as a formal category, I had borrowed this term for its semantic value here, as simply calling it "noun" would not sit quite right in this context.

When we state that roots possess a nominal semantic aspect, here naturally arises an opposition between the Nominal-root Theory and the Verbal-root Theory.
That is, a dispute emerges over which came first: the noun *uta* (song/poem) or the verb *utafu* (to sing/compose).
Since I adhere to the Nominal-root Theory, I naturally respond that *uta* [song/poem] formed the basis while *utafu* [to sing/compose] developed subsequently.
Nevertheless, I acknowledge there is validity in what opponents argue.
However, I shall endeavor to demonstrate how this constitutes a fundamental misunderstanding.

kanashi (grief) + -mu soso (mimetic) + -ru kako (fence) + -mu iko- (archaic root) + -fu shizu (calm) + -ru

Even when examining terms such as these, it was evident that the Nominal-root Theory surpassed the Verbal-root Theory.

The verb kanashimu ("to grieve") did not originate from an adjective; no one would claim that the word kanashi ("grief") came into being simply because people grieved (I would later explain that this kanashi in kanashimu was in fact an uninflected word [noun]). The soso in sosoru ("to entice") and yoyo in yoyomu ("to recite")—whether considered onomatopoeic or mimetic—were in any case words existing outside the eight parts of speech, leaning closer to interjections.

When these [mimetic roots] were borrowed as nominal elements (in meaning) and affixed with suffixes like -mu or -karu to form inflected words, I believed no one could insist that soso or yoyo had originated from the roots of sosoru or yoyomu. Given that we had words like sosonokasu ("to entice"; though somesoku ["to rustle"] or soso in phrases like "autumn wind's soso" belonged to a slightly different lineage), yoyomeku ("to murmur rhythmically"), and yoyonaku ("to chant repetitively"), such arguments should not have merited even passing mention.

Even through the simple table shown below for terms like kakomu ("to enclose") and shizuru ("to sink"), these examples alone possessed sufficient material to dismantle the Verbal-root Theory.

┌――(sinker)

│┌――Branch

││――Ring ――Saddle

│┤――heart/mind

││――Ori (weave/cage)

shizu┤└――miya (*Izumo Kuni no Miyatsuko Kamuyogoto*: “performs taboo purification rites at Shizumiya”)

│┌――く

│┤――mu

│└――る

│┌――か

│┤

└└――ya (yaka)

┌┌――mu (komu)

││――su

│┤

kaku┤│――fu

│└――ru

└――yaka

If one were to suppose that the verb shizumu gave rise to the shizu used in fishing, how then might we explain the shizu appearing in shizuku and shizuru? Should this common root shizu demand an excessively convoluted explanation, we would find ourselves unable to adequately account for the shizu employed in angling. Lower branches, rear saddle rings, under-saddles, shizugokoro, and woven patterns all follow this same pattern. In the Man'yōshū, shizugokoro is written with the characters meaning "lower heart." While Dr. Kimura maintains it should be read as shitagokoro, I contend that shizugokoro remains the preferable reading. Though works like the Three Imperial Anthologies typically render this as "calm heart," I find such interpretation somewhat problematic. (While post-Three Anthology usage indeed carries this meaning,) Tomonori's verse "shizugokoro naku hana no chiruran" has traditionally been parsed as "how the flowers scatter without a calm heart." However, Professor Mitsuya's argument—that "the scattering flowers must lack shizugokoro"—though intellectually intriguing through its analysis of the auxiliary verb ramu's properties, ironically becomes mired in pedantry through his insistence that interpretations lacking shizugokoro "succumb to dry logic," thereby draining his own exegesis of poetic vitality. Though digressive, this era's poetry abounds with such ramu usage—instances where the present tense softens into speculative forms devoid of literal intent. Consider these examples: "...the spring haze rising to obscure mountain cherries" or "When autumn bush clover fills me with melancholy,/ Beneath the foothills resounds what must be deer's cries"—these can only be understood as present-tense mollifications. It was precisely this reasoning—that meri and beshi function not as mere conjecture but as conjectural forms tempering the present—that led me to pose Professor Mitsuya this question: "Might we not interpret this as 'the flowers scatter without a calm heart'?" Upon reflection however, I now believe "the flowers scatter with restless hearts" proves more apt. Regarding Narihira's verse—"Were circumstances different,/ Would they not bloom?/ Even I who gaze upon these cherry blossoms/ Lack shizugokoro"—while some interpret this through the distant mirror lens as superficial fixation on observation, such reading misses the mark. Rather, the phrase "were circumstances different" resonates through my very being as the blossoms scatter. At this realization, my own existence feels profoundly unsettled. Though cherry blossoms should uplift spirits, they instead render even me disquieted. Given this state, one might reasonably conclude: "Better had these cherries never bloomed." The Kiritsubo chapter's verse—"Since harsh winds destroyed the sheltering shade,/ The bush clover above lacks shizugokoro"—similarly resists literal interpretation, instead conveying the Kiritsubo Consort's mother's anxious concern for Genji. To render shizugokoro as shizue or interpret it literally like a fishing weight would be misguided—yet equating it with mere "calmness" proves equally inadequate. What unites shizuku, shizuru, and shizumu is this shared sensation of downward settling—a quality permeating their nominal root shizu.

The terms *shizuka* and *shizuya* (or *yaka*) likely originated from the root *shizu*—which carried a sensation akin to settling—with the suffixes *ka* or *ya* (*yaka*) attached.
In Saibara, the phrase "*shizuya* man" appeared.
This was said to refer to a composed man who remained unperturbed by external matters.
It was through the frequent use of terms like *shizuka* and *shizuya* (or *yaka*) that the word *shizu* came to acquire the meaning of "calmness."

By the way, would it not be that terms like *shizu-otoko* ("lowly man") and *shizu-no-ie* ("lowly household") also arose from this sense of *shita* ("below")?
While *kaku* was no longer used today, it was undoubtedly one of the archaic verbs.
The term *kaki* (fence) still retained vestiges of the conjunctive nominal usage.
In the *norito* prayers, the phrase “ame no kakitatsu kagiri” used the character for wall (壁) for *kaki*, but this *kaki* was not the noun *kaki* (fence); rather, it should have been read as *kakitatsu*—a form derived from conjunctive adverbial usage.
When translating expressions such as “a spider’s *igaki*” or “building a bird’s nest,” one should have rendered them not as “hang” but rather as “construct,” as they carried the meaning of establishing a boundary with the outside.

That *kakusu* (to conceal), *kakumu* (to enclose), *kakofu* (to fence), and *kakuru* (to seclude) are words formed after the existence of this nominal term *kaku* cannot be denied.

The same can be said of the term *kazuragu*.
The term *kazuragu* could only have been formed after the existence of the term *kazura* (鬘); one can never assert that *kazura* originated from *kazuragu*.
Other examples include *kata-gu* (shoulder-related actions), *hara-mu* (swelling), *chika-fu* (as in pledging vows), *ura-fu* (nearly identical in meaning to divining), *aga-fu* (nearly identical in meaning to redeeming), and *agito-fu* (referring to fish biting).
the element that activated *agito*.
Even when examining terms such as the children’s *agito-fu* (to bite) being *aki-tofu*, one cannot possibly conclude that *kata-gu* gave rise to *kata* (shoulder), *hara-mu* to *hara* (belly), *chika-fu* to *chika* (pledge), *ura-fu* or *ura-nafu* to divination, *aga-fu* or *aga-nafu* to redemption, or *akito-fu* to jaw.

