Differential manifestations of subjectification in the development of Japanese and English modal auxiliary systems

Until the Middle Ages, many of the Japanese modality expressions could express both deontic and epistemic modal meanings. For example, -mu had both deontic (intention) and epistemic (probability) meanings, and -beshi had both deontic (obligation) and epistemic (certainty) meanings (Narrog 2002, 2007, 2012). Contrary to the arguably universalist claim of unidirectional development of modal expressions from deontic as in (1a) to epistemic as in (1b) (Traugott 1989, Sweetser 1990, Bybee et al. 1994), Japanese shows no conclusive sign of such directionality (Kurotaki 2005, Narrog 2012). 

(1) a. John must be home by ten; Mother won’t let him stay out any later. 

      b. John must be home already; I see his coat.                                                       Sweetser (1990: 49)                                                                                                      

English modal auxiliaries are known to have evolved from matrix verbs (Traugott 1989). For example, the modal auxiliary must originated from the Old English verb motan, which means “be obliged to”. On the other hand, Old Japanese modal auxiliaries such as -beshi (obligation; certainty) and -mu (intention; probability) had already been present by the time the oldest extant documents were written (Kitahara 2000).  

Furthermore, the form-meaning correspondence of modality expressions underwent a radical change in the early Modern period (around the 17^th century) when the synthetic modality expressions of Old Japanese were replaced by more analytical/periphrastic modality expressions such as nakere-ba-nara-nai (obligation) and kamo-shire-nai (weak possibility) (Kitahara 2000). This process is considered an example of insubordination, the conventionalized main clause use of what appear to be formally subordinate clauses (Moriya and Horie 2024, Evans 2007).  Deontic modality expressions and epistemic modality expressions thus became differentiated and a lot of newer periphrastic modal auxiliaries are available in Modern Japanese.

In this presentation, we will argue that subjectification, which is often considered a concomitant factor of grammaticalization (Hopper and Traugott 2003), is manifest both in Japanese and English modal auxiliaries but in a different manner. We adopt the view that subjectification is “a shift to a relatively abstract and subjective construal of the world in terms of language” (Hopper and Traugott 2003:86). 

In the case of English, matrix verbs changed into auxiliary verbs through grammaticalization, and the process of subjectification affects the extension from deontic to epistemic meanings (Sweetser 1990). In the case of Japanese, where the older  modal auxiliary and the newer periphrastic modal auxiliary coexist, the older one tends to express the speaker’s subjective judgment while the periphrastic ones tend to express more objective judgment. For example, in (2a), -bekida is a modal auxiliary derived from Old Japanese -beshi and expresses the speaker’s subjective, often ungrounded, judgment (Narrog 2002). In contrast, nakere-ba-nara-nai in (2b) is a periphrastic modal auxiliary form consisting of a negation (na), a conditional (kere-ba) ‘become’ (nara),  and the negative auxiliary (nai). This periphrastic auxiliary appeared around 17^th century and it expresses typically necessity according to social rules or customs (Okuda 1999). Since the traffic rules are a kind of social rules, bekida in (2a) sounds odd while nakere-ba-nara-nai (2b) sounds more natural. 

(2) a. ??Nihon-de-wa  kuruma-wa  hidari-gawa-o   tooru-bekida

         Japan-LOC-TOP  car-TOP      left-side-ACC    go-must

     b. Nihon-de-wa    kuruma-wa    hidari-gawa-o   toora-nakereba-naranai

     Japan-LOC-TOP  car-TOP          left-side-ACC   go-have to

              ‘In Japan, cars have to use the left lane.’ 

Similar contrasts can be observed in other pairs of old and new Japanese modal auxiliaries such as –(r)eru and -koto ga dekiru, which both express ability (Kuno 1983). The former is derived from Old Japanese and expresses ability  inherent in the agent while the latter periphrastic auxiliary was born in the middle age and signifies ability or possibility caused by social rule or customs.

We propose these contrasts can be explained if we assume auxiliaries derived from Old Japanese have undergone subjectification as predicted by grammaticalization theory, while newer periphrastic modal auxiliaries have not undergone that process yet to express all the subjective meanings covered by the older ones.