Mechanical music - search for patterns

I’ve been looking for interesting pattern-making devices in mechanical music, but not finding anything. Lots of musical sequences from rotating drums (like music boxes) or punched paper (like pianolas) but I can’t seem to find any historical devices that create variations or interference patterns. I suppose this is partly because it would be difficult to do in a satisfactory way when systems of tonality comes with a lot of ‘rules’ that would be hard to encode in a machine (and indeed are hard enough to encode in software). But where are the machines generating interesting rhythms from interference patterns etc? Designs from the 9th century Banū Mūsā brothers and 12th century Al-Jazari are interesting but as far as I can tell store musical sequences but not perform pattern transformations.

1 Like

Interesting topic ! If you haven’t heard Siegfried Zielinski’s lecture about Ibn al-Haytham, it’s a nice introduction to mechanical media-related devices. I think I remember a book related to the symposium, it might be worth checking it out. This hydraulic organ is rather surprising, although I doubt it matches your research criteria.

1 Like

Thanks that’s a great talk! It was especially nice to watch it after Eglash’s talk, there is an interesting connection between Zielinski’s criticism of Hegel’s “gradation of the world’s spirit” and Eglash’s criticism of extractivist approaches in STEM, with both Zielinski and Eglash advocating for cultural feedback loops of knowledge generation. Good to also learn about Max Bense’s work.

I didn’t see a lot in the talk about mechanical devices, but will have to search out the catalogue for “Allah’s Automata” and Zielinski’s other work.

1 Like

I recently came across these “Ancient Incan Water Whistles”. More on it [here].(

Not quite mechanical music, but incredibly ingenious!


Wow, this is amazing sound synthesis !!

1 Like

Returning to my original question - why no generative patterns in the history of mechanical music?

In 2015, ZKM Karlsruhe had an amazing looking exhibition “Allah’s Automata: Artifacts of the Arab-Islamic Renaissance (800–1200)”. The catalogue is out of print, but while visiting ZKM I found a copy in the library. It includes an interesting essay “Divine water clock: reading al-Jazari in the light of al-Ghazali’s mechanistic universe argument” by Ayhan Ayteş (I took some slightly shaky photos).

Ayteş quotes Iranian philosopher Seyyed Hossein Nasr in pointing out self-imposed limits to technology. Here’s the full paragraph that he quotes from:

Islamic civilization had the means to make complicated machines and apply them to the problems of the daily life of the Islamic community. But like the Chinese who had gunpowder but never made guns, the Muslims never took that step which would mean the creation of a technology out of harmony with the natural environment. Their works on machines dealt with a variety of subjects all the way from agricultural and transportational devices which were actually used in everyday life to complicated clocks which were the joy of caliphs and princes, to other complicated gadgets and devices which at their extreme became combined with magic and magical practices. They did not make practical use of all they knew in this domain, feeling instinctively the danger of the development of a technology which makes use of metals and fire, both elements alien to the natural environment, and which therefore ultimately results in the loss of that equilibrium vis d vis nature which is so central to the Islamic perspective and whose destruction is such a danger for modern man.
From “Islamic Science: An Illustrated Study”, Seyyed Hossein Nasr

A really interesting passage that makes me think about the self-imposed limits in collapse computing.

In his conclusion, Ayteş reflects on Al-Jazari’s water clock:
"… the water clock represents the concept of divine causality by assigning each element of the water clock with a distinct function for the particular aspect of a causal chain, which ultimately subjects the natural chain of events to the “first cause.”

I’ve only skimmed it so far, but in his essay, I think Ayteş is making a point that we should view this technology from the point of view of the philosophy of the time, rather than imposing our own perspective on it through the lens of the history of technology. I feel quite far from being able to do that, but think that in the then contemporary Islamic view of causal links, the mechanical universe and the divine could be where the answer to my question lies.

1 Like

This week went down a little bit of a historical generative music rabbit hole–
the 18th century is just a minute ago compared to the Banū Mūsā brothers, and I’m not sure six-sided dice are correctly classified as machines or devices, but I started looking at the 1700s musical dice games (aka Musikalisches Würfelspiel). A musicologist I met once told me they are relatively understudied and not thoroughly archived. While the ones I’ve read are often stems of melody, harmony, and counterpoint that can be arranged in different ways based on dice rolls, I wonder if in the bowels of some German or Italian library there’s one that was concerned with rhythm that allowed the player to transform the pattern? It does seem possible to me.

There’s a famous one of these attributed to Mozart which has 176 one-bar fragments of music, but contains no instructions as to how exactly dice are supposed to select them or shift the patterns. Just yesterday I found this github page of one musician’s playable experiment with those stems which seems kind of fun, even though it isn’t a super generative pattern maker Musical Dice