I shall add a few more remarks here: to those scholars who insist that the verbal-root theory is correct and the nominal-root theory erroneous, I would urge them to provide an explanation for the following phenomena.

(1) It was indeed a matter of uninflected words (体言), but what was the reason that invoking a certain suffix transformed them into inflected words (用言)? That is—

aki-nafu (akijikori, akiudo)

*oto-nafu* *maka-nafu* (*makadachi*)

*mahi-nafu* (If I were young, I would not know the way; have the lower attendant accompany me and send me forth. —Yakamochi)

ni-nafu ama-nafu maji-nafu (majimono, majikoru)

and others *nafu*

*tata-yofu* (tatamu, tatahashi)

*shirazu-yofu* *moko-yofu* (*mukume ku*, *mukukumukushi*)

and others *yofu*

*saki-hafu* *wasa-hafu*

*iha-fu* (*iha* meaning “taboo,” i.e., related to *yuniwa* and the *yu* in *yuyushiyu*)

*tane-hahi* (chigusa, kusakusa)

*aji-hafu*

and others *-hafu*

*chiri-bofu* *yoro-bofu* *ki-hofu*

and others *-hofu*

*tada·shi* (正): Considering that the noun *tadashi* (“correctness”) has the verb *tazunu* (“to inquire”), there must exist a root *tazu*—likely its nominal form.
(to which the adverbial form 〈-shiku〉 was attached)

*hisa·shi* (“in ‘mizu hisa ni,’” “hisa ni furu”) *kore·shiki* (“kore shiki mono”)

mono-sh(i) monomono-sh(i) ōyake~~sh(i)

woman-like · adult-like || we~~-like (our station)

koma-inu-shiku (to be like koma-inu).
The Pillow Book contains two instances of this.
However, although Dr. Sekine stated that *koma-inu* are lions, when considering the following instances of “dimly” (*kumamajikushiku*) and others, it would be more accurate to say “*koma-inu*-like” (*koma-inu-shiku*).

kuma~~shiku (lacking intensity).
(In the evening faces,) one dimly perceives them here and there, while things clatter noisily underfoot, trampling about.

and others -shi

*naga·rafu* (not the *nagara-fu* referred to as a prolonged form of *nagaru*)

and others *-rafu*

In addition

*-meku* (*toki·meku*, *ugo·meku*)

*tsuku* (*gasa·tsuku*, *uro·tsuku*, *soha·tsuku*)

*-garu* (*maro·garu*, *kura·garu*, *hiro·garu*)

*-garu* (*dislike·garu*, *grieve·garu*)

*-kafu* *-ku (gu)* *-su (zu)* *-tsu (zu)*

*-numu* *-fu (bu)* *-yu* *-ru* *-u (gain)*

How does one explain that these suffixes attach to form verbs?

(2) *kare-su* *tsuki-su* *shini-su*

What are *kare*, *tsuki*, and *shi* if not nominalizations of verbs' conjunctive forms?

(3) How does one explain that *ryōri ga reuru* and *shōzoku ga shōzoku* become verbs, while *ohamoji* ("embarrassing things") and *himojiga* turn into adjectives like *ohamojii* and *himojii*?

(4) As for *wakayaka*, *yawaraka*, *sumiyaka*, and others—

Setting aside whether *wakayagu*, *yawaragu*, *sumiyaku*, and similar forms arose earlier or later, I wish to hear a detailed explanation of what purpose the *-ra* (or *-ka*) in *yawarakara* and the *-ya* in *wakayaka* serve.

Note

In seeking explanations for *ya*, *ra*, and *ka*, I feel it necessary to first provide an explanation of these elements from my own theoretical standpoint.

I consider elements such as the *-ra* in *yawarakara*, the *-ya* in *wakayaka*, the *-ka* in *hokorika*, and the *-ha* in *ateha ka* to be suffixes that form nouns.

Elements such as *ya*, *waka*, *hokori*, and *ate* already possess a nominal quality; however, since they could not fully function as uninflected nouns (*taigen*), suffixes like *-ra*, *-ya*, *-ka*, and *-ha* were attached to render their nominal quality more complete—thereby allowing them to be employed as nouns, adjectives, or adverbs.

Furthermore—and I elaborate here—these elements (*ya*, *ra*, *ka*, and *ha*) were not mere external additions but rather prospective forms of words that originally possessed conjugations.

I shall have more to say on this matter later.

In Princess Sususeri’s poem from the first volume of the *Kojiki*—〽*ayakaki no fuhayagashitani, mushifusuma ni koyagashitani, takubusuma sayagugashitani*—the elements *fuhaya* and *nikoya* correspond to what we would now call *fuhayaka* (“fluttering”), *fuwafuwashita* (“softly billowing”), or *nikoyaka* (“gentle”). However, since merely placing *fuha* below and *niko* below—as if reading the verb *sayagu* (“to rustle”) directly from its attributive noun form with particles like *ga*—would have rendered the expression incomplete, the poet likely employed *ya*. I surmise this *ya* originated from the verbal suffix *-yu* taking the form of *ya* in the prospective form.

The same principle applies to *taoyame* (“graceful maidens”) and similar terms.

While some explain *taoyame* as a phonetic shift from *tayowame*, based on its orthographic rendering as 手弱女 in passages like the Ama no Manai Ukehi section of the *Kojiki*, it would be more accurate to view this term as deriving from roots like *tawoto* or *katato* with suffixes such as *ka* and *ya* attached—a form that developed adjectivally to describe women.

Given that the sounds of *ya* and *ra* are phonetically proximate, one might consider whether this represents a sound shift.

There also exist terms such as *tayora* (*tayoya*, *tayoyaka*) and *sawara matsukaze*.

The *ha* in the term *atehaka* likely constitutes nothing more than the prospective form of *atebu*, does it not?

Though today we distinguish between *atebu* and *atehaka* through voicing contrasts, we cannot thereby disregard their etymological kinship.

The nominal form of *sayagu* was *sayaka* (needless to say, the phrase "takubusuma sayagu ga shita ni" meant "beneath rustling taku-fiber bedding"). Miyabika, nayobika, hokorika, nihohika, and others were thought to be either instances where *ya* had dropped away from *kawayaga* or where they had attached through the conjunctive noun form. That this *ya* did not simply adhere as *ya* but likely transferred from some verb was partially corroborated by how the nominal form of *sayu* became *saya*, and that of *atebu* became *ateha*.

(5) We demand proof that verbs such as *kakofu* ("to enclose") and *kashizuru* ("to attend") existed prior to nouns like *kaku* ("fence") and *shizu* ("stillness"), or that these nouns were merely coincidentally included among several unrelated words.

