The link you gave does not support this statement.
For example suppose a geometer explains something to another geometer just by drawing diagrams with no verbal or textual annotation. You can plausibly claim that "geometry diagrams" are a language, but now you are defining language so broadly that the term is almost useless.
Google, define language:
lan·guage ˈlaNGɡwij/Submit noun 1. the method of human communication, either spoken or written, consisting of the use of words in a structured and conventional way. "a study of the way children learn language" 2. the system of communication used by a particular community or country. "the book was translated into twenty-five languages" synonyms: tongue, mother tongue, native tongue; More
I would say "geometry diagrams" are a language, by the second definition.
Mathematics is often expressed in language, That language us usually a mix of a human natural language English/French/German/etc plus symbolic extensions, and that is very poorly typed.
In the recent failed proof of P!=NP, the flaw was a type error:
https://cstheory.stackexchange.com/questions/38803/is-norber...
Type errors are rather common in math proofs that end up retracted.
But it's hard to call them type errors, because it's not like a computer program where the typechecker detected the collision or failed in some way: the proofs make very little mention of types at all, and it's up to the readers to apply a type-system from their own heads.