Unicode Demystified

Singleton Decompositions

A small number of compatibility characters have one-character canonical decompositions, often referred to as “singleton decompositions.” These characters count as canonical composites, but are fundamentally different from the precomposed characters—the other canonical composites—because they don't actually represent the composition of anything. They're more akin to the compatibility composites, but they have canonical decompositions because replacing one of these characters with its decomposition doesn't lose any data. You see this kind of thing when a character, for historical reasons, is assigned two code point values in Unicode, usually because this duplication had happened in some source standard and Unicode's designers ...