Tashelhiyt permits any segment to act as a syllable nucleus, regardless of sonority. There's lots of theoretical analyses out there, but descriptively, Tashelhiyt consonant syllabification moves left to right in a word, targeting the most sonorous segment first, assigning it an onset (if possible), and then syllabifying the next-most sonorous segment, unless it would create adjacent syllabic segments. Anything left over that can't be syllabified is a coda. The sonorant hierarchy for Tashelhiyt goes like:
low vowel > high vowel > glide > liquid > nasal > voiced fricative voiceless fricative > voiced plosive > voiceless plosive
For example (from Dell & Elmedlaoui 1985):
/rajrˤz/ [ra.jr̩ˤz] 'it will be broken'
/ʜawltn/ [ʜa.wl̩.tn̩] 'make them(m.) plentiful'
/tlbʒt/ [tl̩.bʒ̍t] 'you(sg.) stepped onto'
/tzmt / [t̩.zm̩t] 'it(f.) is stifling'
If multiple eligible candidates are of the same sonority, the leftmost one generally wins:
/rksx/ [r̩.ks̩x] (**r̩k.sx̩) 'I hid'
I found some examples in an acoustic study (Ridouane 2008). Most of the syllabifications given perfectly match the description above, but some of the forms deviated concerning /f/; when syllabification has the chance to target /f/ or a another fricative, it will apparently always choose /f/, even if the other fricative is the leftmost candidate, even though both are voiceless fricatives and apparently of the same sonority:
/tkʃf/ = [tk̩.ʃf̩] (not [t̩.kʃ̩f]) 'it dried'
/sfqqst/ = [sf̩q.qs̩t] (not [s̩f.qːs̩t]) 'irritate him'
I want to know if this phenomenon arises because /f/ is more sonorous than /s/, or if there's some other explanation. Perhaps when presented with otherwise equally eligible candidates, Tashelhiyt will first choose the leftmost one, unless it would create a word-initial onsetless syllable, in which case it will choose to syllabify the rightmost one. That seems to be partly true--that's probably why /tkʃf/ is [tk̩.ʃf̩] instead of [t̩k.ʃf̩]. But this doesn't explain another example given, /sxf/ [s̩.xf̩] 'fade away'; if the syllabification where instead [sx̩f], that would avoid an onsetless syllable. Instead, [s̩.xf̩] seems to show /f/ being syllabified before /x/, suggesting that there is also a sonority hierarchy where /f/ is more sonorous than /x/ as well as /s/.
This seems counterintuitive to me--I'd think sibilants would be more sonorous than /f/, not less. Any insight into this phenomenon would be appreciated greatly!
