What is the extent of /r/-colouring on Dutch vowels?

This project investigates the acoustics of vowels that are followed by /r/ in Dutch. It is well-known that postvocalic /r/ has a ‘colouring’ effect on a preceding vowel; an effect that is important for speech perception, since many speakers of Dutch articulate postvocalic /r/ itself very weakly. This means that minimal pairs such as poot ‘leg’ and poort ‘gate’ are often mainly distinguished by the quality of the vowel. The literature suggests that the nature of /r/-colouring differs from vowel to vowel: for some vowels, it is clear from the start that /r/ will follow, whereas for others, this only becomes obvious towards the end of the vowel.

This project aims to provide a detailed insight into the nature and extent of /r/-colouring in everyday spontaneous Dutch speech. It will use large numbers of instances of each vowel of Dutch, with and without a following /r/, drawn from a corpus of spontaneous Dutch speech. It will generate time-normalised average formant tracks for each vowel (see right, with grey lines representing formants for a vowel without a following /r/, and black lines formants for the same vowel with a following /r/), and run statistical tests on these tracks to establish how far through the vowel the presence or absence of a following /r/ can be reliably predicted.

Relevant fields of linguistics:

• Corpus linguistics
• Phonetics

The RES student does not need knowledge of Dutch for this project.

Start date: negotiable
End date: 6 weeks later

Hours per week: 3

The work could also be carried out over a shorter period if necessary partly on a private computer.

Activities to be undertaken by RES student:

•  For a number of Dutch vowels, run searches in the Spoken Dutch Corpus to identify comparable words with and without postvocalic /r/
•  Process the corpus search results in Excel, Notepad++ and SPSS to produce spreadsheets that can be used for further data selection.
• Extract selected audio files from the Spoken Dutch Corpus using Praat.

Number of possible students: 2

Likely outcomes:

• Expertise in using Corex corpus software, including familiarity with regular expression searches
• Expertise in the use of Excel, Notepad++ and SPSS for data processing, including familiarity with conditional formatting, data filtering, macros and frequency analysis
• Expertise in the use of Praat for audio file management, including familiarity with scripts
• An insight into the data collection phase of the research process

Students who have questions about this project should contact Dr Leendert Plug on [email protected]

If you would like to work on this project, please complete an RES application form. If your application is successful, the supervisor will be in touch in due course.