In this chapter I consider a range of phenomena involving the laryngeal contrast between obstruents in Bothoa Breton. These phenomena provide the strongest support for a substance-free model of the relationship between phonological representations and their phonetic realisation. In section 7.2.1, we saw that the phonetic realisation of the contrast between [p t k] and [b d ɡ] in Bothoa Breton agrees with languages such as French or Russian, in that the ‘fortis’ series of stops is realised with short-lag VOT (as ‘voiceless unaspirated’ stops), while the ‘lenis’ series is realised with full voicing. It has been proposed in the literature (e.g. Honeybone 2005a; Petrova et al. 2006; Beckman et al. 2013) that in such languages the ‘lenis’ series always shows the hallmarks of greater markedness, since it is these segments that bear a laryngeal feature whilst the ‘fortis’ ones remain unspecified. In this chapter I argue that this precisely not the case in Bothoa Breton, and that instead it is the ‘fortis’, short-lag VOT stops (and fricatives) that are the more marked series, on the basis of their phonological behaviour. I examine three areas of Bothoa Breton phonology to provide this evidence:

Final laryngeal neutralisation

As discussed in section 7.3.2, the laryngeal contrast in Bothoa Breton is suspended word-finally, with both fortis and lenis stops being realised in ways that largely depend on the context. Several types of realisations were identified:

I analyse this neutralisation as complete suspension of laryngeal contrast in non-onset position. I formalise it as deletion of the C-laryngeal node with preservation of other features. In other words, there are three possible types of laryngeal specification available in surface representations. Bothoa Breton is therefore a counterexample to Lombardi's (1995a: 28) dictum: ‘There is no phonological contrast between a representation with a bare Laryngeal node and no Laryngeal node at all.’