Leeds Working Papers in Linguistics and Phonetics
ISSN 1747-9339
Edited by Barry Heselwood and Leendert Plug
CONTENTS
- Editorial preface
- Valentina Brunetto
Syntactic diagnostics for referentiality marking in early null objects: Evidence from Italian
abstract - full text - pp.1-13 - Rehab Gad
The role of illi 'that' in Egyptian Arabic
abstract - full text - pp.14-39 - Katie Harris
Native English speakers' production of Italian /t/: The extent of phonetic learning in adult second language acquisition and the effect of speaker input
abstract - full text - pp.40-73 - Manel Herat
A study of child-directed speech in a Sinhala-English bilingual household
abstract - full text - pp.74-91 - Fumio Inoue
Gravity model of diffusion for Tokyo new dialect forms
abstract - full text - pp.92-100 - Leendert Plug
Phonetic correlates of postvocalic /r/ in spontaneous Dutch speech
abstract - full text - pp.101-119
The Leeds Working Papers in Linguistics and Phonetics is a series produced by the Department of Linguistics and Phonetics at the University of Leeds. Its aim is to publicise ongoing research by staff and students of the Department. We are also pleased to include contributions from colleagues in other departments at the University of Leeds, from other institutions and from visiting scholars who have made presentations in our research seminar series. All submissions have been subject to peer review.
The contributions in this volume cover syntax (Brunetto, Gad), L2 phonetic learning (Harris), bilingual child-directed speech (Herat), dialect diffusion (Inoue) and acoustic phonetics (Plug). Languages represented are Italian, Egyptian Arabic, Sinhala, Japanese and Dutch.
Many thanks to all the contributors and reviewers, and to Miho Kamata for assistance with figures.
Barry Heselwood and Leendert Plug (Editors)
Syntactic diagnostics for referentiality marking in early null objects: Evidence from Italian
Valentina Brunetto
This paper discusses the properties of object omission at the early stages providing evidence that children acquiring Italian are sensitive to the universal syntactic conditions for marking referentiality on null objects. Data are gathered from an elicitation task administered to three, four and five year-olds prompting a highly contextually salient object, felicitously expressed as a clitic. From the asymmetries detected in comparison between matrix and adjunct clauses, some speculations are drawn. Firstly, children may not be failing in interpreting referential objects at the syntax-discourse interface, but would rather appear to favour an operator-variable construction universally, which competes with cliticization in terms of computational economy. Furthermore, syntactic tests prove valid diagnostics to discriminate between different derivations of null objects, casting more light on the complex typology of null arguments and the role of the features associated with them.
Back to Table of Contents
The role of illi 'that' in Egyptian Arabic
Rehab Gad
This paper investigates the role of illi in the grammar of one of the colloquial dialects of Arabic; that is Egyptian Arabic (EA). It investigates how illi affects the formation of wh-questions (with initial and in-situ wh-phrases) and relative clauses. Since the classification of illi has been a subject of debate in the literature, the study aims at providing a new analysis for it. The major claim is that illi belongs to the class of functional categories which serves the grammatical function of a relative pronoun. This paper presents data where illi acts as both a relative pronoun and a licensor for wh-fronting. The following questions are addressed:
(1) If illi is analysed as a relative pronoun, how can we account for its occurrence in an initial position within some wh-questions without having to propose a movement analysis?
(2) Can illi be classified as a complementizer that shares some syntactic properties with the complementizer inn 'that'? (3) Within wh-questions, does illi behave as a question particle?
(4) How can we account for the EA data where illi has the dual function of a relative pronoun and a complementizer? The major claim is that illi does not belong to the class of question particles which mark a yes/no question and a wh-question. Though illi and inn 'that' occur as C elements equivalent to the English 'that', illi does not exhibit the morphological or the functional properties of inn 'that', hence it cannot be classified as a complementizer. Within wh-questions, the position of illi depends on the position of the argument wh-phrases which can either occupy the Spec CP position when followed by illi, or remain in situ with illi in an initial position. I conclude that illi shares the structural position of complementizers while carrying out the grammatical function of relative pronouns.
