Generally speaking, ascenders and descenders should not touch each other in the text body. This is self-explanatory as it could lead to text that is difficult to read.
However, Erik Spiekermann points out that there are use cases that profit from partly overlapping letters, e. g. to render headings more forceful (Spiekermann 1986: 43).
He also provides an example (Spiekermann 1986: 42, arrows by me), which says:
There’s a rule according to which descenders and ascenders must never touch. There’s an exception to this rule which states that they may touch if it looks better.
In this example, the letter g touches the letters ü and R in the following lines.
In digital typography, letters are not necessarily connected to the size of their “metal” block (which still exists virtually) anymore, e. g. in the font Amsterdamer Garamont, the lowercase h exceeds its block on the top while the lowercase p exceeds its block on the left and bottom side (Forssman and de Jong 2014: 86). The authors write (transl. by me):
In manual typesetting, this would not be possible; the overlapping parts of the letter would collide with the letters on the lines above and below, and break.
The question:
Before the invention of digital type, in manual typesetting with lead body type, how were overlapping letters produced, and how were the problems described above solved?
Works referenced:
Forssman, Friedrich and Ralf de Jong. Detailtypografie. Mainz 2014 (2002).
Spiekermann, Erik. Ursache & Wirkung: ein typografischer Roman. Erlangen 1986.