In this concluding chapter, I summarize some of the major findings and contributions of this study and discuss implications for future research.
From the start, the goal of the study has been to understand the applicative systems of West Nusantara in typological context, but also on their own terms, in the context of the diachronic and synchronic systems in which they developed and are used. In addition, I have given special attention to broadening the description and cross-linguistic comparison of applicative constructions (ACs) in languages of West Nusantara. Accordingly, the depth and the breadth by which I have described ACs in West Nusantara is a major contribution of this study. This has been driven by two factors.
First, the conceptual framework for applicatives developed in this study allows for broad inclusion in the types of constructions that were considered relevant. In Chapters 1 and 3, I outlined a constructional approach whereby ACs are understood as a conventionalized pairing of a fixed form and consistent meaning, and I showed how elements of form and meaning for ACs might be further elaborated. This focus led to consideration of a variety of clausal constructions with functional similarities or formal similarities to canonical applicatives, including non-canonical ACs of various types, and non-applicative AM-marked constructions with diverse functions.
Second, I have sought to emphasize distinctions that are meaningful in the internal linguistic systems of West Nusantara languages. In Chapter 2, I gave a thorough descriptive account of the applicative system of Sundanese, a language which has received relatively little attention in previous research on applicatives. As part of that case study, I showed that ACs in Sundanese cannot be clearly distinguished from causative constructions, pluractional meanings, and category-changing constructions marked by the same morphemes. Furthermore, in Sundanese the overlap between applicative and non-applicative functions of AMs is observed across diverse forms of AMs and in diverse types of ACs, e.g. goal-selecting ACs with ‑an, theme-selecting ACs with ‑keun, and beneficiary-selecting ACs with the unique applicative circumfix pang- ‑keun.
In keeping with these two driving factors, in Chapter 3, the relationship between applicatives and symmetrical voice in western Austronesian languages was also examined. On the one hand, in languages like Balinese, symmetrical voice and applicatives are two separate dimensions of the verbal system. In languages like Kimaragang, on the other hand, certain symmetrical voice constructions (lv, cv, etc.) can be considered applicatives in cross-linguistic context on the basis of their form and function. But within the internal grammatical system of the individual language, they operate within a single paradigm of alternations that includes constructions that are clearly not applicatives (av, pv). On this basis I proposed the terms pivot-neutral applicative for the Balinese type and pivot-selecting applicative for the Kimaragang type, and showed how this would allow for better integration of Philippine-type languages into a cross-linguistic typology of applicatives.
As reported in Chapter 4, I conducted a large scale typological survey of the languages of West Nusantara, examining data in 85 languages representing extensive geographic diversity and genetic diversity under Malayo-Polynesian. This yielded a breadth and depth of data on applicatives in West Nusantara on a scale far larger than any previous research. Based on the survey, basic distributional facts for applicatives in West Nusantara languages were established, including a number of important patterns that have not been previously identified.
First, while the presence of pivot-neutral applicatives are the norm in West Nusantara, there are conspicuous subareal patterns whereby applicatives are missing in mainland Southeast Asia north of Peninsular Malaysia, and Borneo south of Sabah. This lack of applicatives correlates with other typological changes in the relevant languages, including more reliance on fixed word order to signal grammatical relations, shift to analytic rather than morphological means for modulating voice and valency, and loss of suffixation. In mainland Southeast Asia it is well known that these changes were caused by language contact with Austroasiatic languages. In Borneo, there are also indications that these changes have spread through contact, beginning in southwest Borneo in the area where Land Dayak languages are spoken and gradually affecting North Sarawak languages and Greater Barito languages in adjacent areas, as well as some Malayic languages like Iban and Mualang. While Adelaar (2006) previously identified some such trends in West Kalimantan and southern Central Sarawak, the extent of these patterns is larger than he recognized at the time. Still, the initial trigger for these changes in Borneo is not identified, though one must wonder whether speakers of non-Austronesian languages present in Borneo upon the arrival of Malayo-Polynesian groups did not influence the course of development for the newcomers’ languages.
Second, in West Nusantara not only is the presence of pivot-neutral applicatives the norm, pivot-neutral applicatives are associated with diverse typological profiles, and are found across widely differing systems of alignment and across many types of symmetrical and asymmetrical voice systems. This casts doubt on the association of (pivot-neutral) applicatives with a proposed category of Indonesian-type languages showing two-way symmetrical voice systems and use of special coding for non-pivot actors, among other characteristic properties.
Third, instead of association with the emergence of a particular typological profile, there are indications that the presence of pivot-neutral applicatives in West Nusantara is associated specifically with the decline of Philippine-type voice. Particularly in Central Sama, Totoli, Pendau and Bobongko, we see incomplete stages in the transition from Philippine-type voice to pivot-neutral applicatives, showing how the former might have receded in productivity, while the latter expanded. The extension of applicatives into co-occurrence with both av and pv constructions drove reorganization from an integrated Philippine-type voice system (av, pv, lv, cv) to the separation of symmetrical voice (av and pv) and pivot-neutral applicatives as independent paradigmatic alternations in the verbal system. This probably took place incrementally. For example, in Pendau the goal- and beneficiary-selecting applicatives co-occur with av marking, but locative- and instrumental-selecting applicatives with the same suffixal marking are only found in pv.
