Fractal Vagueness

Fractal vagueness happens when a term or phrase is defined in terms of other words or phrases that are vague or ambiguous. Requests for clarification of those other words or phrases just repeats the same problem in the next level.

Rough example:

Bob: "A glisko is a horse with a nib".

Lew: "What's a nib?"

Bob: "A nib is a plobb without a lock."

Lew: "What's a plobb?"

Bob: "A plobb is a fipple with at least 7 screws."

Lew: "What's a fipple?"

Bob: "A fipple is grommin that can hold at least a pint of water."

Lew: "What's a grommin?"

Bob: "A grommin is a two-sided yatter."

Lew: "Never mind, sorry I asked. I think I've had enough 'yatter' for the day."

--top

I suggest Bob's last line be, "A grommin is a two-sided glisko." See what I did there?

Recursion will break a horse's back!

[It's also unavoidable in any language capable of defining all its words. There must either be circular definitions or an infinite descent of definitions.]

True. In practice as long as the communicating parties agree on a definition or interpretation of a definition, they are communicating "good enough". A lot of concepts come from the mere experience of living as a human. This is something that AI struggles with: AI hasn't lived the life of a human and can't use that experience as a common reference point. The languages of kids these days often confounds elders because it's based on the assumption of social networking experience and usage of new devices or services. "Her likes are all fake" means that somebody's popularity credit (real or online) has been "bought", using the analogy of gaining Facebook "likes" via bribery or bartering of some kind. Perfect definitions don't use analogies, but humans quite often use analogies. Google the origin of "milling around" or "milling about", for example.

The example is linear, but the vagueness can also be a tree, DAG, Graph, etc., such as one term being defined using two or more vague terms, and those paths could eventually circle back on themselves.