Sound change in 20th century RP: monophthongs 1901-1930

Postscript 27 Aug 2018: Added RP09.

Eight examples of 20th century RP (born 1901-1930) are examined for signs of some reported sound changes to RP monophthongs. This post follows on from the similar study for 19th century RP.

All the examples are distinguished people in their own fields – literature, theatre or science. They are anonymous here as some are still living, all personal information being withheld for the sake of privacy.

The monophthong sound changes are the same as for the previous period 1851-1900:

  1. RP STRUT was formerly [ʌ], nowadays [a] or [ɑ]. Fabricius (2007) studied this sound change in a set of published data from RP speakers born in the first half of the 20th century, and concluded it was complete by 1900 (everyone in her corpus had the new peripheral STRUT). All seven 19th century examples in the previous post had the open STRUT and the conclusion there was that this sound change had commenced at least by 1850, and was possibly complete by then. Consequently, all these RP examples born in 1901-1930 are expected to have the open STRUT.
  2. RP TRAP was formerly close to DRESS at the closer IPA [æ] (accompanied by compression of DRESS and KIT towards FLEECE), but is now an open [ae] (with DRESS and KIT spread out). Fabricius found that this change was already in progress in the early decades of the 20th century, both forms continuing side by side. Her oldest example was born in 1900, implying this sound change may have commenced before that. Of the seven 19th century examples in the previous post, just one had the older closer TRAP with DRESS and KIT compressed towards FLEECE, while all the others had the new open TRAP. It was concluded that this sound change had commenced in RP at least by 1850. Based on Fabricius’ evidence, these RP examples born in 1901-1930 are expected to include both forms.
  3. RP THOUGHT is described by Jones (1932) and Gimson (1962) as [ɔː]. But Gimson also reported a possible on-going RP sound change to [oː], supported by Wells (1982). Not one of the seven 19th century examples had anything like this sound change. They all had only the [ɔː]-like THOUGHT as described by Jones and Gimson. The RP examples born in 1901-1930 will also be checked for any signs of this sound change. It is important to be aware that regional Home Counties and LONDON SBE LOT and THOUGHT are [ɔ] and [oː] respectively, following sound changes that spread from 19th century Kent (Wood 2017 and here and here) and Essex. These accents are similar in many respects and easily confused. RP is  distinguished at least by LOT, THOUGHT, GOAT and MOUTH.
  4. RP GOOSE had a relatively dark timbre, although Jones noted an advanced variety following /j/, while Gimson noted increasing centralization. Henton (1983) reported higher instances of F2 since Wells (1962). The 19th century RP examples all had instances of darker [uː]-like GOOSE with F2 as low as 800-1000Hz, but at the same time the higher F2 limit extended to 1200-1300Hz, confirming Jones’ observation. These RP examples born in 1901-1930 will also be checked for any signs of more GOOSE fronting.
Fabricius, Anne H. 2007. Variation and change in the TRAP and STRUT vowels of RP: a real time comparison of five acoustic data sets. Journal of the International Phonetic Association 37:293-320.
Gimson, A. C. 1962. An Introduction to the Pronunciation of English. London, Arnold.
Jones, Daniel. 1932. An Outline of English Phonetics. Leipzig, Teubner. 3rd edition.
Henton, C. 1983. Changes in the vowels of Received Pronunciation. Journal of Phonetics 11:353-371.
Wells, J. C. 1962. A Study of the Formants of the Pure Vowels of British English. MA thesis, University of London.
Wells, J. C. 1982. Accents of English. Cambridge, CUP. Vol. 2.
Wood, Sidney. 2017. A spectrographic study of sound changes in nineteenth century Kent. 2017. In Tsudzuki, Masaki & Masaki Taniguchi (eds), A Festschrift for Jack Windsor Lewis on the occasion of his 90th Birthday 215-246, Journal of the English Phonetic Society of Japan 21.

