Leibniz Manuscript Research
Transcribed and translated unseen Leibniz manuscripts on calculus and binary. Wrote a transcription key for future scholars.
- Research
- Neo-Latin
- OCR
- Calculus
- Binary
- History of Science
Roughly eight months of remote research mentored by Dr. Harry Lewis (Harvard) and Dr. Lloyd Strickland (Manchester Metropolitan University). I delved into the unpublished manuscripts of Gottfried Wilhelm Leibniz, the co-inventor of calculus and the originator of the binary system that runs every modern computer.
The work
I transcribed and translated previously unseen 17th-century Latin manuscripts on calculus and binary. There were two hard parts, and they stack:
Neo-Latin translation. Leibniz’s mathematical Latin is not classroom Latin. It is technical prose written by someone inventing the vocabulary as he goes, in a period before mathematical notation settled, so a term’s meaning has to be reconstructed from how he uses it rather than looked up.
Handwriting. These are working manuscripts, written for an audience of one: Leibniz himself. He wasn’t writing for future scholars; he was scribbling ideas down as fast as they came, in barely legible cursive, abbreviating and correcting as he went. Three hundred and fifty years later, OCR helps, but the characters it fails on are exactly the ones that matter most.
17th-century Leibniz manuscript
[ scan coming: original + my transcription ]
The transcription key
The output I care about most isn’t a single translation. It’s the universal transcription key I wrote to go with it.
Every scholar who works on these manuscripts re-derives the same mappings from scratch: this mark is that character, this abbreviation expands to that word, this scratch-out means a revision. Writing it all down once, consistently, turns that scattered private knowledge into something anyone can use: the next researcher, or an AI model being trained to read the rest of the corpus.
That’s what connects this to the rest of my work. It’s the same problem I ran into on the Physical AI team at Janicki: taking something a skilled person does on intuition (reading a messy manuscript, or spotting and grabbing the right part on a factory floor) and making it explicit enough that a machine can do it reliably, at scale.
Why a robotics student did Latin research
Because Leibniz was chasing the same question I am: how do you get a formal system to capture something real? He was figuring out how to write down continuous change, and how to encode any number using just two symbols. Reading his drafts is watching him invent that language in real time, before the notation existed to make it clean.
His binary is also not abstract to me. I’ve built it with my hands: a 5-bit binary adder on a single breadboard, out of nothing but logic gates, after first building the gates themselves from transistors. The photos below are what his two symbols look like in copper and silicon.