I’ve recently read several contradictory statements about whether a musician can or should be able to write out the score of a piece they are playing from memory. On the one hand, neuroscientist Ray Dolan asserts in Alan Rusbridger’s book ‘Play it Again’ that musicians are unlikely to be able to recall a memorised piece and literally sit down with a pen and paper and conjure it up note by note. On the other hand, pianist Charles Kenyon states quite categorically in his book ‘How to Memorise Music‘ that musicians should be able to write a piece out entirely from memory, and this is crucial for a memorised performance. I would imagine both opinions are commonplace, but which are we to believe?
I decided to test whether I could write a score out from memory, by trying to write down the first few pages of JS Bach’s Fugue in A minor (BWV 944) from memory, which I learnt about a year ago. Quite a few people have highlighted Bach fugues as being particularly difficult to memorise, which is one of the reasons I chose this piece, plus the fact that I can still play it fairly reliably from memory. Written around 1714, in Bach’s early years, the piece comprises a short fantasia followed by one of Bach’s longest and most developed fugues outside the set contained in the Well Tempered Clavier. The furiously driven fugue (which starts about 50 seconds into the recording, at the big dynamic dip) is a whirlwind of continuous, unbroken semi-quaver motion, punctuated with cadences, modulations, and layers of counterpoint which build from the original theme. Here’s me playing the whole thing:
When I finally sat down to attempt to write out the first few pages, I expected to be in for a mind-numbingly dull couple of hours. In fact, I spent a thoroughly enjoyable few hours totally absorbed in the intricate details of this wonderful music. It was a fascinating, though cognitively taxing, experience that I might even consider repeating again as I learn new pieces. Although there were a few bars where I got completely stuck, I was mostly able to recall the notes relatively easily. And when I later went through and checked my score, I was quite pleased to find that most of what I’d written down was correct.
More interesting than the specifics of this exercise, though, are the observations I made while doing it:
- I couldn’t remember the notes without singing! Auditory memory was absolutely essential to writing the score out. I’m not usually a singer, though I have been told frequently how useful it is for pianists. And although I don’t have perfect pitch (which I’m sure would have helped!), when I checked I was less than a semitone off-key.
- I couldn’t remember the notes without ‘playing’ on the table! Motor memory was very helpful for writing the score out. Sometimes I ‘played’ hands separately, sometimes hands together, and sometimes while singing another part on top! Although writing the notes down forces you to consider each stave individually, sometimes my memory required both parts to be present.
- My memory was best at structural boundaries. Repeats of the first fugue subject were easiest to remember, and acted as landmarks on my mental map of the piece. The bars that I simply couldn’t remember tended to be located towards the end of musical chunks, leading up to the next structural boundary.
- Most of my mistakes were chromatic and I found accidentals particularly hard to remember. I’m not sure if this is universally true, or a problem specific to me. But when I do get thrown off course playing from memory, I’ve recently noticed that it’s often due to uncertainty around note recall specifically relating to whether a particular note is sharp/flat or not (i.e. white or black notes on the piano keyboard).
- Despite relying quite heavily on visual memory when I play, I found that I didn’t use visual memory of either the score or my hands at all. I was really surprised by this, and it makes me question whether visual memory is really useful for music. I certainly don’t have a photographic memory (presumably this would have been a rather facile task if I did!), and the published score actually looks quite different from my written one.
Has anyone else tried this? It’s honestly a lot less boring than it sounds, and I’d love to know if others have made the same sorts of observations…