This paper concerns applicative constructions in Sundanese, the principal language spoken in West Java, Indonesia. Like many other western Indonesian languages, Sundanese makes use of applicative morphemes (AM) which affect the syntactic and semantic properties of the verbal clause when they are affixed to the verb stem. In linguistic typology, applicative constructions are widely understood to be marked by overt morphology, and to “allow the coding of a thematically peripheral argument or adjunct as a core argument” (Peterson 2007:1). In Sundanese, when the verbal stem bears apparent AM affixes, a thematically peripheral semantic role is selected to map to a clausal constituent, i.e., the applied phrase. However, syntactic coding of the applied phrase varies across and within AM-marked constructions. While it may show coding associated with core arguments, i.e., an unmarked NP, sometimes it takes the form of a PP instead. Given these facts, how should we understand the function of these affixes and the status of the applied phrase in Sundanese?
Using original primary data, published literature, and corpus resources, I examine three Sundanese constructions: the goal-selecting construction marked with -an, the theme-selecting construction marked with -keun, and the beneficiary-selecting construction marked with pang–keun. Clauses marked with these three AMs show different patterns of possible syntactic coding of the applied phrase and the companion phrase, i.e., the constituent that maps to the semantic role expressed as the P argument of the base verb. Even so, examination of the semantic and inferential properties of AM-marked clauses using tests from Riesberg (2014) demonstrates that these constructions have an identifiable and consistent semantic structure, and that the applied phrase shows properties of a clausal argument, rather than an adjunct, even when coded as a PP. This shows that certain syntactic properties of ACs (e.g., relative syntactic valency, syntactic coding of arguments) often do not correlate with their stable semantic properties. Thus, careful examination of semantic properties is key to developing an adequate typology of the AM-marked verbs in western Indonesian languages and the observed range of structures found with them in usage, especially those that differ from expected forms and accepted definitions for applicatives.
In this presentation, we compare the different functions of applicative morphemes (AMs) across the lexicons of a sample of four western Indonesian languages in order to understand the influence of lexical semantics on their distribution. To show how semantic and syntactic effects of AMs pattern, we adopt the treatment of AM-marked clausal constructions as pairings of form (morphological content and syntactic
structure) and meaning (semantic content and structure) (see Goldberg 1995).
We address the following research questions: Do AM-marked constructions have an identifiable set of core meanings? How are these meanings distributed across languages of western Indonesia and their
lexicons? Using data from original fieldwork, published descriptions, corpora and lexical
resources, we compare unmarked and AM-marked constructions across a common set of roots based on the 80 lexical meanings on the Leipzig Valency Classes Project Questionnaire
(Malchukov & Comrie 2015).
On the basis of patterns identified, we propose an inventory of core constructional
meanings marked with AMs and show how languages of the sample differ with respect to this
inventory. For example, benefactive applicative constructions vary in productivity, with the
Sundanese benefactives being compatible with far more lexical bases than the
others. Across languages, the benefactive construction is most commonly centered around
lexemes that entail acquisition, production or processing of material objects. In languages
with more productive benefactive constructions, many additionally compatible verbs entail a
change of state.
Our results reinforce previous findings that syntactic categories, such as transitivity of the base or the unergative/unaccusative distinction, do not adequately explain the distribution of the functions of AMs across the lexicon (see Kroeger 2007). The study also systematically expands identification of types of semantic information that speakers are likely sensitive to in producing and interpreting verbal constructions formed with these affixes, suggesting pathways along which lexicalization of such constructions has occurred.
In this paper, I show that the argument structure of base verbs and components of lexical meaning
together influence compatibility of bases with functions of applicative suffixes. For example, both
transitive verbs of caused motion that select a theme argument, e.g. mengambil‘take’, membawa‘carry’,
and transitive verbs of creation that select a ‘product’ argument, e.g. memasak‘cook’, menjahit‘sew,’ are
compatible with the benefactive applicative –kanconstruction, as shown below. However there are some
semantic differences in resulting constructions.
Some other transitive verbs are entirely incompatible with benefactive –kan, including verbs of
perception, e.g. there is no verb *menontonkanmeaning ‘to watch something for someone’. Accordingly,
I argue for the incorporation of lexical meaning into representations of applicative constructions, and
propose ways to do so, using corpus data representing one million Indonesian sentences and drawing on construction grammar and frame semantics (Goldberg 1995, Perek & Patten 2019, Goldhahn et al. 2012).
