Should Otto Graf Lambsdorff read family Graf Lambsdorff, as Manfred Freiherr von Richthofen reads family Freiherr von Richthofen? #709

Since 1919 a German noble rank is no longer a title: it is part of the legal surname, written between the given name and the particle (Weimar Constitution, Art. 109). Graf (count), Freiherr (baron), Fürst, Prinz, Herzog and Ritter are all the same kind of word in this respect.

The parser reads only one of them that way. freiherr is in both TITLES and particles, so it chains into the family. Every other rank is a title only.

Measured on the #701 branch (bcea15a4):

Why freiherr could be a particle and the others probably can't, as is. Counts from uv run python tools/name_usage.py <word>, 2026-10-17:

Freiherr is never anyone's plain surname. Graf, Ritter and Herzog are common ones (Steffi Graf, John Ritter, Werner Herzog). A particle joins the words after it into the family, so making graf a particle needs a look at what that does to a name that ENDS in the word (Steffi Graf, P6's trailing-particle attachment) and to one that begins with it.

Questions:

• Should the ranks read as part of the family where a surname follows them (Otto Graf Lambsdorff → family Graf Lambsdorff)? Or should they read as titles everywhere, freiherr included? Today's split is the one reading that follows no principle.
• If they belong in the family, is that a new kind of word (a rank that joins the surname behind it, but is a name word when it stands last) rather than a particle? particles carries P1/P6 behaviors these words should not get.
• The sorted form Lambsdorff, Otto Graf puts the rank after the given name, as German library catalogs (GND/RAK) write it. Is the rank then a middle name, a family word, or a title?
• Austria abolished noble names in 1919 rather than folding them into the surname, so an Austrian Graf is no legal part of the name. Should the reading differ, or is that out of scope like every other language-detection question?

Related: #701 (H6, whose Accepted clause records that the particle stop, not the 1919 argument, is what keeps Freiherr in the family), #606 (graf shipped as a title under the graf-class judgment).