A short sequence from an X-ray motion film: Uvular [yʀa]

This is a short sequence where a speaker of Southern Swedish is saying [ˈyʀa], broken out from a longer sequence, /ˈfy:ra/, [ˈfəyʀa], fyra, ‘four’. This example is illustrated with every other film frame, so there is 26ms between each picture. However, the movements were originally analysed from every frame in the sequence.

This was originally presented in 2004 at the Sound to sense symposium honouring Ken Stevens (PDF here) and in 2006 at the annual Swedish phonetics conference (PDF here).

The standard description of Swedish /r/ is coronal, a retroflexed trill or fricative. In the south, however, roughly south of a line from Varberg on the west coast to Oskarshamn in the east, it’s a uvular [ʀ], rarely trilled, usually fricative or approximant. The example illustrated here is a uvular approximant, for which the IPA alphabet lacks a dedicated character.

The standard account of the production of uvular consonants is unfortunately misleading. Direct visual inspection of articulation is restricted posteriorly by the faucial arches and the oropharyngeal isthmus, hence the traditional focus on the visible velum and uvula as the place of articulation for uvular consonants. However, X-ray motion films like this one show that the critical constriction is located in the upper pharynx, at the level of the superior pharyngeal constrictors. The tongue body was moved directly from the hard palate into the upper pharynx, without deviating towards the uvula, and without involving the uvula in any way. The name uvular is retained now for the sake of tradition, but the constriction is located in the upper pharynx.

The sequence illustrated here starts where the speaker has just reached the palatal posture for /y/ and ends where the speaker has just reached the lower pharyngeal posture for /a/. In between, he moved the tongue body from the palatal /y/ posture to the uvular /r/>[ʀ] posture in the upper pharynx, while withdrawing the lower lip from rounding for /y/. Towards the end he depressed the mandible and moved the tongue body into the lower pharynx, both for /a/, while continuing to withdraw rounding and larynx depression of /y/.

Each articulator gesture has been assigned to its phoneme, and colour coded for ease of identification: red for /y/, blue for /r/>[ʀ], and orange for /a/.

• Picture 1: The tongue body gesture towards the hard palate for /y/ is just arriving, lip rounding for /y/ continues.
• Picture 2: The palatal tongue body posture for /y/ is still being held, lip rounding for /y/ continues.
• Picture 3:  The palatal tongue body posture for /y/ is still being held, lip rounding for /y/ is just complete.

• Picture 4: The uvular tongue body gesture into the upper pharynx for [ʀ] has just commenced; the withdrawal of lip rounding from /y/ has just commenced.
• Picture 5: The uvular tongue body gesture into the upper pharynx continues for [ʀ]; withdrawal of lip rounding from /y/ continues; mandible depression for /a/ has just commenced.
• Picture 6: The uvular tongue body gesture towards the upper pharynx has just arrived for [ʀ]; withdrawal of lip rounding from /y/ continues; mandible depression for /a/ continues.

• Picture 7: The tongue root gesture into the lower pharynx for /a/ has just commenced; mandibular depression for /a/ continues; withdrawal of lip rounding from /y/ continues; withdrawal of larynx depression from /y/ has just commenced.
• Picture 8: The tongue root gesture into the lower pharynx for /a/ continues; mandibular depression for /a/ continues; slight opening of the velo-pharyngeal port has just commenced for /a/; withdrawal of lip rounding from /y/ is completed; withdrawal of larynx depression from /y/ continues.
• Picture 9: The tongue root gesture into the lower pharynx for /a/ continues; mandibular depression for /a/ continues; slight opening of the velo-pharyngeal port continues for /a/; withdrawal of larynx depression from /y/ continues.

• Picture 10: The tongue root gesture into the lower pharynx for /a/ concludes; mandibular depression for /a/ concludes; slight opening of the velo-pharyngeal port for /a/ concludes; withdrawal of larynx depression from /y/ continues.

Coarticulation is seen in various ways. For example, there are bits and pieces of /y/ scattered throughout this sequence from picture 1 to picture 10, each executed correctly and with precision, ingeniously interwoven with movements belonging to other phonemes, to avoid conflicting instructions to the articulators. That sort of orderly behaviour calls for careful organisation with predictive preplanning. There are none of Hocket’s randomly broken eggs here.

Alternatively, see how many phonemes have simultaneous activity in any one picture. For example, in some pictures there is activity for two phonemes. At most, there is simultaneous activity for three phonemes in pictures 5-6. Only the stressed vowel /y/ has a brief moment in pictures 1-3 when there was only its own activity.

Faced with this organisation of articulator gestures, one might wonder what, for example, the expression vowel duration means. Traditionally, we measure the duration of the vocoid segment on say a spectrogram, corresponding to pictures 1-6 here for /y/, or /y:/ since this vocoid segment is also assumed to include the quantity of the vowel. Six pictures here correspond to 12 film frames, so that duration is about 160ms plus a few frames before picture 1 for the CV transition after the initial /f/ hiss (not shown here). On the other hand, the duration of the interval when the speaker was busy articulating /y:/ was all ten pictures, 260ms plus those extra frames before picture 1.

Subsequent frames (after picture 10, not included here) show withdrawal of the /a/ gestures and a breathing pause.

Top

© Sidney Wood and SWPhonetics, 1994-2014

_______________________________________________________________