Why Computer Rock History is so uniquely annoying and hard (and why I want to work on it anyway)

Riffing on my previous research questions post, I wanted to lay out some of the method and limits to taking on the topic of early electronics supply chains in the postwar 20th century United States.

Why I think this is interesting

So in general, it seems to be broadly understood by capital-H history that the United States poured a lot of resources into a) communications technology and b) mineral procurement during the war. Makes sense. With one very legible exception, that minerals procurement isn’t understood to have served or supported communications technology.

This seems unlikely for a couple of reasons:

  • the aforementioned legible exception that we’ll get into later
  • the extraordinary emphasis on materials science and solid-state physics in early electronics history would require a steady supply of materials, i.e. metals and minerals prepared for lab conditions
  • some of the major technical breakthroughs of the period used metals that simply aren’t mined in the United States, like tantalum—not to mention metals that were in commonplace use but aren’t mined in the United States, like tin

Today, the mineralogical/geological materiality of digital devices is a readily-accessed and understood talking point: the Democratic Republic of the Congo is practically synonymous with cobalt and tantalum even to people who don’t pay a ton of attention to this stuff. But the mineral basis of computers before, say, the early 2000s is pretty poorly documented. This seems like an important gap to fill in computer history!

On a more woo-woo theory-ish I Am Literally In A Geography Program note: I think a lot about the relationship of what nerds like me call space-time annihilation—a phrase used starting around the late 19th century to talk about what communications technologies like the telegraph and radio did. Space-time annihilation is really powerful, and very useful if you’re trying to do an imperialism and/or a capitalism (they often tend to go hand in hand). And because I’m a hopelessly provincial American and I grew up in Silicon Valley, I think a lot about the contradictions of the liberatory rhetoric often applied to these technologies and the coercive violence they often enable (while, let’s be fair, also being used sometimes as liberatory tools—containing multitudes, etc.).

I also unfortunately at a formative age heard that Angela Davis quote about how radical simply means grasping things at the root.” Reaching past layers of software abstraction that go into electronics and looking at the coercive violence involved in the physical stuff of it—mineral and petrochemical stuff, it’s important to note—is something I think is important because it conveys the scale and scope of the coercive violence of space-time annihilation. I think of this as similar to making an exploded diagram of a complex tool: when you take apart” the production process of the telegraph cable or the crystal radio or the integrated circuit, its parts actually spread across the whole planet. We’re literally annihilating geological time here, baby, and that’s basically always been necessary for Doing A Space-Time Annihilation.

Why it’s difficult to research this

Gaps in history don’t exist just because people don’t care about them: they exist because historians rely on records, and the records for historical supply chains are kind of shit. Even companies that have publicly accessible archives don’t tend to keep purchase orders for like, where they got copper wire from let alone where that vendor got the copper from. Materials sourcing is generally not a well-studied or documented aspect of the history of materials science. One of the first patents for making silicon crystals in 1941 names two chemical company vendors of high purity quartz sand, but good luck finding out anything about where those companies were buying materials from! Maybe they were using quartz sand from France, maybe from North Carolina, maybe Brazil, who knows? (Would you believe that lab equipment and chemical vendors are not deeply studied industries, ha ha ha.)

With communication technologies in particular, this is also difficult because most of the metals used for communications technology had multiple applications and electronics made up a relatively small sliver of it. Tungsten and tantalum were also used in steel alloys, mica was used as an insulator in like plane windows, tin was used in fucking everything. Certainly some of those materials acquired through World War II procurement efforts were going into labs, but when looking at big-picture estimates tech is probably somewhere in the tiny other” section of the bar chart.

Finally, when it comes to the one Huge Exception I mentioned earlier, even that particular material starts to get hard to trace once you get into the postwar period. During World War II, the United States government basically built a massive Brazilian quartz crystal supply chain from scratch for the sake of making radios (space-time annihilation, see??). There’s some decent scholarship out there on this effort, but the main result of the government’s massive effort to import Brazilian quartz crystals for radios was the development of synthetic quartz, which led to a huge drop in imports. Brazilian quartz was still often used in the making of synthetic quartz as basically the seed crystal and there are some threads to pull on this one, but in general it’s just extremely hard to follow this stuff and even the big exception isn’t totally an exception.

So how do I think I’m going to pull this off

Honestly, I don’t actually know if I will. But here’s my current plan of attack:

  • Picking a few definitely-needed-for-electronics metals/minerals to focus my attention on. Quartz crystal is obviously here, mica and tungsten are likely other candidates; tantalum wasn’t actually considered a critical mineral” until after the war but that in and of itself might be work looking further into. I kind of want to include tin partly because of some important labor history that overlaps with tin mining but it might be too far afield; similar thoughts with copper.
  • Another probably shouldn’t, but thought about it” material to include is petrochemical industry—there’s some pretty good scholarship on how the war accelerated chemical advances in the United States (mostly be seizing and stealing German IP), but connecting it to telecommunications in particular gets into the annoying small-sliver-of-market problem mentioned above.
  • Archival USGS cross-referencing: Historical Mineral Commodity Summaries are honestly pretty hit-or-miss in terms of how much they provide hard numbers on consumption and imports for certain materials, but they do sometimes have names of companies (which can be helpful for looking into, say, who was doing business with the Metals Reserve Company at that time).
  • National Archives research: there’s a lot of records about World War II and postwar minerals procurement, and within those there’s contracts and correspondence. Big needle in a haystack time.
  • There aren’t a lot of corporate archives that I expect will have insight into what I’m looking for (again, seems like most of the electronics manufacturers were ordering from chemical companies, not smelters or the state), but there are some decent oral history interviews that can fill in some gaps here maybe.

The point is not to find the smoking gun receipt tracing one ton of high purity quartz from a Brazilian mine to Fairchild Semiconductor; it’s fiendishly hard to do that kind of tracing today when people actually care about and keep those records. But I think if I can find evidence that electronics applications mattered to the feds and number on imports/consumption, that’s at least a gesture at something. Or I just focus my attentions on quartz or something.

December 9, 2022