With reference to "in many languages nasals tend to velarize in the rhyme position", in Against the sonority scale: evidence from Frankish tones, by Ben Hermans & Marc van Oostendorp, Meertens Institute, Amsterdam www.vanoostendorp.nl/pdf/sonorityscale.pdf , I would like to confirm that it is so in Burmese-Myanmar (Burmese spoken language written in Myanmar script).
U Kyaw Tun: In Yaw (northwestern Myanmar), {na.that} is pronounced as {nga.that}. Thus {gyi: twan:} (to rub off dirt} becomes {gyi: twing:}. This takes place only in the coda. In the onset {na.}, {nwa.}, and {nha.} are quite distinct from {nga.}, {ngwa.}, and {ngha.}. (Info checked by U Aung Myo Maw)
Explanation of {...}: Myanmar is phonetically based and syllables are pronounced as they are written. The adage is "what is written is correct (implying permanace), what is pronounced is just sound (implying transience). {...} stands for Burmese-Myanmar. {na.that} is [n] in coda. {nga.that} is <ng> as in English <king> in the coda. For more on Burmese-Myanmar, please look into Romabama in www.tuninst.net
U Kyaw Tun: In Yaw (northwestern Myanmar), {na.that} is pronounced as {nga.that}. Thus {gyi: twan:} (to rub off dirt} becomes {gyi: twing:}. This takes place only in the coda. In the onset {na.}, {nwa.}, and {nha.} are quite distinct from {nga.}, {ngwa.}, and {ngha.}. (Info checked by U Aung Myo Maw)
Explanation of {...}: Myanmar is phonetically based and syllables are pronounced as they are written. The adage is "what is written is correct (implying permanace), what is pronounced is just sound (implying transience). {...} stands for Burmese-Myanmar. {na.that} is [n] in coda. {nga.that} is <ng> as in English <king> in the coda. For more on Burmese-Myanmar, please look into Romabama in www.tuninst.net