Brachos 21b - Shomeia K'oneh

The gemara discusses when a person can start davening when they come late to shul and says that since we hold that one cannot say kedusha alone, one cannot start shemoneh esrei unless they assess that they will be done by the time the shliach tzibbur reaches kedusha. There is a machlokes between Rashi and Rabbeinu Tam cited by Tosafos here (and in Succah 38b) how to regard the concept of shomeiah k'oneh. Rashi understands that if one were to remain silent during their shemone esrei and listen to the kedusha with intent as if they were saying it, they could be yotzei, and it would not constitute a hefsek in the shemone esrei. Rabbeinu Tam and R"I both hold that it would constitute a hefsek. Clearly, according to Rabbeinu Tam and the R"i, the concept of shomeiah k'oneh is literal, that by hearing it is tantamount to saying. However, Rashi holds that it creates some level of tziruf that allows the listener to combine with the the mitzvah being performed by the reader, but isn't as if the listener is actually saying the words and therefore doesn't constitute a hefsek. The Tzlach explains that according to the R"I and Rabbeinu Tam that shomeia is literally like being o'neh and saying the words oneself, if one were to use shomeiah k'oneh to make a bracha that they were not obligated in, it would be a violation of a bracha l'vatala (whereas the concept of hirhur k'dibur even according to Ravina wouldn't constitute a bracha l'vatala).

Tosafos asks that according to Rashi that one can be yotzei by listening even during their own private shemoneh esrei, why would the gemara require one to wait until after kedusha before starting shemone esrei? Why not start shemoneh esrei and be yotzei kedusha through shomeiah k'oneh? Tosafos answers that according to Rashi although one can be yotzei through shomeiah k'oneh, it is not ideal. It is more of a hiddur mitzvah to actually say it and not be yotzei through listening.

It doesn't seem from Tosafos that we regard the listening as a type of hefsek lichatchila, rather the mitzvah of saying kedusha is not being performed ideally unless one actually says the words. Why is this different from kiddush on shabbos when we use shomeiah k'oneh lichatchila to be motzi others? If the concept of shomeia k'oneh doesn't work lichatchila, we shouldn't use it for kiddush either?

I was at a chasuna last night and posed this question to Rav Asher Weiss. He thought that although shomeiah k'oneh is only a "bidieved", there are other considerations that can override it and that is why we sometimes use it even lichtachila. I asked what is the consideration by kiddush that overrides shomeiah k'oneh, to which he said that since we have kiddush b'makom seduah, the group element is important. In my opinion this is not correct. Although the kiddush may require a makom s'eudah (be it d'oraysa or d'rabonon - see Rosh Pesachim), it does not require a group (or perhaps, as r. warren cinamon suggested to me it is just the regular din of b'rov am hadras melech that would override the lichatchila of saying it oneself). Perhaps by Megillah there is a concept of being yotzei b'tzibbur such as we find shelo b'zmano must be with a minyan, but by kiddush we don't find such a concept.

It seems to me that Tosafos is not saying a rule for all shomeiah k'oneh that it is only bidieved, to which kiddush is an exception. Rather, Tosafos is saying something very specific to the replying to kaddish and kedusha. Tosafos holds that since the goal of answering kaddish and kedusha is to actively participate and show that one agrees and confirms the greatness and holiness of Hashem, it is ideal to do this actively. Generally, the din shomeiah k'oneh is as ideal as doing it oneself, but when it comes to things that are meant to proclaim the holiness and greatness of Hashem, it is a hiddur mitzvah to do it actively by actually saying the words as a declaration.