Design principles of the NAPA

The evolution of the main goal

The initial design goal of the NAPA was straightforward:
To produce a fully anacrophonic phonetic alphabet.

That turned out to be harder to do than one might think, especially given the rules that I settled on. It turns out that the letters F and R are quite difficult, the letter L is tricky given the rules of the game, and the letter S doesn't seem to have particularly satisfying possibilities. (Those I considered included sgraffito and seidel.)

But while I was laboring to meet this initial design goal, I realized that what I was really trying to do was to be misleading. Words like aural and djinn and ewe and gneiss were especially satisfying choices, because people would most likely hear them as their more common homophones. And when homophones were not available, unfamiliar anacrophonic words (like bdellatomy and ctenoid) were good choices.

So after struggling for some time to find appropriate choices for the trickier letters, it finally occurred to me that the true goal should not be perfect anacrophony, but rather:

To produce a maximally confusing phonetic alphabet, using anacrophony wherever possible.
The relative freedom provided by this new design goal suggested a new way to use misleading homophones: Using acrophonic words that native speakers would confuse with more common words starting with different letters. Thus fantasm, rath, and segar (which native speakers will most likely assume are spelled phantasm, wrath, and cigar) gave acceptable solutions for three tricky letters. Note that fantasm and segar are in fact alternate spellings of phantasm and cigar, while rath and wrath have completely different meanings and etymologies.

This goal of being misleading also explains why I chose oneing instead of the simpler one for the letter O. When I hear the word one pronounced, the spelling ‘O-N-E’ appears unbidden in my mind, so even though the word is anacrophonic it does not seem misleading to me. On the other hand, oneing is a fantastic word. It's anacrophonic, it's unusual, it makes sense once you know what it means, and it sounds like you're saying running with a speech impediment.

The rules

I haven't yet mentioned what qualifies as a ‘word’, and how one decides how a word is pronounced. And rules are certainly necessary, or else the game becomes too easy. My primary definition was that a word is an entry in the Oxford English Dictionary that is not marked as foreign or obsolete; my source for pronunciation is the first pronunciation given by the OED. However, I was willing to consider entries in Webster's Unabridged as well. The Webster codicil was only necessary for llareta.

Apart from llareta, there is one other word in the NAPA that stretched this definition: djinn. The OED says that

dj- is not an English combination, but is sometimes put to represent the Arabic letter jim, = English j […]. So far as these [English words coming from Arabic words that begin with jim] come under the scope of this Dictionary, they will be found under J.
Nevertheless, I am allowing the initial dj in djinn, as does Webster.

A number of useful words are disqualified by these rules. For example, the OED indicates that the names of certain genera and species are foreign (presumably the editors count them as Latin). So bdellium, which is used in some other phonetic alphabets (see the examples below), is not an option for us.

Earlier work

Phonetic alphabets involving anacrophony have been considered before. These have elements in common with the NAPA: There is also this discussion on, and this song by Barenaked Ladies, available at the iTunes Store.

If you know of other examples, please contact me or add a comment on the NAPA blog.

HTML acknowledgement

The HTML footnote technique on the main NAPA page was stolen from John Gruber's Daring Fireball blog.

Valid HTML 4.01 Transitional