wyl118 wrote: ↑20 Mar 2020 10:43
The suffixes involved are -ize and -ation, so you should think of terms such as palatalization as palatal-iz-ation ((pal- + -ate + -al) + -ize + -ation), and nasalization as nasal-iz-ation ((nas- + -al) + -ize + -ation). Regardless, the correct term for the loss of voicing is "devoicing".
In twelve + -th = twelfth /twɛlfθ/ [tʰw̥ɛlfθ], /v/ undergoes devoicing to /f/ [f] because of regressive assimilation (or "anticipatory" assimilation, if you will) from the following /θ/.
In eight + -th = eighth /eɪtθ/ [eɪt̪θ], as mentioned by shimobaatar, /t/ [t] becomes dental [t̪] because of regressive assimilation for place of articulation from the following dental /θ/. This doesn't affect the phonemic representation, but there's a phonetic change in there (an allophone).
In ten + -th > tenth /tɛnθ/ [tʰɛn̪θ] (or [tʰɛn̪t̪θ]), you see the same thing as with "eighth" (except possibly with an epenthetic [t̪], but this is likely irrelevant for your homework).
An example of regressive assimilation for mode of articulation in Korean (as in the devoicing of "twelfth") would be the stem-final consonant of 뒤쫓다 dwijjotda 'to chase after sth'. Notice the difference between the original /tɕ/ [tɕ ~ dʑ] in 뒤쫓음 /twitɕ͈otɕ
ɯm] and the assimilated versions of it in 뒤쫓기 /twitɕ͈otɕk
i] (where ㅈ /tɕ/ becomes like ㄷ before ㄱ by losing its original emphasis and released sound) and 뒤쫓는다 /twitɕ͈otɕn
ɯnda] (where ㅈ /tɕ/ becomes like ㄴ before ㄴ by undergoing nasalization and voicing in a full assimilation to that following ㄴ /n/).