Exploring the difference between linguistic borrowing and inheritance...
In areas without written historical records, where archaeological and ethnographic data are absent or sparse, language forms the backbone of our understanding of socio-cultural history. This project investigates one such region, in eastern Indonesia. What can languages spoken in the Lesser Sunda Islands today tell us about the histories of its various population groups? Answering this question requires a productive conjunction of contact linguistics, historical linguistics, and language typology studies. Our methodology includes quantitative cross-validation of qualitative research, and careful control of the variables that is uniquely enabled by the situation of the Lesser Sundas. A fundamental idea in historical and contact linguistics is that similarities between geographically close languages are not accidental, but point to a shared history of their speakers. Either, the speakers descend from a common ancestor, and the similar features were passed down the generations (vertical transmission); or they are, or once were, in mutual contact, and adopted features from one language into the other (horizontal transmission). Classic methods in historical comparative linguistics largely focus on vertical transmission and internally motivated changes, aiming to reconstruct the common ancestor and mutual relationships within groups of related languages. Language contact studies, on the other hand, focus on patterns and constraints in externally motivated changes. Truly unravelling the linguistic history of a region requires an approach which combines the historical comparative method with a constrained theory of language contact (Harrison 2003), investigating both vertical and horizontal transmission. Studying the past through a linguistic lens also implies that we study change and retention in both the lexical and the grammatical domain, as both are influenced by different dynamics of contact and retention, and show different types of traces. Lexicon is easily borrowed; grammar is not (cf. Thomason 2001: 70-71, 2010). Within the lexicon, items have different borrowability, e.g., nouns are more easily borrowed than verbs. Loan words suggest contact in particular socio-semantic domains like religion, politics, or technology at specific moments in time (e.g. abbey, prince < French; dike, dam < Dutch), datable by their spread through a group of languages and level of integration into individual languages. They reflect cultural systems (e.g., kinship) and social networks. Changes in sentence grammar can also point to foreign influence, but syntactic changes follow different paths and are induced by different sociolinguistic contexts than changes in the lexicon, typically involving more intimate and long-term contact. For instance, in a situation where speakers are bilingual from childhood, the intensity of contact is much higher and has different outcomes than in language contact through casual trade. In turn, contact through childhood bilingualism differs in intensity from bilingualism where post-adolescents or adults adopt a second language. Apart from life-stage loci of change and intensity of contact, many other factors determine the linguistic outcomes of contact, including the social status of the languages, the language attitudes of the speakers, the degree of (geographical, social) isolation of the community, and the duration of contact. Given the number and variety of factors that bring about linguistic change through contact, efforts to develop universally valid models for contact-induced changes have been met with skepticism. Thomason claims that: ‘...deterministic predictive theories of contact-induced change [...] are doomed’ (2007:41). Historical linguists have found few if any constraints in language contact: ‘...any linguistic feature can be transferred from any language to any other language‘ (Thomason and Kaufman 1988:14). On the other hand, language contact specialists argue that not all types of borrowing are equally likely to happen (e.g., Matras 2007); that contact-induced transfers may be shaped by universal principles of grammaticalization (Heine and Kuteva 2005); and that specific contemporary contact settings constrain transfer in various ways (Sankoff et al. 1988). In searching for constraints on contact-induced transfer, case studies in this project investigate specific paths of change that occur in the language of multilingual individuals. This provides a bottom-up perspective that is fundamentally different from studies that start from the resulting language situation to retrace the factors that brought it about (cf. Curnow 2001). We will conduct three case studies of ongoing change in individuals in contemporary contact situations, varying in intensity of contact, language status, direction and time-depth, applying the ‘scenario’ approach (Muysken 2010) adduced with evidence about socio-cultural history and cultural contact. The resulting models will be quantitatively validated. Because the contact situations occur between Papuan and Austronesian languages, the structural typology of the languages is kept constant in the comparison, with social context and types of change as variables. These studies will provide a clearer picture of ongoing changes in the languages of the Lamaholot-Pantar-Alor (LPA) region. This will show which social circumstances and types of contact lead to which patterns of change - including changes that do not happen. As such, it will refine (probabilistic) constraints on contact-induced transfer. We also expect to gain detailed insight into the types of grammatical change that occur when languages of different typological profiles are in contact. This is of high theoretical relevance, as it will yield a clearer distinction of vertically and horizontally transmitted language features. Thus, it will expand the scope and the reliability of linguistic findings for reconstructing the past of population groups in the region, and produce a methodology that can be utilised elsewhere in the world. Within the Lesser Sundas, we focus on the region indicated on Figure 1. Covering a latitudinal distance of almost 200 kilometers, it includes east Flores and adjacent islands, Pantar and Alor. These islands are the westernmost place where Austronesian and Papuan languages meet (fn. 1). In the west, the Austronesian languages Lamaholot and Alorese are spoken; in the east, the Papuan languages of Pantar and Alor. Below we refer to this geographical region as the aforesaid “Lamaholot-Pantar-Alor (LPA) region”, as distinct from the linguistic grouping “Alor-Pantar (AP) family”. The AP family consists of some twenty languages, including Western Pantar, Teiwa, Kaera, Blagar, Adang, and Abui (Fig 1.). It is related to the five Papuan languages spoken on Timor and Kisar (Schapper et al., to appear).
Figure 1. Languages under investigation in their geographical and genealogical context. The Austronesians are commonly assumed to have arrived in the area ~3,000 Before Present (BP) (Pawley 2005:100, Spriggs 2011). The origin and age of the AP family, which is located some 1000 kilometers away from the Papuan mainland and surrounded by islands with Austronesian languages, is less clear. One hypothesis holds that they are descendants of immigrants from New Guinea who arrived in the Lesser Sundas 4,500-4,000 BP (Bellwood 1997:123, Ross 2005:42, Pawley 2005). However, recent bottom-up historical comparative research (Robinson and Holton 2012, Holton and Robinson, to appear b) argues that there is no lexical evidence to support an affiliation with the Trans New Guinea languages (cf. Wurm, Voorhoeve, McElhanon 1975, Ross 2005). Another hypothesis holds that the Papuans in the Lesser Sundas descend from arrivals 20,000 BP (Summerhayes 2007). While this possibility cannot be excluded, the level of lexical and grammatical similarity in the AP family does not support an age of more than several millennia, and the reconstructed vocabulary of proto-AP appears to contain Austronesian loan words (e.g., ‘betel nut’, Holton et. al. 2012). Ancient Austronesian loans found across the AP family following regular sound changes suggest that the AP family split up after being in contact with the Austronesian languages in the area, which would give it a maximum age of ~3,000 years.
Resolving this issue requires independent evidence dating proto-AP relative to Austronesian – which requires detailed information on the linguistic and socio-cultural history of population groups in the LPA region and their interactions. This project will seek to provide such information.
(1) ‘Austronesian’ is used here as shorthand for the subgroup of Malayo-Polynesian languages spoken in the Lesser Sundas, steering clear from the debate about the internal structure of the MP subgroup (Blust 1993b, Donohue and Grimes 2008, Blust 2009b). ‘Papuan’ conventionally refers to non-Austronesian languages spoken in New Guinea or its vicinity. Unlike ‘Austronesian’, ‘Papuan’ refers to a cluster of unrelated language families.