I was listening to a recording of a story in Welsh the other day and it occurred to me that one of the sounds I was hearing doesn’t really have an accepted phonetic description. This is the sound which is written with the digraph <ch>.

This sound is usually labelled as a voiceless uvular fricative and has the phonetic symbol χ. The sound produced by the speaker in the recording was certainly voiceless, uvular and fricative, but the label misses something important about the sound I was hearing. I’d better not use a clip from the recording for copyright reasons, but below is a sentence in Welsh spoken by yours truly in which I have done my best to produce the type of fricative the Welsh speaker used. The sentence means What is your name? Click on the sentence to hear my utterance.

Beth ydy’ch enw chi?

The only term I have ever heard to describe this particular sort of χ is scrapy, but I don’t think I have ever seen the term in print. I heard the term many, many years ago from my former teacher and colleague John Baldwin. It’s perhaps not surprising that the IPA chart makes no mention of scrapiness. It’s a pretty fair bet that no language contrasts scrapy and non-scrapy fricatives. The ExtIPA chart also seems to be silent on the matter.

Here is my take on scrapiness for what it’s worth. Given the evidence of the speech pressure waveform below, which is taken from the first <ch> sound in my recording, scrapiness involves some sort of quasi-periodicity. My proprioceptive impression is that something is vibrating and it is not the uvula itself, as would be found in a voiceless uvular trill . I think the scrapy fricative is a laxer articulation than its smoother counterpart and that it requires a stronger airflow.

7 thoughts on “Scraping

  1. Luciano Canepari calls these constrictive trills. In case you’re interested download the PDFs titled Consonants and contoids (1) and …(2) in the Natural Phonetics & Tonetics (A Handbook of Phonetics) section of this webpage.

    These sounds seem the typical realisations of /x/ in Dutch, and many Germans (including myself) also use them.

    Occasionally voiceless approximant trills [ʀ̥] also occur in German, though usually not as realisations of /x/, but as allophones of /r/ after voiceless obstruents.

  2. There’s anecdotal evidence that south Walian speakers are less scrapy (maybe even velar) in those fricatives. Do you know whether the speaker you listened to was northern or southern?

  3. Doesn’t copyright allow you to quote, with the customary source acknowledgment.

    Regarding uvular consonants, the action is in the upper pharynx. For a fricative like [χ], the back of the tongue forms a constriction against the posterior pharyngeal wall. For a stop like [q] you have occlusion there instead. For a trill like [ʀ] you probably have a Bernoulli membrane vibration, the ideal conditions are at least available. I have X-ray motion films of uvular consonants in Inuit and South Swedish.

    I’ve never seen an explanation of how you get an occlusion against the uvula, there are open air passages either side. I’ve heard believers suggest the tongue squashes the uvular flat and thin for stops. Nor have I heard where enough energy comes from for a wobbling uvula to be heard as a trill.

  4. Martin, Sydney

    Thanks for the info. What I was listening to can be found on the itunes store, where one can listen to sample clips. If you’re interested, search for Cyfres Pen Dafad and listen to the first clip (Jibar Pennod 3).

    I have listened a number of times, but can’t spot any obvious clues as to whether the speaker is from the North or the South.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.