I now find -ctví to be a better analysis than -tví, e.g. in otroctví (=otrok + -ctví) and sobectví (=sobec + -ctví). The analysis as -tví lets the sibilant of -ství disappear from the suffix representation; by constrast, the analysis as -ctví lets the sibilant s transform into c. Admittedly, c can lay claim to be part of the base word as well as the suffix, consistent with "kde c vytváří morfematický uzel, neboť patří k základu i k sufixu" of Slovník afixů. Therefore, both analyses have some strength, but then again, the -tví analysis prevents the suffix representation from revealing there is a sibilant, and I find that inferior. See also Talk:-cký. --Dan Polansky (talk) 12:12, 11 September 2019 (UTC)