This dissertation investigates the prosodic structure of French and the Oceanic language Drehu, spoken by a small bilingual community on the island of Lifou, in the South Pacific. Lifou is a remote island belonging to the archipelago of New Caledonia, localised more than 16000 km away from mainland France. Although officially a French overseas territory, live in Lifou is to a large degree organised according to customary tradition of the indigenous population, the Kanak people. There is no obvious societal majority language on the island and French and Drehu are commonly spoken by the indigenous population, who make up the majority of the inhabitants. The aim of this examination is to develop a phonetic prosodic model for the two languages and determine whether there are effects of prosodic transfer between the two languages of bilingual speakers. Of particular interest is the the phonetic description of prominence and phrasing of the two languages for a categorisation of their prosodic typology. This thesis presents five studies dealing with (i) the acoustics of Drehu word prosody, (ii) the acoustic correlates of intonational structure in Lifou French, (iii) the acoustic durational properties of Lifou French, (iv) the acoustics of prominence marking and phrasing in Drehu, and (v) acoustic cues used in word recognition in Drehu and French. The speech and perception data for these studies were collected during four field work trips to the island of Lifou, where more than 100 adult and teenage speakers participated. To investigate processing in the French language and incorporate a monolingual control group, additional experimental work was conducted in the facilities of the Laboratoire Parole et Langage, Aix Marseille University, CNRS, in Aix-en-Provence, in metropolitan France. A variety of methods, typically used in laboratory phonology, such as controlled reading tasks or a forced choice word identification experiment were employed. For exploration and interpretation of the data, all five studies include a statistical analysis. This work puts forward a revised model of Drehu word prosody and postulates an intonation phonological account of the language. In addition, the intonational phonology of Lifou French is documented, providing the first description of this previously undocumented variety. Building on this descriptive work and taking into consideration previous phonetic research on the speech production and prosody of bilingual speakers, the role of sociolinguistic motivations and functional constraints is discussed. This dissertation highlights the relevance of applying detailed acoustic descriptions to under-documented languages which are poorly understood regarding their prosodic systems. It contributes to the documentation of the languages in the Oceanic region and advances our understanding of bilingual speech processes.

The orchestral ensemble exists as a group of people who come together to prepare for public performance of music and has done so for several hundred years. In this thesis I examine the interactions which occur during this process in a current day professional orchestra. My focus is on analysing how members of the orchestra, the orchestral organisation and the conductor use their bodies, artefacts, time and space. My approach to examining these behaviours is informed by social interaction methodologies and theories of distributed cognition. Chapter 5 presents an ethnographic account of the construction of space and delineation of time for rehearsal. I examine how the City Symphony Orchestra (CSO) and their management use both space and time to prioritise and privilege the work of the orchestra. Chapter 6 focuses on conductor gestures and I use this analysis to argue that the gestures are complex with components occurring simultaneously as well as sequentially. I argue that conductor gesture creates its own context as it is deployed interactionally and is deeply embedded within social and cultural context. I use the theory of composite utterances to demonstrate that conductor gesture is more than a simple single sign per semantic unit. Chapter 7 considers how orchestral musicians organise their cognition within the physical and social environment of the rehearsal. I show that orchestral musicians distribute their cognition across their bodies, other interactants and culturally constructed artefacts. I further argue that understanding musician cognition in this way allows us to see that the very purpose of orchestral rehearsal is to transform the internal, individual cognition into the external and shared. Chapter 8 shifts the focus of analysis onto the talk-based interaction between conductor, concertmaster and other players within the rehearsal. I approach this talk using analysis which allows me to focus on the epistemic stance taking that occurs. I show that musicians are highly aware of sources of knowledge and knowing within the rehearsal process. I argue that musicians use their own bodies as sources of knowing and orient to them as important to the rehearsal interaction. Chapter 9 presents an ethnographic account of a CSO performance and considers the orchestra as a social situation. I argue that observability and monitoring occur across the social situation in both visual and aural modalities but that the access to others is asymmetrically constructed by the social roles of the orchestra. I focus on the first violin section using the leadership gestural actions as an example of this asymmetry. Chapter 10 discusses my analyses and proposes several novel contributions to existing research on how stance taking occurs in group interactions. This research is based on original fieldwork with an orchestra referred to by the pseudonym ‘City Symphony Orchestra’ (CSO) within this thesis.

