Prototypical adverbs: from comparative concept to typological prototype

Abstract

While adjectives and their potential universality have been much debated, adverbs remain rather neglected in the typological and cognitive literature. From a typological perspective, adjectives can be dealt with using a comparative concept: rather than assuming from the outset the existence of a class of adjectives, a particular language-independent definition of adjectives is used as a heuristic for examining recurrent form-meaning combinations. In the present article, adverb is addressed as a comparative concept in the same vein: an adverb is a lexeme that denotes a descriptive property and can be used to narrow the predication of a verb. This comparative concept is applied to a sample of 41 languages from the whole world. The results show that although there are diverse structural possibilities in terms of different adverbial constructions of varying spread and productivity, simple adverbs are found in a considerable number of unrelated languages, even in some cases where adjectives cannot be found. Clear adverb subtypes reminiscent of semantic types of adjectives further emerge, leading to a discussion of whether the comparative concepts in this case allow us to uncover a substantial cross-linguistic prototype.

Keywords:

• Comparative concept
• adverb
• adjective
• prototype
• parts of speech

Notes

¹ The notions prototype and prototypical are here used in the sense of Rosch (1978).

² Part of speech labels are capitalized throughout the article when referring to language-specific categories (such as the English Adjective), but not elsewhere.

³ It should be noted that our interpretation of the term reference here is restricted to the discourse function as defined by Croft (1991, 2001, 2003).

⁴ Dixon (2004, 5) also proposed a third set, consisting of semantic types in large adjective classes, but being on the marginal side of prototypical adjectives, they will not be discussed here.

⁵ The 41 languages are: Sango (Atlantic-Congo) (Samarin 1967), Swahili (Atlantic-Congo) (M. Thanssoula, p.c.), Lango (Nilotic) (Noonan 1992), Nama (Khoe-Kwadi) (Hagman 1977), Maltese (Afro-Asiatic) (Borg and Azzopardi-Alexander 1997), Dutch (Indo-European) (Donaldson 1997), Lithuanian (Indo-European) (Ambrazas 1997), Basque (Isolate) (Hualde and de Urbina 2003), Georgian (Kartvelian) (Hewitt 1995), Kham (Sino-Tibetan) (Watters 2002), Marathi (Indo-European) (Dhongde and Wali 2009), Turkish (Turkic) (Kerslake and Göksel 2005), Abkhaz (Abkhaz-Adyge) (Hewitt 1979), Lezgian (Nakh-Daghestanian) (Haspelmath 1993), Tsakhur (Nakh-Daghestanian) (Schulze 1997), Ainu (Isolate) (Tamura 2000), Lahu (Sino-Tibetan) (Matisoff 1973, 2003), Kiowa (Kiowa-Tanoan) (Watkins 1984), Yimas (Lower-Sepik Ramu) (Foley 1991), Abau (Sepik) (Lock 2011), Alamblak (Sepik) (Bruce 1984), Bukiyip (Nuclear Toricelli) (Conrad and Wogiga 1991), Imonda (Border) (Seiler 1985), Bininj Gun-Wok (Gunwinyguan) (Evans 2003), Mian (Nuclear Trans New Guinean) (Fedden 2011), Kewa (Nuclear Trans New Guinean) (Franklin 1971), Kambera (Austronesian) (Klamer 1998), Kilivila (Austronesian) (Senft 1986), Paamese (Austronesian) (Crowley 1987, 2002), Tagalog (Austronesian) (Schachter 1972; Himmelmann 2005), Koasati (Muskogean) (Kimball 1991), Slave (Athapaskan) (Rice 1989), Ojibwa/Nishnaabemwin (Algic) (Valentine 2001), Lakota (Siouan) (Boas and Deloria 1941; Van Valin 1977), Cherokee (Iroquian) (Montgomery-Anderson 2008, Pulte and Feeling 1975), Yakima Sahaptin (Sahaptian) (Jansen 2010), Tariana (Arawakan) (Aikhenvald 2003), Bora (Boran) (Thiesen 1996; Thiesen and Weber 2012), Mam (Mayan) (England 1983), Mapudungun (Auracanian) (Smeets 2008), Waiwai (Cariban) (Hawkins 1998).

⁶ As commented by a reviewer, this example does not illustrate a clear adjectival use of gehppa: it is not clear that it narrows the reference of a noun. After extensive searching, this is still the only in-context example that I can find for Kham, which is why it is kept, even though its value in this context can be questioned. Still, Watters treats the items discussed as Adjectives, an analysis which I follow in this article. However, the fact that these items can only questionably be classified as Adjectives does not go against my main point, which is that Kham has (a few) Adverbs.

⁷ Glosses follow the Leipzig Glossing Rules, with the additions of aor = aorist, cnj = conjunction, emp = emphatic, exc = exclamation, tag = question tag. For Mian: n2 = neuter 2. For Swahili: 7 = class marker 7, 9 = class marker 9, 11 = class marker 11.

⁸ Kambera does not allow monosyllabic roots as independent prosodic words, which is why in (7)b a default consonant /ŋ/ is attached to the root together with /u/, to avoid a closed syllable, which is not allowed in Kambera (Klamer 1998, 284). In (7)a, where the clitic = nya already is attached to the Verb, adding ngu is not necessary to form a prosodic word. In (7)c below, the full form pingu is used even though a clitic is attached, as ya could not form a prosodic word with the root pi.

⁹ As commented by a reviewer, this could potentially also be a resultative usage of sioní “bad,” comparable to e.g., beat someone blue and green.