1. The mishna lists certain things that can be said in any language and others that we insist on loshon hakodesh. Tosafos asks why we leave out Hallel, kiddush, and brachos from the list. Tosafos answers that the mishna only lists things that can be said in a language that one understands, but these 3 things can be said even in a language that one does not understand.
It seems that Rashi avoids Tosafos question differently. Rashi writes that there is a hava amina that parshas sotah would need to be said in loshon hakodesh because we may learn from chalitzah (rashi implies that for viduy ma'aser there is also a hava amina of loshon hakodesh). Apparently rashi understands that the mishna does not intend to give a exhaustive list, rather only picks on items where there is a chiddush that they can be said in all languages. Therefore, hallel, kiddush and brachos aren't listed since it is pashut that they can be said in all languages.
Perhaps Rashi and Tosafos go lishatasam. Tosafos Shantz (33a) [see also tosafos brachos 13a] explains that the hava amina that we would require loshon hakodesh for everything is the idea thatכל התורה כולה בלשון הקדש נאמרה and without a specific list of exceptions, i would have assumed that everything needs loshon hakodesh. Tosafos would apply this principal to literally everything, and therefore there would be a hava amina that hallel and kiddush require loshon hakodesh. That is why Tosafos asks that the mishna should list hallel and kiddush also, since there is a chiddush that they can be said in any language. However, Rashi goes lishitaso that the statement כל התורה כולה בלשון הקודש נאמרה is limited to Torah readings in a shul and therefore has to come up with specific rationale to explain why there is a hava amina that the items in the mishna would require loshon hakodesh, but hallel, kiddush and brachos are poshut that they don't need loshon hakodesh so they are not even listed in the mishna.
2. Pri Megadim (psicha kolleles, Tefilla) points to a machlokes whether tefillah can be accomplished through hirhur, without speaking. With this possibility he answers the question as to why we don't make a birchas hamitzvah on tefillah, since it can be achieved without speech similar to bitul chametz that we don't make a bracha on devorim sh'bleiv (rather we make a bracha on the biur). Our mishna that lists tefillah as something which can be fulfilled in any language implies that without speech one cannot be yotzeh, because if hirhur would be enough then it is obvious that there isn't any kepeida on loshon hakodesh.
3. Pri Megadim (Psicha kolleles, ch. 3:6) writes that Tosafos holds that the items listed in the mishna one has to understand their meaning even if they are reading in loshon hakodesh [to the exclusion of the mishna berura (62) who writes that when done in loshon hakodesh one does not have to understand what he is saying]. This is clear from Tosafos when they write that megilla is not listed in the mishna because one can be yotzei in "ashuris" without understanding, which implies that everything else listed in the mishna one must understand even if it is written in loshon hakodesh. Based on this, he asks what is the kavana needed specifically by the first pasuk of shema, since one must understand what they are saying for the entire shema. Pri Megadim answers that for the list in the mishna of things that can be said in any language, one must understand that language, even if is being read in hebrew because it only qualifies as a "language" for them if they understand it. But, they are not required to concentrate on the meaning as they are reading it, whereas for the first pasuk of shema one must concentrate on the meaning as he is reading it.
Aside from having to understand what is being said the Biur Halacha writes that one can only read in other languages when that is the spoken language of the country, but it is not sufficient if just a small group of people speak that language.
4. Pri Megadim (P.K. 3:7) offers a simple answer to Tosafos question. The mishna only lists d'oraysa mitzvos, and therefore doesn't list Hallel, kiddush on a cup (based on tosafos shita in pesachim) and brachos and even megillah which are only d'rabonon. Therefore, the pri megadim argues on the magen avrohom (62) who brings Tosafos l'halacha that for the hallel, kiddush, brachos one does not have to understans, since if we answer tosafos question simply by saying that the mishna leaves out d'rabonons then we are not forced to be mechadesh tosafos answer. Therefore it is possible that even in loshon hakodesh, one cannot be yoztei hallel, kiddush and brachos without understanding what they are saying. Based on this it works out better that there is no distinction between brachos and birchas hamazon, that birchas hamazon you need to understand and brachos you don't need to understand, rather כעין דאורייתא תקנו that even brachos must be understood.