The "error" in Brahms 76/2

[(poco) Sforzando]

That's a good question, it's plausible as written -- V/V with flat 5th, or an incomplete French +6 in 5/3 position, is conceivable Brahms, and the voice-leading makes sense with G^# in the preceding bar, inflected to to G^nat in the target bar , continuing as expected to F^# in the following bar.

But I too have always read G# at this point.

Possibly because: in preparing a return to the tonic key and main theme, the simple V/V is the more likely possibility at this point. Possibly too the C# major at this point may be felt as an extension of the C# major chord an octave lower and five bars earlier.

[Luke]

Exactly.    And I'm sure I've always played that G#.

But how do we know it's a misprint and not just a wonderful dissonance?

Ha! Wish I'd PM-ed you now, because that was what I saw straight off - in fact, I can't see what other possible misprint there is! I thought Paul had got it before this point (I haven't read on in the thread yet), because he talked above of how the harmony the misprint created would have been atypical for Brahms, and I guess that answers your second line.

[Ten thumbs]

I must say that I have rather neglected Opus 76, so it does me good to draw attention to it. I do sight read a good deal but do so in what you may consider an ingenuous way. That is, I play what I see and if it shock's me I will stop and check the score. Just playing this passage therefore produced the G natural. It is not that shocking but I do think it is 'wrong'. I put it that way because historically there was a convention that an added accidental is carried over to the first note of the next bar unless it is cancelled; Brahms may have been relying on that.
I have come across quite a few errors in printed scores, most of which are misplaced accidental markings in chords. I don't keep a record of these but if I do recall one, I'll post it here. In the meantime, here is one that goes back to my youth:
SCHUBERT: Impromptu Opus 142 No.3
The 4th variation, in six flats, ends with an unmodified C flat in the right hand ( this is urtext). It most patently should be C natural. 

[(poco) Sforzando]

I must say that I have rather neglected Opus 76, so it does me good to draw attention to it. I do sight read a good deal but do so in what you may consider an ingenuous way. That is, I play what I see and if it shock's me I will stop and check the score. Just playing this passage therefore produced the G natural. It is not that shocking but I do think it is 'wrong'. I put it that way because historically there was a convention that an added accidental is carried over to the first note of the next bar unless it is cancelled; Brahms may have been relying on that.

As far as I know, that convention applied only with a tied note (see an example 7 bars farther in our piece). What's more, in the measure in question, there is a completely new harmony, which would make the G# more necessary. Looking through the edition I have, I see a number of cautionary accidentals, including naturals, apparently intended to make the notation unmistakable. Goldovsky's point, however, seems to be that most educated musicians use their sense of harmony and tonality to "correct" the error intuitively, despite what's written.

[Ten thumbs]

As far as I know, that convention applied only with a tied note (see an example 7 bars farther in our piece). What's more, in the measure in question, there is a completely new harmony, which would make the G# more necessary. Looking through the edition I have, I see a number of cautionary accidentals, including naturals, apparently intended to make the notation unmistakable. Goldovsky's point, however, seems to be that most educated musicians use their sense of harmony and tonality to "correct" the error intuitively, despite what's written.

It is indeed quite rare. One of those little things I was taught to look out for years ago. I don't think you'll find it in any modern editions and it does depend on the way the composer notates his manuscript. Cautionary accidentals are the answer and, as you point out, Brahms uses them quite often. It is interesting that in this case playing the unmodified G as written feels wrong when followed by the E#C#. This does not match any other 'shape' within the piece.

[jochanaan]

Exactly.    And I'm sure I've always played that G#.

But how do we know it's a misprint and not just a wonderful dissonance?

If Brahms had really wanted the G natural, he would have added a cautionary accidental sign.  He was that detail-oriented.  So we can safely assume that the lack of a sharp is an error.

[Brahmsian]

I guess Brahms just forgot to toss this one in the fireplace.