The foregoing has been my attempt to outline—however tentatively—the nominal-root theory.
I now wish to examine noun forms corresponding to the five conjugation stages of inflected words.
First, regarding the prospective form:

■Prospective Noun Form

The formation of nouns beginning at this stage had recently been addressed by Dr. Kameda in a university lecture.
According to Dr. Kameda’s theory, oya represented the prospective noun form of oyu ("to age"), while the prospective forms of nau ("to twist") and naru ("to sound") became nara ("fart").
By borrowing and elaborating on this theory, we see that the nominal form of *tsuru* became *tsura* ("row") and *tsure* ("series"), the nominal form of *sayu* became *saya* ("—ni"—"—ka"), and from the prospective form of *chiru* emerged uninflected words like *chira~chira* ("scattering") and *chira-tsuku* ("to scatter"). There could be no doubt that *yura* in phrases such as "swaying foot jewels and hand jewels" represented the nominal form of *yuru*.
The roots of all adverbs were uninflected words.
When the prospective forms of inflected words became uninflected words, they could not fully transform into nouns; it was not difficult to imagine that many remained as adverbial roots with nominal qualities.
Regarding adjective roots as well, we observed the same phenomenon.
Waka ("young"), taka ("high"), yū ("refined"—as in yū-otoko ["refined man"] and yū-kata ["graceful form"]), asa ("shallow"), and fuka ("deep") also originated from the prospective forms of verbs with the adjectival suffixes -shi (-shi, -shiku) attached.

The term *waka* ("young") would have been the prospective noun form of the ancient verb *waku* (though no extant literature currently attests to this), with examples including *wakiiratsuko* (likely not a phonetic shift from *wakairatsuko*) and *wakugo*.
There also existed *iwakinashi* and *iwarenashi*—while *itokinashi* and *itokenashi* might be interpreted as *i-waki-nashi* ("without discernment") and *i-wake-nashi* ("without reason"), which could not be entirely dismissed, considering that the *ito* in *itoke* and *itoki* carried connotations of youthfulness (as seen in "young princesses" from *The Diary of Lady Murasaki* and the Kyoto-Osaka dialect term *ito*), it seemed more plausible to view these as formations where suffixes like *-ke* or *-ki* attached to this root, thus retaining a sense of immaturity.
I maintained that it was more accurate to consider *i-wa-ku* had not acquired the meaning of "excessive" through attachment of [-shi] to this verb extant today.
That *i* functioned as a verbal prefix was evident from examples like *i-yu* ("go"), *i-sa* ("depart"), *i-ha* ("crawl") (*ihahi motohori utite siyamamu*—*Kojiki*), and *i-no* ("proclaim"). Thus, there could be no doubt that the verb *waku* had indeed existed.
People found it perplexing that *oyu* was a verb while *wakashi* was an adjective.
Verb-adjective monists adopted this contrast as one robust foundation for their theories.
However, contrasting with *oyu*, there existed the verb *wakayu*.
In opposition to *wakashi*, there was the term *oshi*, meaning "large."
The ancients—lacking systematic logical concepts—had contrasted *ōkii* ("large") with *wakai* ("small").
Simultaneously, this term *oshi* ("large") was also contrasted with *woshi* (an archaic spelling of *oshi*).
(That people expressed size through sound symbolism went without saying.) Some might argue that no adjective like *oshi* existed.
Yet examining the *Kojiki*, one found numerous instances of terms like Oshikorowake, Oshikuma-ō, Oshihoui, and Oshisaka.
Regarding this *oshi*, the *Kojiki-den* interpreted these occurrences as meaning "large."
Tachibana Akemi criticized this interpretation, contending that if *oshi* truly meant "large," one ought to simply use the character 大 (*ō*) instead—
—this being impossible due to the *Kojiki*’s stylistic prohibition against varying characters for words of identical meaning.
Those arguing that names like Oshihito and Oshikatsu used 押 (*oshi*) to signify valor or martial virtue overreached. I found the *Kojiki-den*’s theory more convincing: it cited cases like *Kumano Oshisumi no Mikoto* appearing as *Kumano Ōsumi no Mikoto* in variant texts, and *Oshikauchi* being phonetically rendered as *Ōkawauchi*, thereby equating 忍 (*oshi*) with 大 (*ō*).
Oshisaka meant Ōsaka ("Great Slope"). Along the route from Shiki District in Yamato to Akino in Uda—where an extraordinarily steep incline still called Hansaka ("Half Slope") remained—this was the name ancients bestowed when hunting in Akino, later transferring to the village below.

Another instance: the *oshi* in Kadashi-no-Miya of Oshinumi was likewise likely to mean *ōshi* (that is, *ōkishi*), I thought. The insertion of the particle の between the adjectives *oshi* and *mito* was akin to expressions like “a beautiful person” or “a pitiful child,” or to place names such as Kaminami (literally “divine south”). The phrase "〽Yamato e ni migahoshimono wa Oshinomi no kono takaki naru Tunusano Minoya" found in Korean texts, I believed, provided an excellent explanation of its geographical features. Moreover, in Soga no Emishi’s poem—〽*Yamato no Oshino Hirose o wataran to ayoi tazukuri koshizukurau mo* (*Kōgyoku-ki*)—the term *Oshino Hirose* did not denote a place name called *Oshi* but rather signified “large Hirose.”

※（Hangul characters: “ロ／亅／一”）※（Hangul characters: “○＋｜”）＞※（Hangul characters: “「ロ／亅／一”＋｜”） was indeed a mountain term, but in our country it was frequently used to signify small hills, knolls, or elevated areas.

Upon the lofty Omure (*Saimei-ki*).

In the *Wamyōshō*, 培 (土偏+婁) was annotated as "a small hill amidst fields."

Mori (forest).

However, there were also instances where it carried the meaning of "mountain."
Kii's Muro District likely signified a "mountainous district," while Miyoshino's Omura (small village) belonged to the classification of peaks.

I believe that the fact that *wakayu* corresponds to *oyu* and *wakashi* should correspond to *oshi* can also be clearly demonstrated through this method.

The same applied to taka ("high"). That taku constituted the prospective nominal form of a verb was beyond doubt. Of course, there remained aspects where the semantic content of present-day taku and takashi did not fully align. However, as the conceptual frameworks contained within these two words had diverged over time, when considering their relationship through this lens, it proved by no means difficult to conclude that takashi had originated from taku.

The term yū, while serving as the root of a shiku-adjective, displayed unusual nominal qualities; this yū undoubtedly adopted a form containing the a-vowel from the lower bigrade verb yasu. If one accepted the theory that quadrigrade verbs formed the root of various verbs, there would be no objection to designating this as the prospective form. (Though I held differing views on this matter, I deferred discussion to avoid digression.) What intrigued me was how expressions like "graceful man" or "graceful form" still struggled to fully shed vestiges of yasu's original meaning.

Next, *asa* ("shallow") takes the form of the a-vowel from the verb *asu*, which should be considered its prospective form. When describing rivers as shallow or waters as shallow, this contains the notion of water receding (*aseru*). As for mountains being called shallow or hearts being deemed shallow, these are merely analogical extensions applied through inference from water's shallowness.

As for the term *fuka* ("deep"), whether its original sense derived from water being deep or night being deep could not be easily determined; in either case, it undoubtedly originated from the term *fuku*. Today, while people spoke of "the night growing deep," they did not speak of "the water growing deep." Some argued that *yo no fukai* ("the night grows deep") had been derived from the Chinese characters for *shinya* ("midnight"), but phrases like *"ubatama no yo no fukeyukeba"* ("as the jet-black night deepens") did not sound nearly so literal in their phrasing. When explaining *fukashi* ("deep") by taking this *yo fuku* ("night deepens") as a basis, one could provide an extremely concise explanation. However, matters could not be explained solely in those terms. I believed there might have been an ancient verb that expressed water’s depth as *fuku*, but at present, this could not be definitively concluded. (Furthermore, could there have been a relationship between the terms *fuku-mu* and this *fuku*?) In the former Sanra District of Kawachi Province, there existed a place written as Fukano but pronounced Fukuno [< fuku]. This area had been marshland in ancient times due to river water, so it went without saying that this place name reflected the depth of that water and mud. However, such matters were dismissed as mere sound changes, so there was no need for such elaborate analysis.)

