Passive head only selects for agentive Voice in Japanese: a reply to Jo and Seo (2023)

Journal of East Asian Linguistics https://doi.org/10.1007/s10831-025-09294-4 Passive head only selects for agentive Voice in Japanese: a reply to Jo and Seo (2023) Daiki Asami1 Received: 25 April 2024 / Accepted: 6 January 2025 © The Author(s) 2025 Abstract Developing a comprehensive theory of passives has been a central goal in Japanese theoretical linguistics. A recent study by Jo and Seo (J East Asian Linguist 32:91– 132, 2023. https://doi.org/10.1007/s10831-023-09253-x) represents one such attempt. They propose a uniform syntactic analysis of direct and indirect passives in Japanese. Building on the syntactic head Pass(ive) introduced by Bruening (Syntax 16(1):1– 41, 2013. https://doi.org/10.1111/j.1467-9612.2012.00171.x), they argue that, unlike in English, where Pass selects exclusively for an agentive Voice, the Japanese Pass head can take as its sister a projection of either an agentive or an expletive Voice. While (Jo and Seo in J East Asian Linguist 32:91–132, 2023. https://doi.org/10.1007/ s10831-023-09253-x) make a valuable contribution to the theory of Japanese passives, I argue that their analysis encounters empirical challenges and adopts unmotivated theoretical assumptions. Specifically, I demonstrate that the Pass head in Japanese, like its counterpart in English, selects only for agentive Voice. Furthermore, I contend that no syntactic or semantic differences in the Pass head need to be posited between Japanese and English. If my arguments hold, this paper supports the view that the defining characteristic of passives is the demotion or removal of the external argument, and this is made possible by the Pass head as one of the functional heads provided by Universal Grammar (UG). Keywords Passive · Agentive voice · Expletive voice · Syntax · Japanese 1 Introduction Since Kuroda’s (1965) seminal work, passives in Japanese have attracted much attention from formal linguists (see Hoshi (1999) and Ishizuka (2017) for overviews). Japanese passives are marked by the morpheme -(r)are. Prior studies typically clas- B 1 Daiki Asami Department of Linguistics and Cognitive Science, University of Delaware, Newark, USA 123 D. Asami sify them into two types: direct and indirect passives, as exemplified in (1) and (2), respectively.1 (1) Hanako-ga sensei-ni home-rare-ta. Hanako-nom teacher-by praise-pass-pst ‘Hanako was praised by the teacher.’ (2) Hanako-ga sensei-ni musume-o home-rare-ta. Hanako-nom teacher-by daughter-acc praise-pass-pst ‘Hanako was affected by the daughter’s being praised by the teacher.’ One key characteristic distinguishing the two types of passives is the interpretation of the subject. In the direct passive, the subject is directly involved in the event denoted by the verb. For example, in (1), Hanako is the patient of the praising event. In contrast, in the indirect passive, the subject is indirectly affected by the event described in the rest of the sentence. For instance, in (2), Hanako is an experiencer of the teacherpraising-daughter event but is not necessarily a participant in it. The study by Jo and Seo (2023) (hereafter JS) represents one of the most recent attempts to propose a uniform analysis of the two types of passives in Japanese. Their analysis builds on the notion of a functional head Pass(ive) proposed by Bruening (2013) (see also Bruening 2014, 2019). According to Bruening (2013), passives in English are derived by the Pass head, which selects a projection of an agentive Voice, a functional head responsible for introducing an external argument (Kratzer 1996). JS argue that, in contrast to the English Pass head, its Japanese counterpart-realized as (r)are-can take as its sister a projection of either the agentive Voice or a semantically contentless expletive Voice (Schäfer 2008; Alexiadou et al. 2015). The expletive Voice occurs in constructions that lack an external argument, such as unaccusatives. Abstracting away from details, JS propose that both direct and indirect passives in Japanese share the following partial structure, where “ag” and “expl” stand for “agentive” and “expletive,” respectively. (3) ... VoiceP Pass . . . Voiceag/expl JS’s syntactic analysis successfully accounts for a wide range of facts about Japanese passives. However, I argue that their approach makes incorrect empirical predictions and adopts theoretically unmotivated assumptions. Specifically, I propose that the Pass head in Japanese, like its counterpart in English, selects exclusively for agentive Voice. Additionally, I highlight that some of the theoretical assumptions made by JS are unnecessary. In my revised version of JS’s analysis, the partial structure for both direct and indirect passives can be represented as (4a) or (4b), depending on the presence of 1 The acceptability judgments for the empirical data in this paper are based on the author’s (a native speaker of Japanese) intuitions and those of three to six additional native speakers of Japanese, depending on the example. While most data receive consistent judgments, any disagreements are noted along with the number of speakers who accept or reject the relevant examples. 123 Passive head only selects for agentive Voice… a -ni ‘by’ phrase. These structures are essentially head-final counterparts of the partial structures proposed for English passives by Bruening (2013, 2014, 2019). (4) a. b. PassP VoiceP PP Voice Pass (r)are PassP Pass XP NP P . . . Voiceag aaaa -ni by VoiceP Pass (r)are . . . Voiceag If my argument is correct, this paper supports the view that the defining property of passives is the demotion or removal of an external argument (e.g., Comrie 1977; Keenan 1980, 1985; Shibatani 1985; Bruening 2013; Bruening and Tran 2015), and this is a property made possible by Universal Grammar (UG) through the availability of the Pass head (Bruening 2024: 5). Section 2 provides an overview of the relevant aspects of JS’s proposal. In Section 3, I present counter-arguments to their claims about the syntactic selectional properties of the Pass head in Japanese. Section 4 outlines a revised version of JS’s analysis and explores its implications. Section 5 examines the argumenthood of -ni ‘by’ phrases in passives. Finally, Section 6 concludes the paper. 2 Jo and Seo (2023) This section begins by introducing Bruening’s (2013) syntactic analysis of English passives and then goes over JS’s extension of this analysis to Japanese passives. 2.1 Bruening’s (2013) analysis of English passives Bruening (2013) focuses on passives in English like (5). (5) The dog was chased (by the cat). Building on Keenan (1980, 1985), he argues that the defining property of the English passive is the demotion or removal of an external argument. To formalize this, he proposes a functional head Pass(ive), defined in (6). Semantically, this head either (i) takes an open predicate of type <e,st> and existentially closes an unsaturated argument of that predicate or (ii) takes a predic (...truncated)