Publications

This paper provides a survey of apprehension constructions in the Ngumpin-Yapa (Pama-Nyungan) languages of northern Australia. As in many Australian languages, there are morphosyntactically distinct strategies for encoding apprehension in Ngumpin-Yapa languages: a verbal apprehensional morpheme and a nominal aversive morpheme. Languages differ in terms of the extent to which these morphemes uniquely encode apprehension, some also encoding physical location or potentiality. We survey apprehensional morphemes in nine Ngumpin-Yapa languages: Walmajarri, Ngardi, Jaru, Wanyjirra, Gurindji, Bilinarra, Mudburra, Warlmanpa and Warlpiri. In these Ngumpin-Yapa languages, the verbal apprehensional morphemes are found in the auxiliary complex as a base for bound pronominals. The nominal aversive morphemes are oblique case suffixes that relate nominals and nominalised predicates to main predicates. The auxiliary bases form part of the system of finite tense–aspect–modality, whereas the case suffixes act as complementisers for non-finite clauses, with dependent tense features. Both classes of morphemes can appear in independent and dependent syntactic environments. Additionally, both classes of apprehensional morphemes can have non-apprehensional functions, for example in Wanyjirra the comitative case suffix introduces concomitant participants, which can also be used to express aversion to particular participants. This chapter makes a substantive contribution to our understanding of apprehension in Australian languages – a topic that hitherto has received little attention, despite the prevalence of apprehensional morphemes in Australian languages. This survey highlights the fact that even in a group of closely related languages, the grammatical encoding of undesirability can vary greatly.

This chapter addresses two main questions: (a) To what extent is the cultural intelligence framework (CQ) and its elements utilized in the design and implementation of training programs; and (b) What factors facilitate or hinder the effectiveness of these training programs? We conclude by generating practical implications and opportunities for the future development of these training programs.

The vowel inventories of Kamu and Larrakia each consist of peripheral /i/-/u/-/e/-/a/-/o/ and another non-low vowel that has not been phonetically characterized. Descriptions of similar six-vowel systems in Western Top End languages offer inconsistent accounts of the non-peripheral vowel, but no instrumental analysis has yet been conducted. Acoustic and phonological analysis reveals that the sixth vowel is best characterized as /ʉ/ in Kamu and /ɨ/ in Larrakia. Height and backness are the critical phonetic properties that differentiate the non-peripheral vowel from the other vowel phonemes in each language. These results enhance our understanding of the structure and properties of higher cardinality vowel systems in Australian languages, and the typology of six-vowel systems more generally.

Warumungu is a language of Central Australia with a reported contrast for stops. There has been no formal distributional or acoustic analysis of this contrast; this study addresses this gap. Analysis of 439 stops produced by three speakers reveals that the consistent phonetic correlates of the manner contrast are duration and voicing.

About one third of the Pama-Nyungan languages of Australia employ pronominal cross-referencing, yet systematic typological patterns of non-subject argument registration remain unexamined. We analyze this variation from two perspectives by surveying 22 Pama-Nyungan languages. Firstly, we survey which kinds of case-marked arguments can be cross-referenced by these pronominal systems. From this perspective, we find that a number of nominal expressions marked with so-called ‘local’ cases (e.g. locative, allative, ablative, etc.) can be cross-referenced when instantiating certain argument relations. Secondly, we find striking cross-linguistic predictability in how such relations, which we descriptively group as ‘locational’, are morphologically integrated into the pronominal paradigms. We show that the variation can be captured by two major parameters: firstly, whether locational cross-referencing utilizes the same form as another non-subject series, or whether locational cross-referencing is serviced by a unique series formally built off another non-subject series. In this latter case there is further variation as to which other non-subject series provides the base for the dedicated locational series. These parameters result in six surface pattern types, and we show that each of the patterns is instantiated in languages of the survey.

This was my PhD thesis, which has been superceded by the published monograph (ANU Press, 2024).

Warlpiri and Warlmanpa (Ngumpin-Yapa languages of Australia) exhibit a complex predicate construction in which a class of preverbs introduces a single argument that is not shared by the argument structure of the inflecting verb, nor is there necessarily any shared event structure. This is problematic for many theories of linking structures of complex predicates, since no arguments or events are shared between the predicative elements of the complex predicate. The same grammatical relation is instantiated by a beneficiary adjunct. In light of new research in event and argument structure, I propose a lexical rule which introduces an applicative argument to account for the beneficiary construction; and that the preverbs take another predicate as one of their arguments to account for the complex predicates. The applicative rule and the preverbs both introduce an argument of the same grammatical relation, leading to interesting interactions, given that two grammatical relations of the same type are not expected to co-occur within a single clause.

Pre-trained speech representations like wav2vec 2.0 are a powerful tool for automatic speech recognition (ASR). Yet many endangered languages lack sufficient data for pre-training such models, or are predominantly oral vernaculars without a standardised writing system, precluding fine-tuning. Query-by-example spoken term detection (QbE-STD) offers an alternative for iteratively indexing untranscribed speech corpora by locating spoken query terms. Using data from 7 Australian Aboriginal languages and a regional variety of Dutch, all of which are endangered or vulnerable, we show that QbE-STD can be improved by leveraging representations developed for ASR (wav2vec 2.0: the English monolingual model and XLSR53 multilingual model). Surprisingly, the English model outperformed the multilingual model on 4 Australian language datasets, raising questions around how to optimally leverage self-supervised speech representations for QbE-STD. Nevertheless, we find that wav2vec 2.0 representations (either English or XLSR53) offer large improvements (56-86% relative) over state-of-the-art approaches on our endangered language datasets.

The clitic =lku in Warlpiri and Warlmanpa (Ngumpin-Yapa, Pama-Nyungan, Australia) has been previously analyzed as a ‘change of state’ marker indicating some proposition is false at an earlier time, and true at a later time. In this paper, I examine a number of uses of =lku, which require expanding and refining the ‘change of state’ analysis, in order to argue for a new analysis of =lku as a temporal and discourse clitic. The temporal function locates the event or state being modified by =lku in the temporal consequence of another event or state. This function interacts with lexical aspect, triggering an inference of the cessation of an ongoing state or event, if there is one, and encoding the inception of the state or event being modified by =lku. The discourse use of =lku frames new information which should be evaluated against the ongoing discourse. I also demonstrate some striking cross-linguistic similarities, showing that temporal morphemes commonly interact with discourse structure, just as =lku does in Warlpiri and Warlmanpa.