The term “near” appears to derive from *tsu-ku*.
Chika (“near”), tsuku (“attach”), tsukizukushi (“thoroughly”), and tsugu (“connect”) all share meanings related to closeness or similarity.

Incidentally, I believe the phrasing from the *Gosenshū*—"Though not a path crossing borders, yet hindered by years, I await spring"—merits particular attention.

If we posit that the *bera* in *beranari* constitutes a phonetic shift from *mera*’s prospective form, this too could support the theory of uninflected words.
We must note that *-nari* attaches not only to verbs’ terminal and attributive forms but most frequently to nouns.
Though ordinary grammarians subsume adjectives’ prospective forms under conjunctive forms, terms like *yoke*, *ashike*, and *nake* appear in both realis and prospective usages.
However, given existing theories about fusion with *ari*, I shall set this matter aside temporarily.

From what I have discussed above, I believe it has become clear that inflected words possess nominal properties in their prospective forms. I think that if one were to bring forth various inflected words and examine their roots, this would become even clearer.

The term *ukaru*, much like *ukauka*, involves *uka* being treated as a noun with the passive suffix *-ru* attached. Even when interpreting this in the causative sense as *uka-su*, it still ultimately signifies causing [someone] to perform *uku*. When *naku* becomes *nakaru* and *nakasu*, it still signifies either that *naku* is performed or that *naku* is caused to be performed. Similarly, since *kudaru* ("to descend") and *kudasu* ("to cause to descend") have *kuda* as their root, I believe this too must likewise be an instance of the prospective noun form. That is to say, there ought to be a term called *kuzu*. However, this [term] remains so unfamiliar that it is unclear whether such a word ever existed or not. Yet if we transpose this into phonetically related terms within the ya-row [of kana], *kyu* (崩, collapse) clearly indicates something positioned above descending downward; thus it becomes evident that *kuda* shares phonetic kinship with *kyu*, regardless of whether *kuzu* actually existed. When considering the term *kutsu*, people tend to think only of its meaning "to rot." However, we can understand the existence of *kudashi* for rain through *kutashi* applied to deutzia flowers—that is to say, while *unohana kutashi* ("deutzia-withering") has traditionally been explained as deriving from *kutasu* ("to wither") deutzia flowers, reports indicate that in Shōnai regional dialect (*Shōnai Dialect Study*), *kutasu* means "to get wet by rain"; thus *unohana kudashi* would essentially mean "deutzia rain."

The forms *oharu* (負) and *ohasu* derived from *ou* functioned in both passive and causative roles through *-maru* and *-su* respectively. In cases such as *yuka-mu*, *yuka-ji*, *yuka-zu*, and *yuka-mashi*, this *yuka* might initially appear to lack any nominal meaning; however, upon closer examination, some nominal significance seems present there.

The auxiliary verbs *keri* and *ken*, sharing elements with *ri* and *mu*, express temporal distinctions—since *keri* inherently contains past meaning (with *ri* being a demonstrative marker), *keri* denotes definite past time, while *mu*, being conjectural, allows imagining a past moment from the present perspective. The manner in which *ri* and *mu* connect with *keri*, as seen in pairs like *ramu* and *rashi*, *meri* and *beshi*, or *nari* and *namu*, each carries its own significance through these combinations. Yet why did they become connected? If we posit this as conceptual activity operating between them, the same logic should apply to forms like *yuka-mu*, *yuka-ji*, *yuka-zu*, and *yuka-mashi*. In cases such as *yuka-nan*, it is hardly unreasonable to state that the suffix *-nan* attached to *yuka* (the act of going) conveys a meaning of desiring to go.

I believe that sawa-gu (clamor), naya-mu (worry), tata-mu (strike), a-gu (raise), kaka-gu (soar), saka-ru (prosper), uma-ru (be born), tsuga-fu (join), yuka-shi (desirable), and itoha-shi (precious) can likewise be said to follow this same principle in their roots.

■Conjunctive Noun Form

That the conjunctive form possesses a noun form goes without saying.
I wish to discuss here how words in the conjunctive noun form combine with other suffixes.

Forms such as shini-su ("to die") and yuki-su ("to go") demonstrate that shini, kare, and tsuki are in the conjunctive form.
I believe it goes without saying that these forms—shini-su (die-su), yuki-su (go-su), kare-su (wither-su), oi-su (age-su), and tsuki-su (attach-su)—receive *su* as uninflected words.

Forms such as *yogi-ru* ("pass by"), *wabi-shi* ("lonely"), *koi-shi* ("yearning"), colloquial *yure-ru* ("waver"), *uke-ru* ("float"), and *oki-ru* ("arise") were also thought to belong not to the prospective form but rather to the conjunctive form.

■Terminal Noun Form

The notion that the terminal form could function as a noun had not been recognized by most grammarians up to that time.
However, within historical kana usage—though one might argue that *sumafu* and *kagerofu* represent euphonic forms of *sumahi* and *kagerohi*—I had never encountered anyone who amended them to *sumau* and *kagerou*.
Honored Motoori maintained that in the application of phonetic kanji for Chinese readings, the *fu* in terms like *asachifu* and *kakashifu* constituted an abbreviation of *ōfu*.
Even were we to interpret this as an attributive form, it would remain incongruous.
I believe we must ultimately regard Honored Motoori’s position as having tacitly permitted the noun form of the terminal form.

The division between terminal and attributive forms corresponded to quadrigrade and bigrade conjugations; the distinction between terminal and attributive forms arose from differences in grammatical roles—verbs changed their forms accordingly, meaning this was merely a nominal distinction based on form. Taking what was simply a nominal distinction based on form and directly applying it to certain verbs that showed no actual formal distinction, then assigning separate labels like "terminal" or "attributive," constituted an odd practice. If one posited that all conjugations developed from the singular origin of quadrigrade conjugation, then during that era when conjugation remained solely within the simple quadrigrade system, there naturally had existed no distinction between terminal and attributive forms. Mr. Chamberlain stated that in old quadrigrade conjugation, terminal and attributive forms maintained distinct shapes; however, as the overall tendency of verbs was to conflate attributive and terminal forms, quadrigrade conjugation no longer distinguished between them. Bigrade verbs too had lost this distinction in colloquial language. Hence, while some claimed (as cited in *Japanese Grammar Theory*) that bigrade verbs preserved older forms than quadrigrade verbs solely in this aspect, many facts could be found within relatively ancient extant documents demonstrating how attributive forms shared identical shapes with terminal forms—proving that terminal and attributive forms originally bore no distinction. These facts might not have represented the oldest forms of Japanese verbs, but among traceable evidence available today, they had to be acknowledged as preserving the most ancient vestiges. (The *Kojiki*'s *isachiru* [泣伊佐知流]—where the conjunctive form *isachi* was clarified by the gloss "啼伊佐知伎也" ["to weep, *isachiki*"], and *isachiru* itself, given the preceding phrase "何由以" ["by what means"], was likely the attributive form—might it not preserve within the *Kojiki* a solitary remnant from an era predating even quadrigrade conjugation, currently regarded as the oldest attested form? Yet this could not be easily asserted with certainty.) Instances requiring interpretation as attributive forms—such as *ikutachi*, *iku yumiya*, *naguya*, *shikasugani* ("indeed"), *yukigatenukamo*, *koyonashi* (likely *koyunashi*), and *oyosuku* (where *oyu* showed traces of meaning not merely "old age" but possibly "growth")—all adopted the same form as terminal forms, did they not? Should someone argue that in these cases the attributive ending *ru* had simply been omitted, let me note these examples amounted to but a single hair from nine oxen—such instances abounded in older texts. Were one to assert complete omission of [attributive endings], then—since grammar must be grounded in factual evidence rather than speculative constructs—even accepting that forms conceding one step might represent ancient attributive forms, the fact remained: one could not dismantle attributive forms identical to terminal forms in ancient texts lacking any accompanying *ru* [attributive marker]. Rather, the ancient form of the attributive—to the furthest traceable extent—had shared identical structure with the terminal form. To state plainly, I believed this could be called a terminal form (?) that encompassed both terminal and attributive functions.

