ABSTRACT. In sign languages, mouth actions regularly accompany manual movements and play various linguistic functions. For example, mouth actions can function like adverbs and adjectives, disambiguate manual homonyms, and constitute an obligatory part of lexical items.

One particular type of mouth action is mouthings, which take the form of articulation of all or part of a related word in an ambient spoken language. By contrast, mouth gestures are not related to spoken language and emerge from within a sign language itself (Boyes-Braem & Sutton-Spence 2001).

The growing availability of data from diverse sign languages has enabled cross-linguistic comparison and the emergence of the new field of sign language typology (de Vos & Zeshan 2012:8-9). A study by Crasborn and colleagues (2008) indicated that mouthings constitute the majority of mouth actions in British Sign Language, Swedish Sign Language, and Sign Language of the Netherlands. A later study found this pattern to be consistent for Australian Sign Language (Johnston et al. 2016). Yet, both these studies have focused on largely urban deaf community sign languages, with Crasborn and colleagues themselves (2008:52-53) noting the need to compare their results to sign languages with less similar sociolinguistic contexts.

Kailge Sign Language is an undocumented language used in and around the Kailge community in the Western Highlands province of Papua New Guinea. Initial fieldwork suggests that it is a shared sign language; that is, one used extensively by both deaf and hearing community members (Nyst 2012).

A preliminary analysis of elicitated material from four unrelated Kailge Sign Language signers (two deaf, two hearing; 4M) has suggested that mouth gestures are the most frequently and consistently occurring type of mouth action in Kailge Sign Language, not mouthings. Table 1 indicates some mouth gestural sign components in Kailge Sign Language.

The data suggest consistency in manual signs across all signers. However, in terms of signs which have an accompanying mouth gesture, the data suggest individual signer consistency but cross-signer variation. The two deaf signers in the data are from different clans, which suggests possible clan-related variation.

Coincidentally, other researchers have recently begun studying another undocumented Papuan sign language, Sinasina Sign Language in nearby Chimbu province (Rarrick & Asonye 2017). Their data thus far suggest some shared manual signs with Kailge Sign Language but again, variation in mouth actions (Samantha Rarrick, p.c., 21 July 2017).

Indigenous shared sign languages are particularly fragile and susceptible to endangerment (Nonaka 2004). Their study enriches not only sign language typology but our understanding of the human language capacity in general. Furthermore, the study of mouth gestures is more broadly relevant as it has been suggested that they provide clues as to how human communication may have leapt from iconicity to abstraction (Woll 2014). Hence, this research has benefit not only for sign language linguistics, but for the study of human language evolution.

ABSTRACT. uploaded

ABSTRACT. Processing foreign-accented speech is an essential skill in multicultural contexts. However, the effect of foreign-accentedness on how listeners understand the meaning of utterances (i.e. semantic processing) is largely unknown. The few relevant studies have tested this by examining brain responses to a semantic-violation paradigm using event-related potentials (ERPs). Specifically, they compared neural responses between semantically-correct and -incorrect sentences (e.g. The lady bites the cupcake/*suitcase), and between native and foreign-accented speech [1, 2, 3]. Compared to semantically-correct sentences, incorrect sentences elicited an N400-effect (a negativity that peaks approximately 400 ms post-stimulus with a centro-parietal maximum, typically associated with semantic processing difficulty). This N400-effect occurred for all semantic violations, but with differing amplitude and/or topography across speakers. However, these studies were all conducted in cities with small migrant populations. This suggests that the listeners were relatively unfamiliar with foreign-accented speech, which could influence their neural responses. We therefore aimed to replicate these findings in a multicultural context (Sydney, Australia). As previous findings have been inconsistent, we did not make strong predictions about the characteristics of the N400s we expected. Additionally, N400-effects elicited by native speech are subject to extensive individual variability which has been linked to listener characteristics such as age, working memory and empathy [4, 5, 6]. Such variability has not been investigated in the processing of foreign-accented speech. Our second aim was therefore to investigate potential sources of individual variability in any ERP effects observed.

Data were collected from 30 monolingual Australian-English-speaking adults as they heard semantically-correct and -incorrect sentences produced in native Australian and Mandarin-accented English. Data were analysed using cluster-based permutation tests. Listeners’ age, gender, working memory, empathy, evaluation of the foreign-accentedness of the stimuli and accuracy at identifying the speaker’s accent were included in a linear mixed-effects model of effect amplitude to examine individual variability. An N400-effect was present for both native (p < .001) and foreign-accented speech (p = .03; Figure 1), but there were no significant between-speaker differences in amplitude, topography, latency or duration (all ps ≥ .22). Consistent with previous findings [5], listeners with greater working memory had greater overall N400 amplitude (β = -0.13, SE = 0.05, p = .02). Listeners who were unable to correctly identify the accent used (a proxy for accent unfamiliarity) also showed a difference in N400 amplitude between the two speakers, but this did not occur in those who could identify the accent (β = 0.58, SE = 0.19, p = .006).

In contrast with previous studies, these results showed no differences in the N400-effect between native and foreign-accented speech. This may be due to the more conservative statistical method used, but also the familiarity of the listeners with Mandarin-accented speech. The majority of listeners correctly identified the speaker’s accent (i.e. more familiar with accented speech) and showed no difference in N400s. In contrast, those who were incorrect (i.e. less familiar) showed amplitude differences like in [1]. This suggests that unfamiliarity with accented speech modulates the differences in semantic processing between native and foreign-accented speech.

ABSTRACT. Bilingual speech perception research has found evidence of first language (L1) biases in discriminating and categorizing second language (L2) phones, as predicted by the Perceptual Assimilation Model (PAM: Best, 1994, 1995; Best and Tyler, 2007) and other research perspectives (Sebastián-Gallés and Soto-Faraco, 1999; Navarra, Sebastián-Gallés and Soto-Faraco, 2005). These studies have involved bilinguals whose two languages are independent of each other. However, this study examined Jamaican Creole (JC), an English-lexified contact language that developed in Jamaica, and Jamaican English (JE), which present overlapping phonological systems with just a few critical phonological differences. We examined Jamaican how Creole dominant pre-primary (5-6 years) and primary school children (aged 8-9 years) acquiring JE as an L2 perceptually distinguish JE-only /θ/ and /h/ phonemic contrasts. In an AX discrimination task, children judged nonsense word pairs involving L2-JE phonemic contrasts that do not exist in L1-JC: /θ/ in contrast with /f/ and /t/, and the presence versus absence of /h/. In adult JC speech, /θ/ is replaced with /t/, and /h/ is deleted. We also included baseline contrasts that exist in both JC and JE, which should be easily discriminated. Children had to indicate whether the two “words” in each pair (each played 3 times) were the same or different. The analyses were conducted on d’ transformations of the raw data, with higher d’ values indicating greater the ability to detect the distinction. The analysis of variance (ANOVA) compared contrast type as a within-subject factor (experimental, baseline contrasts) and age (pre-primary, primary) as a between-subject factor. A significant main effect of contrast type indicated that baseline contrasts were more easily discriminated than experimental contrasts, F(5, 62) = 408.884, p = 0.031). The results showed that while the children had difficulty discriminating the JE phonemic contrasts overall, they had greater difficulty discriminating /h/-/∅/, followed by /θ/-/f/, and the least difficulty discriminating /θ/-/t/, as evidenced by the mean d’ scores (respectively, M = 0.16, M = 0.53, M = 1.85). A contrast*group interaction was observed for /θ/-/t/ versus /θ/-/f/ contrasts. The primary school children discriminated the /θ/-/f/ contrast better than the pre-primary children, while the opposite was true for the /θ/-/t/ contrast (see Table 1). Overall, despite formal exposure to JE in school, the children did not discriminate JE phonemic contrasts that do not exist in JC. The fact that the children perceived both members of the /θ/-/f/ and /h/-/∅/ contrasts as the same, confirms PAM and other research perspectives predictions regarding L1 biases in discriminating L2 phones. These JC-L1 children were expected to assimilate JE-L2 /θ/ to /t/ and /h/ to /∅/ because the L1 makes these correlations. However, the older children appear to have assimilated /θ/ to /t/ more strongly than the younger children, who did not have any difficulty discriminating /θ/-/t/ like the older children did. Instead the younger children showed stronger assimilation of /θ/ to /f/. Thus, the older children performed similarly to adult JC speakers for /θ/-/t/, but the younger children had not yet made the correlation seen in JC adult speech.

ABSTRACT. Studies have found that /s/ production in /stɹ/ onset clusters in words like street and straw has undergone retraction in many varieties of English (Rutter, 2011; Gylfadottir, 2015; Wilbanks, 2016). This change results in a production closer to [ʃtɹ]. A recent study by Stevens & Harrington (2016) could not find evidence of this sound change in the production of Australian English (AusE)-speaking participants over the age of 29. The question remains as to whether /stɹ/ retraction might be found in younger AusE speakers. Younger speakers are typically considered leaders in sound change.

This study therefore investigated the possibility of sound change in AusE in /stɹ/ onsets by comparing equivalent content from two data sources collected at different points in time. Selected sentences from the ANDOSL dataset collected in 1993/4 (Vonwiller et al., 1995) were identified as containing /stɹ/ clusters, as well as singleton onset /s/ and /ʃ/ in comparable vocalic environments. Audio data were extracted from ANDOSL for young (<30 yo) speakers (f=16, m=18) and middle age (31-45 yo) speakers (f=16, m=18). The same sentences were elicited in 2017 from a group of speakers under 30 years (f=22, m=15). This design provides a ‘real time’ comparison between two generations of speakers, approximately 25 years apart, as produced by ‘young’ speakers in each dataset (ANDOSL and present-day data). It also allows an 'apparent time' study comparing speakers of different age groups (i.e. two ANDOSL age groups).

Selected sentences included productions of: a) ‘see’ and ‘saw’, to establish a canonical /s/ for each speaker in unrounded (/iː/) and rounded (/oː/) vowel contexts; b) ‘she’ and ‘sure’, to establish a corresponding canonical /ʃ/; and c) ‘street’ and ‘straw(berries)', to exemplify the /stɹ/ cluster. The sampling frequencies of the audio files were aligned using SOX Resample software and the following measurements were extracted using Praat after manual annotation: 1) Centre of Gravity (COG) representing the acoustic ‘trajectory’ of the sound, based on 20 windows of 5msec each, across the consonant; and 2) a single ‘core’ COG value based on a set of windows, each 5msec with 5msec frame-shift, over the central 50% of the fricative.

The core COG values were used to calculate a ‘retraction ratio’ for each speaker, based on the scale of their canonical /s/ and /ʃ/ (raw data in Fig1). A mixed effects regression model with vowel context, speaker gender, and cohort as predictors shows that present day young speakers were significantly more retracted in /stɹ/ production than their peers born between 1960-79 (p<.05). An interaction between vowel context and cohort (p<.01) also shows that these peers were already more retracted in the /stɹoː/ context than older speakers born between 1940-59 (model predictions in Fig2).

By showing a significant difference between the retraction ratio across the three ‘generations’ of speakers, these results fail to find support for Stevens & Harrington, suggesting that sound change has already taken place in Australian English in respect of s-retraction, starting with /stɹ/ in the /oː/ context and now extended to the /iː/ context.

ABSTRACT. Colloquialization, defined as the drift of formal written registers towards a more informal spoken style, has been identified as an important factor in short-term language change (Collins 2013; Leech et al. 2009). At the same time, some studies have observed a contrasting trend of densification or increased economy, where information is condensed into fewer words (Leech et al. 2009). This is a characteristic of written language unlike the more diffuse expression expected in spoken styles, and is therefore an anti-colloquializing trend.

There is limited research into the dynamic between these partly competing patterns of language change (see Biber & Gray 2012). A recent study by Kruger and Smith (2018, in press) addresses this gap by applying the multidimensional method of Biber (1988) to the Australian Diachronic Hansard Corpus (1901-2015). In this paper, we extend the same method to a comparative analysis of the Australian and British Hansard over the period 1901-2015, motivated by research suggesting that the drift towards colloquialism is more pronounced in Australian English than British English (Collins 2013). We present a refined model of the five factors that Kruger and Smith (2018, in press) identify as potentially affecting the interplay between colloquialization and densification in the Hansard: the social context, editorial policy, production mechanisms, audience construal, and the aims of the register. Against this background, Biber’s (1988) analysis was replicated for the corpus. Regression analysis of the dimension scores (with Variety and Period as predictors) and interaction plots to visualize the normalized scores of individual features were used to compare trends of colloquialization and densification in the two varieties over time.

