Nik Taylor: > jesse stephen bangs wrote: > > where # indicates a morpheme boundary. Thus, "sing" is underlying > > /sing/; "ringer" is ['riNr=] because of the morpheme boundary > > /ring#+er/; "finger" is [fiNgr=] because it's a single morpheme. Does any > > one (read: Dirk) care to disagree? > > Hmm, interesting view. Ah! Just found an exception: longer = (at least > in my dialect) /lANgr=/, same with "longest" = /lANgIst/. Also, there's > a theater in my home town called /sejNr=/ (can't remember the spelling), > monomorphemic, but does not have /Ng/. By your theory, it should be > pronounced [sejNgr=] It's not just Jesse's theory: it's a very well-known generalization. _longer, longest, stronger, strongest_ are exceptions; their exceptionality can unproblematically be handled as irregular inflectional morphology. /sejNr=/ is a much more interesting exception. It's exceptional also because there's a long vowel, which is prohibited before /N/. How is it spelt? Do you know where the name comes from? --And.