Which Rabbi Eleazar?

While we are pleased with the current status of Mi vaMi, one big item on our agenda is better resolution of ambiguous Tannaim and Amoraim. There are several Tannaim and Amoraim who bear the same name. There are four Amoraim who go by "Rav Kahana", in the first, second, third, and fifth-sixth generation. There is Rabbi Yehuda the Tanna, and Rav Yehuda the Amora. There is Rabbi Yitzchak bar Avdimi, the fifth generation Tanna, and Rav Yitzchak bar Avdimi, the third generation Amora. Is it Ravina I, or his nephew Ravina II? How do we distinguish between them? Right now, we often don't. Most names are not ambiguous, because

• the name is unique
• the name includes the name of the father, such as Rav Nachman bar Yitzchak (with the unadorned Rav Nachman being Rav Nachman bar Yaakov)
• the title is either Rabbi or Rav. Rabbi would imply a Tanna or else an Amora ordained in Israel, while Rav would imply an Amora from Bavel.

Rabbi Eleazar ben Shamua or ben Pedas?

However, some names remain ambiguous, and we are exploring various avenues to better decide among candidates.

One such example, partly resolved, is that of "Rabbi Eleazar". Most often, when Rabbi Eleazar appears in a Mishna or brayta, it is Rabbi Eleazar ben Shamua, a fifth-generation Tanna. Most often, when Rabbi Eleazar appears in the gemara, it is Rabbi Eleazar ben Pedat, a second or third-generation Amora. Both are referred to as "Rabbi", and so we cannot use the title to (automatically) disambiguate. And it is looks quite silly to mark a Rabbi Eleazar in the Mishna as an Amora.

There are several avenues of attack. But we recently have made some nice progress. Consider Zevachim 93b. In the Mishna, Rabbi Eleazar and his contemporary, Rabbi Yehuda, are both marked in cyan. Success!

We accomplished this by adding some awareness of context. The named entity recognition module is now aware of whether it is operating on a Mishna or the gemara.

There is still more to be done. In the gemara which follows, repeated reference is made to Rabbi Yehuda and Rabbi Eleazar, and this time, the recognizer mistakes Rabbi Eleazar for the Amora.

How would you go about this?

(We have some plans in the works. Some features we can use to help make the determination, in the general case are:is the context a citation of a brayta?

• language of the speech verb or of the actual statement: Aramaic or Hebrew?
• with whom does Rabbi Eleazar (or any other ambiguous figure) directly interact? Try to minimize the generational gap. Who discusses the statement?
• does the word 'Mishna' or 'brayta' occur in the English or Hebrew text?

We'll hope to get around to adding and considering many of these features eventually.)