Shared as well as divergent colloquialization and anti-colloquialization effects are evident. We find strong evidence of densification in both varieties. While there is also a shared colloquialization trend, the Australian data are characterized in particular by colloquialization that may be ascribed to more permissive norms for reflecting spoken-language or informal features in formal writing. Combined with this is another trend (again more strongly evidenced in the Australian Hansard) which we argue relates to an anti-colloquial change in rhetorical strategies, reflecting less emphasis on overt interactive debate and a greater emphasis on information presentation.

References Biber D & B Gray 2012 ‘The competing demands of popularization vs. economy: written language in the age of mass literacy’ in T Nevalainen & EC Traugott (eds) The Oxford Handbook of the History of English Oxford: Oxford University Press. pp. 314–328. Biber D 1988 Variation Across Speech and Writing Cambridge: Cambridge University Press. Collins P 2013 ‘Grammatical variation in English worldwide: the role of colloquialization’ Linguistics and the Human Sciences 8(3): 289–306. Kruger H & A Smith 2018 (in press) ‘Colloquialization and densification in Australian English: A multidimensional analysis of the Australian Diachronic Hansard Corpus’ Australian Journal of Linguistics 38(2). Leech G, M Hundt, C Mair & G Smith 2009 Change in Contemporary English: a grammatical study Cambridge: Cambridge University Press.

ABSTRACT. In this introduction we will present the results of a preliminary typological research on body parts in emotion expressions across the Australian continent. Across the ≈40 languages surveyed, 25 body parts occur in figurative emotional expressions. The most common body part is (by quite far) the belly, followed by the head and the heart, as well as the eyes and other facial body parts. We will discuss the distribution of the figurative expressions we found, the emotions each body part tends to relate to, and the specific metaphors involved. We will also introduce some methodological principles relative to metaphors and their cognitive significance.

ABSTRACT. This talk focuses on a description of expressions of emotion in MalakMalak. Bodily comfort or discomfort is associated with happy or sad emotions respectively as seen in (1) and (2).

(1)      menwunetjed

men-wu-net-tjed

stomach-3sg.n-bad.-stand

noun pronoun-adjective coverb

`be unhappy, be sad, lit. have a bad stomach'

(2)      menwunypainwai

men-wuny-pain-wai

stomach-without.sense-good-take.a.long.time

noun-coverb-adjective-coverb

`be happily surprised, lit. have an uneventfully good stomach for a long time'

Additionally, expressions of emotion are structurally unusual in MalakMalak in forming compounds of up to three different parts of speech.

ABSTRACT. Previous research found that the two main ethnolects of New Zealand English differ with regard to pitch and voice quality settings. Māori English mean pitch has been shown to be significantly higher than standard Pākehā English (Szakay 2006), while at the same time also exhibiting more creaky phonation as suggested by lower H1-H2 values (Szakay 2012). The question whether these Māori English suprasegmental features are historically transfer effects from the indigenous Māori language, or are alternatively innovative ethnolectal features remains to be answered. It has been previously suggested that the syllable-timed rhythm of Māori English might be a result of contact with the Māori language (Holmes 2005), however, to our knowledge there have been no systematic studies to date on Māori language voice quality and pitch.

As a first step we examine potential cross-linguistic suprasegmental transfer effects within the bilingual individual. Our data comes from Gruber et al. (2016) where four young male ethnically Māori bilingual speakers were recorded in three different language settings: English with a Pākehā interviewer, English with a Māori interviewer, and Māori language with the same Māori interviewer. In addition, six monolingual young male Pākehā English speakers were also recorded in the two English language settings. Each interview lasted 30–60 minutes and consisted mainly of directed conversation. The data for our analysis is taken from one part of the interview where participants retold the events of a Tweedy cartoon they had just watched.

With this experimental set-up we are able to examine (i) the differences in voice quality and pitch in Māori English and the Māori language within a bilingual individual; (ii) the effect of the interviewer’s ethnicity on pitch and voice quality for both Māori English speakers and Pākehā English speakers; and (iii) whether Māori English suprasegmentals are more similar to Pākehā English or the Māori language across individuals. For the analysis F0 and H1-H2 measurements were taken at the mid-point of each vowel in Praat. We investigate mean pitch, pitch range, standard deviation of pitch, as well as creakiness and breathiness for each of the ten participants.

Studies on fundamental frequency generally agree that F0 is language specific (e.g. Loveday 1981), however, only a handful of studies have examined F0 in two languages spoken by the same speaker, and the results have been inconsistent (e.g. Goldman 2016, Voigt et al. 2016). No research has previously investigated bilingual voice quality transfer effects. Notably, the present study looks at prosody in a Polynesian language, in a field dominated by research on Indo-European languages. Finally, this study contributes not only to the literature on the suprasegmental properties of New Zealand’s language varieties, but also to the broader research agenda investigating the extent to which prosody is shaped by cultural and social factors.

ABSTRACT. Language policy is, in many cases, developed by governments to regulate and standardise languages used by its citizens. In the case of Belgium, policy has been developed with the intent of mitigating tensions between the Dutch and French communities, but in turn may have bolstered the linguistic divide between the monolingual regions of Flanders and Wallonia (Blommaert, 2011; Deschouwer, 2005; Hooghe, 2012). As the sole bilingual region of Belgium, Brussels is a unique context in which both language communities come into contact. Previous studies have shown that language tensions are tangible in Brussels, but also acknowledge the city’s vital role both as a capital and a shared linguistic space (Mettewie & Janssens, 2007; O'Donnell & Toebosch, 2008; Treffers-Daller, 2002). Despite its pivotal role in Belgium’s national framework, a current and comprehensive analysis which considers government policy alongside the attitudes and practices of residents has yet to be given. This study aims to determine the role that Brussels plays within the language policy framework of Belgium: a battleground for linguistic communities, or a point of unity despite outside divisions. This study uses Spolsky’s model for language policy as a framework for the Brussels context. This model (Spolsky, 2004) considers the overt government policies alongside community practices and individual attitudes towards language use, giving a comprehensive view of any given country or region. The major data source for this study is the Corpus of French as Spoken in Brussels (Dister & Labeau, 2016), which provides extensive interview data from francophone Brussels residents. The methodology of Qualitative Content Analysis (Haukås, 2016) is used to bring the attitudes expressed by respondents to come to light. The analysis of this corpus data, alongside previous sociolinguistic research and the analysis of current government policy, paints a compelling picture of Brussels as a point of strength for the Belgian union. There is no evidence of explosive tensions in the attitudes and practices of Brussels residents; francophone parents nevertheless entertain the notion of sending their children to Dutch-medium schools, and view competency in Dutch as a very real advantage for their children’s futures in the city of Brussels. There are reports of marriages between language groups, and children growing up in a Dutch-French bilingual environment. In short, the language policy, attitudes, and practices within Brussels seem to be aiding the linguistic tensions of Belgium rather than worsening them. In an age in which supranational structures such as the European Union are increasingly considered endangered, and in some cases abandoned, it is necessary to understand linguistic identity as it intersects with regional and national identity, and how this is negotiated by policy. This study demonstrates a successful case, in which Brussels’ bilingualism allows a flexibility for Belgium’s various linguistic communities, and allows a space in which they can exist in parallel, diffusing the tensions between them.

ABSTRACT. In the latter part of the 19th century, Frege put forth the idea that no theory of meaning in languages complete unless it accounts for the connection between truth and the meaning of the words andlarger linguistic units, and their reference to things in the world. Since the days of Aristotle,logicians and philosophers had appealed to logic to explain the truth conditions of propositions innatural language; truth in semantics is defined according to the truth in a proposition. Building on Frege and Aristotelian logic, 20th century model theoretic semanticists developed a system ofexplaining meaning by applying techniques from mathematical logic to the semantics of natural language Dowty, et al (1981). Kamp & Reyle (1993) further developed this in DiscourseRepresentation Theory (DRT). This paper Investigates meaning and its connections to referents in the real world of IL discourse by using notions from DRT and model theoretic semantics. A task-based corpus of twenty-four hours of nonnative to nonnative learners (NNS-NNS) of English IL speech was analysed using this empirical data rather than pairs from semantic modeling. Utterances with non-referring referencewithin the discourse and between the real world of the four tasks were disclosed. Constraining the context using the information in the IL discourse, the tasks, and the set of presuppositions within the discourse and real world of the tasks, the truth of the utterances and their meaning is investigated following Frege. This reveals patterns of truth values in IL speech which can be used for the foundation of the systematic study of presupposition in the study of the development of meaning for second language learners. Specifically, types of presuppositions are isolated in the data based on their relation to the context of the discourse and information in the tasks. This requires information that goes beyond the presuppositions embedded in the lexical and sentential properties.

References Dowty D, R Wall & S Peters 1981 Introduction to Montague semantics v. 2. Dordrecht: Holland: Reidel Publishing Company. Frege G 1892 ‘Uber Sinn und Bedeutung’ Zetschrift fur Philosophie und philosophische Kritik 100: 25-50. Kamp H & Reyle U 1993 From Discourse to Logic Vol 2. Dordrecht: Kluwer Academic Publishers.

ABSTRACT. Zhangzhou is tonally interesting. It makes good use of eight tones to distinguish lexical meanings. The tonal realsations are observed multi-dimensional, involving a variety of segmental and suprasegmental parameters that include pitch, duration, vowel quality, voice quality and syllable coda types. Nevertheless, the realisations can be modified in utterances beyond monosyllables, reflecting significant trans-disciplinary interactions among phonetics, phonology and morph-syntax. For example, before the diminutive/nominative morpheme /ʔɐ51/, the number of pitch contrasts is sharply reduced to only two as a consequence of two different processes of neutralisation. The falling pitch contours are all neuralised to a high level; while the non-falling pitch contours are all neutralised to a rising contour. Therefore, in this morphological environment, the pitch of the eight tones is either a high level [55] or a rise [35] depending on its pitch contour shape in citation. This finding expands the understanding of tone sandhi in Zhangzhou, but also contributes to the knowledge of trans-disciplinary interactions in tonal languages in general.

ABSTRACT. A general assumption in the adult second language (L2) literature is that the younger the age of acquisition (AoA) and/or greater length of exposure (LoE), the better the chances of later grammatical morpheme acquisition [1]. However, findings from recent studies of younger L2-learners suggest that non-AoA factors, such as vocabulary size, short-term memory, quantity and quality of the input as well as the type of first language (L1), better predict grammatical outcomes [2,3,4]. In general, there is plenty of literature describing how different aspects of L1 grammatical structure modulate the L1 transfer effects, which in turn correlate with L2 outcomes. However, very little is known about what modulates other non-AoA factors, for example, L2 vocabulary size. Understanding these factors will inform L2 acquisition theories and have implications for L2 teaching and language assessment programmes for school-age children [5]. The aim of the current study, therefore, was to explore what factors correlate with vocabulary size in Mandarin L2-English-speaking children with varying AoA and LoE to English.

Participants were 20 young Mandarin L2-speakers of English (12 girls, mean age 10; 3, range 9-11;6, mean AoA 3;3, range 1;3-8;0) from high SES suburbs. Their vocabulary knowledge was measured using the Peabody Picture Vocabulary Test (PPVT) where, out of an array of four pictures, they were asked to point to one which best matched the word said by the experimenter. Data were also collected to measure children’s L2 grammatical performance (elicited production of present & past tense inflection), short-term verbal memory (the visual forward and backward Digit Span) and non-verbal intelligence (the Test of Non-verbal Intelligence (TONI)). Parental reports were used to collect data on language-related activities that the children engaged in, such as the range of English print material that children had access to reading and the frequency of communication in English with parents, siblings or other caregivers. Raw scores for the PPVT, TONI, and Digit span were converted to standard scores to normalize for age. Children’s tense inflection productions were scored as number correct. Parental reports on children’s frequency of communication were ranked 1-5 while access to print media coded and scored as a percentage.

