Gittin 10A

איכא בינייהו שמות מובהקין.

The Talmud explains: It is possible that even the first tanna holds in accordance with the opinion of R' Elazar that the witnesses who observe its transmission effect the divorce, and he does not dispute R' Shimon on this point. Instead, the difference between the first tanna and R' Shimon concerns a case where the signatures on the bill of divorce are unambiguous [muvhakin] non-Jew names. The first tanna holds that although a bill of divorce is valid if it was transmitted in the presence of valid witnesses, there is always a concern that it might have been transmitted in the presence of the same non-Jew witnesses who signed it. Therefore, it is rendered invalid by rabbinic law. Conversely, R' Shimon holds that if it contained names that clearly belonged to non-Jews it can be assumed that the bill of divorce was transmitted in the presence of two valid witnesses, and therefore it is valid.

והא חזרה דאורייתא, וקתני!

The Talmud raises another difficulty: But the halakha of retraction applies by Torah law, as according to the opinion of R' Meir the husband can retract his instruction to give the bill of divorce and the master can retract his instruction to give the bill of manumission by Torah law, thereby canceling the agency. And yet the baraita teaches it among the ways in which bills of divorce are equal to bills of manumission. This indicates that the tanna does not distinguish between a case that applies by Torah law and one that applies by rabbinic law.