Pernilla Hallonsten Halling

Sampling procedure and data collection

The language sample of this study contains 60 languages from around the world. The first principle for selecting sample languages was whether any encoding characteristic of the adv function could be found. This principle comes from the focus of the thesis on adverbs specifically, and the fact that many languages do not have any adverbs. The 60 languages of the sample are presented in table 4.2. As indicated in the leftmost column, the languages are grouped into five linguistic continents, with around a dozen languages for each continent. The boundaries for linguistic continents were roughly based on Dryer (1992), with the difference that Dryer groups Southeast Asia with Oceania, and Australia with New Guinea. In my sample, Southeast Asia is placed with Eurasia, and Australia and Oceania have been collapsed into one continent which includes Papua New Guinea. The language names are found in the middle column, and the rightmost column presents their family affiliation. Families follow the classification of Glottolog (Hammarström et al. 2017), except where this clashes with a well-recognized reference grammar classification. Language names are primarily used based on reference grammars. Only in a few cases where this clashes with Glottolog is the name used according to the latter, if that name is more established.
The sample has 48 different families and three isolate languages. Two families have a representation of three languages each, and five families contribute two languages each. In these cases, languages have been included because they display encoding patterns remarkably different from the one or ones that already represented the family in the sample. There is no intention behind the exact distribution of the families that are represented by more than one language. Rather, these languages were selected based solely on the fact that their data contributed something new to the study. To achieve areal spread and a representative distribution in this worldwide sample, twelve languages from each of the major five linguistic continents were selected. Unrelated languages have thus been selected as far as possible, aiming for maximal genealogical diversity. The geographic spread of the languages is captured in the map in figure 4.7.

Continent