Pernilla Hallonsten Halling

Language
Affiliation
Africa	Acoli	Nilotic
	Bambara	Mande
	Ewe	Niger-Congo
	Hdi	Afro-Asiatic
	Jamsay	Dogon
	Koyra Chiini	Songhay
	Krongo	Kadugli-Krongo
	Ma’di	Central Sudanic
	Maltese	Afro-Asiatic
	Nama	Khoe-Kwadi
	Sango	Niger-Congo
	Swahili	Niger-Congo
Eurasia	Abkhaz	Northwest Caucasian
	Ainu	isolate
	Basque	isolate
	Dutch	Indo-European
	Estonian	Uralic
	Georgian	Kartvelian
	Kham	Sino-Tibetan
	Marathi	Indo-European
	Lahu	Sino-Tibetan
	Lezgian	Nakh-Daghestanian
	Turkish	Altaic
Australia & Oceania	Abau	Sepik
	Alamblak	Sepik
	Bukiyip	Toricelli
	Gooniyandi	Bunaban
	Bininj Gun-Wok	Gunwinyguan
	Imonda	Border
	Kambera	Austronesian
	Kewa	Nuclear Trans New Guinea
	Kilivila	Austronesian
	Mian	Nuclear Trans New Guinea
	Tagalog	Austronesian
	Yankunytjatjara	Pama-Nyugan
	Yimas	Lower Sepik-Ramu
North America	Cherokee	Iroquoian
	Cuicatec	Otomanguean
	Jamul Tiipay	Cochimi-Yuman
	Kalaallisut	Eskimo-Aleut
	Kiowa	Kiowa-Tanoan

Koasati Muskogean

Lakota Siouan
Mam Mayan
Nishnaabemwin Algic
Nuu-chah-nulth Wakashan Sahaptin Sahaptian
Slave Athapaskan-Eyak-Tlingit
South America Bora Witotoan
Cavineña Tacanan
Guaraní Tupian
Hup Nadahup
Imbabura Quechua Quechuan Mapudungun Auracanian
Paumarí Arawan
Piraha –
Urarina isolate
Yagua Peba-Yagua
Wai Wai Cariban

Warekena Arawakan
Table 4.2. Language sample

Africa
Australia & Oceania Eurasia
North America South America

Figure 4.7. Language sample with linguistic continents

The data used in this study comes primarily from language descriptions in the form of reference grammars. Secondarily, language experts and native speakers have been consulted in unclear cases wherever possible.

As discussed in sections 4.2, a qualitatively thorough analysis is necessary to investigate the functions of attr, pred, and adv in a way that can answer the research questions (cf. chapter 1). This motivates the average size of my sample – with too large a sample, detailed analysis would not be possible. Since I have selected languages partly based on

whether they have any characteristic features in adv, my sample is in accordance with the characteristics of a variety sample (Bakker 2011; Veselinova 2012). Bakker describes a variety sample as suitable “when linguistic variables are explored about which not much is known in advance” (2011: 105). As discussed in chapter 2, adjectives have been thoroughly studied typologically, but adverbs have received far less attention. The presence of the phenomenon in focus is a requirement for a language to be included in a variety sample. In my sample, I have not exclusively selected such languages, with the aim of balancing the sample with languages that have other encoding patterns as well (see, e.g., section 5.3 in the next chapter). However, an implication of primarily selecting languages in which a phenomenon is in some way attested is also that languages for which useful documentation can be found must be chosen. This is in line with what Bakker discusses as a bibliographic bias (2011: 107), by which all documented languages already belong to a biased sample of all the languages of the world. Furthermore, language descriptions vary in terms of what they include, both when it comes to focusing on different aspects of a language as such, and in cases where theoretical frameworks are used as a basis. In my case, the fact that encoding characteristic of the adv function does not appear to be that widespread made it all the more important to find as reliable sources as possible, resulting in further restrictions on the availability of appropriate languages for the sample. For an overview of sampling methods and the problems surrounding them, see Bakker (2011).

