supersede biological indicators of race in some applications and not others?

> so, a black man who has altered his hair and skin w/features
>typically associated w/the white race (or any non-black race)
>but who lives and identifies as a black man is not transracial
>for the purpose of this discussion. and that's why i say MJ
>and Sosa don't apply.

For example, according to what has been championed throughout this post, the following would be true:

Michael Jackson, in spite of presenting with the physical features of a race different from the race he was born, is of black race simply because he said so. Rachel Dolezal, in spite of presenting with the physical features of a race different from the race she was born, is not of black race simply because she said so.

^How is this not logically fallacious?

And in the case of Sammy Sosa, we do not know whether Sammy Sosa identifies as a black man or did when he presented as a black man physically. We do know that many individuals in/from the Dominican Republic, with significant African ancestry, do not identify as black. So perhaps a better question is would YOU have considered Sammy Sosa, in his former physical state, a black man if he did not identify as such? What about other Dominicans that look similar to Sammy's former physical state but do not identify as black. Would you consider them black people?