A sincere question - what about the Carpenter version do proponents find convincing/compelling?
From a general standpoint, I like that the orchestration is more substantial. Even if some of it is not particularly Mahlerian (especially certain parts of the second movement), I still feel that it is well done. Litton states in his liner notes that he made numerous changes to Carpenters' version, so I am never certain what is strictly Carpenter and what is Litton. I should also tell you that I think Litton is a fine Mahlerian and that the Dallas SO is an excellent Mahler orchestra.
I typed the following in a recent thread (which I am cutting and pasting):
Despite my love of the overall Carpenter/Litton (“C/L”) version, their second movement is very much over-orchestrated – although at least it is FULLY orchestrated. That would be my main complaint about the Cooke version – it is too bare – even though the entirety of his effort is monumental, worthwhile, and highly enjoyable. Mazzetti has some interesting ideas, but is not generally preferable to the above two. I have very little praise about Wheeler or Barshai in any of the final four movements (and have not re-listened to S-M yet).
Ironically, it seems that much of the critical complaining is about the second scherzo (fourth movement), which I think is fabulously done by either C/L or Cooke (especially Cooke II with the snare drums).
Barry, many of your recommendations are easy to imagine (hear in my head)– and I very enthusiastically agree with those. If I had time, money, and any ability, I would love to start with the C/L version and scale back or replace (i.e. with ideas from the other existing versions) some of the things that don’t work for me. For example, in the C/L finale I am perfectly thrilled with the last 6+ minutes through to the end. However, I would add back the Cooke II treatment of the music after the dissonant trumpet music dissipates. C/L goes directly to the strings from this point to the end, but Cooke starts out with some brass and winds first before shifting to the strings – which makes for more of a transitional development. So I like Cooke’s idea for this first spot but much prefer everything Litton does through the end: the very Mahlerian use of dynamic and tempi changes, the use of the bass drum in critical spots (and I think that there is a tam-tam underneath that final bass drum thump about 30 seconds before the end). C/L gives M10-5 the Mahlerian ending I would expect.
I also like the changes C/L make to the first movement as well as the efforts to make the M10-3 Purgatorio repeat into something special (i.e. not a literal repeat).