Short vowel placements in RP past and present: an acoustic and sociolinguistic study of the TRAP/STRUT configuration Anne H. Fabricius SCALPS1 Research Group, Language and Culture, Roskilde University [email protected] Exegesis: Queen Elizabeth II’s Christmas messages revisited Results 1 This study addresses diachronic change in the short vowel system of RP. While TRAP lowering and backing in RP over the 20th century has been reported previously (e.g. Wells 1982:291), the identification of STRUT’s movements has proven more difficult (see e.g. Bauer 1985). The present discussion compares RP vowel data from a range of published and unpublished sources and a range of age cohorts. It goes further than previous work in uncovering a possible interaction between the TRAP and STRUT vowels, whose positions vis-à-vis each other show variation across individuals and an identifiable trend of change in their relative alignment. The analysis identifies and names a ‘TRAP/STRUT rotation’ which seems to be the observed ongoing result of approximately half a century of TRAP backing and lowering, in response to which the STRUT vowel can be seen to have been in the process of rotating upwards into a midcentral position in the vowel space. Methods 1 • Interview corpus data originally taped in soundtreated room, Phonetics Lab, University of Cambridge in 1997/1998 •Digitized at 22kHz •Analysed using Speech Analyzer, with measurements taken of tokens of primary stressed vowels KIT, DRESS, TRAP, STRUT, LOT, and FOOT with stop or fricative syllable codas •F1 and F2 read off at points indicating the vowel’s main tendency •Additional data in Hertz from Wells (1962), Hawkins and Midgley (2005), and Deterding (1997) and converted to Hertz from Bark data in Harrington et al (2000) Angle measurements on standardised Excel vowel plots: Queen Elizabeth's short vowels; based on Harrington et al (2000) F2 in S Hawkins & Midgley data: Speaker number: Independent comparisons Source: Birth year: 1 2 3 4 5 pre 1928 -28 Deterding 1997 male h born 1928-36 (group 1) -13 10 8 6 5 5 Deterding 1997 male c -12 Queen recorded 1950s 0 Queen recorded 1960s -6 Queen recorded 1980s 6 Wells 1962 born 1946-51 (group 2) “Break group” 7 30 -12 -15 51 born 1961-66 (group 3) “Break group” 45 -15 -2 24 15 34 Cambridge male b. 1966 137 Cambridge male b. 1973 born 1976-81 (group 4) 44 80 37 70 41 56 Cambridge male b.1980 2 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 0,5 0,75 KIT FOOT 1 DRESS 1,25 F1 in S Introduction LOT 1,5 TRAP 1950's QEII 1960's QEII 1980's QEII STRUT 1,75 2 2,25 Table 1: The angle measurements grouped by age cohort The figure above shows a plot in S of the three sets of data from the Queen’s Christmas messages, based on Harrington et al. (2000:71) The chart above shows the data arranged into age cohorts based on Hawkins and Midgley’s (2005) four birth-year based divisions. The oldest and youngest age cohorts have majority patterns which are very different from each other, whereas the two middle groups exhibit a variational range (reminiscent of Hawkins and Midgley’s break groups).Independent support for the systematicity of the patterns shown by Hawkins and Midgley’s age cohorts can be seen in the patterns displayed by the data from Deterding’s (1997) two male speakers, the Queen’s Christmas message data (Harrington et al 2000), Wells’ (1962) data from speakers born in 1944 or before, and data from my own Cambridge interviews. It is significant that similar configurations can be traced across the quite different speech events in these data sets. Harrington, Palethorpe and Watson’s (2000) comparison of the Queen’s vowel formant values across real time concludes that significant changes in formant values on the F2 or F1 axis can be identified, suggesting that the Queen’s speech has gradually moved closer in phonetic terms towards a more mainstream and less upper-class RP through an opening of the vowel space. The figure above is a combined chart based on the figures given in Harrington et al. (2000:71). From a sociolinguistic perspective, however, the TRAP/ STRUT configuration does not show a radical ‘movement’ in the vowels’ positions, if we understand relative position as being within the individual’s short vowel system at any one point in time. The lines plotted on the graph take the example of the relative