Overall, children’s mean vocabulary scores were above the PPVT standard score for monolingual (i.e. 100). To explore which factors predicted vocabulary size, data were analysed using a backward linear regression model, in SPSS, with vocabulary score as the dependent variable and LoE, short-term memory, non-verbal intelligence, tense scores, English communication: parent-child, child-child, caregiver-child and access to print media as predictors. The analysis revealed that children’s vocabulary size was significantly explained by the amount of print media that children read (β = .55, t(19) = 4.93, p < .01) and their grammatical knowledge (β = .43, t(19) = 3.51, p < .01) (see scatterplots in Figure 1). This suggests that extensive L2 reading may override AoA and LoE effects in predicting L2 vocabulary size. These findings highlight the importance of considering non-AoA factors in explaining the different L2-grammatical learning outcomes.

ABSTRACT. See attached PDF.

ABSTRACT. Attached PDF.

ABSTRACT. Associated motion is a grammatical category encoding by affixes or other grammatical elements whose main function is to specify various kinds of motion associated with activity of the verb. (Koch 1984: 26), a phenomenon is firstly discovered in Australian languages (Koch 1984, Wilkins 1991) and also found in large number of Amazonian languages (Guillaume 2006, 2016)and some other languages families. Only two languages in Tibeto-Burman (Jacques 2013, Konnerth 2014: 232, 2015), but little attention is paid to Sinitic languages ( Lamarre 2016, Lamarre et al 2017). The present paper provides a preliminary investigation of associated motion system of Baoding dialect and it shows that despite of the simple system of AM in Baoding with only ventive (lɛ) and itive (tɕhi) markers, Baoding differs from most of the previous description of AM system in the way of that the motion is determined by the situation type and the TAM values encoded by the predicates. For instance, with the same itive marker tɕhi, meaning ‘to go’, on account of VP’s type, the interpretation of ‘VP+tɕhi’ could be either Go&do (motion-CUM-purpose) as in (1) or Go&do&return as in (2) or even Go&do&maintain the state as in (3). (1)mɛ213tshɛ51-45 tɕhi ! (2) tʂhɤ̃22 tɕhi ! buy.food go fill.bowl.with go ‘go and buy food!’ ‘go and fill the bowl (with rice) and come back!’ (3) thɑ̃213-tʂo tɕhi ! lay.down-STA go ‘go and lay down (and keep laying)!’ The motion is always deictically anchored. However, ‘VP+tɕhi’ cannot be considered as deictic directionals in which the directional complement is expressed by a bimorphemic component in Baoding: (4) na22 tɕhi! (5) na22 kuɤ51tɕhi! take go take over.go ‘go and take it and bring it back !’ ‘take it away!’ Traditionally ‘VP+ tɕhi or lɛ’ is considered as a serial verb construction in Sinitic languages, however, in Baoding, tɕhi and lɛ lose the ability to receive a tone. As in many Northern Mandarin dialects, in Baoding, a tone sandhi occurs in the syllable preceding an unstressed syllable. Whereas in serial verb construction in Baoding, tɕhi51 keeps its original tone value. The unstressed tɕhi and lɛ trigger the sandhi of previous syllable. tɕhi and lɛ are grammaticalied. By apply Guillaume’s parameters, we find that in Baoding, the grammatical fuctction of the moving arguement is the subject of the VP; in ‘VP+ tɕhi’ motion is subsequent to the main event, in ‘VP+ lɛ’, motion is prior to the main event and the motion is deictically anchored. This paper deals with the system of Baoding dialect and tries to challenge the traditional serial verb construction in Sinitics by showing that the prosodic feature of ‘VP+ tɕhi or lɛ’ and its semantic make-up. It focuses on the predicates that may enter this pattern, the semantic value of ‘VP+ tɕhi or lɛ’ and its discourse feature, in particular the interaction with the situation types and the TAM values. This paper provides a basic framework for further research on the typology of simple AM systems.

ABSTRACT. Cross-linguistically, animate beings are more likely than inanimate objects to fill agentive thematic roles (Comrie 1989; Klaiman 1991; Yamamoto 1999; Becker 2014). However, it is unclear whether thematic role selection is sensitive only to a referent’s animacy at the reference time, or whether the continued existence and animacy of the referent also affect role selection. For example, are descriptions of a deceased person’s life affected by the fact that the individual is no longer present?

The current study compared 181 clauses written by a single author, describing well-known Australian painter Margaret Olley, before and after her death in 2011. To keep the study as theory-neutral as possible, only the relative agency of roles in the clauses were considered. That is, roles were annotated according to whether they were the most agentive in the clause, but were not labelled as themes, or patients, etc. Clauses with only one participant and with symmetrical predicates (Dowty 1991) were not considered.

The study found that Olley was mentioned equally often in the earlier and later texts, and was equally likely to be a grammatical subject. However, the later descriptions assigned Olley the most agentive role in a clause significantly less often. Although Olley was of course an animate being throughout her existence, her passing appears to have decreased her agency in texts written after this time.

More unexpectedly, the study also found that Margaret Olley’s artworks were assigned more agentive roles following the artist’s death. For example, a description of Olley’s painting process in a pre-2011 text represents her as the agent, as in (1), whereas a post-2011 text describes the process almost as if the artwork is forming without Olley’s intervention, as in (2).

(1) She has been resurrecting a few early compositions …

(2) One by one, her chalk lines disappeared beneath a vigorous underpainting of acrylic, followed by the oil glazes ...

In sum, when artists die, not only do they appear to be assigned fewer agentive roles, but their artworks seem to take over the agency that the artists have lost. This result has two implications. First, it suggests that thematic role assignment may be affected not only by the animacy of a referent, but by the referent’s status at the time of speaking or writing. Second, it supports art historians’ arguments that artworks have a “life” of their own, independent of the artists who created them (Mitchell 1994). In fact, it would seem from our study that artworks do not just have an independent life, but that they feed off their creators’ deaths. When an artist dies, some of the life of the artist transfers to the life of the artwork. The study also invites speculation as to whether this change in perceived agency might contribute to the greater value and prestige ascribed to works by deceased artists.

ABSTRACT. This papers deals with evidentiality in Amri Karbi, a Tibeto-Burman language spoken in North-East India. The study is based on the natural data collected during author’s fieldtrips to the areas where the language is spoken. Evidentiality in Amri Karbi is expressed by several means, including modality, aspect, reportative and quotative pusi ‘they say’ among others. Amri Karbi uses realis mode and habitual aspect to mark information as direct witness (Aikhenvald 2004) and indefinite for neutral, epistemic mode and tag questions for inference. Amri Karbi uses quotative pusi ‘they say’ to specify the information source as indirect evidence and reportative tangho for similar function in a traditional narrative discourse. Quotative pusi ‘they say’ is used to mark information as valid, per the logic that a fact which was witnessed by many is regarded as more valid than the one that was witnessed by only a speaker. In a narrative, tangho is used for backgrounding as shown in (1) and both tangho ‘REP’ and pusi ‘they say’ exemplified in (2) is used for foregrounding a part of a narrative. (1) Isi asarpe sarpepo he mane hen kiedam tangho. isi a-sar-pe sar-pe-po he mane hen ki-e-dam tangho. one POSS-old-F old-F-husband INTERACT means yam NMLZ-plant-go REP One old woman, old woman's husband, means, they were growing yam.

(2) Kuwang ayoqke akithema akithe aphurul pusi. ku-wang a-yoq-ke a-ki-the=ma a-ki-the a-phurul pu-si. NMLZ-come POSS-OBJ-TOP POSS-NMLZ-be.big=Q POSS-NMLZ-be.big POSS-snake QUOT-NF When he came out, he was a big snake.

When the indirect evidence information has become part of one’s knowledge, as in the example (3), then after indirect pusi ‘they say’ another speech word with deontic than uno ‘(I am) able to say’ is used resulting in pusi than uno which means ‘I know’ and literally ‘I am able to say what they say’. (3) Tene nang arkeng nang sal katiki arleng pusi than uno. tene nang arkeng nang sal ka-tiki arleng pu-si than un-o. then 2 bettelnut 2 work NMLZ-work man QUOT-NF say be.able-RL Then I know you are the person who grows bettelnut.

In everyday conversations pusi ‘they say’ is often used after statements, that are regarded by a speaker as not part of her/his general knowledge or as a strategy to undermine self’s notion, due to the culturally restricted territories of information or due to politeness. References Aikhenvald, A.Y., 2004. Evidentiality. Oxford University Press, Oxford

ABSTRACT. One of Milroy and Milroy’s three tenets of prescriptivism is a belief ‘[t]hat people ought to use the standard language and that it is quite right to discriminate against nonstandard users, as such usage is a sign of stupidity, ignorance, perversity, moral degeneracy, etc.’ (Milroy & Milroy 2012:33). Linguists have argued that nonstandard usage ‘[i.e. variation] does not necessarily show that a user is ignorant or stupid’ (Chapman 2012:193); however, Chapman (2012:193) notes that, ‘to date the complaints continue to be as popular as ever’. Yet this view of prescriptive behaviour ignores the fact that human beings – including speakers of standard and nonstandard languages alike – are often wont to insult people who behave differently or who hold different views to them. Cameron (2012) proposes that there is little difference in the way in which speakers approach language norms regardless of where they sit on the prescriptive/descriptive divide – speakers on either side ‘struggle to control language by defining its nature’ (Cameron 2012:8). Importantly, though, while Lukač (forthcoming) found similarities in the argument types made by prescriptivists and descriptivists in comments made underneath usage blog posts, she also noted marked differences in both the types of argument made and the rates at which similar arguments occurred across the two groups. In this talk, I will draw on data collected from reddit (www.reddit.com) over a two-year period. Reddit is made up of many individual forums (known as subreddits) where users discuss particular topics. I examine the discourse sourced from a linguistics subreddit to explore the nature of prescriptive and descriptive arguments which appear online. In particular, I am interested in the similarities and differences across the ways in which reddit users (known as redditors) who display different normative language stances - i.e. prescriptivism or descriptivism - use the words stupid and dumb when discussing English usage. Through detailed discourse analysis of interaction between redditors, I ask: how do redditors who display different normative language stances use the words stupid and dumb to refer to language in interaction? How do they use the words in reference to others and their opinions? I then compare ways in which these two groups use these words and address the similarities and differences which exist, the values which underlie these uses, and the ways in which the similarities and differences can provide insight into the values people on either side of the prescriptive/descriptive divide hold. Finally, I consider the impacts these behaviours may have on the potential to bridge the prescriptive/descriptive divide.   References Cameron, Deborah. 2012. Verbal hygiene. 2ed. London, New York: Routledge. Chapman, Don. 2012. You Say Nucular; I Say Yourstupid: Popular Prescriptivism in the Politics of the United States. In Carol Percy & Mary Catherine Davidson (eds.), The languages of nation: attitudes and norms, 192–207. (Multilingual Matters 148). Buffalo: Multilingual Matters. Lukač, Morana. forthcoming. From usage guides to language blogs. English Usage Guides. Oxford: OUP. Milroy, James & Lesley Milroy. 2012. Authority in language: investigating standard English. 4ed. London; New York: Routledge.

ABSTRACT. This poster provides an introduction to three open access corpora of Tibeto-Burman languages of Nepal, archived with Paradisec. The collections under discussion include both digitised historical materials and natively digital content. This poster serves to highlight lessons learnt during the process of archiving these materials, to assist others who are working with collections of existing materials. Issues include gathering metadata for historical collections and wrangling the outputs of SayMore projects for archiving. The poster also demonstrates what is to be found in these collections for people interested in including these materials in their own research projects, and are particularly relevant to researchers interested in tone, case marking, narrative structure and evidentiality. The first collection is Hari (1980) (AH1), which includes digitised materials from 20 cassettes and is the only known collection of materials from the Helambu variety of Yolmo. The 18 and a half hours of audio are part of the basis of Hari (2010). The materials include wordlists, language learning drills, stories, conversations, and songs. The second collection is Höhlig (1972) (MH1), which includes digitised materials from 3 reel-to-reel tapes and 6 cassettes of Syuba from 1972-1976, as well as typed transcriptions, slide film and Super8 footage. The 4 hours and 50 minutes of recordings include elicited wordlists, phonetic tokens in carrier phrases, conversation, narratives and songs. These recordings form the basis for the analysis in Hohlig & Hari (1976) and provide the opportunity for longitudinal comparison with newer corpora including the MTC1 collection (discussed below). The third collection is from the Mother Tongue Centre Nepal (2013) (MTC1). This collection was created by the small NGO with a large number of Syuba community members using the Basic Oral Language Documentation (BOLD) method (Reiman 2010) in SayMore (Hatton 2013). This collection includes 28 hours of original recordings, including songs, stories and conversation from over 50 participants, with oral translations and careful respeaking for an hour of materials and written translations for over 6 hours. The presentation will include interactive access to the corpora through the Paradisec portal, and audio examples of the content in the archive.