All the recordings were found on the web, mostly from the BBC program archive or the British Library. All had been subjected to lossy MP3 compression at some time during their lives and were saved at once in WAV format to prevent further degradation. They were analysed using Praat, by LPC formant tracking where possible, by measuring directly on the spectrogram where that failed, or by measuring formant peaks on FFT slices. Some instances of lower back vowels had to be abandoned and there are consequently fewer examples of FOOT, THOUGHT or LOT in some cases. Vowel formant frequencies were collected from fully focused syllables, to minimize spectral variation due to vowel reduction. Vowels were sampled at the moment where vowel articulation was least affected by surrounding consonant articulations, selected by comparing VC and CV formant transitions, to minimize spectral variation due to coarticulation. Instances for each phoneme are enclosed in ellipses, fitted freehand, ignoring outlyiers (that are plotted and can be seen).

The spectral criterion for an open vowel is F1 higher than about 600 or 650 Hz, depending on the individual speaker, even higher for shorter vocal tracts. This would be achieved with a low pharyngeal constriction that is maximally wide together with a large mouth-opening (Wood 1979).

Wood, Sidney. 1979. A radiographic analysis of constriction locations for vowels. Journal of Phonetics 7, 25-43.

Speakers born in 1901-1910:

Figure 1. F1/F2 diagram for RP14 born 1901-1910; the arrow
shows one instance of an [ɛæ]-like diphthong for TRAP.

Figure 2. F1/F2 diagram for RP15 born 1901-1910; the arrow
shows one instance of an [ɛæ]-like diphthong for TRAP.

Figure 3. F1/F2 diagram for RP03 born 1901-1910. Recording (a).

Figure 4. F1/F2 diagram for RP03 born 1901-1910.
Recording (b) five years later than Figure 3.

Speakers born in 1911-1920:

Figure 5. F1/F2 diagram for RP16 born in 1911-1920.

Figure 6. F1/F2 diagram for RP17 born 1911-1920; the arrow
shows one instance of an [ɛæ]-like diphthong for TRAP.

Speakers born in 1921-1930:

Figure 7. F1/F2 diagram for RP04 born 1921-1930.

Figure 8. F1/F2 diagram for RP19 born 1921-1930.
Figure 9. F1/F2 diagram for RP09 born 1921-1930.

General comments on Figures 1-9

All seven are definitely examples of RP (LOT is [ɒ] partly overlapping or very close to  BATH, while THOUGHT is [ɔː], but see also the discussion on THOUGHT below). There is no other accent of British English that has this particular combination of vowel timbres.

There are two recordings for RP03 (Figures 3-4) made five years apart. They are very similar.

Example RP04 (Figure 7) stands out because his F1 frequencies are generally lower than expected, and hardly extend to 700Hz at most. This is possibly due to a less energetic style of delivery, resulting in smaller mouth openings and generally lower F1 (the extreme case of this would be mumbling, but RP04’s speech was still clear). One consequence of this is that his open vowel boundary is shifted to 550Hz. An alternative explanation might be that his vocal tract was longer than average, but his F2 for FLEECE (2250-2450Hz) suggest instead that his vocal tract was shorter than average. In contrast, RP03 (Figures 3-4) appears to have stretched his F1 range to 900 Hz, although his F2 range for FLEECE was low (1800-2200Hz) reflecting a longer vocal tract. This is presumably an example of energetic delivery with larger mouth-openings generally and higher F1 frequencies.


All the examples had high F1 and an [a-ɑ]-like timbre, confirming Fabricius’ conclusion that this sound change was complete by 1900. There is possibly a second sound change here from a brighter [a]-like timbre (F2 roughly 1250-1500Hz, Figures 2, 5, 8-9) to a darker [ɑ]-like timbre (F2 roughly 1000-1250Hz, Figures 1, 3-4, 6, 7). Only one informant had the darker variety in the 19th century examples, the youngest born in the 1890s. Now, this group born in 1901-1930 was divided equally between the two timbres. This is something to look out for in other more recent examples of RP.

