We are here concerned, with assimilations of nasalisation in consonant clusters at boundaries, where an underlying stop is followed by a nasal.
The stops of Nunggubuyu are /b dh d d j g/ Of these, /b dh/ do not occur at the end of morphemes (except for a couple of ininflectable verbal root forms or interjections ending in /b/ , so they are irrelevant. Velar /g/ is common at the end of stems and other morphemes, but is always deleted as nonfinal element in a consonant cluster by Stop-Deletion P-29, so it is also irrelevant. This leaves /d d j/.
Of these, /j/ is by far the most common morpheme-finally. It is regularly nasalised to /ny/ before a nasal consonant. For example, Pl prefix //mij-// is often followed by a nasal consonant, resulting in forms like //mij-ngalanyji// → /minyngalanyji/ 'girls'. For the underlying form //mij-// cf. /mij-gulmur/ 'young circumcised men' from /w1ulmur/ where the prefix must end in underlying //j// rather than //ny// (there being no denasalisation rule which could apply to this prefix). Similarly, compound initials like /-lhaj-/ 'firewood' and /-ngaj-/ '(solid) fat' show final /ny/ in compounds like /-ngany=ma-/ 'to get (solid) fat'. Nouns ending in /j/ show /ny/ before Instrumental suffix /-miri/ as in /wu-dhany-miri/ 'by means of firewood' from //uG-lhaj-miri// There are no consistent counterexx.
Relevant combinations involving underlying morpheme-final /d d/ are uncommon, since these are not high-frequency stem-final consonants. There are, however, a few noun stems ending in them which can be followed by Instrumental /-miri/ In these combinations Nasalisation is possible but not obligatory. Thus from /yi:mid/ 'tree sp. (Planchonella)' we get /yi:mid-miri/ or /yi:min-miri/ and from /yimbid/ 'cypress tree' we get /yimbid-miri/ or /yimbin-miri/.
| (P-22) | Nasalisation | |
| j → ny | // ____ | - C |
| d → n | [+nasal] | |
| d → n | ||
| usually obligatory for /j/ optional for /d d/ | ||
There is no strong argument against a more general formulation of the rule, to the effect that (any) stop is nasalised before a nasal. As noted above, /b dh/ do not occur in the relevant context, so there is no reason to exclude them specifically. As for the remaining stop /g/ (and the stop archiphoneme /G/ which is indistinguishable in its effects from /g/ , we have mentioned that it just disappears in the relevant contexts. This can be taken care of by Stop-Deletion P-29, which deletes /g/ and /G/ before any consonant. However, it would be possible in principle to allow P-22 to apply to combinations like //g-n// giving intermediate //ng-n// since /ng/ is also subject to deletion before another nasal by Nasal-Deletion P-30. In other words, we could allow P-22 to apply to //g// as well as the other stops, though in the case of //g// the application would be vacuous so far as the eventual surface output is concerned.