References Hari, Anne-Marie (collector). 1980. Hyolmo songs, stories and grammar drills (AH1). Digital collection managed by PARADISEC. [Open Access] DOI: 10.4225/72/56E9795C3C78B Hari, Anna Maria. (2010). Yohlmo Sketch Grammar. Kathmandu: Ekta books. Hatton, John. 2013. SayMore: Language documentation productivity. Paper presented at the 3rd International Conference on Language Documentation and Conservation (ICLDC). University of Hawai‘i at Mānoa, 28 February. http://scholarspace.manoa. hawaii.edu/handle/10125/26153 Höhlig, Monika (collector). 1972. Monika Hoehlig's Syuba (Kagate) materials (MH1). Digital collection managed by PARADISEC. [Open Access] DOI: 10.4225/72/56E9795C3C78B Höhlig, Monika, & Hari, Anna Maria. (1976). Kagate phonemic summary. Kathmandu: Summer Institute of Linguistics Institute of Nepal and Asian Studies. Mother Tongue Centre Nepal (collector). 2013. Syuba audio recordings from the Mother Tongue Centre Nepal (MTC1), Digital collection managed by PARADISEC. [Open Access] DOI: 10.4225/72/56E9795C3C78BReiman, D. Will. 2010. Basic oral language documentation. Language Documentation & Conservation 4:254-268. Reiman, D. Will. 2009 2010. Basic oral language documentation. Language Documentation & Conservation 4: 254-268.

ABSTRACT. According to Brugmann (1904) and Lyons (1977), the prototypical environment in which the temporal deixis emerged was the oral communication. However, there are no extensive studies to date that examine historical linguistic evidence to support this claim. In this paper, I try to fill this gap in the literature by looking at the use of the temporal adverb "nu" (now) in Old Saxon and Old High German. To achieve this goal, I analyzed all the instances of “nu” in the Old Saxon poem Hêliand and the Old High German Gospel Harmony Evangelienbuch. Textual analyses of these works reveal that the Hêliand and the Evangelienbuch exhibit a similar use of the adverb nu and demonstrate that it can be used in combination with Präsens, Präteritum, and Perfekt. The majority of instances occurs in examples with the present tense (105 out of 120 in the Hêliand and 249 out of 314 in the Evangelienbuch) in dialogic and narrative parts. While the instances in the dialogic parts suggest the use of this adverb in the spoken language, the examples from narrative contexts suggest that suggest that “nu” was used to reinforce the shift the author's perspective from an objective to a more subjective one. In the Evangelienbuch this change was made more explicit through personal pronouns, such as “ih” (I). The combinations of the adverb “nu” with the Präteritum, on the contrary, can be explained through Bühler’s theory of the Deixis am Phantasma (1934, 1982). In this process, the Origo, which is the point from which temporal reference are made, is shifted to a fictive word, in which the author narrates the events. Moreover, the instances of the Perfekt suggest that the resultative reading of the past participle combined with the verb haben or wesan in the present tenses could have favored the combinations with the adverb "nu'. In conclusion, these findings suggest that such temporal deixis was well established during this historical period and common in the Germanic oral tradition and provide empirical evidence for Brugmann (1904) and Lyons (1983) claims. In the final section of the paper, the implications for this analysis will be discussed in terms of future research on historical morphosyntax and temporal deixis.

ABSTRACT. Abstract

Founded in the evolutionary work of Richard Dawkins and situated at the nexus of language, society, and digital communication, Internet memes reflect “artifacts of participatory digital culture” (Wiggins & Bowers, 2014: 1886) that extend opportunities for political expression, engagement and participation which otherwise might not have been accessible. Central to the concept of the Internet meme is the notion of contagion or virality that fosters the “spread, distribution, replication, and propagation of memes in digital networks” (Wiggins & Bowers, 2014: 1890). Shifman (2012) further highlights how Internet memes require human agency, as they are only able to spread as a direct result of actions taken and choices made by people. Although the exchange and spread of memes may appear chaotic across intersecting virtual networks, Nissenbaum and Shifman (2017: 486) point out that “the exchange of Internet memes is to a large extent a product of societal and communal coordination” whereby meme creators depend upon the content and stance of the meme fitting with the worldview of a digital community in order for the meme to survive insertion into that community and thus spread and be reformed into new iterations and manipulations. This poster presentation adopts a sociolinguistic perspective to examine the sociointeractional function and visiolinguistic form of Internet memes created in relation to the candidates for the 2016 U.S. election – Hillary Clinton and Donald Trump. These memes, serving as an example of multimodal political participation, are processed in accordance with Van Leeuwen’s (2007) framework for the analysis of legitimizing discourse with emphasis placed upon how the memes worked to de-legitimize each of the candidates. The analysis shared reveals that the (de)legitimization strategies of moral evaluation, rationalization and mythopoeisis are all commonly observable in the sociointeractional function and visiolinguistic form of the Internet meme.

References

Nissenbaum, A. and Shifman, L. (2017). Internet memes as contested cultural capital: The case of 4chan’s /b/ board. New Media and Society,19(4): 483-501 Van Leeuwen, T. (2007). Legitimation in discourse and communication. Discourse and Communication,1(1): 91-112. Wiggins, B.E. and Bowers, G.B. (2014). Memes as genre: A structurational analysis of the memescape. New Media and Society, 17(11): 1886-1906.

ABSTRACT. Research topic: PRAGMATICS INPUT IN ENGLISH TEXTBOOK SERIES FOR VIETNAMESE UPPER-SECONDARY SCHOOL STUDENTS UNDER THE NATIONAL FOREIGN LANGUAGE PROJECT 2020

The growing recognition of the importance of pragmatic competence and its central place within overall language competence, together with the appeal of bringing a focus on pragmatics into classrooms has made research on the teaching and learning of Pragmatics in ESL/EFL contexts gain momentum during the last two decades. The critical review of the relevant literature shows that there is still a gap between what research in pragmatics has found and how the English language is taught and learnt in EFL classrooms. Additionally, recent research studies in EFL contexts generally and in Vietnam particularly also find that pragmatic knowledge is under-represented in EFL textbooks. Therefore, the present study is conducted to explore whether the newly-published national EFL textbook series for Vietnamese upper-secondary school students are successfully improved to overcome the limitations found in the former to include adequate pragmatic input that could facilitate students’ development of communicative competence in English as set out in the goal of the National Foreign Language Project 2020. On the completion of this study, it is hoped to serve as an evaluation of the new textbook series for upper-secondary school students in Vietnam in terms of pragmatics so that implications and suggestions for Vietnamese English teachers, Vietnamese textbooks writers and policy-makers could be made for the success of the renovation of English teaching and learning in Vietnam of the 2020 project. The page-by-page analysis of all student’s books reveals a paucity of general pragmatic information on politeness, appropriacy, formality, register, and cultural knowledge, as well as of meta-pragmatic descriptions of explicitly-mentioned speech acts in the books. Also, the number of pragmatic tasks is quite small in comparison with the total number of tasks included in the books, and these tasks themselves are not pragmatically-oriented. These findings imply more efforts from textbooks writers and policy-makers in providing supplementary pragmatic materials to assist teachers in their teaching practices so that the ultimate goal of the NFLP 2020 could be reached.

Appendix Framework for textbook analysis (Adapted from Vellenga (2004) and Vu (2017)) Pragmatic information Pragmatic tasks General pragmatic information Politeness Pragmatically oriented tasks Appropriacy Culture-oriented tasks Formality Register Cultural knowledge Metalanguage style Description Instruction Introduction Task-related Speech acts Explicitly mentioned Meta-pragmatic descriptions of speech acts

References Abushariefeh, A. (2016). The importance of the inclusion of pragmatics in the English as a second language (ESL) classroom: Maximizing the communicative competence of Arabic-speaking English language learners (Doctoral thesis, The University of San Francisco, The United States of America). Retrieved from http://search.proquest.com/docview/1876958125/ Nguyen, T. T. M. (2011). Learning to Communicate in a Globalized World: To What Extent Do School Textbooks Facilitate the Development of Intercultural Pragmatic Competence? RELC Journal, 42(1), 17–30. Retrieved from https://doi.org/10.1177/0033688210390265 Vellenga, H. (2004). Learning pragmatics from ESL & EFL textbooks: how likely? TESL-EJ 8(2): 1-18. Retrieved from http://www-writing.berkeley.edu/TESLEJ/ei30/a3.html …

ABSTRACT. In table-top roleplaying games such as Dungeons and Dragons, participants portray a fictional character that exists in a world that is separate from the real, but is still inextricably linked to their player. This seemingly impossible dichotomy is reflected in the way players refer to both real and fictional people during play. To add to this already complex system of reference, the dynamic of players’ reference choices changes when the game is played for the entertainment of a non-present audience rather than in a private setting. This paper presents a comparative study of the reference choices of groups of table top gamers with and without an audience. Data from a group of private players is compared to data from a group who livestream for an online audience to determine the impact audience has on the choice of player or character reference. Both player index forms and character index forms can refer to either entity. Using conceptual integration networks (Fauconnier and Turner, 2001; Tea and Lee, 2004), I investigate the balance created between selves and characters when players refer, particularly how much the speakers bring themselves into blends with their characters. The findings suggest that the balance tips toward the fictional when a game is played for an audience, while remaining with the player in a private setting. This investigation not only contributes to the field of cognitive reference, but also moves the investigation of audience accommodation to an online platform, where bystander participants, and in some cases interlocutors, in a discourse event are not present. With social and interactive entertainment making ever-increasing moves into the online environment, it is increasingly important to understand the affect this movement has on ourselves and our use of language.

References: Fauconnier, G. and Turner, M., 2001. ‘Conceptual Integration Networks’ in Cognitive Science, 22(2), 133-187 Tea, A. and Lee, B., 2004. ‘Reference and Blending in a Computer Roleplaying Game’ in Journal of Pragmatics 36(9), 1609–1633

ABSTRACT. Suppletion is when lexeme displays different roots or stems in its paradigm; as in English good > bett-er (*good-er; vs. smart > smart-er) (Corbett, 2007; Čumakina et al., 2004; Mel'čuk, 1983). Suppletion has, as a matter of cause, been related to several factors, among which:

a) Suppletion is a way of filling gaps in otherwise defective paradigms. b) Suppletion is an 'armistice' between two forms that at a point in history acquired the same meaning and were in competition. c) Suppletion aligns with a major semantic category (e.g. number: singular takes root X, and plural Y) (Bybee, 1985).

However, data from recent fieldwork in West Polesian (an Eastern Slavonic variety) present challenges to those traditional assumptions. In Slavonic languages the noun 'year' can be expressed by three different lexical roots (ɦod, lit, rɪk). And in some, it displays a suppletive paradigm (combining two of these roots), which is mostly conditioned by the type of numeral governing the phrase (i.e. Polish [1] rok; but, [5] lat). The case of West Polesian is more striking, since some varieties concentrate all the roots; especially, when it comes to the GEN PL cell [Tables 1a-c]:

Moreover, data from my corpus reveal a high degree of intra-speaker and inter-speaker variation (within the same variety) regarding the choice of the suppletive roots and their distribution. This results in an overabundant (Thornton, 2011) paradigm, where there can be up to three different roots in one cell (as in [2]).

Thus, a detailed analysis of the corpus of West Polesian calls for an extension of the traditional claims about suppletion. First, suppletion is not filling any gap in [Tables 1a, 1b], but it is rather an intrusion. Hence, there is no lexical motivation; neither synchronically or diachronically. Second, the lexical roots involved can co-exist in the same cells without necessarily having “to split the cake”. And third, there are no morpho-semantic patterns that can predict which root is to be employed because there is overabundance.

For this talk, first, I briefly introduce West Polesian, and how its sociolinguistic situation could partly explain the aetiology of such an unusual paradigm. Second, I introduce the general pattern of numeral phrases in Slavonic languages and compare it with the one found for the noun 'year' (in the Slavonic family, as well as in West Polesian). Third, I provide a summary of all the tokens of the noun 'year' in the recently collected data. I propose some general patterns (i.e. distribution of the roots) for suppletion, and I show how West Polesian speakers can combine several of them in different syntactic contexts. And finally, I use that as empirical evidence to address some of the traditional assumptions about suppletion. In sum, I present how two typologically uncommon phenomena (suppletion and overabundance) come together in the West Polesian noun 'year'; and how that can challenge some of the traditional postulates about suppletion in Morphology.