In late 1920s’ Spain, Elena Fortun (pseudonym of Encarnacion Aragoneses, 1886-1952) started publishing “Celia y su mundo”, considered the best children's books series of the time. Her innovative character, Celia, tries to make sense of a world dictated by grown-ups and continually attempts to escape their impositions. Through the voice of a character belonging to two traditionally marginalized groups, children and women, Fortun found a way to transmit progressive messages to her readers. However, as she gets older, the character who broke the mould with her transgressive behaviour and convincing speech, gradually adapts to what is expected of her. The historical events reflected in the books range from the final years of Miguel Primo de Rivera's dictatorship, to the historical advances of feminism during the Republic and their loss during Franco's dictatorship after the Civil War (1936-1939). This thesis contends that the Celia series planted the seed of postwar bildungsroman for Carmen Laforet (1921-2004), Carmen Martin Gaite (1925-2000) and Ana Maria Matute (1925-2014), who read Celia in their childhood and whose novels featured teenage girls fleeing their oppressive households. The trace of Fortun’s Celia is analysed in the works: Nada (1944) and La isla y los demonios (1952), by Laforet; Entre visillos (1952) and El cuarto de atras (1978), by Martin Gaite; and Los Abel (1948) and Primera memoria (1959), by Matute. By breaking the rules of children’s literature, usually didactic and moralistic, Fortun created a character that paved the way to arguably the first generation of Spanish women writers. Celia was considered a rebel just because she could not make sense of her status quo, which continually limited her existence to that of a silent secondary character. Frustrated, Celia spoke to her girl readers who considered themselves raras for pretty much the same reasons and who created brave, nonconformist female characters years later.

Amidst accelerating globalisation, international students currently occupy a sizeable percentage of university enrolments in English-speaking countries. Of these countries, Australia has the highest ratio of international student tertiary enrolments at 21.5% (OECD, 2019). Due to this ongoing trend, the validity of high-stakes language tests has become more vital than ever before. For language tests to be valid for university entrance, they should elicit language that is relevant and applicable to successful university participation. This relates to the extrapolation of language elicited by tests to university contexts (Chapelle, Enright, & Jamieson, 2008). The purpose of this study is to investigate the relevance of language elicited by IELTS Academic Writing Task 2 (WT2) to successful graduate-level university writing produced for the Master of Applied Linguistics at the University of Melbourne. To date, studies investigating IELTS extrapolation have yielded limited evidence with little relevance to stakeholders. The present study addresses this issue by interviewing stakeholders about language needs and using the interview results to inform functional move-step analyses (Swales, 1990, 2004) of test and assignment writing for comparison. The results suggest that IELTS Academic WT2 is only somewhat relevant to successful university writing and highlight the need for students to develop skills in defining disciplinary concepts and interpreting study results. The results also suggest that IELTS scores should be interpreted with caution, given that claims pertaining to score use currently rest on weak extrapolation evidence. The methodological approach of the study demonstrates the potential to provide stronger extrapolation evidence if applied to broader contexts.

Aimed at identifying learners’ strengths and weaknesses on specific skills or contents, diagnostic assessment can provide fine-grained information to formatively promote teaching, learning, and language development in an ongoing language classroom (Alderson, et al., 2015, Elder, 2017; Jang, 2012; Knoch & Macqueen, 2017; Lee, 2015). While much research has developed diagnostic tools for large-scale standardised assessment, few have constructed diagnostic instruments for low-stakes formative classroom assessment. To contribute to the existing knowledge of diagnostic language assessment (e.g., Alderson et al., 2015; Jang, 2012; Knoch, 2007, 2009a, 2009b, 2011; Lee, 2015), this PhD research aimed to (1) develop a diagnostic rating scale for a formative diagnostic assessment to diagnose students’ strengths and weaknesses in academic writing products and support ongoing teaching and learning in an EFL university classroom, and (2) explore the validity of the assessment claims following an argument-based approach to validation (Chapelle et al., 2008, 2010; Kane, 1992, 2006, 2011, 2012, 2013, 2016a, 2016b; Knoch & Chapelle, 2018). To this end, this research employed a multistage exploratory sequential mixed-methods design (Creswell & Plano Clark, 2018) to undertake the scale development and validation over three study stages: scale construction, scale trialling, and scale implementation. Following the line of a multisource-driven approach to scale development (e.g., Banerjee et al., 2015; Knoch, 2007, 2009b; Montee & Malone, 2014), the scale was constructed and revised on the basis of theories of L2 writing ability, existing scales, expert intuition, and classroom curriculum. The scale was operationally implemented over the course of one semester in four writing classrooms, in which 80 English-major undergraduates used the scale to write, self-diagnose, and revise their assignment essays, and five teachers applied the scale to diagnose the students’ essays and use diagnostic results to support teaching and learning. The teachers and twenty students were interviewed regarding their perceptions of the scale and assessment. The diagnostic scores were analysed using Classical Test Theory, Many-Facets Rasch, correlation, regression, and ANOVA statistics, and the perception protocols were analysed following a qualitative content analysis. Overall, findings offered reasonable support for the overarching validity argument for the scale-driven assessment system. Yet, the different writing tasks to which the scale was applied over the course of instruction made it difficult to reliably gauge student progress, highlighting the need for stronger evidence relating to the consequence inference. This limits the usefulness of a measurement-driven assessment approach in detecting learning progression over the course. In addition, the current validation framework, driven by Kane’s argument-based approach, appeared not to well capture the dynamic and varying evidentiary sources of learning and writing development in the classroom assessment. The present study provides implications for developing a diagnostic rating scale for diagnostic purposes in a formative assessment, and examining the validity of the assessment within the context of EFL language classroom.