Back to Table of Contents
Native English speakers' production of Italian /t/: The extent of phonetic learning in adult second language acquisition and the effect of speaker input
Katie Harris
This study examines the production of Italian /t/ by native English participants who began acquiring Italian as adults. It investigates the extent of phonetic learning in Italian /t/ for individuals who have been living in Italy for many years (late bilinguals) and the effect of quantity of native speaker input on phonetic learning in Italian /t/ for the late bilinguals and three groups of students who differ in terms of amount of received exposure to native Italian speech. In addition, it investigates whether phonetic learning in Italian /t/ coincides with modifications to the way in which English /t/ is produced. Voice-onset time (VOT), burst amplitude and closure duration were measured in English and Italian monolingual productions of /t/ in order to establish phonetic norms. Only VOT exhibited enough variation to be used as an indicator of phonetic learning. Late bilinguals showed partial phonetic learning in the form of compromise values which were intermediate to English and Italian monolingual VOT. Phonetic learning in student realisations of Italian /t/ increased in line with amount of received native speaker input. Contrastingly, the late bilinguals did not display the most phonetic learning, despite having received the most native speaker input. It was proposed that native speaker input influences phonetic learning in the early stages of acquisition, but that phonetic learning for the late bilinguals had fossilised. In the later stages of acquisition, age of learning (AOL) was deemed to be a more accurate predictor of phonetic learning than input. Neither the late bilinguals, nor the students exhibited modifications to VOT in their English realisations of /t/ which may have occurred as a result of a lack of plasticity in the L1 phonetic system, or a high level of L1 use in participants' daily lives.
Back to Table of Contents
A study of child-directed speech in a Sinhala-English bilingual household
Manel Herat
This study investigated parents talk to a young child in a Sinhala/English bilingual household in Sri Lanka. The household consists of two children but the recorded data mainly consists of interactions between the parents and one child. I was interested in exploring the type of language used in speaking to children in a bilingual household where English is the mother's second language. Using spontaneous language data collected by the parents over one month, the language used by the mother, especially Sinhala child directed speech and the type of grammatical features and phonological modifications made to English were examined. The mother's language was important in determining the potential influence of the input on the output of the children. The findings suggest that there is a tendency to use code-switching when the children switch to Sinhala. Although at times the father appears to encourage a more monolingual context, overall, code-switching is seen as a positive phenomenon.
Back to Table of Contents
Gravity model of diffusion for Tokyo new dialect forms
Fumio Inoue
In this paper it will be advocated that a simplified gravity model is applicable to linguistic change (lexical change) in progress. New dialect forms are born and are spreading all over Japan even in this age of standardization. 'Tokyo new dialect' exhibits new dialectal forms recently introduced among young people in Tokyo. A new kind of database of geographical distribution of the Tokyo new dialect forms has been constructed. The result shows an inverse correlation between usage rate and distance from Tokyo. After making a congregate map, a correspondence analysis was applied and the whole pattern of geographical distribution was analyzed. The large city of Osaka has shown a great influence on this Tokyo new dialect. Thus, it is clear that the 'geographical proximity effect' has basically influenced the quantity of words introduced into Tokyo. The whole picture is clearly explained when population density is also taken into consideration, and when the simplified gravity model of geographical diffusion is applied. All the processes of geographical diffusion to and from Tokyo fit with the 'umbrella model' of diffusion which is proposed for new dialect formation and standardization in Japan. This umbrella model should be applicable to many languages of the world.
Back to Table of Contents
Phonetic correlates of postvocalic /r/ in spontaneous Dutch speech
Leendert Plug
This paper reports on a study of the realisation of postvocalic /r/ in stressed syllables followed by /t/ or /d/ in Dutch. Two recurrent observations regarding /r/ in this context are that /r/ can be 'deleted', and that /r/ has an effect on the quality and duration of a preceding vowel and a following alveolar consonant. Most recently, Plug and Ogden (2003) have presented auditory and acoustic analysis of controlled, read speech by four speakers of Standard Dutch, focussing on /r/'s segmental status and non-segmental correlates. This paper presents results of an attempt to replicate Plug and Ogden's study using uncontrolled and unscripted speech by 20 speakers. One might expect that in spontaneous speech, 'deletion' of /r/ is relatively common. The main question addressed in this paper is whether the non-segmental correlates of /r/ described by Plug and Ogden (2003) are as robustly observed in spontaneous speech as they are in laboratory speech. The findings suggest that this is not the case, although a subset of Plug and Ogden's correlates is clearly observable.
Back to Table of Contents