The "which Latin?" issue rears its head, relative to the Mahler 8 first movement (Veni, Creator Spiritus). When I was in college, this issue did not exist, at least in Iowa: we pronounced the language one way in Latin class (weni, widi, wiki), and another in choir (veni, vidi, vichi). The choral pronunciation we termed "ecclesiastical Latin;" only later did I learn that we were using an Italiante pronunciation, greatly influenced by our upper Midwestern vowels. I first encountered German Latin several years later, under Margaret Hillis, and became accustomed to singing the Latin settings by German composers with this pronunciation. Through a rather intense immersion in early music, especially under the influence of Belle Bouche, Bel Parleur ( I have lost track of this book, even of the proper spelling of its title), I learned about German, French, English, even Swedish Latin, as well as appropriate pronunciation shifts related to dates and religious influences. The options, as well as the consequences of making a wrong choice, mounted alarmingly. I found myself wallowing in a morass of strongly held opinions, lacking the background to do anything with confidence. About this time I landed a teaching job back in Iowa---and happily returned to the "ecclesiatic" pronunciation people out there were using. I didn't stay in Iowa for long; but, because I did not return to a specifically early music-oriented milieu, I continued with my Iowa Latin: I had enough trouble teaching my college-age choirs the proper notes, and had little energy left to hound them about their Latin. Before long, I began singing with Robert Shaw, who used the Iowa Latin for everything. Mr. Shaw was anything but passive about pronunciation, however: his concern was always for the sounds of language, and their utility in producing the same level of rhythmic precision and articulation in the chorus, that he got from the orchestra. He demanded an extraordinary amount of precision from us, and would experiment with various emphases and substitutions to get our sound over the orchestra and out to the listeners. Within the parameters of his chosen Latin pronunciation, he experimented with substituting unvoiced for voiced consonants, depending on pitch and volume level (Klawdia in place of Gloria, for instance), to clarify our attacks and rhythmic precision. He would mix the two sounds, G and K, in an attempt to find just the right balance between attack and correct listener perception of the sound; he also combined S and Z in our attack on the word Sanctus, for the same reason. He was a very verbal and expressive person; but his use of language as the vehicle for voices was remarkably objective and utilitarian.
My years with Helmuth Rilling present an entirely different situation. One has only to listen to the Oregon Bach Festival recording of Messiah, to understand how completely German Rilling is---his innate requirements, deeply based in his native language, are so specific, that he manages to get an all-American choir, singing in English, to sound like a German ensemble singing in a second language. Like Shaw, he is deeply committed to rhythmic precision and articulation; but his model and motivation seem to lie in the language itself, rather than in the more abstract realm of orchestral playing. His Bach Passions, in German, are absolutely riveting---he is completely committed to telling the story, and bends every element of the performance toward expression of the narrative. And his B Minor Mass profits from an idiomatically German pronunciation of the Latin; one has no problem imagining that this is precisely the pronunciation Bach had in mind.
Rilling's Mozart performances present a different situation, however. I have sung Robert Levin's reconstructions of both the Requiem and the Mass in C Minor several times with Mr. Rilling, often with Mr. Levin present during rehearsals; twice I have heard Levin state that, in his judgment, Mozart---Austrian, Roman Catholic, influenced by Italian singing---would have assumed an Italianate pronunciation of the Latin, rather than a German one. By extension, this pronunciation, through utilizing more open vowels and softer, more insinuating, less precise consonants-- magnificat pronounced man-yificat, rather than mak-nificat, for instance---would encourage a fuller, richer, more legato approach than that produced by the German pronunciation. Mr. Rilling's Mozart is very satisfying; but I often think of Mozart's own words, in a letter to his father (written in Italian): Prima la musica, poi le parole. First the music, then the words. I can imagine him preferring the Italian sound.
I have heard arguments about the sung pronunciation of Latin from several points of view. What does the composer assume? and by extension, what does his contemporary audience assume? What does a modern audience assume, and prefer? I remember hearing, when singing under Margaret Hillis: CSO recordings were marketed internationally, not just to Chicago audiences, and nationally appropriate pronunciation would be appropriate to a world-class group. American choral educator Don Moses has written that we should base our pronunciation upon the preference and comfort of our target audience---American (i.e., "ecclesiastical") for American audiences. Helmuth Rilling takes yet another position: German pronunciation just works better for him, and accomplishes his musical goals more efficiently.
On to Mahler and his 8th symphony later.