This thesis is a conversation analytic and poststructuralist study of discourses and social categories. In particular, it analyses how discursive categories of race, sexuality and gender operate in talk among a group of Chinese lesbians, and how the categories produced by these discourses are implemented in, and potentially reshaped by, interaction. The data which is presented and analysed in this thesis comes from a corpus of approximately 16 hours of audio recordings of conversations between and with Chinese lesbians who live in Melbourne, Australia. The thesis has two main aims. Firstly, it analyses how discourses operate in interaction. It does this by locating moments in interaction where discourses of race, sexuality and gender are oriented to. It shows how categories of race are resisted in ambiguous interactional projects; categories of sexuality are shown to operate and potentially alter in repair sequences; and categories of gender are shown to operate and be normalised in storytelling sequences. Secondly, the thesis aims to develop a critical conversation analysis methodology. In order to achieve this aim, the thesis builds on feminist and critically-oriented conversation analysis (CA) to develop and implement critical CA. This methodology finds points of compatibility between CA and poststructuralism. The implementation of this methodology also contributes to the long-standing debate between conversation analysts and critically-oriented discourse analysts about the compatibility between poststructuralist and conversation analytic epistemologies. The thesis concludes that critical CA can indeed be used to show the operation of discourses in mundane, everyday interaction. This can improve our understanding of how social categories are produced, sustained, resisted, and even potentially altered. Such findings may allow us to contribute an understanding of the operation of discourse to a political project of reducing discrimination based on social categories.

Understanding the roles of spoken and sign phonological code in reading processes is important for educators of deaf and hard-of-hearing students. However, the pool of data on this topic is limited and has mostly centered on readers of alphabetic languages. In places like Hong Kong, where deaf signers are relatively few, the shortage of research on phonological processing during reading is even more severe. This thesis addressed this problem by investigating the cognitive processes underlying Chinese reading in Hong Kong deaf readers using two methodological approaches, eye movements and event-related potentials, across four separate studies. Studies 1 and 2 used the error disruption paradigm with eye-tracking to investigate the patterns of orthographic and phonological activation in hearing and deaf readers. The hearing reader data suggested that they rely mostly on orthography to access word meanings in early processing. However, early phonological activation was found to be facilitated top-down by semantics when targets were predictable. The deaf readers were also found to rely primarily on orthographic information to access word meanings, but phonological code played a role in late processing and was modulated by contextual predictability and reading level. Study 3 used a parafoveal preview paradigm with eye-tracking to investigate how sign phonologically related previews affect reading processes in Hong Kong deaf readers. The pattern of results suggested that these readers activate sign phonological representations when reading Chinese words and that different sign phonological parameters (i.e., handshape, location, and movement) have different effects on parafoveal processing. Study 4 investigated orthographic, spoken phonological, and sign phonological processing in Hong Kong deaf readers using two error disruption paradigms with ERPs, focusing on the P200 and N400 components. The results for Experiment 1 revealed that N400 amplitudes were reduced in the orthographic condition, which suggested that orthographic representations were facilitating lexical access. In the homophonic condition, N400 amplitudes were increased in the right central eleci trodes. In Experiment 2, P200 amplitudes were significantly reduced in the left anterior electrodes in the sign phonological condition. In sum, the results of Experiment 1 and Experiment 2 suggest that the early P200 component is modulated by sign phonology, and the later N400 component is modulated by orthography and spoken phonology in Hong Kong deaf readers. In sum, these studies suggest that while deaf readers tend to activate word meanings directly through orthography, they can also activate spoken phonological and sign phonological codes. Consistently across the eye movement and ERP studies, the effects of sign phonological activation emerged in early stages of processing, and the effects of spoken phonology emerged in later stages of processing. The different time courses of spoken and sign phonological activation may be an indication that deaf readers tend to use sign representations to activate word meanings and spoken phonological representations for later integrative processes. In conclusion, these findings can be taken to suggest that both types of phonological code are important for deaf readers.

