Schubert Impromptu G flat, op 90 no 3: measure 5 variations?

[johsm]

Hi everyone - I have just joined here and this is my first!
I am an amateur dabbling at the piano as a hobby that makes it easier to deal with my stressful and very technical day-to-day job, so my knowledge is small and my technique is close to non-existing 

My question is about a piece I started to practice (without a teacher) recently: Schubert's Impromptu in G flat, op 90, no 3. I have several scores of this piece and there is a quite big difference in measure 5 between some of those: the chord played or the last half-note is the same as in measure 1 in some editions (G flat) but is different (B flat dominant 7th) in other editions.
However none of the editions comment on this - does anyone know where that dominant 7th comes from? It is written in my Edition Peters edited by Niemann, probably from 1979.
How does one go about finding out these things?

I am not sure which version I like better: I had been used to the dominant 7th version when I was young, so I kind of expect it, but I like not doing the dominant 7th there better because it fits better the calm pp flow at that point and leaves more room for building up later.

[j_menz]

I don't have the Peters edition, but that sounds ridiculous. Bar 5 reproduces bar 1 exactly, and I can't see any reason why it shouldn't (the developmental change happening at the end of bar 6), nor have I seen an edition where it didn't.

EDIT: You can check Schubert's manuscript here. There are some interesting aspects to it in a number of ways, but your chord isn't amongst them. I have no idea where it might have sprung from.

[iansinclair]

Both versions I have (the better being the Buonamici 1897) have the G flat.  And that is what belongs with the way the piece goes.  IMHO...

[johsm]

Thanks for the pointer to the manuscript and for pointing out that the other editions just repeat measure 1 in measure 5.

By accident after listening to a few recordings of this on Youtube, I just found a video of Horowitz playing the piece, and he actually does play the B flat dominant 7th in measure 5:

From hints I have read on the internet it seems that the changed measure may have been introduced by Liszt in one of his editions, but I am not sure.

[j_menz]

Liszt did a fairly extensive reworking of it (S.565b/2 - score here). He changes the key to the more reader friendly G, and, taking that into account, introduces your chord. Horowitz may have been influenced by it (or may have come up with the change himself).

Liszt also uses the rather more usual cut common time rather than the peculiar double version adopted by Schubert (retrospectively once he got a few bars in, judging from the ms).