5. Summary

In this chapter, the functions of attr, pred, and adv have been discussed. The relations between these functions have been examined as involving property words in modification for attr and adv, and property words at some level of predication for pred and adv. Different potential encoding patterns for the three functions were examined along with examples from English and Lakota. It was shown that analysis of encoding must be car- ried out at three, or perhaps even four, different levels: the lexeme, root, word form, and construction. The constructional-typological approach following Koch (2012) was intro- duced and discussed in detail, in order to be employed in the analysis of the results that follow in the next part of this dissertation. Finally, the language sample was presented, and the procedures that were followed in sampling and data collection were discussed.

Part II.

Results

2. Results I: adverbs and adverbials

   1. Introduction

In this chapter and the two that follow, the results of examining the encoding of attr, pred, and adv are analyzed. In chapter 6, this analysis is carried out at the root, word form, and lexeme levels, while in chapter 7, it is done at the level of the construction. The present chapter is devoted specifically to how the adverbial function (adv) is encoded in the languages of the sample. The aim is to primarily examine encoding characteristic of adv, i.e. encoding that cannot be used in exactly the same form in any other function. The most important findings are the following:

• 
  The majority of languages within the sample have at least a few simple adverbs.
  
• 
  A considerable number of languages with simple adverbs do not have simple adjec- tives.
  
• 
  In the languages without simple adverbs, various encoding strategies are attested in
  

adv.

Languages with simple adverbs are described in section 5.2. A number of examples are presented that show how classes of simple adverbs vary across languages. Section 5.3 is devoted to other encoding strategies: derived adverbs, case-formed adverbs, adverbials formed by reduplication, ideophonic adverbs, adverbial affixation, etc. The chapter is summarized and concluded in section 5.4.

2. Languages with simple adverbs

Just over two thirds of the languages in the sample of this study have simple adverbs. We may recall from chapters 1 and 4 that simple here means single-word, monomorphemic lexemes. The languages with simple adverbs are listed in table 5.1, and are displayed on the map in figure 5.1.
Table 5.1. Languages with simple adverbs
Acoli (Nilotic) Koyra Chiini (Songhay)
Ainu (isolate) Krongo (Kadugli-Krongo)
Alamblak (Sepik) Lahu (Sino-Tibetan)
Bambara (Mande) Lakota (Siouan)
Bininj Gun-Wok (Gunwinyguan) Lezgian (Nakh-Daghestanian) Bora (Witotoan) Ma’di (Central Sudanic) Bukiyip (Nuclear Toricelli) Maltese (Afro-Asiatic) Cuicatec (Otomanguean) Mam (Mayan)
Estonian (Uralic) Mapudungun (Araucanian)
Ewe (Atlantic-Congo)¹ Marathi (Indo-European)
Gooniyandi (Bunaban) Mian (Nuclear Trans New Guinea)
Guaraní (Tupian) Nishnaabemwin (Algic)
Hdi (Afro-Asiatic) Paumarí (Arawan) Imbabura Quechua (Quechuan) Sahaptin (Sahaptian) Imonda (Border) Sango (Atlantic-Congo) Jamul Tiipay (Cochimi-Yuman) Turkish (Turkic)² Kalaallisut (Eskimo-Aleut) Waiwai (Cariban) Kambera (Austronesian) Warekena (Arawakan) Kewa (Nuclear Trans New Guinea) Yagua (Peba-Yagua)

Kham (Sino-Tibetan) Yankunytjatjara (Pama-Nyungan) Koasati (Muskogean)

simple adverbs no simple adverbs

Figure 5.1. Languages with and without simple adverbs


The number of adverbs in each language varies remarkably. For instance, Fedden (2011) lists 38 Adverbs in Mian (Nuclear Trans New Guinean), whereas Koyra Chiini (Songhay) has only two Adverbs, according to Heath (1999: 253). Following the comparative concept
¹ The Ewe simple Adverbs constitute a marginal example, since they can also be used adjectivally. However, the adverbial use seems to be primary (Ameka 1991: 55, Yvonne Agbetsoamedo, p.c.), which is why they are included here.
² Turkish is a marginal example, on similar grounds as Ewe (Hatice Zora, p.c.).

that I use, the scope for adv is delimited to property words that are modifiers within predicating expressions. Only 13 of the Mian Adverbs are property words, as illustrated in table 5.2 (2011: 21).

Table 5.2. Mian Adverbs (Fedden 2011: 21)