ABSTRACT. This paper investigates the syntactic location of pronouns in Standard Fijian (SF). Early work on the Determiner Phrase (DP) located pronouns in head of DP, i.e. in the syntactic position D (Abney 1987:281-284; Longobardi 1994 etc). Subsequent work (e.g. Ritter 1995) demonstrated that pronouns may occur in other locations, leading to hypotheses on intermediate positions between D and N. Cowper and Hall (2009), refining work by Déchaine and Wiltschko (2002), proposed four positions for pronouns: in D (thus pro-DP), in N (pro-NP), or as one of two intermediate phrase types: φP (as pro-φP), and #P (usually associated with NUMP) (as pro-#P), each with its own syntactic characteristics.

I investigate the locus of SF pronouns by comparing their syntactic behaviour with that of articles, nouns, quantifiers, and proper nouns, finding that SF pronouns display the same syntactic behaviour as proper names, concluding that SF contains a Proper Noun Phrase (PNP), and that SF pronouns are pro-PNP. I then examine these findings in terms of Cowper and Hall’s four pronoun loci, concluding that PNP is best associated with φP, being below the level of the DP but above that of NUMP. SF pronouns therefore have the status of pro-φP.

Pronouns in SF are demonstrably not located in D, as they are occur with the personal article o (1b), itself occupying D. This fact lead Alderete (1998) to assume that pronouns in Fijian are nouns because they head an NP that carries a determiner. However, SF pronouns are not located in N: the internal syntax of phrases with an N head differs from that of phrases with a pronoun head: Ns may be accompanied by a dependent AP (2a-b), while pronouns are accompanied by a DP (2c). A comparison of the syntactic behaviour of pronouns with that of quantifiers on the other hand reveals shared phrase-internal syntax: both are accompanied by a DP dependent (compare (3) with (2c)). However, their external syntax differs. Pronoun objects (4a) are located inside the VP (defined by the right-edge particle tiko in (4)), while objects with a noun head (4b) must occur outside the VP, coreferential with a VP-internal pronoun. Phrases with a quantifier head (4c) also occur outside the VP (4c), like those with an N head, and unlike pronouns. Pronouns therefore do not behave in the same way as quantifiers so are not pro-NUMP.

Having demonstrated that SF pronouns are not pro-DP, pro-NP or pro-NUMP, I then demonstrate that they do behave in the same way as proper nouns. Phrases with a PN head have the same internal and external syntax as phrases headed by a pronoun: both have a DP dependent (compare (5a) with (2c)), and both occur inside the VP (compare (5b) with (4a)), justifying positing a distinct PN phrase. SF pronouns are therefore pro-PNP.

I conclude that PNPs in SF correspond to Cowper and Hall’s φP. SF pronouns are pro-φP, a finding consistent with Cowper and Hall’s claim that the feature ‘person’ is associated with the φ head.

ABSTRACT. Frames of reference (FoRs) are coordinate systems for expressing the spatial relationship between one object and another (Levinson 2003:24–25). Previous research has identified correlations between FoR(s) used in linguistic and non-linguistic behaviour, suggesting a possible Whorfian effect (e.g., Levinson 2003; Majid et al. 2004), while other studies have challenged this position (e.g., Li & Gleitman 2002; Li et al. 2011). Relatively few studies, however, have identified substantial variation in FoR choice amongst speakers of a single language, though a preference for egocentric strategies (e.g., The man is to the left of the tree) among urban communities has been reported (Dasen & Mishra 2010; Majid et al. 2004; Pederson 1993).

This paper presents the results of a comparative study on spatial language in Dhivehi (Indo-Aryan, Maldives) and Marshallese (Oceanic, Marshall Islands), based on primary data collected during fieldwork in the Maldives, the Marshall Islands, and a Marshallese-speaking community in Arkansas, USA. Several task-based elicitation techniques were employed at all field sites. Here we report on the results of the ‘Man and Tree’ photo description task (Senghas version: Terrill & Burenhult 2008), performed by 59 Dhivehi and 48 Marshallese-speaking pairs, producing 5,014 and 3,383 FoR descriptions respectively.

Our results show some correlation with environment. For both languages, for example, speakers living in urban environments use egocentric strategies at significantly higher rates than speakers living in less urban environments (p<0.05, Mann-Whitney U tests), corroborating previous studies. However, other results crosscut environment, correlating principally with language. Despite living in similar topographic environments, Dhivehi speakers on Laamu Atoll and Marshallese speakers on Jaluit Atoll show important differences in FoR patterns. While some strategies, such as the use of cardinal directions, are shared, Dhivehi speakers show a significantly higher preference for intrinsic FoR than Marshallese speakers, and reference to atoll topography (‘lagoonward’, ‘oceanward’) is a major strategy in Marshallese but not Dhivehi. Crucially, the results also reveal significant variation according to demographic factors, including occupation, age, gender and level of education, suggesting that while the environment plays a role, sociocultural factors intervene. For example, in both languages speakers with only primary education used cardinal directions at significantly higher rates than more highly educated speakers (p<0.05, Mann-Whitney U tests). The Dhivehi data in particular shows systematic variation according to demographic factors: egocentric strategies are used more frequently by indoor workers, speakers in predominantly white-collar communities, younger speakers, women, and more highly educated speakers, while cardinal directions are favoured by fishermen, other speakers in fishing communities, older speakers, men, and less educated speakers. We argue that occupation-based variation points to influence of habitual interaction with environment, and underpins much of the other demographic variation (e.g., via traditional gender roles, and intergenerational shifts in subsistence patterns). We discuss implications of these findings for our understanding of the relationship between language, culture, cognition and the physical environment, arguing that a full picture of the spatial strategies preferred by speakers of a given language can only be obtained through an understanding of speakers’ cultural interaction with the world around them.

ABSTRACT. The usage-based study of language initiated by the theoretical works of Langacker (1987, 1988), has led to a wide range of linguistic studies (e.g. Bybee, 2010; Kemmer & Barlow, 2000). Views on linguistic structures (i.e. as the reflections of a cognitive and a communicative system) lead usage-based language studies in different directions, with one focusing on the close relations between usage and mental structures, and the other investigating the connection between usages and communicative needs. Following Coussé and von Mengden (2014), we believe that the cognitively-oriented and communicatively-oriented perspectives complement each other and both contribute to usage-based models. This study investigates the diachronic developments of 2749 instances of spatial word shang in historical texts of Archaic Chinese (AC, 12th c B.C.- A.D. 220), Medieval Chinese (MC, A. D. 220-1368 c.), Modern Chinese (MOC, 15th -mid 19th c.) and Contemporary Chinese (CC, mid-19th-20th c.) from Chinese corpora. Tracing the diachronic developments based on these instances, we show how various meanings and grammatical functions of shang emerged and developed in systematic ways based on the cognitive processes of metonymy (Lakoff 1987), metaphor (Lakoff & Johnson 1980) and construal (Langacker, 1987), and pragmatic inferences (Traugott & Dasher, 2002).We also demonstrate how shang acquired its meanings by fitting into the sentential contexts where it occurred. Our study fills a research gap by focusing on the earlier stages of shang and illustrating how language uses interact with linguistic structures. It is thus a contribution to the history of Chinese, usage-based language studies, and to grammaticalization studies.

(1) Wen-wang zai shang Wen-king be-located above ‘(The divinity of) Wen-king is in a high area of the sky.’

(2) shi-yi mei shan zai shang Therefore beauty kindness be-located above ‘Therefore, beauty and kindness belong to the highest-ranking officials.’

The examples demonstrate the usages of shang from denoting the spatial meaning ‘a higher area of the sky’ to the abstract meaning ‘people in the highest social positions’. The latter meaning of shang was motivated by the pragmatic inference ‘the highest’, the metaphor HIGH SOCIAL POSITION IS UP and the metonymy PEOPLE AT THE POSITION FOR THE POSITION.

Bybee, J. (2010). Language, Usage, and Cognition. Cambridge, New York: Cambridge University Press. Coussé, E., & von Mengden, F. (2014). Introduction: The Role of Change in Usage-based Conceptions of Langauge. In E. Coussé & F. von Mengden (Eds.), Usage-based Approaches to Language Change (Vol. 69, pp. 1-19). Amsterdam/Philadelphia: John Benjamins Publishing Company. Kemmer, S., & Barlow, M. (2000). Introduction: A Usage-Based Conception of Language. In S. K. Michael Barlow (Ed.), Usage Based Models of Language (pp. vii-xxvii). California: CSLI Publications. Langacker, R. W. (1987). Foundations of Cognitive Grammar Stanford, Calif.: Stanford, Calif. : Stanford University Press 1987. Langacker, R. W. (1988). A Usage-based Model. In B. Rudzke-Ostyn (Ed.), Topics in Cognitive Linguistics (Vol. 50, pp. 127-161). Amsterdam/Philadelphia: J. Benjamins. Traugott, E. C., & Dasher, R. B. (2002). Regularity in Semantic Change. New York: Cambridge University Press.

ABSTRACT. Comparatives of superiority have been well studied for many decades. Representative literature on this topic based on the languages of the world includes Ultan (1972), Stassen (1985, 2013), Heine (1997), Dixon (2012), inter alia. Lolo-Burmese languages, a sub-branch of Tibeto-Burman, are rarely represented in typological studies outside of China due to the lack of data, yet they employ an unusual coding strategy, when seen in cross-linguistic perspective.

Apart from locational (including from-type, to-type, and at-type [Stassen 2013]) and exceed comparatives, several interesting phenomena are revealed on the basis of our study on 26 Lolo-Burmese languages (see Table 1): 1) The spatial orientation terms ‘upside’ and ‘downside’ or their derivational forms are used to code the comparatives of superiority. 2) Partial semantic bleaching is observed in 8/26 languages, i.e. positive grading (e.g. ‘big’) associates with ‘upside’, while negative grading associates with ‘downside’ (e.g. ‘small’). This indicates that the semantic nature of these spatial orientation terms still predominates—they are not fully grammaticalized standard markers. 3) The terms ‘upside’ and/or ‘downside’ remain as nominal in 14/26 languages. The two examples of Hani in (1) (Dai 1991:541) demonstrate both the partial semantic bleaching and the retention as nominal. In this case, the standard is in its pronominal form of possessive modifying the spatial orientation term ‘upside’ or ‘downside’. Even though ‘upside’ and ‘downside’ are semantically related to location, the comparatives formed by them cannot be classified as locational type.

It can be observed in Table 1 that above/under-type is singled out from at-type. This is because the standard markers in from, to, and at types in our sample are characterized by multi-functionalities. Take the marker ciə˞21 in Naxi as an example. Derived from ‘upside’, it serves not only as a standard marker , but also as allative, benefactive, dative, causee and differential object marker. By contrast, above/under appear relatively uni-functional. As pointed out, their semantic meanings are not bleached out when forming comparatives. In addition, the terms ‘upside’ and/or ‘downside’ in these languages show certain divergence of degree of conventionalization. Some languages employ the reduced forms of ‘upside’ and/or ‘downside’, such as Nasu, while some of them deploy the full forms, for example Ximoluo. Note that Rouruo even uses both full and reduced forms of ‘upside’ when forming comparatives, which shows that the spatial nominal strategy is shifting to the locational comparative. Furthermore, the term ‘upside’ in some languages is taking the place of ‘downside’ to form the comparatives of both positive and negative grading. For example, the term ‘downside’ in Kucong is retained in only a few cases.

Therefore, we first propose that using spatial orientation terms ‘upside’ and/or ‘downside’ should be treated as a non-grammatical strategy to form comparatives of superiority in Lolo-Burmese, since their grammaticalization remains at a young stage. Second, the on-going process of grammaticalization gave rise to some new reflections on the development of a conventionalized standard marker, i.e. the spatial orientation term ‘upside’ might develop into a standard marker without passing through an oblique case marker stage.