Numerous studies have shown that native phonotactic constraints influence listeners’ perception of non-native speech. These studies typically show that speech which violates the phonotactic rules of the listener’s native language is often misperceived as speech that adheres to native phonotactics, behaviour which is often characterised as phonotactically conditioned perceptual repair. The present thesis extends these findings by examining the effects of language-specific listener expectation on non-native speech perception. The central hypothesis presented within is that sequences of speech that are unexpected by the listener are often misperceived as sequences that are expected. This suggests that phonotactically conditioned perceptual repair is not necessarily caused by an innate understanding of the listener’s native phonotactic constraints: sequences of speech that contain phonotactic violations may be considered as one extreme of a continuum of expectedness. Additionally, the present thesis explores the relationship between predictability and attention. The studies presented show that listeners pay less attention to incoming speech input when they can make predictions regarding upcoming input based upon preceding materials. The present thesis tests these hypotheses in six studies conducted on native Japanese speakers. These experiments use allophonic variability, transitional probability, and word frequency as measures of expectation. The results confirm the central hypothesis, they show that listeners misperceive sequences of speech that are improbable in Japanese. The way that listeners misperceive speech is also found to be dependent on expectation, when presented with unexpected speech that may be repaired in multiple ways, listeners experience the more probable illusion.

The 2014 Nobel Prize in Literature was awarded to the French writer Patrick Modiano (b. 1945) for work that ‘uncovered the life-world of the Occupation’. The jury recognised the author’s significant role in addressing the heavy silences surrounding this uneasy period in French history. Modiano’s novels, however, are full of silences, the unsaid, and the unsayable. While his work is an indictment of certain sociohistorical silences, it is also a literature which performs, and is permeated by, a multitude of silences. This thesis seeks to elucidate the nature of this tension through an analysis of silence across twenty-seven of Modiano’s texts, from his first novel, La place de l’etoile (1968), to Souvenirs dormants (2017). A defining aspect of silence is its fluctuating semantic value and the plurality of its applications, from language and noise, to art and ethics. Its meaning is not fixed; it is, rather, an amorphous concept and its denotation and connotations shift depending on the context in which it occurs. The way in which silence signifies also changes over time; what might be silenced in 1945 could become sayable a generation later. Given the intrinsic semantic plasticity of the concept, this thesis develops an evolutive analysis of silence, which exists in multiple guises and occupies multiple spaces in Modiano’s works, from the awkward silences to those that are more eloquent, to the salutary, recuperative and respectful silences, to the silences that condemn and denounce, and the silences that gesture to the very limits of language, thought and logic. This thesis is arranged in three parts, each one based on a particular interpretation of silence. Firstly, reading silence as a generative force which, paradoxically, can be productive of narrative, the multiform silent beginnings and origins of Modiano’s texts are investigated. Secondly, silence is considered as a place of forgetting, loss and erasure, which is opposed by Modiano’s texts. This shifts the discussion to the ethical tensions associated with biographical writing and testimony in his works. Finally, silence is interpreted as something that can be created, both through the text and through other artistic media. Here, silence is positioned as a destination in its own right and the trajectories made towards and around silence by Modiano’s characters in their various fields of expression are explored. Through these differing interpretations of silence, this thesis examines how Modiano’s writing comes from and engages with silences, resists meaning and evades conclusion, creating a kind of literary silence unique to his work.