Listen to the brighter [a]-like STRUT by RP15 (Figure 2):
“much month other sun suddenly subject luck subject something”

Listen to the brighter [a]-like STRUT by RP16 (Figure 5):
“husband result some discovered young another discuss”

Listen to the brighter [a]-like STRUT by RP19 (Figure 8):
“muscle suddenly discovery”

Listen to the brighter [a]-like STRUT by RP09 (Figure 9):
“number destruction repercussions production suddenly done”

Listen to the darker [ɑ]-like STRUT by RP14 (Figure 1):
“trump become done touch doesn’t”

Listen to the darker [ɑ]-like STRUT by RP03 (Figure 3):
“does love guns glove tough blunt”

Listen to the darker [ɑ]-like STRUT by RP17 (Figure 6):
“others up other”

Listen to the darker [ɑ]-like STRUT by RP04 (Figure 7):
“lucky but countryside”


Two of these seven examples had the earlier closer TRAP complete with compression of DRESS and KIT towards FLEECE. They are RP15 and RP19, Figures 2, 8. F1 for compressed DRESS is usually about 400-500Hz, so that the FLEECE-DRESS F1 range is shortened to about 250Hz. F1 for uncompressed DRESS is about 500-600Hz and the FLEECE-DRESS F1 range is larger, about 350Hz. Previously, in the 19th century group, there was just one example who fulfilled all the criteria for closer TRAP. Like him, these two also had both closer instances of TRAP with F1 roughly 500-600Hz and open instances with F1>600Hz.

Additionally, three had occasional instances of a diphthong [ɛæ], not previously found in the 19th century group. This diphthong was never mentioned by Jones, but it is described by Wells (1982, vol. 2). The three examples are RP15 (Figure 2, together with close TRAP) and RP14 and RP17 (Figures 1, 6) with open TRAP. Based on these examples, there was no obvious systematic rule for producing a diphthong or a monophthong.

Listen to close TRAP by RP15 (Figure 2), ordered by rising F1 from close to open [æ]:
Close [æ]: “Cavendish mathematics natural scattering sabbatical mechanics”; Open [æ] “that planet hand”

Listen to close TRAP by RP19 (Figure 8), ordered by rising F1 from close to open [æ]:
Close [æ]: “that understanding”; Open [æ] “action flashes”

Listen to open TRAP by RP14 (Figure 1):
“fascinated natural patchy packed pack family understand”

Listen to open TRAP by RP03 (Figure 4):
“background valuable man fashion fantastic passage fascinating”

Listen to open TRAP by RP16 (Figure 5):
“family caterpillar enthusiastc accident perhaps badly”

Listen to open TRAP by RP04 (Figure 7):
“characters admirable attitudes chat acts”

Listen to open TRAP by RP09 (Figure 9):
“tragic famines can charismatic animals”


All these examples have the expected RP [ɔː] for THOUGHT, but they also show some variation extending towards [oː]. Roughly speaking, an [ɔ]-like timbre would have F1 around 500-600Hz, and [o] around 350-450Hz. Three examples in particular have instances close to 400Hz: RP16, RP17 and RP19 (Figures 5-6, 8), all in the younger half of this group. None of these is what Gimson was describing in (1962), an RP change to [oː]. But there was a possible tendency starting in the early 20th century to extend the THOUGHT variation towards [oː] while still anchoring [ɔː].


A hallmark of earlier RP was a dark [u]-like GOOSE, with F2 below 1000Hz. Only one example in this group has all GOOSE F2 below 1000Hz, RP14 (Figure 1), the oldest member of the group. A second type has some F2 below 1000Hz and some above (RP03, RP17, RP19 and RP09, Figures 3-4, 6-9). Finally, a third type has no instances with F2 below 1000Hz (RP15, RP16, RP04, Figures 2, 5, 7). These three types are themselves a measure of GOOSE fronting in RP, the gradual loss of dark [u] allophones. A further measure is the highest GOOSE F2: 1800Hz (RP16, Figure 5), 1500Hz (RP03, RP04, RP09, Figures 4, 7, 9), 1400Hz (RP19, Figure 8), 1200-1300Hz (RP15, RP17 Figures 2, 6), representing the gradual acceptance of fronted [u]-like allophones into RP in the early 20th century.

This entry was posted in Accents, Articulation, English, Pronunciation, RP, sound change, Vowels and tagged , , , , , , , , , . Bookmark the permalink.