ABSTRACT. Glottalisation functions as a cue to coda voicelessness in many varieties of English. However, varieties differ according to which stops are glottalised and how frequently they are glottalised. In British English all voiceless stops can be glottalised, though /t/ is most frequently glottalised in London (Tollfree, 1999), whereas in Tyneside /p/ is more frequently glottalised (Docherty et al. 1997). In American English glottalisation commonly occurs for /t, p/, though more frequently and in a greater range of contexts for /t/ (Pierrehumbert, 1994). Another important cue to coda stop voicing in English is preceding vowel duration; vowels preceding voiced codas are longer than vowels preceding voiceless codas (Fowler, 1992; Lisker, 1978). Penney et al. (2016) found evidence of a trading relationship between glottalisation and vowel duration as cues to the voicing status of coda stops in Australian English (AusE), with increased glottalisation co-occurring with a decrease in utility of the preceding vowel duration voicing cue. However, this has only been examined for alveolar coda stops and was based on observations of stressed syllables; questions remain about whether the same effects are present in unstressed syllables, where voicing related durational differences may already be reduced. This study extended the examination of coda stops to all three places of articulation (POA) in AusE to investigate dialect-specific distributional patterns of glottalisation and preceding vowel duration in unstressed syllables.

Twenty-eight female AusE speakers (aged 18-38; mean = 24) read /(C)VɹəC/ trochees where the final consonant was a voiced or voiceless stop at bilabial, alveolar or velar POA. This enabled an examination of unstressed final syllables with voiceless coda stops at each POA and a comparison of voiced and voiceless codas at bilabial and alveolar POA. Note English lacks voiced velar codas in this form. 755 items were extracted. The tokens were automatically aligned through WebMAUS and hand-labelled for subsegmental components including vowel onset and offset; coda stop release burst; and presence and duration of glottalisation. We fitted logistic mixed effects models to examine effects of voicing and POA on the presence of glottalisation, and a linear mixed effects model to analyse unstressed vowel duration between voicing contexts.

Analysis of glottalisation related to coda stop voicing found voicing to be a significant predictor, as expected: glottalisation occurred at high rates for both POAs in the voiceless context compared to the voiced context. Analysis of tokens in the voiceless context showed no significant difference between POA: glottalisation occurred at high rates at all three POAs. We found no significant difference in vowel duration across voicing categories indicating coda voicing induced vowel duration effects are minimised in unstressed syllables. The results show glottalisation cues coda stop voicelessness in unstressed syllables at all POAs. As coda voicing-related vowel durational effects were minimised in the unstressed environment, glottalisation may be applied to support the coda-voicing contrast when the vowel duration cue is compromised.

ABSTRACT. Large-scale regularities both at synchronic and diachronic levels of language description are often argued to be the result of extra-linguistic factors such as ease of production or processing, learnability or simplicity, among others. Yet such concepts are often invoked without specific indications about how can they be evaluated empirically, which hinders effective progress in our understanding of the dynamics of languages.

Here we provide a fleshed out experimental and observational evaluation of the claim that certain articulatory gestures might have entered later in the history of our species as a result of widespread cultural changes that took place in many independent human populations starting from the Early Pleistocene (Hockett 1985). We focus on the development of labiodental segments in relation to changes in dental occlusion configuration. The angle and the positioning between upper and lower anterior teeth and their lifelong changes are connected and affected by tooth wear: an increase in wear (due to the consumption of tough foods or the usage of teeth as a third hand, a widespread feature of many hunter-gatherer populations) leads to uprighting movements of the anterior teeth, so that eventually upper and lower teeth align with each other, forming edge-to-edge bite (Kaifu et al. 2003, Margvelashvili et al. 2013 ). By using computational simulations created in ArtiSynth (www.artisynth.org; Lloyd et al. 2012) of labiodental articulation with both regular and edge-to-edge bites we show that the latter implies an increase in the articulatory cost of labiodentals.

This finding predicts that labiodentals would be systematically dispreferred in languages whose speakers are more likely to have edge-to-edge bite. This is confirmed by our data to the extent that hunter-gatherer lifestyles are more likely to support edge-to-edge bite than agriculturalist lifestyles. Using large-scale descriptions of both segment inventories and ethnographic records covering thousands of languages of all continents (Moran et al. 2012), we show, both synchronically and diachronically, that languages spoken by populations that have been described as hunter-gatherers in the last century have less or no labiodentals in relation to languages spoken by agricultural or industrial societies.

ABSTRACT. The Yam Family (aka Morehead-Maro Family) comprises around twenty languages of Southern New Guinea, straddling the border between Indonesia and Papua New Guinea. Apart from possible remote connections to the Eastern Trans-Fly languages, they currently appear unrelated to any other Papuan family. Though little-known until the last few years, a number of recent studies have now raised our level of documentation to the point where a reconstruction of the language’s phonology becomes possible: we now have reliable data for several languages in two of the family’s branches (Tonda to the west and Nambu to the east) though material on the third branch, Yei, remains scanty.

In this paper we sketch the main phonological changes that have occurred at subgroup and language level, and reconstruct the phoneme inventory, syllable structure and basic lexicon of the family. We will particularly focus on the loss of initial /ŋ/ and its consequences in many languages of the group, loss of final vowels, palatalisation in the environment of front vowels, the very variable descent of one reconstructed fricative whose ancestral point of articulation remains problematic (descending variously as /s/, /h/, /ɣ/, /ɰ/, /ð/ and /d/), the development or extension of coarticulated labial-velar phonemes (/k͡p/,/g͡b/,/ŋ͡mg͡b/) by the absorption and loss of rounded vowels, the reduction of phonemic schwa in some languages so that it becomes a predictable epenthetic insert, the defricativisation of some languages under contact influence, and on denasalisation of prenasalised stops to create voiced stops in some languages.

We will also introduce the workings of an online lexical database, Yamfinder, which we are developing to study the historical phonology and lexicon of a number of language families in Southern New Guinea.

ABSTRACT. As with many languages (cf. Wierzbicka 1999; Ogarkova 2013), Anindilyakwa (Gunwinyguan, north-east Arnhem Land, NT) makes extensive use of body-parts in the expression of emotions and feelings.

Visible and internal body-parts occur in both descriptive and figurative expressions in Anindilyakwa. While in some expressions body-parts act as a literal denotation, often a body-part instead denotes a ‘locus of emotional/intellectual activity’ (Enfield 2002, p. 86), with resulting emotion meanings expressed through different kinds of semantic extension.

Body-parts are involved in the description of emotions in Anindilyakwa in the form of simple predicates, body-part compounds, and productive incorporated nominals. Both free and bound nominal forms can occur in emotion descriptions (examples (1) and (2)), sometimes both occurring in the same expression (example (3)).

1. na-larrə-na     alyelyikba
   3a-fall-npst     neut.lips
   they are sulky (Waddy n.d., p. 13).

1. na-lyirrkburrkwu-larrə-na
   3a-lips-fall-npst
   they are sulky (Waddy n.d., p. 13).

1. nu-werrikə-lada-Ø                  yukwudhukwudha
   3a-chest-be.satisfied-pfv         masc.chest
   they’ve had enough/ they are fed up (Anin. database)

Body-part compounds involve lexicalised constructions, with speakers not generally overtly aware of the body-part in the emotion description, and being surprised when they perceive the connection (Waddy n.d., p. 1). For example, the Anindilyakwa compound -werrikə-ngwanji- ‘chest+rest’ expresses the meaning ‘to be discouraged’ which, while non-compositional, can be interpreted metaphorically, with the chest representing an emotional state.

This paper thus examines the metaphorical and metonymic associations of body parts (both free body parts and lexical compounds) with emotion descriptions, and considers the principled ways in which particular body parts are associated with particular emotions and feelings in Anindilyakwa.

Of particular interest to this paper is the chest, -werrik-/ yukwudhukwudha; the body-part conceptualised as the primary site of emotion in Anindilyakwa, and the principle and most productive body-related morpheme used to describe emotions in the language. I examine -werrik-/ yukwudhukwudha in compound forms, considering the figurative expressions this body part conveys, and demonstrate how these expressions can be seen to stem from regular semantic extension from the description of a somatic state. For example, for -ngwurrkbali- + yukwudhukwudha ‘be.open + chest’, ‘be overjoyed/ feel encouraged’, the feeling of being overjoyed is represented metonymically with being open in the chest.

In addition to the chest, other frequently encountered body parts associated with emotion description include the arəngka/ -lyangk- ‘head’, mangma/ -marngk- ‘mind’ and arndərnda/ -lyikarrk- ‘heart’, and these are also examined and compared in some detail.

Finally, Anindilyakwa contains a small number of emotion nominals that are not associated with body parts (e.g. awinyamba ‘anger’, akwiyekirrerra ‘happiness’). I briefly examine these nominals in comparison with more commonly encountered emotion descriptions containing body-parts.

This paper builds on research by Waddy (n.d.) and van Egmond (2012), as well as incorporating new data collected by the author in 2016-17.

ABSTRACT. This paper is concerned with the question of how emotions are expressed in Kunbarlang, a Gunwinyguan language in western Arnhem Land. It outlines the particular formal constructions used to express emotions, including simple predicates, compounds (lexicalized noun incorporation constructions, which include both body-part and non-body-part nouns), coverb constructions and noun-verb idioms. We note some of the similarities and differences of these constructions to those of other Gunwinyguan languages, and to other neighbouring languages from other language families (Mawng (Iwaidjan) and Njébbana (Maningrida)), and how this sheds light on our understanding of Kunbarlang and its relationship to languages of the region.

ABSTRACT. Gija is a non-Pama Nyungan language from the east Kimberley in Western Australia. It is a member of the Jarragan family along with Miriwoong and Gajirrabeng. This paper will explore the role of the body and verbal structure in construing emotion in Gija. The stomach is frequently expressed as the site of strong emotion in Gija. For example, the noun liyan is often translated as meaning ‘a good feeling in the gut’ which expresses a deep sense of contentment and wellbeing. Events or the actions of people are evaluated as ‘making the stomach good or bad’.  Related to this concept is a phenomenon described by Miriwoong and Gija as a throbbing sensation that acts as a signal to the senser of some significant event associated with their kin. Specific relations are associated with particular body parts. Most complex affectual processes are realised verbally and can be construed as agentive happenings or states of being. Gija shares a bipartite verbal structure with several languages across the Kimberley, Victoria River and other parts of the top end. In Gija it consists of a ‘verb’ expressing general lexical meanings and inflecting for tense, mood and aspect which in compounds combines with an uninflected ‘coverb’ that carries the main semantic load. Processes of cognition are most often expressed with the simple verb SAY/DO. In contrast, coverbs concerning affect combine with a wider variety of both transitive and intransitive inflected verbs including SAY/DO, BECOME, GO/COME, BURN/BITE, FALL/GO_DOWN and HIT. This paper will explore examples from the corpus of contrasting coverb and verb combinations in order to reflect on what this suggests about the social and conceptual construction of emotion and agency in Gija.

ABSTRACT. This paper presents a considerably more detailed description of Associated Motion (AM) constructions in Kaytetye than previous accounts in Author (1984, 2006), Turpin (2000), and Turpin & Ross (2012). The semantic distinctions made by the various AM constructions are justified by a careful study of text examples. The revised analysis is informed by comparative data from other Arandic languages (e.g. Wilkins 1991, 2006, Moore 2013) as well as typologically comparable findings from non-Australian, especially South American, languages (Guillaume 2016, Rose 2015). In addition to the parameters of Direction (go, come, return, approach) and Relative Timing of the main and the motion event (Prior, Subsequent and Concurrent Motion), we recognise that of the Moving Argument (Subject for most subcategories, but Object for one Prior Motion subcategory. Speed is recognised as an additional dimension (GO.QUICKLY&DO). An additional value within the Relative Timing dimension is GO&DO&RETURN, with both prior and subsequent motion. Within Concurrent Motion a quasi-aspectual dimension is recognised: do once, intermittently, or continuously on the motion path. We show further how AM forms contribute to Respect Language—implying either doing something on one’s way or doing it speedily. We note as well the lexicalisation of particular AM forms, non-canonical usages of the forms—to mark metaphorical motion, to indicate spatial notions with static verbs, and to mark purely directional usages with inherently motional verbs.

