Kasha (Eruvin 40b)

The gemara on daf 40a asks if we mention Rosh Chodesh in davening on Rosh Hashana. The gemara brings a braisa as proof. The gemara going on to daf 40b brings a proof from a braisa that you don't mention Rosh Chodesh because Beis Shammai says that you only make 10 brachos on Rosh Hashana and not 11 so you see that Rosh Chodesh doesn't get its own bracha. The gemara concludes the proof and says "kasha." Then five lines later the gemara asks what the conclusion is - do we or do we not mention Rosh Chodesh. This question seems strange because we just answered it a few lines earlier.

I think that the simple answer is like the Rosh says on the gemara in Shabbos 22b (the Rashash on our gemara pointed me there) that even though the gemara says kasha that's not irrefutable proof. It just means that it seems to be a question. Now the gemara wants to know if that proof is accepted l'halacha. This fits in with the explanation (I think it's the Rashbam in Baba Basra) that tyuvta means that it's a real proof but kasha just means that we don't know the answer. The Ritva explains why it's not a clear proof. All this braisa proves is that you don't mention Rosh Chodesh in the bracha but it does not proof that it doesn't get any mention in shmone esrei.

However, this explanation doesn't work with Rashi in d"h kasha on our gemara because he says that "shma mina" from the question that the halacha is that you don't mention shma. So if Rashi's right then why does the gemara ask what the halacha is. The Rashash explains that the next line after kasha is "kollel atzmo tanai hi" - the entire concept of kollel (from that braisa) is actually a machlokes tanaim. The gemara's point is that the entire braisa which is clear proof (acc. to Rashi) is only a daas yachid because it's assuming kollel but in reality others argue. So after the gemara says that it goes back and says that the proof isn't a proof according to everyone so then what is the halacha.