The authors are the famous Kernighan and Brailsford (!). They spent their life on stuff like this, and know him personally very well. They are talking about spacing much finer than half-character.
There's vertical spacing of 1/24″ in some places, which is twice the usual Selectric precision (6lpi with half-line motion). To me the most likely explanation is a Selectric variation with twice the detents, or even a third-party add-on, made specifically for technical reports.
It is too late for me to edit my other comment, but IBM offered ratchet wheels with double the normal number of teeth (54 rather than 27), providing quarter-line motion, for the Selectric.
Some other comments¹ have said that some Selectrics had a half-space control, so that would provide 1/24″ horizontally. I think that's the finest in the paper, unless I've missed a case.
The authors are the famous Kernighan and Brailsford (!). They spent their life on stuff like this, and know him personally very well. They are talking about spacing much finer than half-character.