References Guillaume, Antoine. 2016. Associated motion in South America: Typological and areal perspectives. Linguistic Typology 20(1): 1-97. Author. 1984. The category of ‘associated motion’ in Kaytej. Language in Central Australia 1: 23-34. Author. 2006. Languages of the world: Kaytetye. In Keith Brown (ed), The encyclopedia of language and linguistics, vol. 6. Elsevier: Oxford.170-172. Moore, David. 2013. Alyawarr verb morphology. MA thesis, University of Western Australia: Nedlands, W.A. Rose, Françoise. 2015. Associated motion in Mojeño Trinitario: Some typological considerations. Folia Linguistica 49: 117–158. Turpin, Myfany. 2000. A learner's guide to Kaytetye. Alice Springs: IAD Press. Turpin, Myfany and Alison Ross. 2012. Kaytetye to English dictionary. Alice Springs: IAD Press. Wilkins, David P. 1991. The semantics, pragmatics and diachronic development of ‘associated motion’ in Mparntwe Arrernte. Buffalo Papers in Linguistics 91(1): 207-257. Wilkins, David P. 2006. Towards an Arrernte grammar of space. In Stephen C. Levinson and David P. Wilkins (eds), Grammars of space: explorations in cognitive diversity. Cambridge; Cambridge University Press. 24-62.

ABSTRACT. In many languages, determiners, especially demonstratives, are said to encode whether the referent is visible (Anderson and Keenan 1985; Matthewson 1998). But very little is known about the semantics of visibility contrasts. I argue that the demonstratives of Ticuna (isolate; Brazil, Colombia, Peru) do systematically encode perceptual meanings, but that these meanings are not reducible to literal visibility. Data come from experiments, elicitation, and recordings of informal conversation during recent fieldwork.

Data. Ticuna has four exophoric demonstratives. They define a person-centered demonstrative system (1). Two of the demonstratives, ŋe3a2 and ɟe4a2, can be used only for referents that are visible to the speaker. Another demonstrative, ɲa4a2, can be used for invisible parts of the speaker’s body and invisible objects in contact with her body. Otherwise, ɲa4a2 is acceptable only for visible referents. ŋe3ma2 is used for visible and invisible referents.

(1) (Table in PDF abstract not reproducible here)

Analysis. It has been suggested that demonstratives privilege spatial or attentional meanings, and that visibility meanings are often epiphenomenal (Enfield 2003:96). The perceptual meanings of ɲa4a2, ɟe4a2, and ŋe3a2, however, are encoded. Speakers use only ŋe3ma2 to index smells and other necessarily invisible referents, and they reject ɟe4a2 and ŋe3a2 in reference to (contingently) invisible objects, no matter the spatial and attentional context. Additionally, productive morphology can derive -- from ɲa4a2, ɟe4a2, and ŋe3a2 -- demonstratives that can refer to invisible referents. These data show that the perceptual meaning component of these demonstratives cannot be canceled, and can be manipulated by morphosyntax; therefore, it is encoded.

I then provide a semi-formal analysis of the perceptual meaning component. ɟe4a2 and ŋe3a2 encode that the speaker has visual access to the referent. ɲa4a2 encodes that the speaker has either visual or tactual access to the referent. Consequently, the form is (a) acceptable with literally invisible referents in contact with the body, but (b) unacceptable for invisible referents beyond the speaker’s body space, even if metrically very close to her. For ŋe3ma2, I argue -- from anaphora, recognitional uses, and interactions with TAM -- that the addressee-centered use and the invisible use represent separate lexical items. Addressee-centered ŋe3ma2 is a deictic. Invisible ŋe3ma2 is not a deictic, but a familiar definite compatible with both weak and strong familiarity (Roberts 2003).

Implications. Ticuna challenges theories that treat demonstratives as primarily indexing the location of the referent in space. In this language as in Yucatec Maya (Hanks 1990), mode of sensory access meanings are central, encoded features of demonstratives. Though vision is prominent in this domain, touch is relevant as well, eliminating an analysis of mode of access as narrow visibility. The results support views of demonstratives as potentially encoding perceptual and attentional meanings (Hanks 2011; Peeters and Özyürek 2016, a.o.) as well as spatial ones.

ABSTRACT. Noun phrases have long been of interest to grammarians of Australian languages (eg. Hale 1983, Austin & Bresnan 1996, Austin 2001, Nordlinger 2014), because in many of these languages nominals and other co-referring elements do not have a fixed order, nor must they occur contigiously , as example (1) from the Australian language, Garrwa. In (1), the nominals walkurra ‘big (one)’ and mali ‘flood’ both point to the same referent , but they are separated by a 2nd position clitic cluster (Mushin 2012) and a verb.

Like many other Australian Languages, Garrwa nominals and demonstratives which are co-referent within clauses share the same case marking, as in (2) where the distal demonstrative na(nda) ‘that’ and the nominal kulabajarra ‘hat’ are both datively marked as the transferred object. Because of these properties, Mushin (2012:255) avoided ascribing phrasehood to co-referential nominal expressions, instead defining ‘nominal groups’ as “…forms that refer to the same referent and share the same grammatical and semantic role”. However Louagie & Verstraete’s (2016) recent typology of noun phases in Australian Languages proposes that phrasehood may be an appropriate classification for certain kinds of nominal expressions in languages like Garrwa.

Examples like (1) and (2), where reference is achieved through multiple nominal expressions, are rare in Garrwa conversation: turn construction units (TCUs) generally include one referential nominal expression (or none). This is because most referents in conversation are already established for participants and therefore indexed with clitic pronouns or left unexpressed. In a corpus Garrwa conversation (6 conversations), the most common multiple nominal expression was a demonstrative followed by a noun, as in (2), (3) and (4). In (2) and (3) the demonstrative + nominal occur within the same prosodic unit which is integrated into the overall prosody of the clause while in (4), a prosodic break occurs between the demonstrative and nominal, suggesting a more incremental turn design.

In this paper I take an emergentist perspective (e.g. Ford, Fox & Thompson 2013) on the analysis of the status of noun phrases in Garrwa, showing how their production is contingent on their interactional context, and coordinates with prosodic and gestural features of the design of conversational turns. Here I focus on the prosodic and syntactic properties of the demonstrative + nominal expression, and argue that when prosodically integrated, they constitute a construction that serves particular discourse functions, while in other contexts they do not form a construction, and this is reflective of the local contingencies of the context in which they were uttered. The results support Louagie & Verstaete’s (2016) classification of Garrwa as a language which shows evidence of noun phrasehood, albeit under certain discourse conditions.

(1) walkurra nurr=i yabimba mali nana-ba big 1plExclNOM=PAST make floodwater that-DEIC We had a big flood over there

(2) yanka ja=nga wajba ngana na-nkanyi when FUT=1sgACC give 1SGACC that-DAT kulabajarra-nyi hat-DAT ‘When are (you) going to give me that hat?’

(3) jakajba=yi nanda mudika start.up=PAST that car ‘The car started up.’

(4) wanya nanama, kulajbajarra what that-IDENT hat ‘What’s that one, hat?’

ABSTRACT. Studies on French varieties in contact with other languages find that intonation patterns differ to those documented for Standard French (Bullock 2009, Bordal 2012). For French in contact with Occitan, a language with lexical stress, Sichel-Bazin et al. (2012) find Accentual Phrase (AP)-internal rhythmic prominences that are interpreted as remainders of Occitan lexical accent in the bilingual's French. Studies of French L2 in contact with Swiss German and Wolof (Bordal et al. 2012, Obin et al. 2012), suggest the contact varieties share features with word prosodic systems. However, the realization of the AP by early bilinguals of French has not been studied systematically yet. The AP represents the lowest intonationally marked prosodic constituent in French (Jun and Fougeron 2002). Its intonation is typically characterized by a rise in fundamental frequency on the last syllable of a phrase that is not utterance final and an optional early rise occurring before the late rise. The purpose of this study is to investigate the realization of the AP in a variety of French spoken by bilinguals of French and Drehu, a language described as having lexical stress. It is hypothesised the contact language has an impact on the realization of the French AP of the bilinguals. Lenormand (1954) and Tryon (1968), state stress has a demarcative function indicating the division between words in Drehu. Stress always falls on the first syllable of a word, regardless the number of syllables, optionally having a secondary stress, on the third syllable of polysyllabic words. Recordings of five female speakers (aged 29-47) were made in Lifou in New Caledonia. Participants responded to a questionnaire (Gertken 2012) evaluating language dominance and four were found to be dominant in Drehu and one nearly balanced. Elicitation materials consist of a set of utterances in French, with 17 target words counting 2, 3 and 4 syllables in the content word and 1 or 2 monosyllabic preceding function words, taken from Welby (2006). The position of the target phrase in the sentence was initial or medial. The materials ensure comparable data regarding the position of the early and the late rise, or eventual variation in patterns. Sound files were segmented using WebMAUS using a parameter model based on SAMPA. The f0 at the beginning of the utterance, early L tone (for initial phrases) or late H of preceding phrases (for medial phrases), early H tone, and late H tone were tagged by hand. Speech rate in phones per second was calculated. Results show that overall bilinguals make use of patterns similar to those of Metropolitan French speakers and that the realization of the AP does not vary as dramatically as initially suspected. The speakers are early bilinguals educated within the French system which explains the difference to L2 speakers of French. In future, it is of interest to investigate the status of Drehu as a lexical stress language and to evaluate the stress patterns initially reported by Lenormand and Tryon.

ABSTRACT. This talk presents results from an ongoing research project on the intonational grammar of Fataluku, a Papuan language of East Timor, along with some observations on its contributions to prosodic typology. Most of the best intonational descriptions are of large languages from a few major language families; this study contributes a more diverse perspective by describing the intonational patterns of a language that is spoken by a relatively small speaker population in an underrepresented linguistic area. Analyses are based on first-hand fieldwork and experimentation and situated within the Autosegmental-Metrical approach to intonational phonology (Pierrehumbert 1980, Ladd 2008).

Fataluku shows little evidence for stress or tone; intonation is rather organized the accentual phrase (AP), a prosodic unit that contains either a single word or a few closely connected words. Each accentual phrase bears a rising-falling intonational melody that peaks around its first or second syllable, aligning with the edge of the first metrical foot (Hft). The grouping of words into APs is sensitive to both phonological and syntactic criteria, and the alignment of intonational contours may in some cases be important for disambiguation.

Accentual phrases are grouped into intonational phrases (IP), units that bear a complete intonational contour and may be bounded by silence. The most pervasive IP-final boundary tone in Fataluku is a simple low f0 target (L%), which indicates completeness or finality. Questions are indicated with a low f0 on the penultimate syllable of the IP with a rise-fall over the course of the final syllable (L+HL%; see figure 1). The same contour is used for both wh- questions and polar questions, and is crucial for disambiguating polar questions and declaratives. Fataluku also has two alternatives for showing continuation, a high level (Hː%) and an extra high rise (^H%), both of which are also associated with lengthening of the final syllable.

To sum up, Fataluku intonation is characterized by a rising-falling intonational pattern over each accentual phrase, and an IP-final boundary tone that conveys information about how the utterance should be interpreted. Typologically, word- or AP-level intonational contours such as those of Fataluku are very common in languages which lack phonological stress and tone (Jun & Fletcher 2015). Fataluku’s use of IP-final boundary tones is also common cross-linguistically (e.g., Lindström & Remijsen 2005). One unique feature of Fataluku’s intonational system is the importance of foot edges as alignment sites for intonational pitch targets, a possibility predicted by the Autosegmental- Metrical theory, but not to my knowledge elsewhere described. Another typologically unusual feature of Fataluku’s intonation is the use of a question contour which falls, rather than rises. A similar intonational pattern for questions is also reported in Kalamang, a Papuan language spoken on the Bomberai peninsula of New Guinea (Visser 2016), but more work is needed on other languages of the area to determine the extent of this pattern, and more generally, the extent to which the intonational systems of Papuan languages differ from intonational patterns found elsewhere in the world.

ABSTRACT. In certain phonological accounts of rhythm (e.g. Dauer, 1983; Bertinetto, 1989), languages are positioned along a continuum between stress- and syllable-timing. Durational variability is linked to speech rhythm (Ramus, Nespor & Mehler, 1999). Stress-timed languages (e.g. Germanic languages) tend to display higher durational variability in their vocalic and consonantal intervals, and a lower proportion of ‘vocalic-ness’ overall. Conversely, syllable-timed languages (e.g. Romance languages) display lower variability but higher ‘vocalic-ness’. Rhythm metrics quantify such durational variability across (e.g. DeltaV, VarcoV, %V; Ramus et al., 1999; Dellwo, 2006) or within (e.g. nPVI-V; Low & Grabe, 2002) a speech sample, and successfully discriminate rhythms of different languages and varieties (e.g. White & Mattys, 2007).

