Introduction

The preceding chapters dealt with the behaviour of moraic onsets with regard to stress (Ch. 2), compensatory lengthening (Ch. 3) and word minimality (Ch. 4). All these phenomena provide evidence to a varying degree for the existence of coerced moraic onsets, i.e. onset weight enforced on the surface. The foremost of the three is of course stress. Recall that stress systems sensitive to the quality of the onset exhibit certain and predictable patterns according to which onsets may be moraic, as discussed extensively in §1.3.3, §2.2, §2.4–5. The word-minimality facts relating to onsets in Bella Coola seem to work as predicted by the coerced onset weight conditions (§1.3.3.5 and Ch. 4), although the data currently available are too scarce to state this with certainty. Finally, the analysis of compensatory lengthening was modelled in such a way so that it is framework neutral. More concretely, it is compatible with both the underlying presence and the absence of moras on constituents that eventually get to be syllabified as onsets on the surface, in other words, it does not care about input onset moraicity. Consequently, CL does not provide much evidence for either coerced or underlying onset weight (§3.1).

This chapter clearly shifts its interest to consonants that are underlyingly moraic, i.e. /Cμ/, and that on the surface constitute geminates (for the difference between geminate and geminated consonants, see the discussion in §1.3.2 and fn. 11 of this chapter for completeness).