Unraveling the Un-unravelable

Tuesday, August 27, 2019 · 3 min read

Say what you will about The Big Bang Theory, it has its moments, and this is one of my favorites (watch it to the end of the clip!):

Putting aside the wonderful meta-dramatic irony of this moment, the word “un-unravelable” is the kind of thing you can sit and think about for hours. I know because I have.

See, you might think that the “un-“ prefixes cancel, such that “un-unravelable” means the same thing as “ravelable.” This is kind of like how “un-unforgettable” probably means “ordinary.” And indeed, Wiktionary agrees with this assessment, suggesting instead “ravelable” as the “simpler and more immediately logical choice.”

But clearly that’s not what Dr. Cooper meant in this case! “Ravelable” means can be raveled whereas Dr. Cooper meant cannot be unraveled, which are at least intuitively separate concepts. So what’s going on?

First of all, we need to figure out what “ravel” means: it means to entangle. Something that is “raveled” is entangled and knotted like the HDMI cords behind your cable box. The sleep that Macbeth murders is the “sleep that knits up the raveled sleeve of care” (2.2.37). Actually, that’s not the full story. “Ravel” also means to disentangle, making the word a so-called “Janus word” (read more at Merriam-Webster). But let’s stick with the former definition for the moment.

Okay, so now we can start tacking on prefixes and suffixes. But we have to be careful! The dynamics of affixes are tricky — the order in which we tack them on matters.

For example, “ravel-able” would mean can be entangled, so spaghetti is ravelable. Then “un-ravelable” could mean cannot be entangled, so ravioli is unravelable.

On the other hand, “un-ravel” would mean to disentangle, so Sherlock unravels mysteries. Then “unravel-able” could mean disentanglable, so the mysteries Sherlock solves are unravelable (by him, at least).

To summarize:

Alas! The associative property has broken down! This means we need to examine each of the five possible paranthesisifications of “un+un+ravel+able” separately (why five?).

Here they are:

  1. (un+un)+(ravel+able) and
  2. ((un+un)+ravel)+able are easy: neither of them typecheck, because “un+un” is meaningless.
  3. un+(un+(ravel+able)): “ravelable” means can be entangled, so “un-ravelable” means cannot be entangled, and thus “un-unravelable” means is not such that it cannot be entangled, i.e. can be entangled. For example, your new favorite organizational scheme might seem perfect at first, but after a few weeks you will find that no scheme can fend off the entropic tendencies of the universe, and that your files are by nature un-un-ravelable.
  4. (un+(un+ravel))+able: “unravel” means disentangle, so “un-unravel” means re-entangle (by analogy to “un-undo” meaning “redo”) and thus “ununravel-able” means re-entanglable. For example, a jigsaw puzzle that you mix up again after solving is ununravel-able.
  5. un+((un+ravel)+able): “unravel” means disentangle, so “unravel-able” means can be disentangled, and thus “un-unravelable” means cannot be disentangled. For example, Dr. Cooper’s web of lies is un-unravelable.

So what’s really going on here is that the prefix un- has many faces (a Janus prefix?); it reverses verbs and negates adjectives. Perhaps this is clearer in the language of first-order logic, where -able introduces an existential quantifier over a way to do the thing (and thus do(w, x) does w to x).

  1. N/A
  2. N/A
  3. (un+(un+(ravel+able)))(x) means ¬¬∃w. raveled(do(w, x))
  4. ((un+(un+ravel))+able)(x) means ∃w. ¬¬raveled(do(w, x))
  5. (un+((un+ravel)+able))(x) means ¬∃w. ¬raveled(do(w, x))

Only (5) puts an odd number of negations outside the existential quantifier, flipping it to a universal quantifier, and expressing that you cannot unravel whatever needs unraveling.

Why doesn’t this effect happen with “un-unforgettable”? It’s because “unforget” isn’t a common word, and so we reach for interpretation (3) rather than (5). Though now that I think about it, “unforget” is an excellent replacement for “remember.”

◊ ◊ ◊