In this project we a) provide a preliminary description of the rhythm of Barunga Kriol - a variety of Kriol local to Barunga Community, NT – and b) compare it with a relatively standard variety of Australian English. Kriol is an English-lexifier creole language spoken by over 20,000 children and adults in the Northern parts of Australia (NILS 2005), yet much about the prosody of this language remains unknown. Although recent research has started to examine the phonetics and phonology of Kriol varieties (e.g. Bundgaard-Nielsen & Baker, 2015; Jones, Meakins & Buchan, 2011), nothing is yet known about its rhythm.

The phonological and grammatical system of Kriol varieties often display patterns derived from Australian Aboriginal languages of the area (Munro, 2000). We hypothesised that Barunga Kriol would be more ‘syllable-timed’ than Australian English, as at least some areal languages (e.g. Bininj Gun-wok) are more syllable-timed than English (Butcher and Fletcher, 2008) and as Australian English tends towards the stress-timed pattern. 10 young adult female participants (5 English, 5 Kriol) were recorded telling a story to a familiar peer using picture books as stimulus. Recordings were orthographically transcribed then automatically aligned (webMAUS). Speech samples were divided into vocalic and consonantal intervals, and rhythm metrics (VarcoV, nPVI-V, %V) were calculated for each utterance, and mixed effects models fitted.

As hypothesized, the rhythm metrics quantified Barunga Kriol as more syllable-timed compared to Australian English. For nPVI-V, Barunga Kriol was lower than English (χ2(1) = 12.91, p = 0.00033) (Figure 1), reflecting less variability in vocalic durations. Barunga Kriol was also higher in ‘vocalic-ness’ (%V) than Australian English, (χ2(1) = 7.35, p = 0.0067) suggesting more syllable timing (Figure 2).

Our results suggest that Barunga Kriol is more syllable-timed than English. However, rhythm may well be supported by parameters other than those reflected in durational rhythm scores. In future, spectral analysis of the patterns of vowel reduction - another specific source of rhythmic variation (Dauer, 1983; 1987) - will enrich this description and help determine the extent of rhythm differences between Kriol and English.

Rhythm is a major contributor to intelligibility (see Dellwo, 2010 for review). Beyond furthering the description of Kriol, our findings also have potential application in education and second language learning.

ABSTRACT. There has been increasing interest, across varieties of English, in like and its numerous functions (e.g. D'Arcy, 2007, 2008). Based on her extensive research, D’Arcy (2007) has analysed these into nine functions, exemplified here with examples from the Australian English data (excluding (5)): 1. verb: I like the Australian accent. 2. noun: Oh, for example the like. 3. adverb: It’s just like learning to ride a bike. 4. conjunction: I always feel like it's forced. 5. suffix: something stroke-like. 6. quotative complementiser: And they hear me swear and they're like, what was that? 7. approximation adverb: I've got like three or four options. 8. discourse marker (initial [or final]): Like it’s just amazing how much I don’t know. 9. discourse particle (medial): just `cause there's so many, like, people that have come over. Whilst it has been argued that further divisions are possible (Dinkin, 2016), the list and examples above account for like’s high frequency and potentially, via this, attention in folklinguistic comment. This comment particularly relates to the vernacular uses, in D’Arcy’s terms, of functions (6) to (9). In such comment, these uses of like are often seen as having no function other than making the speaker seem inarticulate and are associated with younger speakers and particularly young women (Cameron, 2007, pp. 149–150). Previous research in Australian English has examined (6), quotative uses, (e.g. Rodríguez Louro, 2013) and there has also been a corpus-based analysis of discourse uses as in (8) and (9) (i.e. Miller, 2009). Rodríguez Louro (2013) found the highest usage of quotative be like amongst younger participants but also males over females. Miller (2009) states that discursive uses are not restricted to young speakers but does suggest that usage may decline with age. It has recently been suggested that analysing like in its separate variables misses the ‘conspiracy’ of increased use of like across functions (Dinkin, 2016). This study therefore aims to examine the big picture of like in Australian English, allowing for the possibility of some shared meaning of like (aside from in the verb). This paper firstly overviews like usage, detailing the relative frequency of each form, and then explores if there is a relationship to age and/or gender for each function. It then focusses on uses (7) – (9), comparing findings to Miller (2009) where appropriate. Finally, it addresses if there is apparent support of the notion of interrelated increase in like across functions. The study is based on over 5,650 tokens of like from 87 Australian English speakers in sociolinguistic interviews. These participants represent four age groups: adolescent-, young adult-, middle-aged- and older participants. The study therefore contributes to understanding possible language change or age-based patterns in usage in Australian English, as well as providing the better understanding of like and its visibility.

ABSTRACT. Western Sydney and the Northern Beaches are two areas of Sydney that are geographically and socially distant. Western Sydney is highly diverse and historically maligned while the Northern Beaches is predominantly white and characterised as the ‘insular peninsula’. This study consists of interviews with 20 native speakers of Australian English between the ages of 16 and 26 years old from Western Sydney and the Northern Beaches (5 boys and 5 girls from each area). This study aims to determine whether there is a relation between a speaker’s area of origin (Western Sydney or the Northern Beaches), their use of like and the assignation(s) of social meaning to the form. This also known as indexicality (Silverstein 2003).

The interview had a unique structure designed for pairs of participants. The first section is a brief one-on-one discussion with the interviewer. The second section consists of the participant reading a word list and providing a running about a video for the purpose of prompting ING tokens. These two sections are done individually. The third section attempts to overcome the ‘observer’s paradox’ (Labov 1972: 113) by recording pairs of participants (who are familiar with one another) as they discuss a set of questions about specific features of Australian English without the interviewer present. One of these features is the perceived linguistic trademark of this generation: like.

While like has long been the object of linguistic research and analysis, studies have focused on its pragmatic/syntactic function(s) and its sociolinguistic distribution (Romaine and Lange 1991; Ferrara and Bell 1995; Tagliamonte and D’Arcy 2004). Although some studies have focused on speaker attitudes and ideologies regarding like they are limited in number and have been for varieties other than Australian English (D’Arcy 2007; Dailey-O’Cain 2000). This study indicates that like in Australian English has not only acquired social indexical meaning(s), some specific to the Sydney context, but that these meanings and perception of use do vary in relation to the area in which a speaker grew up.

Speakers from the Northern Beaches almost universally say they (and their friends) use like, regardless of gender. They also make a connection between Northern Beaches culture (relaxed, surfie, wealthy yet alternative) and the use of like. This appears to be an extension of the common association of like as spoken by Californians and Valley girls, in that the stereotypical Northern Beaches inhabitant is essentially a Valley Girl or a surfie. Speakers from Western Sydney in contrast explicitly distance themselves from like on the basis of its usage by ‘blonde prancy-ass girls’ and its association with ‘wealthy people’. Notably, the speakers from Western Sydney who did say they used like were all female. Despite these ideological differences, no discernible difference in speaker use of like has been identified. In fact, the speaker who used like the most in terms of frequency was a male from Western Sydney, while the two speakers who used like the least were from the Northern Beaches.

ABSTRACT. Variously labelled as a hedge, downtoner, adjustor (Aijmer, 2002), de-intensifer or magnitude stretcher (Zhang, 2011, 2013), sort of performs a range of functions in discourse. These functions are both evidential and interpersonal (Holmes, 1988; Aijmer 2002), including hedging and mitigating (Miskovic-Lukovic, 2009), as well as supporting rapport building (Aijmer, 2002). The last function is achieved via the notion that claims are de-intensified and the interactants understand one another so there is no need to be precise (Aijmer, 2002).

This paper examines the range of functions of sort of in Australian English, utilising discourse from the Australian Broadcasting Corporation’s television program Q & A from 2016 and 2017. These data are particularly valuable as the panel program contains discussion of ‘touchy’ topics, known to increase the use of ‘vague’ language (Zhang, 2013). Whilst Q & A’s format brings together opposing views, it has a studio audience and takes questions from the general public so there is a desire for respectful discussion. This context potentially leads to both a need to mitigate claims and the importance of rapport building. Indeed there is a high usage of this form: the corpus is close to 620,000 words and there are 481 total instances of sort of.

In this study, the analysis is restricted to uses from speakers of Australian English and pragmatic marker uses (as opposed to the ‘type of’ meaning) in the discussion of social issues. For these tokens, the paper overviews the frequency, syntactic position, discourse location, and scope of sort of. From this, and examining the instances in their context, it provides a schema which synthesises the described pragmatic functions. As sort of has not previously been the focus of research in Australian English, this allows comparison to research in American English (Miskovic-Lukovic, 2009), British English (Aijmer, 1984, 2002; Miskovic- Lukovic, 2009; Lin, 2010), Irish English (Kirk, 2015) and New Zealand English (Holmes, 1988; Stubbe & Holmes, 1995).

In sum, the paper provides a theoretically up-to-date understanding of the pragmatic functions of sort of in Australian English, contributing to our broader understanding both of this pragmatic marker across varieties of English, and pragmatic markers more generally.

ABSTRACT. Kaurna, the language of the Adelaide Plains, is an awakening language undergoing revival since 1989 (Amery, 2016a). Though little knowledge of Kaurna remains in the oral tradition and no sound recordings of the language as it was spoken in the 19^th century exist, a surprising number and range of emotion terms were documented. A great many of these involve the tangka ‘liver’ followed by kuntu ‘chest’, wingku ‘lungs’, yurni ‘throat’ and yurlu ‘forehead’, whilst mukamuka ‘brain’ and yuri ‘ear’ are involved in cognition. The role of pultha ‘heart’ is minimal. But these are not the only means to talk about emotions. Muiyu, a more elusive term, which may or may not be located in a body part and yitpi ‘seed’ are also central to emotions. These three terms tangka ‘liver’, muiyu and yitpi ‘seed’, appear to be viewed by Teichelmann and Schürmann (1840) and especially Teichelmann (1857) as seats of emotion. There also appear to be some stand-alone emotion terms such as turla ’anger’, marngu ‘jealousy’, wayi ‘fear’ etc, a range of verbs, many of which involve the above body parts, but others which do not, an endearment clitic, suffixes to express ‘inclination towards’ or ‘aversion’ and some exclamations to facilitate the expression of emotions.

In the context of re-introducing a reclaimed language, such as Kaurna, how to talk about emotions can become the topic of serious and sometimes unresolved debate. The title of a book of poetry (Proctor & Gale, 1997) ended up having two translations, one involving tangka ‘liver’ and the other pultha ‘heart’. Historical phrases expressing emotions are often co-opted for specific purposes in names, speeches, poetry and written texts.

This paper will discuss in detail the historical documentation, its interpretation and the ways in which this documentation is used today.

ABSTRACT. In this paper I compare 5 mono-morphemic emotion terms in the Arandic dialect-chain; and their prevalence in traditional songs. All of these terms are associated with feeling located in the stomach. Drawing on examples from speech, I suggest that there are three broad concepts, two of which are semantic equivalents or near equivalents restricted to the northern or southern lects, as follows:

1. arawerrnge                          ‘worry, love-sick; strong emotion that overtakes you’
2. irrare,  althere (N)              ‘longing, homesickness’
3. amperrnge, alhwarrpe (S)  ‘grief’

Arawerrenge ‘worried’, is an overwhelming emotion that makes it difficult to think of or do anything else. In other contexts arawerrenge can mean ‘power, spirit, vitality, life force’. Arawerrnge occurs in seven Arandic songs, many of which are of the ‘love song genre’(ilpentye). In some it is interpreted as referring to angry and strength in fighting; for others it is romantic or sexual love; and for others; confidence and power.

Irrare ‘homesickness’ typically refers to a feeling for one’s country or family. It occurs in six Arandic and Warlpiri land-based songs; but is not found in any love songs.

Amperrnge ‘grief’ also often refers to family; but also to animals and other people. It is a kin to empathy and compassion. Amperrnge occurs in eight songs. Interestingly. the more restricted equivalents althere ‘homesickness’ and alhwarrpe ‘grief’ do not occur in any songs.  Neither does a similar verb atham-arre- ‘to feel sorry for’.

One finding is that certain emotions concepts are associated with certain genres of song: arawerrnge ‘worry’ is more likely to occur in the love song genre, while irrare ‘homesickness’ in the land-based songs. I suggest that these three emotions may be associated with certain imagery in song and song genre; and raise the question of whether these images alone, may be enough to connote these emotions.