The universality of categories and meaning: a Coserian perspective

Abstract

Studies in linguistic typology have challenged the idea that languages can be analyzed in terms of a set of preestablished universal categories. Each language should instead be described “in its own terms,” a view consistent with the ‘old’ structuralist paradigm in linguistics. The renewed orientation toward differences between languages raises two questions: (i) How do we identify the meanings which are assumed to be crosslinguistically comparable? (ii) What is the relationship between language-particular categories and comparative concepts commonly used in linguistic typology? To answer these questions, this article focuses on a number of distinctions advocated by Eugenio Coseriu (1921–2002). Coseriu distinguishes three levels of meaning (designation, “signifiés,” and sense) and three types of universals (essential, empirical, and possible universals). Their relevance for linguistic typology is discussed with regard to the expression of possession and a particular diathesis in Japanese, viz. ukemi or “indirect passive.” As well as relating language-particular categories and comparative concepts, Coseriu’s approach offers a promising avenue to account for the ways language-specific meanings interact with extralinguistic knowledge and contents of discourse and texts, which are the object of translation.

Key words:

• Linguistic typology
• categorial particularism
• possession
• Japanese passive
• Eugenio Coseriu

Acknowledgment

I am grateful to Leonid Kulikov (Ghent University), Hartmut Haberland (Roskilde University), Jeroen Van Pottelberge, and two anonymous reviewers for their constructive comments. Special thanks go to Ritsuko Kikusawa (National Museum of Ethnology, Osaka) for help with the Japanese examples discussed in Section 3. Of course, any remaining errors are mine alone.

Notes

¹ “Idiomatic” should not be confused with “idiom” in its more standard usage of a formally and semantically fixed expression. In Coseriu’s model, the “idiomatic level” encompasses everything related to the structural, grammatical, and semantic character of a specific language, which also includes knowledge of diachronic phenomena, dialect variation, and register variation (Coseriu 2007, 133).

² The constraints Seiler assumes for Vater appear to be too strong. Although a sentence such as X hat einen Sohn is more likely to occur than X hat einen Vater for reasons related to elocutional knowledge, the latter does not violate any conditions imposed by the verb’s valency; compare examples such as “Elisa sagt, jeder hat einen Vater” and Damit ist der erste Wunsch von Tochter Susan schon erfüllt: Sie hat einen Vater (examples extracted from the German Referenzkorpus, Mannheim; http://www1.ids-mannheim.de/kl/projekte/korpora/).

³ Compare Dixon (2004) and the recent discussion triggered by Chung (2012) about the alleged universality of noun, verb, and adjective; see also Croft (2003, 183–188), Haspelmath (2012), and Rijkhoff and van Lier (2013).

⁴ “Japanese” here stands for the standard variety of the language, which requires a certain amount of abstraction, in particular from social and dialectal variation. See Coseriu ([1952] 1975, [1958] 1974, § 2) for an extensive discussion of the ontological and analytical status of an “abstract language” in linguistic enquiry. Cf. also Coseriu (2007).

⁵ Compare: Tarō wa eigo ga hanas-eru, Tarō wa eigo wa hanas-eru, Tarō wa eigo o hanas-eru, Tarō ni eigo wa hanas-eru, all roughly translatable as ‘Tarō can speak English’ as well. By contrast, the following sentences are ungrammatical: *Tarō ga eigo ni hanas-eru, *Tarō wa eigo ni hanas-eru, *Tarō ni eigo ni hanas-eru, and *Tarō ni eigo o hanas-eru.

⁶ Contrary to standard practice (e.g. Toyota 2011; Iwasaki 2013), I adopt this procedural perspective, which is still very common in the literature on diathesis in general, for expository purposes only, without implying that the judō and ukemi sentences are in any real sense “transformations” or “conversions” of a basic active sentence. I consider diatheses (sensu latiore, including active, passive, middle, judō, ukemi, causative, benefactive, applicative, antipassive, inverse, non-actor focus etc., cf. Kulikov 2011; Siewierska 2013) to be constructions in their own right, in line with the constructional view of grammar and linguistic knowledge. As Coseriu (1987) recognized, the procedural perspective is a metalinguistic one, which means that “promote” and “demote” are in the first place explanatory devices (even though they can refer to real processes even in “ordinary” language under certain circumstances).

⁷ Shibatani (1994) is an illuminating analysis of adversity against the backdrop of crossconstructional similarities in several languages. For an altogether different analysis of ukemi in terms of a multiple polysemous construction, s. Wierzbicka (1988); compare also Toyota (2011).

⁸ This might partly explain why Japanese speakers had apparently no problems integrating the originally un-Japanese “direct passive” typical of European languages such as English and Dutch into their diathesis system, apparently as an extension of the original ukemi (cf. Coseriu 1987, 112, 118 and Kishitani 1985, 230, pace Shibatani 1985, 842, who analyzes the “indirect passive” as an extension of the “direct passive”). Cf. also Kinsui (1997, 773) who makes a similar remark concerning the introduction of niyotte as a variant of ni in Japanese as a result of language contact with Dutch.