Main question related to this bug:

Can a localized version of Thunderbird have access to both the localized string and the original english string, too?

Intention:

With OpenPGP, when creating a personal key for the user, we want to prepare for a key loss scenario. This might happen, if the user sets a master password and forgets it, and can no longer access the secret key. In this scenario, it's advisable to send a revocation statement that can be sent to correspondents. To prepare for that scenario, we automatically create a file on disk that contains the revocation statement (because it cannot be created any more, once the corresponding password is lost).

That file is a plain text file with some technical data, that can be automatically processed by a correspondent's OpenPGP software.

It's allowed to have some additional human readable text in that file. This can be used to explain what the remainder of the file contents is about.

Because that file might be viewed by both the local user and other correspondents, not everyone might speak the same language. Therefore, I recommend that the human readable description in the file is contained in the english language.

It has been suggested in the review for bug 1621128 that the text should be included in the localized Thunderbird language used by the person that automatically created that file.

This might be helpful, but IMHO, the localized text shouldn't replace the english text, it sould be in addition to the english text.

How could we implement that?

It only knows one language at a time. You'd have to write the English in a non-localized way, have a duplication (localized) that you don't use at all when language is English.