The term *mitama no fuyu* carried the meaning of *fuyu* as "to increase," but even if one considered it as *mitama no fuyuru*, this did not hold together.
It was precisely in not expressing *fuyu* as *fuyuru* ("to increase") but rather stating it as *fuyu* that its force resided.

While I cannot definitively determine whether *shizuku* ("droplet") derives from the terminal or attributive form of *shizuku*, would it not be appropriate to consider that it nevertheless originates from a terminal form encompassing both terminal and attributive functions?

The presence of phrases such as "the path of *yamau*" and "for the sake of *yamau*" in old Jōruri works like *The Tale of the Four Heavenly Kings' Renown* and others—might they not have arisen from such origins? This stems from how attributive nouns containing /i/ vowels in the Keihan dialect avoid shifting to /u/ vowels (though ordinary conjunctive usage most frequently adopts /u/ vowel shifts). For instance, what Tokyo calls *Omukō* becomes *Mukaisan* in Osaka. Since this tendency appears in works chronologically close to Ko-jōruri as well, this must represent *yamau* rather than *yamaū*.

However, since conjugation forms that took separate shapes through having distinct attributive and terminal methods are themselves ancient, I believe we must still disentangle them separately when advancing this inference.

Even today, when bigrade, monograde, and s-irregular verbs lacking /a/ vowels combine with other suffixes to form inflected words and take forms containing /a/ vowels, one might dismiss this as merely resulting from sound changes or analogy to quadrigrade and na-/ra-irregular conjugations—yet this remains a matter demanding attention when discussing ancient verb conjugation forms.

Forms such as *yoshimu* ("to desire") and *kanashimu* ("to grieve"), which derive from adjectives, are cases where *yoshi* ("desire") and *kanashi* ("grief") have become uninflected words; they carry the same meaning as *yo-mu* ("to desire"), *yomi-su* ("to recite"), *kanashi-garu* ("to manifest grief"), and *kanashiku-su* ("to cause sorrow"), thus constituting instances of the terminal form's noun usage.

When we shift our perspective slightly to examine verbs, we find that the term *tsuru* (釣る), formed by attaching the suffix *su* (す) to the terminal form, becomes *tsuru-su* (釣るす).
The term *furu* in upper bigrade conjugation combines with *su* to become *furu-su*.
The verb *yuru-su* derives from the lower bigrade root *yuru*.

Does the verb nayu in lower bigrade conjugation—forming derivatives like nayu-mu, nayo-ru (the theory that this meant “to approach familiarly” was untenable), or nayo-meku—not originate from the terminal form?
Agumu was aku-mu; osuhi and osohi likely derived from osu with su added (as seen in ※（○／六）)—as in “The child of Omi folded the hemp hakama sevenfold and laid it out in the courtyard” (Nihon Shoki)—and then further added fu to form terms like osofu or osufu, which were presumably conjunctive noun forms.
Umore was umu (the terminal form) with ru added; the term obō in the Osaka dialect was considered to be umu with fu added.
If kuzu—the root in takuzuru and kuzusu previously discussed—was indeed a verb sharing the same meaning as kuyu, then this further corroborated the existence of terminal form nouns.
Similarly, the root of sakumu would be the terminal form noun saku.

It was evident that *sugusu* (to rinse), *okosu* (to raise), *okoru* (to occur), *haruku* (to stretch), *komoru* (to seclude), and *oyobosu* (to affect) similarly consisted of terminal-form nouns—*sugu*, *oku*, *haru*, *komu*, and *oyobu*—with various suffixes attached.

The terminal form of *itsuku* (meaning “to revere”) took *shi* to become *itsukushi*, which subsequently transformed into *utsukushi*.
The terminal form of *osoru* received *shi* to become *osoroshi* (fearsome).
The term *samoshi*—though *samu* was not seen in usage—when observing the verb *samisu* formed by attaching *su* to *sami* (a conjunctive-form noun), undoubtedly originated from *samu*’s terminal form taking *shi*. The expression *nitsuko-rashii* had existed in old Osaka dialect.
This did not appear to carry the conjectural meaning of terms like “foolish” or “disgusting,” but rather had attached to what was called the terminal form *nitsuku*.
Could it not be that *tsugu-nau* was formed by attaching *nau* to *tsugu*, a terminal-form noun?
*Hikotsurafu* would probably have been formed by attaching *tsurafu* to the terminal form of *hiku*.
From such semantic associations, one might feel that suffixes—though most frequently employed with terminal forms—tended to manifest in forms like using *suguru* where *sugu* sufficed, *tokuru* for *toku*, *suru* for *su*, or *kuru* for *ku*.
Moreover, I believe it likely that attributive endings unrelated to waka came to be widely employed.

■Attributive Noun Form

The noun forms of the attributive form within our previously discussed framework should naturally exist without surprise; however, they frequently blur with terminal forms, and I could find none that decisively demonstrate transition from attributive-form nouns to inflected words.
(Those used as participles remain exceptions.)

At first glance,*medurashi* might appear derived from the attributive form *meduru*, but this proves incorrect.
The *ra* in *medurashi* belongs to the same category as earlier discussed terms like *yato ka ha* and *kato ka to*, where *medu* functioned as a base noun with *shi* appended.
In conjugations maintaining sharp terminal-attributive distinctions like bigrade verbs,I found no evidence of suffixes attaching to attributive forms to create inflected words.
For conjugations lacking terminal-attributive differentiation—quadrigrade being primary—seeking attributive nouns becomes futile.
While participles exist within attributive conjugations as hybrid entities straddling inflected/uninflected words,the participle method strictly employs *teni o ha* particles rather than inflected-word suffixes.
Though classified under compound noun methodology here,the conjunctive form now predominantly fulfills this role.
True compound nouns involve noun-noun combinations rather than inflected-uninflected pairings.
Considered through modern lenses,*taru-ki*,*shiranu-hi*,and *kuruma-ki*(with its contested “wheel wood” interpretation) parallel formations like *sashimi*,*uki-fune*,and *yori-udo* rendered as *tari-ki*,*shirazu-hi*,and *kuri-maki*.