Calls to conduct research on evaluation have been made to learn about its practical realities, in order to ensure a quality process and to improve future evaluation experiences. A key methodology addressing this need is metaevaluation, the evaluation of an evaluation. Studying seven Australian university case studies of reviews of language programs, metaevaluation is used as both the conceptual framework as well as a methodological tool to explore and assess the evaluation practices and to discover practice patterns, using the findings to make recommendations about how such evaluations should be conducted. The research in this thesis demonstrates that New Public Management (NPM) which operates as a common organizational and policy framework in Australian Higher Education and has been defined as a national and local control system which “models higher education simply in economic terms” (Marginson 2010, p. 6) is the most influential evaluation factor in reviews. In this policy context, a gap is found between the tenets of evaluation theory and the character of Australian Language Program Evaluations (aligned with NPM ideology). As a consequence, most initiators or sponsors of evaluations are not focused so much on the quality of the process but on using reviews as a top-down instrument to implement a cost-efficiency agenda. As the analysis of data progressed, it became clear that the majority of cases reveal a process of reviewing empty of any theoretical foundations based in the evaluation discipline, thus putting in jeopardy the quality of the evaluation process, which resembled more a sponsors’ top-down exercise of checking or imposing compliance with University priorities. In light of these broad findings, this thesis claims that metaevaluation has a key role in raising the quality of the practice of reviewing by ensuring reviews are informed by evaluation theory and practice, thus following principles of good practice, and that it can also help mitigate a purely cost-driven process.

Mangarla is a Pama-Nyungan language of the Marrngu subgroup, originally spoken in the Great Sandy Desert of Western Australia. Today, the language is severely endangered with a small number of speakers living in disparate communities outside of traditional lands. This work describes the core grammatical features of Mangarla and examines its linguistic connections to other languages in the region, both related and typologically unrelated, providing insight into the fluidity of individual language varieties in contact. The analysis is based on notes and audio recordings of narratives, elicited data and spontaneous conversations, recorded between 1990 and 1994, in the Kimberley communities of Bidyadanga (formerly La Grange) and Jarlmadangah Burru (formerly Mt. Anderson Station), Derby and Fitzroy Crossing. It is also informed by the work of Kevin McKelson collected between the 1960s and 1980s. All aspects of the language are impacted by its distinctive location, surrounded by three different Nyungic subgroups and the prefixing non-Pama-Nyungan languages in the north. Mangarla’s phonological inventory and lexical classes are similar to those of related suffixing languages, but unlike those further south, Mangarla’s lexicon includes many consonant-final preverbs and particles, which are often monosyllabic. Morphologically, it is a split-ergative system, with nominal arguments (including free pronouns) marked in an ergative-absolutive system while the pronominal clitics in agreement with them are split along nominative-accusative lines. Unlike other members of the Marrngu subgroup where these clitics attach to the verb, Mangarla’s bound pronouns typically encliticise to the first element of the clause or to an optional post-initial catalyst, although pragmatics also impacts on their placement. Other interesting features include the loss of the widespread dative marker and lateral-initial ergative/locative forms, its movement away from complete case-concord to free marking within the NP, and the reduction of verb conjugation classes to three. A large number of complex predicates, formed with a preverbal element and a relatively small number of inflecting verb roots, augment monomorphemic members of the verbal category. Argument structure also displays a degree of flexibility not generally recognised in Australian languages. Clause combining strategies include coordination of full and reduced clauses by parataxis, subordination of both finite and nonfinite clauses, typically employing a small number of case markers as complementisers, and unusually, clause chain cosubordination. The work adds to the knowledge of Pama-Nyungan languages in this remote region and leaves a detailed record of the language for the future use of both the Mangarla and academic communities.

This dissertation presents an in-depth analysis of size and shape specifiers in Russian Sign Language (RSL). Size and shape specifiers, or SASSes, are signs widely used in sign languages of the world to describe visual characteristics of objects. Highly iconic, they have been noted to possess a range of peculiar features, such as combining categorical and gradual properties, composing the meaning of the whole out of the meanings of parts, and denoting different characteristics of the object simultaneously (location, orientation, length, width, overall shape, and some others). Despite their extensive use, SASSes received surprisingly little attention in the literature. This dissertation fills this gap in the existing body of knowledge by conducting a thorough analysis of SASSes in RSL. It (1) defines the properties of SASSes that set them aside as a separate group of signs; (2) determines the status of SASSes in the sign language lexicon; (3) compares the way SASSes describe objects with the way(s) spoken languages do. The method of the study entails collecting a corpus of video-recordings with semi-spontaneous signing, which is followed by a detailed phonetic transcription of SASSes occurring in the corpus (overall, 625 tokens) in ELAN. A subsequent feature by feature analysis of structural elements in these signs (namely, handshape, place of articulation, movement, and mouth actions), allows us to establish SASSes as a distinct group with its unique set of properties. This is pioneering research that sets the grounds for the future typological analysis of these signs cross different sign languages.