In this study, we investigate the extent to which components of lexical semantics consistently explain the distribution of functions of western Indonesian applicative morphemes (AM) in combination with roots. Using a sample of six western Indonesian languages (Besemah, Indonesian, Sasak, Javanese, Balantak, Sundanese) that contain at least one AM, we compare patterns in function of these across roots that express a set of common meanings. These meanings are selected to represent classes of words with shared semantic components including transfer of possession, e.g. ‘buy’, ‘sell’, application of force, e.g. ‘hit’, locomotion, e.g. ‘walk’, directed motion, e.g. ‘throw’, and sensory perception, e.g. ‘see’, ‘hear’, among others. The meanings of words formed by addition of all available AMs in each of the languages are coded for function and resultant patterns, both those common and variable across the sample are presented. Download slides ♦ Download abstract
This paper presents the results of a dialect survey of the Cia-Cia language of southern Buton Island in Southeast Sulawesi, Indonesia. This survey also encompassed the closely related Kumbewaha and Lasalimu languages, that is to say, all three languages belonging to the Butonic branch of the Muna-Buton group of languages. The principal results of this survey are as follows: (a) the Cia-Cia language comprises two primary dialect areas, a western dialect and a central-eastern dialect complex; (b) the small (and previously undocumented) Wasambua lect is recognized as a third, outlier dialect; (b) Kaisabu is elevated to the status of a separate language; (c) Kumbewaha is more dialectally complex than has hitherto been recognized; (d) the unity of the Butonic subgroup within Muna-Buton is confirmed. Link to paper
The verbal affixes which form applicative constructions in western Indonesian languages are extremely polyfunctional, but this fact is not widely discussed nor well-accounted for. In this paper, we survey the functions of applicative affixes in eight western Indonesian languages. We pay close attention to the varied syntactic properties and semantic nuances of apparent applicative constructions. We demonstrate that the non-prototypical functions are both prevalent in individual languages and broadly distributed across languages of the region. Download slides ♦ Download abstract
This project is an international collaboration with researchers at Universitas Pendidikan Indonesia, teaming up to document the language of the Baduy Dalam ‘Inner Baduy,’ a small group of about 1170 living in a remote area on the island of Java in Indonesia. The team will record natural speech (narratives and conversations) and lexical items to produce an audio and video transcribed corpus of Baduy Dalam speech, a dictionary (with special focus on culturally distinctive concepts), and a grammar sketch. Materials are archived at Pacific and Regional Archive for Digital Sources in Endangered Cultures, where they will be accessible to other researchers and the general public. Broader impacts include producing print and video materials for direct use by the Baduy as linguistic and cultural education resources.
This paper presents evidence from Behoa (Austronesian; Indonesia), showing that vowel harmony developed through phonologization of earlier vowel allophony which was enhanced through vowel-to-vowel coarticulation. The steps of development seen suggest that other morphological, lexical, and prosodic factors favored the rise of VH, including the shape and stress patterns of roots and suffixes, and the contrastive load of low vowel phonemes. Cross-linguistic examples of vowel phenomena showing similar steps of development are also discussed. This study represents new descriptive work on VH in a lesser-known language and contributes to the relatively small body of research on the diachrony of VH. Download slides
Christina L. Truong Paper presented at Workshop on Austronesian voice and related phenomena, University of Hawaiʻi at Mānoa, 26 November 2019.
In this paper, I investigate the extent to which western Indonesian verbal suffixes involved in applicative constructions also perform non-applicative and non-valency increasing functions. I survey forms and functions of applicative morphology in a sample of western Indonesian languages of different types of voice systems, including Karo Batak, Sundanese, Pendau, Balantak, Tukang Besi.
A number of observations emerge from the study. First, the properties of a base are not sufficient to predict which which affix it will combine with, nor what the resultant meaning will be. Thus the constructions are not purely compositional. Second, the function of these affixes cannot be equated with bringing a participant from the periphery of a clause into the core in many cases. In some applicative constructions, the applied object is not a participant in the event, e.g. purpose applied objects. In other cases, no applied object with a peripheral role is present, as in many non-valency-increasing constructions. Furthermore, for many derived verbs, there is no base clause to speak of, and thus no peripheral roles can be identified. A satisfactory analysis of the so-called “applicative affixes” must take into account their non-applicative functions. Download full abstract
Christina L. Truong Working paper based on original data, May 2019
The Sundanese verbal system includes a substitutive benefactive construction which indicates that the agent performs the action on behalf of, and instead of, a beneficiary. This construction is formed with the prefix pang– and the causative/applicative suffix –keun. In this paper I describe the morphosyntax of this construction, including the morphological components found on the verb, the morphophonemic processes involved, and the syntactic properties of various types of clauses with substitutive benefactive meanings. Finally, I compare Sundanese substitutive benefactives with other Western Austronesian benefactive constructions and discuss their historical origin.
Based on the study, the Sundanese pang– prefix appears be derived from the agentive nominalizer *paŋ, which when used in applicatives came to have the meaning ‘to serve as an agent of an action for s.o.’ Sundanese substitutive benefactive verbs can be transitive or ditransitive, and ditransitive argument structure appears to be fully grammatical in both active and passive voice. However, Sundanese also has ditransitive instrumental and simple benefactive applicatives, but I present some evidence that these are not fully grammatical in active voice. The association between passive voice and such three-place verbal constructions might be linked to the four-way voice system of Proto-Austronesian, with earlier undergoer, instrumental, and beneficiary voice functions being subsumed into modern passive voice. Download full text