However, when we turn to adjectives, their character underwent a slight shift.
While terms like *yoki* ("good") and *ashiki* ("bad") became uninflected words, I found it unlikely that these *yoki* and *ashiki* could call upon other suffixes to further create inflected words.
However, Dr. Kanazawa argued that forms like *yokari*, *ashikari*, *yokere*, and *ashikere* should be interpreted as *yoki ari* ("good exists"), *ashiki ari* ("bad exists"), *yoki are* ("be good"), and *ashiki are* ("be bad").
This might have adopted Mr. Aston’s theory regarding roots, but my humble opinion differed somewhat in its approach.
The notion that roots—as posited by Mr. Aston—ended with vowels close to /i/ (as in *yuki*, *uke*, *iki*, *mi*) aligned with the nominal-root theory; however, this approach projected later conceptions backward through time. Upon examining extant documents, I tentatively arrived at the following conclusion through my still-developing analysis.

(Let me briefly clarify: while one might consider the term *ōkishi* to have formed through an adjective suffix attaching to the attributive noun form *ōki*, it actually belongs to the same lineage as *ōnike* seen in constructions like *ōnaraba* or *ōni to ka*—just as with *saya-ke-shi* or *shizu-ke-shi*.)

┌┌む（まし）

│┤ぶ

むつ┤└る

│┌Mutsu-ki

│┤

└└Mutsubi-goto Sumera-agamutsu Kamurogi Kamuromi

┌ me (Sagume)

さぐ┤┌る

│┤

└└（が）す

(ho)┌ hi (Aoi)

afu ┤┌ru

│┤ぐ

└└（ほ）つ

┌――（Fishing Sinker）

│┌Branch

││Wheel, Saddle

│┤

││gogoro

しづ┤└おり

│┌く

│┤む

│└る

│┌か

│┤

└└や

┌┌しね

│┤

│└ち

うる┤┌せし

│┤fu (hashi, hosu, hō)

└└む

┌┌（が）せを

│┤zuchi (Kagutsuchi-no-Kami) (*Kojiki*: Also called Hi-no-Kagabiko-no-Kami)

kagu┤└Yahime

│ ┌や（やか、やく、やかし）

│（が）┤

└ └yofu

┌ふ（はし）

たゝ┤

└yofu (hashi, hasu)

┌づ

│┌ぬ

│┤

い ┤└（ゆ）く（ありくはあい〈ゆ〉くなり、あるくはあゆく也、あゆむのゆむ如何）

└る

┌か（―し）

おろ┤┌おろ

│┤

｜｜└└おぼえ

｜｜┌つく

うろ┤たへる

│┌ga comes (Osaka dialect)

│┤

└└oboe

┌ふ

│(so)gu (kashi, kawashi, kasu)

（いすゝ）│（そ）し（しむ）

いす ┤（そ）はく

││ │sugu

｜｜ │（そ）ばふ（＜いそぶ？）

││ └rokofu (

┌┤moru┘

（ご） ││めく

ugu ┤└-nafu (haru; *Shūden*, *Ōharae norito*, et al.)

むく │┌と

(moko)││tsuku (keshi)

└┤meku

│mu

│mukushi

└（もこ）よふ

┌（ば）る

あぶ┤

└る

┌なふ

うづ┤

└なし

┌た――＜とをゝ

wawa┤┌ku

│┤

└└ra (raba)

Roots such as *mutsu*, *sagu*, *afu*, *shizu*, *uru*, *kagu*, *tata*, *isu* (including variants *isusu*, *susu*), *ugu* (including variants *muku*, *moko*), *abu*, *uzu*, and *wawa*—which hover between nouns and verbs, and among the eight parts of speech should first be regarded as uninflected words closer to interjections—become verbs, adjectives, adverbs, or even nouns either in their original form or through successive suffixation. From such observations (akin to how verb terminal forms ending in the *u* rhyme reveal many *u* sounds at the conclusion of shared roots),

┌ ┌su (whether a conjectural suffix or sound change?)

│（や）┤

│ └mu

なゆ┤┌む

│┤（yo）bu

└└(yo)ru: To explain *nayoru* as “to become accustomed” is inferior to interpreting it as *nayu-ru*.

（ぶ）┌す

のぼ ┤

└る

┌む

kaku┤su (＜safu)

│（こ）ふ

└る

（ぐ）┌む

nago ┤

└る

Terminal forms such as *nonayu*, *nobu*, *kaku*, and *nagu* appear to have become uninflected words that then received suffixes.
I have previously discussed that these should be treated as uninflected words, yet consider further examples: *katagu* (shouldering), *agitofu* (gathering), *akinafu* (trading), and *tokimeku* (throbbing; when used in phrases like *kokoro tokimeku*, it is now voiced as *dokitsuku*).
(The *toki* here differs from the former, representing instances such as *tokimeku* [throbbing]), and this should be evident from examples like *haramu* [to swell] and *kagu* [to emit fragrance].

I had intended to discuss the attributive form but unexpectedly strayed into a digression; to put it concisely, the notion that adding other suffixes to attributive forms creates inflected words is questionable. However, while my teacher does acknowledge this regarding the attributive form of adjectives, I find it more convincing to consider that forms like *yoku ari* and *ashiku ari* derive from the conjunctive form taking *ari*. It is true that we understand how words like *ari*, *su*, and *u* follow conjugation stages deeply tied to the i-vowel—yet extending this to posit forms like *yoki ari* and *ashiki ari* seems overly contrived. Though it is my teacher’s theory, I find it difficult to accept.

Beyond the compound suffixes—such as *ari*, *u*, *su*, and *mu*—whose etymological meanings remained somewhat discernible even today, there also existed monosyllabic ones: *ku*, *tsu*, *nu*, *fu*, *mu*, and *yu*. While one could dimly perceive traces of etymology for suffixes like *yu* and *mu*, for *ku*, *tsu*, *nu*, and *fu* there was currently no choice but to analyze their phonetic and semantic components. I could not help but feel it was rather difficult to assert that verbs used for compounding were limited solely to *ari*, *u*, *su*, and *mu*. How much less could one avoid concluding that each of these various suffixes—which I had initially aligned alongside *zutsu*—must also be considered to have originally been independent inflected words?

In conclusion, the roots of inflected words were not originally based on the i-vowel but rather on words ending with the u-vowel. Since the terminal form resembles this structure, these roots sometimes come to serve as such. These u-vowel roots, as I have previously stated, are neither verbs nor nouns but are words that can flexibly function as both; in formal terms, they should likely be classified as uninflected words.

Dr. Ōya Tōru had argued that the word *iku* (生) originated from *iki* (breath) being activated; however, I believe there existed a word *iku* that was neither noun nor verb, which then immediately underwent conjugation.
I tentatively named this the *Chaos Era*.

Iku — chaos — [Iku-tachi] (*Kojiki*) — [Noun... iku... iki]

Era └Iku-yumiya┘ ├ └─ikofu

Birthday no Ashihi (*Izumo Kuni no Miyatsuko Kamuyogoto*) │ Form: Ikashiho

Iku-i (common in norito) └Verb: iku (to live)

As shown in the table above, the word *iku* in the Chaos Era could be seen functioning adjectivally, verbally, and nominally.
The term *ikashi* in *ikashiho* is commonly interpreted as meaning “strict,” but this actually refers to a place brimming with vitality—it is likely that the sense of “strictness” emerged from such vigor-filled locations.
Regarding this matter, I believe that if one consults terms such as *ikimu* (to strive) and *ikihofu* (to be vigorous), the underlying connections may be somewhat discerned.

