Learn English – Most common consonant sound (token frequency)

Right,

This happens to be a very interesting question, unfortunately I think I can only help in part, but we'll see!

You're right, there is a historic reason for the differing orthography; if we look at the etymologies of the words, we find two very different historical roots. We will start (for no specific reason) with fur, whence inflected to form furry. We get our "fur" from the middle English furre and furren, which in-turn we obtained from the Old French "fuerrer" (where earlier writings replace the "ue" with an "o", but thats irrelevant here), the French may well have obtained this from the Old High German "fotar" (the most probable explanation for the earlier use of an "o" in the French) of little importance yet great interest, this comes from the Gothic for lining "fodr", anyway, the crux of it is: furry was never borne out of a linguistic rule upon inflecting fur, instead it was transported via its root from the Old French — fuerrer, I'm afraid I can't give you an answer as to why the Old French chose to duplicate the "f" found from the Old High German — only regional adoptions of OHG dialect will likely be responsible.

As for Fury, we again look to the French / Middle French, this time the word is "furie" but their scavenging is not from Old High German, as in the case of fur — instead from the Latin furia, this all boils down neatly to the Greek stem "thor" whence "thorub" (the former to hurry or rush, the latter a tumult). The transportation of the pronunciation along with the single "r" construction explains the classic rule's retention. fury {fjʊri}

Hopefully this goes some way if not all the way to answering your question.

I’m not sure that there is a chain shift, though it is possible. Ultimately, you would need to do careful measurements of though, doe, roe as said by:

1. members of your childhood cohort who pronounce though with [dð] not [ð],
2. members of your childhood cohort who pronounce though with [ð] not [dð],

I’m mildly doubtful of a chain shift because apico-dental and apico/lamino-coronal sounds can coexist in the same language, and, to make things easier, you have affrication help to differentiate voiced th from d.

That’s the best answer I can give on the available information!