The Goldberg Variations: in General and in Particular When Crossing Hands
March 15, 2021
The Goldberg Variations: in general and in particular when crossing
hands. The hands always start in a more remote position from each
other than is dictated by where the next notes are.
I held my arms in front of me, floating above the keyboard, the
forearms rotated so that the palms of the hands faced each other
across an empty horizontal space. In that space I imagined there
being a tangible object. The object was well defined but not quite
I imagined the feeling that my hands were two gentle but strong
clamps* on opposite ends of an object that occupied the horizontal
space between them. I stayed aware of the imaginary object tightened
the clamps pushing inwards towards each other and that somehow the
object, with some resistance, without losing its integrity, could be
compressed along its horizontal dimension. I gently compressed it
until my hands were the same distance from each other as the two keys
on the piano that represented in each hand the most current sixteenth
note in the piece. One of the notes being in a voice that was in the
right hand, the other in a voice in the left hand.
I would repeat that procedure by advancing to the notes that lay on
the next sixteenth note of the piece, and so on. The distance between
the palms and the size of the object they held was constantly changing
in size as the piece evolved in sixteenth notes. OFten one of the two
notes was a note longer than a sixteenth, and remained in its place,
while the note in the other voice changed. In other words, things
were constantly changing.
Regardless of the distance between the hands, when articulating the
The next pair of notes ended up being greater than, lesser than, or the
same**, their mutual distance between when articulating the current
pair of notes, I nonetheless re-expanded the distance between my hands
so that I would have to begin again by compressing the imaginary
the object between my hands until the distance between the palms
corresponded to the distance between the new pair of notes.
This technique is also surprisingly effective even at those moments and
episodes in the Goldberg when the hands cross each other. But even
when the left hand took up a position to the right of the right hand,
as the right hand correspondingly took up a position to the left of the
left hand, the position of the arms just prior to the ‘squeezing
together’ of the imaginary object that lay between the palms, was with
uncrossed hands, the left hand being as usual to the left of the right
hand, and the right hand to the right of the left hand. The palms
would then squeeze the imaginary object towards a point when the two
palms would converge in space even though in reality that continued in
their motions until they crossed each other. But the procedure for
the following notes, even if it still involved crossed hands, began
from a position when the arms were not crossed but moved into a
the position that was crossed. This was most easily forgotten if a series
of pairs of notes got further and further away from each other; the
palms should always be moved towards each other to land on the next notes
Try to avoid feeling that adjacent fingers are setting themselves down
on the keyboard one after the other. Each finger comes from afar and
converges centrally, towards the other hand, to get to its note.
One should never feel that they are simply setting adjacent fingers
down, side by side, on adjacent notes – without first drawing the
hands further apart and playing the next sixteenth note via an action
which brings the hands back towards each other.
This means that the action in the finger causing a key to sound is not
a motion that is strictly speaking downwards, since it doesn’t begin
from a static position directly above the key. The motion of the
finger is always swept along the hypotenuse of a long skinny right
triangle. From a physicist’s point of view, motion along a
hypotenuse can always be thought of as having a down component and a
lateral component. It is like the path of a boat crossing a swift-flowing stream. One does not steer directly towards the nearest point
on the opposite side of the stream, but always somewhat to a side so
as to compensate for the flow of the water.
What I am saying today seems to me to be an extension of things I have
said in the past take had the form of “never aim directly at the next
key you are playing, but overshoot it, and then settle back easily to
its location. That this can be more accurate in the long run, instead
of betting all the money you have on getting to the goal through a
direct unwavering motion.
It is also similar to what I’ve said in the past about “damping
curves” where a series ever diminishing but accelerating motions,
first one way and then the other converges on the exact location you
have in mind on the keyboard. I also realize that this is in
contradiction to the method I have often proposed for skipping from
one place on the keyboard to another more remote location, by using
“opposing vectors”: the current finger tries to stay on its key even
though the arm is already pulling as hard as it can towards the more
remote key; that tension builds up between the two opposing motion;
until like an overwound spring, the spring suddenly explodes, and the
hand is propelled in a single, unbroken motion, to the remote key.
* Each side was separately adjustable – not like a vice or c-clamp.
** Thus, even if the hands remained in the same place on the keyboard,
covering the same group of notes, with the five fingers of each hand,
during a series of notes, I would still oscillate the hands apart from
each other and then close up the distance between them.