┌*toyosaka nobori* (Morning Sun’s—) Ōharae, Izumo Kuni no Miyatsuko Kamuyogoto

┌さか┤

│ └*saka*-tree

│ ┌*saki* (幸)┐

│ Laugh │ ――hafu │

saku — Chaos (栄井) ┴saku┴sakayu │

Era ├？

┌Bloom │

…………┤ │

└sakaru ┘

In the Saka era, these [terms] had not yet become verbs, adjectives, or nouns; however, by reading the vowel *i* as *saki* (“fortune”), they evolved into verbs like *emizaku* and acquired conjugations.
Regarding both the previously cited *mutsu* and *kaku*, one can imagine this Chaos Era.
When I consider *iku*, *saku*, and *mutsu* once more, it seems that the words of the Chaos Era were perhaps formed with consonantal rhymes.
That is, when *u* combines with *soharu*, it becomes a verb; when *i* combines with *tsuku*, it becomes a noun.
When that vowel is attached, it primarily becomes an adverb or adjective.

┌masu + ra + o Victory

*masu* Chaos┤Ama-masu-hito Increase

└maso + ke/shi (masa + ki/ku) Justice

The reason *Masurao* was written in *Kojiki* and *Man'yōshū* using phonetic loan characters—which through their glyphs were assumed to carry the meaning of “valor”—was that Oda’s explanation of *masurao* became untenable; yet if one interprets it as *masu* + *ra* + *o*, signifying “a capable man,” there arises no inconsistency.
Does not the term *maso + ke/shi* itself hint at this sense of capability? (This connection may also be observed in how the Hyōbu Ministry’s edict designated *chikarabito* [strong men] as *kenji* [elite guards].)
Even regarding cases like *Ama-masu-hito*, during the events of Yomotsu Hirasaka,

"In your land, the people-grass strangle a thousand in one day. Amananihimemiko: 'If you act thus, I shall erect one thousand five hundred birthing huts in one day.' Therefore, in one day, a thousand must die; in one day, one thousand five hundred must be born."

Fixating on this passage, some interpret the phrase 'those who increase divinely throughout the land' from the Ōharae ritual as indicating that *masu* denotes numerical superiority of births over deaths—an interpretation I found unconvincing. It would be more appropriate to understand this as expressing how Ōmitakara labored diligently each day through the gods' sacred covenant.

The above constituted merely one hypothesis. Regarding the sequence and manner in which these words emerged from their Chaos Era, I recognized there might well be incomplete or erroneous aspects in the table I had presented.

Another matter to consider regarding this attributive form concerns those cases where it is accompanied by what are termed various types of prolonged forms as suffixes.
While some may posit that these forms derive from the prospective form, this is more likely a case of sound change.
The evidence that the *ku*-prolonged form derives from the attributive form can be discerned from poems such as those in the *Man'yōshū*: 「わが背子を何地ゆかめとそきたけのそかひにねしく今しくやしも」 and in *Seigo*: 「桜花ちりかひくもれおいらくのこむといふなる道まがふがに」.
These examples—*neshi koto* and *oi* (*oyuru koto*, to grow old)—are entirely attributive in form; while one might not be entirely unable to view them (as *neshi*, *oyura*) as uninflected words, I find this interpretation rather strained.

To add to this point, it appears that what are termed *ku*-prolonged forms can also be employed as verbs when derived from bases containing the *u* rhyme. For example, *isohaku* is the so-called prolonged form of *isofu*. This is akin to how it shifted into the quadrigrade conjugation.

■ On the Realis Form

Regarding the realis form, I have yet to identify any noun-like elements.
As can be observed at a glance through adjectives' examples, both the realis form as a whole and the imperative form appear distinctly unique among all conjugational patterns of inflected words.

Prospective-Conjunctive-Terminal [Attributive] [Realis]

If quadrigrade unification were factual, the terminal and attributive forms would merge into one. When examining the upper/lower bigrade and monograde conjugations, one observes in both the prospective and conjunctive forms a relationship akin to that between the terminal and attributive forms in quadrigrade conjugation. Regarding old verb conjugation forms, I will await the day when my considerations coalesce; for now, I shall merely present my humble views on the presence or absence of uninflected words (nouns) in each conjugation stage of verbs and adjectives, and further explain how suffixes form inflected words from these nouns.

Now, I returned to the beginning of my discussion.

ito-ha-shi itoho-shi yorokoba-shi yorokobo-shi

*yura-gu* *yuru-gu* *oyowa-su* *oyoho-su*

When I considered these examples, they could not be thought of as merely sound changes.
To some extent, I had to consider sound change; however, it seemed more plausible—as I had previously argued—to posit that the prospective and terminal forms each called upon their respective suffixes to create new inflected words.

*koi-shi* *sabi-shi* *wabi-shi*

*yuki-su* *shini-su* *kare-su*

*yogi-ru* *yuri-ru* *yure-ru*

Forms such as these derived from conjunctive-form nouns (連用法体言), their nominal meaning appearing deeper compared to those emerging from prospective forms (将然言) or terminal forms (終止言) discussed earlier.
If it had been the case that neither prospective nor terminal forms became inflected words by summoning distinct suffixes—but rather one arose through sound change—I would assert against those prioritizing prospective forms that terminal forms constituted the fundamental basis.
Were one to take prospective forms as primary basis, how could *ku*-prolonged forms like *neshiku* or *oiryaku* be explained?
One must further acknowledge peculiar conjugations like *neshimu*, *neshi* (prospective form), *oyuramu*, and *oyuri*.
I had previously addressed kinship-like relationships between terminal and attributive forms.
Even considering this framework, would not terminal forms surpass prospective ones?
While stating terminal forms carried attributive meaning posed no issue here, formal conflation remained strictly impermissible.
Formally speaking—rather than attributive forms becoming nominalized to summon suffixes—it was terminal forms combining both attributive-terminal stages that summoned suffixes.
That is: conjugation unification required forms like modern quadrigrade conjugation—encompassing terminal-attributive within terminal forms.
Otherwise—attributive forms in current upper/lower bigrade/irregular conjugations receiving suffixes from terminal forms rather than becoming inflected words themselves—would demand recognition as peculiar indeed.

For these reasons—up to a certain point—the conjunctive form could also be considered encompassed within the prospective form.
In such circumstances,the problem would largely remain confined to matters concerning both prospective and terminal forms.

Let me reiterate: I acknowledge the phenomenon of sound change.
However, I cannot extend this consideration to extremes.
While I myself do not claim that every single one [of these forms] individually originated from the terminal and prospective forms, I cannot express agreement with the opinion that seeks to attribute them all to a single origin.

Thus,

Tsukurofu is formed by receiving -fu from the terminal form of tsukuru.

Katarafu is formed by summoning -fu from the prospective form of kataru.

Kakofu is formed by attaching -fu to the terminal form of kaku.

*tatakafu* is formed by the prospective form of *tataku* receiving *-fu*.