positions of TRAP and STRUT over the three decades in question. The angle of the line in each decade is similar to the minority pattern observed for the contemporary age group (Hawkins and Midgley’s group 1) described earlier in this paper. While the later TRAP and STRUT vowels are indeed lower and more back relative to the earlier ones, the juxtaposition of the two vowels at each period of time (whether it be the 1950s, 1960s or 1980s) remains similar. That is, in this respect, the Queen’s data fits with her age cohort in Hawkins and Midgley’s (2005) data, (and with a minority of speakers from subsequent generations) and it is not the case that there is an obvious radical realignment of TRAP and STRUT such as we see in some speakers in Hawkins and Midgley’s age groups born after 1946, and especially in those born after 1976, and in my own Cambridge data from younger speakers born after 1965. In conclusion, in configurational and sociolinguistic terms, at least, the Queen’s vowel data can be interpreted as stable and generationally typical, rather than radically variable and generationally untypical. Results 2 The configurational patterns seen diachronically Methods 2 Conclusions • All Hertz values normalised using the S-procedure (Watt and Fabricius 2002) • S-procedure is vowel extrinsic and formant intrinsic, thus fitting the “best-procedure” criteria given in Adank et al. (2004) • One normalised average value for each short vowel group plotted on individual Excel charts. • All charts scaled identically to enable comparability • On printouts of the charts, a line drawn from the TRAP average to the STRUT average was measured as an angle with reference to the horizontal. • Angle measurements provide the basis for comparison of different age cohorts • A ‘two-dimensional’ analysis reveals changing vowel configurations better than a ‘one-dimensional’ analysis • the TRAP/STRUT relative positions over time make sense as a coordinated ‘rotation’ movement 1)The ‘early triangular’ pattern (sp. born 1909) S1-3 Hawkins & Midgley 2005 F2 in S/ START 2 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 References 0,5 0,75 Adank, Patti, Roel Smits and Roeland van Hout. 2004. A comparison of vowel normalization procedures for language variation research. Journal of the Acoustical Society of America. 116, 5: 3099-3107. 1 1,5 KIT DRESS TRAP STRUT LOT FOOT F1 in S/ START 1,25 Bauer, L. 1985. Tracing phonetic change in the Received Pronunciation of British English. Journal of Phonetics. 13: 61-81. Deterding, David. 1997. The Formants of Monophthong vowels in Standard Southern British English Pronunciation. Journal of the International Phonetic Association. 27: 47-55. 1,75 2 Harrington, J, Sallyanne Palethorpe and Catherine I. Watson. 2000. Monophthongal vowel changes in Received Pronunciation: an acoustic analysis of the Queen’s Christmas broadcasts. Journal of the International Phonetic Association. 30, (1-2): 63-78. 2,25 Male b. 1980 Short vowel averages in S 2) The ‘quadrilateral’ pattern (sp. born pre WW2) F2 in S/ TRAP 2 1,8 1,6 1,4 1,2 1 0,8 0,6 Hawkins, Sarah and Jonathan Midgley. 2005. Formant frequencies of RP monophthongs in four age groups of speakers. Journal of the International Phonetic Association. 35,2: 183-199. 0,4 0,5 S4-4 Hawkins and Midgley 2005 F2 in S/ TRAP 0,75 2 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 Watt, Dominic and Anne Fabricius. 2002a. Evaluation of a technique for improving the mapping of multiple speakers’ vowel spaces in the F1~F2 plane. Leeds Working Papers in Linguistics. 9: 159-173. 0,5 1 1,5 KIT DRESS TRAP STRUT LOT FOOT Wells, J. C. 1982. Accents of English. Volume 1: An Introduction, Volume 2: The British Isles; Volume 3: Beyond the British Isles. Cambridge: Cambridge University Press. 1 1,25 1,5 1,75 F1 in S/ TRAP Angle +56 F1 in S/ TRAP 1,25 0,75 KIT DRESS TRAP STRUT LOT FOOT 1,75 2 2 2,25 2,25 3) The ‘late triangular pattern’ (sp. born after 1976) Figure 1: An example of the angle measurement method used on a chart of normalised average values for six short vowels. 1. SCALPS: Research group for Sociolinguistics, Conversation Analysis, Language Pedagogy and Sociocultural Issues Acknowledgements: Thanks for assistance and suggestions to David Deterding, Peter Roach, Simon Arnfield, Paul Kerswill, Eivind Torgersen, Dominic Watt, Paul Foulkes.