Rarely comest thou, Spirit of Delight

Elgar prefixed his second symphony with the quote from Shelley's poem saying "I have put my soul in it". The music certainly breathes with the depth of feeling which is unmatched not only in his output, but all but the best music. A friend once said to me that this kind of music sounds "sad". I know what she means, and yet, I don't think sadness is the right word. It is the feeling one gets from a deep sigh - and that is as much a feeling of connection with everything as it an expression of transient emotion. Shelley's poem is the same:

Rarely, rarely, comest thou,

Spirit of Delight!

Wherefore hast thou left me now

Many a day and night?

Many a weary night and day

'Tis since thou are fled away.

Really, it's about breathing, I think. If music or poetry doesn't breathe, it's no good - it has no life. 

This afternoon I'm giving a talk to the Liverpool University Music Theory Club (see Music Theory Research Group (chromatic-harmony.com)) about breathing in a Haydn piano sonata.  Haydn is definitely one of the most cheerful composers - but his music breathes too. As I've been preparing this, I've been really interested in the question as why, despite it's amazing technical potential, AI doesn't breathe (obviously it doesn't have lungs - but that matters, doesn't it??)

I think the questions about music and the questions about AI are related, and they have to do with the nature of our 3-dimensional world, and the poverty of our description of that world. Breathing is of course 3-dimensional - as is everything in biology. The basic issue is that we struggle to grapple with 3 dimensions - only in actual practice or performance can the concrete experience of the dimensioned world be appreciated, but it cannot be codified without losing its life.

Most of the time in science, we attenuate dimensions. So, for example, we will choose to measure vaariable x, and ignore variable y. Statistical normalisation, correction for confounding, etc are all ways of attenuating the living world. 

However, if we understand the dimensional relations at the heart of the 3-dimensional world, then it is possible, rather than attenuate, to reduce dimensionality in a way which preserves the 3-dimensionality, but presents a 2-dimensional representation of it. This is what happens in a hologram: a two-dimensional pattern formed by interference between (say) light is an encoding of 3 dimensions. All those dimensions are preserved, but what we see, and can analyse, is a surface. This is why the holographic theory of the universe talks of a black hole being represented at its surface as a hologram - a reduction of the dimensionality which encodes the nature of matter. 

For a long time I've been interested in fractal analysis of music (fractals are basically like holograms, and like them, a dimensional reduction) - see onlinelibrary.wiley.com/doi/full/10.1002/sres.2738, and Music, cells and the dimensionality of nature - PubMed (nih.gov). But today I'm going to talk about the normal distribution of noise as another approach to dimensional reduction. Haydn's C major piano sonata looks a bit like this... Normal distributions are very interesting analytical approaches to dimensional reduction - although we tend not to see them like this. 

The interesting thing about this video is that it does look like "breathing". So here's a little experiment:

Relax. 

Take a deep breath. 

Clear your mind. 

And breathe out... 

here are a few questions about your experience: