The dataset contains the F1 and F2 values (in Hz) of each vowel phoneme in the speech of 66 speakers from Manchester, representing different genders, generations, social classes and ethnicities, i.e. White British, Pakistani, and Black Caribbean. The goal of the project is to elucidate of the mechanisms of language change and dialect differentiation by examining changes currently in progress in Manchester's vowels. This will result in the first systematic description of Manchester's accent, and will provide the basis for comparisons with other large urban centres of the UK. The study investigates the patterns of variation within the community, ie, the ways in which Mancunians' speech varies depending on age, gender, social class, neighbourhood, and ethnicity. It is based on the acoustic analysis of 66 speakers representing different generations, genders, social classes, and ethnicities, i.e. White British and Pakistani and Black Caribbean. The informants' speech is analysed acoustically in terms of F1 and F2 (i.e. first and second formant), corresponding to the position of vowels in phonetic space. The measurements obtained through computer-assisted acoustical analysis are subjected to a range of multivariate statistical analyses. Their results will help to assess the role of extralinguistic factors, such as ethnicity, gender, socio-economic status, neighbourhood, and age, in the creation of the linguistic variation found in Manchester.

The dataset comprises the acoustic measurements of vowel phonemes for 66 speakers of Manchester English. For the purposes of the study, Manchester is defined as the area within the M60 motorway and areas immediately south of the M60, i.e. Sale, Wythenshawe, and Stockport. The speakers represent different generations, genders, social classes, and ethnicities, i.e. White British, Pakistani, and Black Caribbean. The speakers were found by student assistants local to the area through their personal networks; the speakers selected for the study must have been born and raised in Manchester, as defined geographically above. They were interviewed using the classic method of data collection used in sociolinguistics, first developed by William Labov in the 1960s, the so-called sociolinguistic interview. This is an unstructured interview, that is, a free-flowing conversation intended to elicit speech that is as relaxed and as informal as possible. The general topic of the conversation was growing up and living in Manchester, and the specific topics depended on whatever each speaker found most interesting and was willing to talk about, e.g. the games they played when they were young, their most memorable experiences growing up, their school experiences, their experiences with football in Manchester, etc. For more information about this method of data collection, see e.g. “The sociolinguistic interview”, Chapter 3 of Tagliamonte, S. 2006. Analyzing sociolinguistic variation. Cambridge: Cambridge University Press. The spontaneous speech part of the interview was followed by the reading of a word list including different vowels of English. The first part of the word list included single words, arranged in four columns, read out by the informant. The second part, called “Minimal pairs”, included pairs of words sounding similar or identical in different dialects of English; the speaker read out each pair and said if the words sounded the same or different to them. Copies of both parts of the word lists are attached. The interviews were recorded on Sony PCM-M10 recorders as uncompressed wav files, at a sampling rate of 44.1kHz and a depth of 24 bits. The sound files, i.e. the informants’ vowels, were then analysed acoustically in Praat, a computer program commonly used for the acoustical analysis of speech, available online at www.praat.org Copies of publications arising from the project will be accessible on the project website as they become available: http://mlm.humanities.manchester.ac.uk/resources/manchester-accent-project/