In Chapter 6, distributional patterns for the forms of AMs, and their functions were presented. The predominant pattern in West Nusantara is for one form of AM to be associated with selection of locative and goal applied phrases, and another form of AM to be associated with selection of beneficiary, theme, and instrument applied phrases. This pattern of functional distribution across AMs is pervasive and broadly distributed across the region. It also shows marked similarities to the pattern of functional distribution across the lv and cv constructions in pmp and pan. Furthermore, the association of benefactive and instrumental ACs with the same AM to the exclusion of locative and goal ACs is cross-linguistically unusual.
Based on evidence from synchronic allomorphic alternations between ‑i and ‑an with locative- and goal-selecting ACs in some languages of West Nusantara, I have argued that AMs marking locative-/goal-selecting ACs with either the form ‑i or ‑an are derived from lv morphology. For AMs that select beneficiaries, instruments, and themes, we see more variance in the form of the suffix. In South Sulawesi languages, Balinese-Sasak-Sumbawa, and Sama-Bajaw the morpheme ‑an or some regular reflex of this is used exclusively. Elsewhere, a variety of forms like ‑akən, ‑kan, and ‑ako are used, but some irregularities in expected sound correspondences for these complicate the picture. In addition, there is some evidence that older forms of benefactive/instrumental AMs have been replaced with newer ones, e.g. Javanese ‑ʔən replaced with ‑akən, and Proto Malayic ‑an replaced with ‑kan. Therefore, I have argued that the most likely explanation is that these constructions are derived from earlier cv constructions marked with a reflex of the pmp imperative/negative cv suffix *-an in earlier stages, with later replacement of the form in many languages.
In Chapter 6, I also present evidence for a number of subareal patterns in the properties of AC and AMs in West Nusantara. With respect to polyfunctionality, AMs in languages of Sulawesi are less likely to show causative functions than AMs in other western Indonesian languages. With respect to properties of ACs, a number of languages of Sulawesi show unexpected patterns of indexing and syntactic properties for the applied phrase. These patterns appear to be driven by animacy effects whereby beneficiaries are treated differently than instruments, themes, and circumstantial roles like reason and purpose. More animate participants are more likely to be indexed on the verb, more likely to show distinctive forms of such indexing, and more likely to show access to syntactic operations and to map to subject or pivot in P-oriented clauses (passive, pv).
In Chapter 7, I present a comprehensive functional typology of ACs and other AM-marked constructions based on a sample of 24 West Nusantara languages. ACs are categorized by the semantic role selected as the applied phrase and then described according to their observed range of syntactic and semantic properties across the languages. Beneficiary-selecting ACs again show different properties than theme- and instrument-selecting ACs, despite being marked with the same AMs. The former almost always show maximally ditransitive structures, while the latter typically show monotransitive structures with remapping of the companion phrase to an oblique relation. Locative- and goal-selecting ACs to a lesser extent also show some differences in properties. Some goal-selecting ACs are maximally ditransitive, but both locative- and goal-selecting ACs show remapping structures in a good number of languages.
Several directions for further research are implicated by this study. Regarding diachronic development, more historical-comparative research is needed, particularly on the forms of pivot-neutral AMs that mark beneficiary, instrument, and theme applied phrases. This research should take into account many more possible cognate forms and detailed analysis of sound correspondences than were possible to consider in this study. Research on languages of Borneo may also be of particular interest. Few languages in Indonesian Kalimantan and parts of Central and North Sarawak could be included in this study due to lack of basic documentation and grammatical description. Patterns of language contact in Borneo are also particularly complex, and it is clear that more research is needed to show how and why Borneo languages have changed structurally over time.
For applicatives in synchronic use, the influence of lexical semantics is a promising direction for further research. Here it would be of value to investigate the extent to which the subareal patterns shown for Sulawesi versus other western Indonesian languages hold. Also, adopting a constructional approach to applicatives of course implicates further studies of applicatives in usage, especially usage in natural discourse and larger scale corpus studies from which frequency patterns might be investigated. At present, it is difficult to conduct these types of research because nearly all corpus resources for West Nusantara languages are concentrated in a very small number of languages, including Indonesian and Javanese. Even smaller documentary linguistic records, collections of textual material, and lexical resources are lacking for the large majority of languages of West Nusantara. Thus, descriptive linguistic research based on field study and language documentation must go hand in hand with deeper studies of particular features of language in this region, including applicatives.
Lastly, study of applicatives in natural speech and interactive speech events is needed. This study has touched on some functions of applicatives related to semantic transitivity and information structure, like use of ACs to indicate higher specificity or affectedness, higher topicality, and given status in the discourse context for the referent of the applied phrase. These types of functions can only be reliably identified and described through study of natural connected speech. At present, the only research of this type for West Nusantara is Donohue’s (2001) study of Tukang Besi applicatives in narrative texts. Interactional and pragmatic uses of applicatives have also received little attention. But several authors mention marking of the predicate with a benefactive AM as a means of expressing a request or polite imperative, and a number of speakers with whom I have worked find that benefactive imperatives are among the most natural examples of ACs in daily usage. This could explain how the pan/pmp imperative cv suffix *-an came to be the general marker of pivot-neutral benefactive/instrumental ACs. Moreover in languages that lost morphologically-marked mood distinctions, frequent use of benefactive ACs in requests could subsequently lead to grammaticalization of the benefactive AM as a general imperative marker. Therefore, new research based on natural connected speech and interactional speech events is of great importance for further investigation of the synchronic functions of applicatives with implications for their diachronic development.