Gordon Pask was an eccentric inventor, among other things, born in Derbyshire but later living in London. Out of all his weird experiments, the one that intrigues me the most is known as "Pask's Ear".
This was an attempt to grown a kind of analog computing device which could choose its own method of relating to its surroundings.
Surviving information about Pask's Ear seems to be fragmentary, but as far as I've been able to determin, Pask used a solution of acidified ferrous sulphate as his medium.
For instance, one such recipe may be found in this paper: https://www.researchgate.net/publication/221798446_One-dimensional_ferromagnetic_dendritic_iron_wire_array_growth_by_facile_electrochemical_deposition
By passing electric currents through this solution he was able to persuade it to grow an "ear like" structure which was able to differentiate between two different tones, at 50 Hz and 100 Hz.
The structure supposedly resembled an ear insasmuch as it grew cilia-like structures. I have carried out some experiments myself with ferrous sulphate and I have been completely unable to get it to grow the kinds of long wires that Pask describes. My thoughts at this point are that Pask must have used considerably higher voltages than would normally be used for electrodeposition.
Tin readily produces long branching crystal structures at low voltages; iron, not so much. But tin salts seem a little unstable in solution and tend to degrade over time, as well as smelling perfectly awful. I've turned to exploring zinc solutions. At least they smell OK! Zinc acetate readily produces dendrites, and there must be a way to persuade these dendrites to grow long and thin instead of short and stubby.
The evolution of hydrogen that accompanies these experiments is rather annoying, as is the tendency of the electrodes to deteriorate rapdily. I've recently switched to using tantalum as the electrode material; it resists corrosion and is cheaper than the platinum that Pask used. Of course, the idea that not only a simple "ear" but an actual supercomputer may be grown like this, is intensely fanciful, but many researchers in the 1950 believe that one day, we would do exactly this.
Some research along these lines continues.
See for example: https://theconversation.com/we-built-a-brain-from-tiny-silver-wires-it-learns-in-real-time-more-efficiently-than-computer-based-ai-216730
The research of Hubler is also relevant and interesting: https://www.nature.com/articles/srep15044 Pask's Ear: https://www.researchgate.net/publication/227722393_To_evolve_an_ear_Epistemological_implications_of_Gordon_Pask's_electrochemical_devices https://en.wikipedia.org/wiki/Gordon_Pask
Share this post