interfaces for creating new ways of thinking

At it's core particle physics is very simple.  You smash to tiny particles into each other hard enough that something dramatic happens and then sift through the debris.  The first major experiment I worked on was a clean example.  A beam made of a tiny negatively charged particles - pi minus mesons in this case -  booking at over 99.99% the speed of light - was directed into a flash filled with liquid hydrogen.  Hydrogen is as proton rich as it gets .. each atom is just a proton and an electron.   There are some collisions with the debris and surviving bits of beam continuing along.  The cloud of stuff flies through a powerful magnet about the size of a three story house.  Charged particles bend in one direction or another depending on the sign of their charge and pass though an sandwich of position detectors along the way.  These  were stacked grids of wires surrounded by a mostly argon atmosphere.  The moving charged particles deposit a bit of their energy in the argon which, in tern gets excited enough to lose electrons.  The  little clouds of electrons attracted to the closest wire where a signal is generated.  Each wire grid, we called them chambers, can only detect along one dimension.  So the first chamber is an array of thin wires spaced a couple of millimeters apart running in, say, horizontally - the x direction.  It's followed a centimeter downstream by another array of wires strung vertically in the y direction.  This repeats over and over giving you a region about five meters long where you can follow the track of a moving charged particle.  The amount of curve in the track adds critical information needed to sort out the particle's momentum and ultimate what it is.  But this isn't quite enough.  To improve accuracy we used another set of arrays rotated 45° to the first.  Think of just rotating a x-y plane by 45°.   These are our u and v detectors.  So the real sandwich was x,u,y,v,x,u,y,v,... and so on .. like a pile of magazines the masthead of the first pointing North, the next North-east, then East, South-east and repeat...

 

It's easy to calculate the u,v position of a point in the x,y plane ... that's just a coordinate transformation.  Just use the formula u = y+x, v= y-x .. I'll leave the transformation in the other direction to you.  In the drawing x=1, y= 3 becomes u = 4, v=2.   The important thing to note is the actual position of the point hasn't changed .. just how we label it.

 

Physics and math are full of transformations.  A few years ago Superstorm Sandy hit our area.  I wanted to see if it was shaking our house enough that I could find the resonant frequency of the walls.  My iPhone, an iPhone 4 I think, had a sensitive three axis accelerometer.  I wrote a little  program for the phone that would dump and email a file raw data to me.   When the wind had the house shaking I taped the iPhone to various walls upstairs as well as in the basement.  Then I analyzed it my Mac using another type of transformation ... the Fourier Transform. (unfortunately 10 days later as we lost power)

 

The Fourier Transform is special.  It allows you to move from looking at the world in a domain our senses and brain are tuned for ..  the time domain with events taking place one after the other .. and move to one that looks looks at the frequency of events that took place during that time -- the frequency domain.  It's just a fancy-pants version of the coordinate change we looked at, but instead of transforming the coordinate in a two dimensional space, we have coordinates on an infinite dimensional space - namely the space of all functions.  (hang on... I'll try to make this clearer without resorting to math - the mathy types among you can make fun of what is probably not the best description).

 

Here's a way to think about it.  To specify a  function of x ... say sin(x)  .. you find the value of the sin for every value of x ... an infinite string of numbers.  But it's not that different from giving the location of the our point in the plane, which was specified by two numbers.   So what if we take the information contained in f(x) [sin(x) in this case, but it could be any function] and express it in some other way.  Sort of like rotating the axis.  We leave it unchanged, but look at it in a different way ..  we mange to label it differently.

 

The trick is to find the right transformation.  I won't get into the math..  either it would look completely intimidating or you've been there already.  Now instead of specifying the function at each point we consider the frequencies of the signal the points represent and come up with a function that is the variation at all possible wavelengths.  If the function is a pure sine wave at 10 hertz you get zero for every frequency except 10 hertz where there is an enormous spike.  If I put it on something that's shaking with a jumble of frequencies it tells me what the frequencies are and just how strong each is.  The sketch shows  a funny looking signal in blue - the way you'd see it that really turns out to be two sine waves of different frequencies and amplitudes.  A Fourier Transform picks out the frequencies easily showing you wants up..  

 

The calculation is really easy with a computer -- you can do it in real time on your phone using an approximation called the Fast Fourier Transform (FFT), but that's a calculation detail - one that led to rather enormous advances in electronics and especially communications.

 

This gives us a new way of looking at the world.   In fact some of Nature is most easily groked in the frequency domain and certain bits of biology work in that domain - hearing and parts of sight for example.  When you're exposed to it and play around enough, at some point your thought processes change.  Your mind gets very comfortable in understanding the world that would normally be abstracted from your senses.  You're mind has been enhanced with a bit of mathematical thinking.  In a way it is an interface to thought itself.

 

Many domains develop powerful ways of thinking about problems.  Many types of problems become easy and you even begin to ask questions and think thoughts that would be impossible without this new approach.  It can poison teaching.  The basics can become obvious.  When you're teaching you try to search for a common language and flail around when you can have a smooth conversation with another physicist.  People claim that quantum mechanics is complicated an incomprehensible, but it's really not - at least at the calculation level.  People think visually and in terms of translations where Nature makes sense. Just say a couple of words and it's clear you're operating in the frequency domain - you don't have to slow down.  You were probably taught about particle wave duality, but that's not true.  There's something more abstract and weirder a bit deeper down that ads enormous clarity.  And a hundred other things like that...  Combined with looking for questions rather than answers it is collectively called "physical thought".. Like I said -  other domains have similar specializations.¹  I have enough math that I can sort of keep up, but that just makes me sort of bicultural. 

 

Now what if you could teach the intuition of how a physicist really thinks (or an electrical engineer, or mathematician, musician or..)..  trust me... this is rarely done.  What if you could build a computational interface that acts as a set of training wheels?  People have played around with this since the early 90s.  It goes way beyond dataviz and data science in general and tends to be domain specific.  A physicist doesn't augment her thought process by accessing the cloud or, but rather changes the way she thinks.  Even baby steps towards that are a big break with conventional models of computer augmented thinking.  Here's a lovely post by Michael Nielson on the subject..

 

Again I'm not horn tooting.. this is just an interface of sorts that allows new ways of thinking.. a different sort of augmentation. It is wonderful being around people in other fields who do this.  Some in the arts are particularly powerful .. composers in particular.  The possibility for hybridization of these approaches is fascinating. 

 

__________

¹ As is true with learning some neural reconnections are being made.  It can create physical differences in some areas of the brain.  Musicians have developed their brains such that you can spot one in a scan or even autopsy .. same for physics, but in different regions.  Also true in London cabbies
__________

 

Recipe corner ... 

 

 

You've probably had too much food so a fine secular tale from Salt Lake City read on Selected Shorts:  Ron Carlson's The H Street Sleddiing Record ..  The opening line is amazing.

 

Comments

[Jheri]

Wowzers!
I love how you think.