I intend to advance this argument.
(I find it dubious to interpret *katarafu* as *katariau* [to converse] or *tatakafu* as *tatakiau* [to fight mutually], if I may say so.)
That the *fanqie* phonetic method saw application across multiple domains remains evident; even within texts like the *Kiki-Manyō* corpus [*Kojiki*, *Nihon Shoki*, and *Man'yōshū*], we observe traces of this practice.
Yet there persists a tendency to employ ateji substitutions—such as writing 'uranau' [to divine] as 占合 or 占相, or rendering 'tawoyame' [graceful woman] as 手弱女—in these very contexts.
Does not the *-fu* in *urafu*, *katarafu*, and *tatakafu* fundamentally lack any inherent connection to *au* [to meet] from its inception?
Might those who insist on viewing all such *-fu* elements not as suffixes—but rather as roots fused with an *-a-* from *-afu*—have fallen prey to the ateji conventions found in works like the *Kiki*?
To claim that 'tawoyame' signifies 'delicate woman' surely stems from errors originating in texts like the *Kojiki*.
(Though the old annotated *Kojiki* employs lower bigrade conjugations for terms like 占合 rendered as *ura-ae*, I remain unconvinced by their treatment of its meaning—akin to *harai* [ritual purification]—as a term transitioning from transitive to intransitive usage.)

In his *Study on Prolonged Forms*, Dr. Kanazawa argued as follows: inferring from the fact that Korean verbs and adjectives possessed two types of nominalization (※ [Hangul characters resembling "fu + to"], ※ [Hangul characters resembling "ro"]) and that only Japanese adjectives retained the *ki* and *mi* nominalizations, verbs too must have historically had an *m*-form nominalization. Just as *hiroki* ("wide") became *hiroku* and *shiroki* ("white") became *shira-gu*, forms like *hiro-mu* ("to broaden") and *shiro-mu* ("to whiten") had shifted from the nominalizations *hiro-mi* and *shiro-mi* into verbs. It was through this transformation of the *m*-form that what we term "prolonged forms" arose.

However, upon consideration, prolonged forms were by no means limited solely to the *ha*-row or *a*-row suffixes. Given that verbs arose from a nominalization method bearing the *ki mi* rhyme’s *i*-sound, this conclusion appeared natural when viewed from Dr. Kanazawa’s position—one that determined verb roots based on their deep association with the *i* sound. Yet how was one to explain forms like *yoso-ru*, *furu-su*, *masa-ru*, and *ukomo-tsu*?

In cases like *yosuga katarafu* or *kamimaku toka ni fu*, *ku ga tsuku*, they received *-ru* in the same manner and became *yoso-ru*.

The formation of *masaru* by attaching -*ru* to the prospective form of *masu* showed no difference from cases like *mimaku* or *katarofu*. Similar phenomena manifested in instances such as *ugomotsu* and *ugomochi*. *Ugomu*, sharing etymological roots with *mukumu*, presented no contradiction when analyzed as deriving from these bases through attachment of -*ru* and -*tsu* suffixes.

Forms like *furu-su* emerged through suffixation of -*su* to *furu*, while *taru-mu* developed via -*mu* addition to *taru*, and *yuru-bu* through -*bu* attachment to *yuru*. This analytical approach revealed that so-called 'prolonged forms' extended beyond mere -*ku* and -*fu* suffixes.

What should be further noted is that there are instances where this process operates doubly.
That is,

*Yoso-ho-fu* is

*yoso-ho-fu* < *yoso-fu* < *yosu*

*Hiko-zuro-fu* is

*hiko-zuro-fu* < *hiko-zuru* < *hiku*

These belong to this category.

Furthermore, upon consideration—whether it be *yuka-ru* or *yuka-su*, *uka-ru* or *uka-su*—these too should indeed be classifiable as what are termed prolonged forms.

The inadequacy of the designation "prolonged form" lies in how they are formed by treating a conjugated stage of inflected words as uninflected nouns and attaching suffixes to create inflected words; thus it becomes evident they are not formed through mere elongation of endings.

In conclusion—though appended last—if we examine the meanings of those suffixes hitherto called prolonged forms (*-fu*, *-ku* [not the *ku* of *miryaku* or *kōfuraku*], *-su*, *-tsu*, *-nu*, *-ha*, *-yu*, *-ru*, *-u* [including *-gu*, *-zu*, *-zu*, *-bu*], along with disyllabic or polysyllabic suffixes), I believe we would obtain intriguing results.
Of course—while *-ru* may suggest existence, *-su* doing, *-u* obtaining, and *-mu* seeing—explanations of *-ru*, *-su*, and *-mu* remain incomplete when based solely on these [meanings].
I believe there remains a need to consider other suffixes too—beginning with *-ku*, *-tsu*, *-nu*, *-fu*, and *-yu*—alongside polysyllabic ones.

-ku

tsuna-gu (to link [綱ぐ] or to align [列ぐ])

*kaga-ya-ku*, *odoro-ku*, *ugo-ku* (< *muku-ku*), *una-ku*, *saya-gu*, *soyo-gu*, *soso-ku*, *sesera-gu*, *yoro-keru* (> *ku*), *era-gu* (*erara*).

Is *kohorogi* [cricket/katydid] the nominalization of *kohoro* (an onomatopoeic *gu*)?

*hara-ra-ku*, *todo-ro-ku*

-su

(go)

｜｜

sugu-su, tada-su, haya-su, kaku-su

nobo-su (延ぶ, ※ [Hangul: 「└／┴／─／｜｜／─」] 〈high〉)

Other auxiliary verbs -su

-tsu

taki-tsu (the sound of falling—, the boiling of water)

momi-zu (Is "momi" related to color?)

i-zu (iru [to be], inu [to go (archaic)], iku [to go])

-nu

fusa-sa-nu, kata-ta-nu

tsura-nu, tsuka-nu

The formation of *na-fu* occurred through this twofold development of *-nu*, its path having necessarily passed through *-nu*.

*kaka* (屈, bend)・*nau*, *tata*・*naha*・*ru*

-mu

shizu-mu, naya-mu, kuda-mu (Osaka dialect)

kaku-mu, mutsu-mu, shiwa-mu, soba-mu, uru-mu, se-mu (semu, "to narrow")

aga (上) -mu

(*yomi-su* and *sami-su* also follow the same nominalization method)

-yu

kogo-yu, muku (向)-yu, obo-yu, mi-yu, taka-yu, ama-yu, ni-yu

-fu

chika-fu, nega-fu, kako-fu, tsuta-fu, and others

-ru

araha-ru, soso-ru, maku-ru, aka-ru, komo (< komu)-ru, kaku-ru, yoso (< su)-ru, mutsu-ru, masa-ru

*Uha* is 'to be' (居).

In the *Shinsen Shōjiroku* and Ban Nobutomo's commentary on the *Takahashi Ujibumi*, there was mention of Wakayuramuraji.
*u...su-u*, *uma-u*—all contained the meaning of determining position.
In the entry regarding the passing of Yamatototohimomosohime no Mikoto in the *Chronicles of Emperor Suinin*, it was recorded: Here, Yamatototohime no Mikoto looked up in regret, hastily sat down (急居, here read as *ugiu*), and then perished when a chopstick struck her private parts.

In concluding this line of reasoning, should there have been any instances in the course of this treatise where I inadvertently showed discourtesy toward our predecessor scholars, I earnestly beg their gracious understanding.