G.A. English Phonetic Inversion
12.02.2011 § 1 Comment
I don’t like to dance with my body a whole lot, but I do like to dance with my mouth. I enjoy coming up with tongue-twisters and sound poems in large part because they’re fun to speak aloud, and I’m always on the lookout for new, seriously goofy strings to articulate.
Sometimes I seek inspiration in backwards words; they’re a quick and cheap way to come up with utterances with unusual consonant clusters and stuff. Other times I just remove the e’s from the ends of words that end with them, and add e’s to the ends of words that don’t, and see what funny new syllable sequences I get. Still other times I change each letter of a word to the letter it mirrors across the middle of the alphabet, i.e. a <-> z, b <-> y, c <-> x, etc. But there are only so many nugohs, bing eatinge disorderes and uozelfirav’s out there. It’s time for a new transformation I can apply to any word I see on a freeway sign or in a waiting room. It is time for phonetic inversion.
What I mean by phonetic inversion is not merely flipping words backwards with respect to phonemes rather than letters (e.g. talk instead of thguac). That’s what a Google search for phonetic inversion presently turns up, but I think this effect would be much more appropriately referred to as phonetic reversal.
Phonetic inversion is a substitution code for phonemes, pairing each with another that could be said to be its opposite in some phonological sense. The resultant phonetically inverted words will therefore bear some resemblance to the original, while still being utterly different. I will give one example to whet your appetite: ZEUR-wee-rull DAWR-eh-vame. This is “cellular telephone” phonetically inverted.
Now I’m not going to claim that my set of oppositions are scientifically definitive. They are mostly logical, as you will see, but not every phoneme has a convenient opposite, and my whim with the leftovers may well be quite different than another’s whim.
First of all, I took the plosive and fricative consonants, and paired the voiced and unvoiced versions of the same sounds.
Plosive & Fricative Consonants
boo <-> poo
do <-> to
vie <-> fie
gay <-> kay
zoo <-> sue
thy <-> thigh
Jew <-> chew
jus <-> shoo
Then I just paired consonants by category.
Lou <-> rue
moo <-> new
woo <-> you
The two leftover consonant phonemes were less neat: The h sound and the ‘ng’ sound, ŋ. They are not phonetically related, as h is voiceless, articulated in the glottis, and is somewhere between a fricative, approximant and pure aspiration, while ŋ is voiced, articulated against the velum and is a nasal. So that was a major loss for them getting paired. Also, the h sound naturally occurs in English only at the beginnings of syllables and the ŋ sound only at the ends, so any time I phonetically inverted either sound I would get an unnatural utterance! I considered resorting to the voiceless velar fricative, which is only used in the Scottish dialect of English (famously in the word loch), since it shares qualities with each of these two sounds and could thereby bridge them and serve as the inverse for both. I finally decided, however, that I’d rather subject people to initial ŋ’s and terminal h’s than introduce a foreign sound. As soon as I stepped out of G.A. English, the vowels were going to get a lot fishier; it was best to stick to what I knew.
Speaking of vowels, they essentially have three aspects: height, backness and roundedness. G. A. English has five pairs of vowels that share the same approximate height but are opposites in backness and roundedness, and this was by far the cleanest starting point for an inversion. These correspond closely with Daniel Jones’s idea of matched and equally spaced cardinal vowels.
Front + Unrounded <-> Back + Rounded
teak <-> tuque
tick <-> took
take <-> toke
tech <-> talk
tack <-> tock
There are two diphthongs that each begin with the otherwise unused a sound then move into one of the two near-close vowels that have already been paired, so these form a natural pair:
file <-> fowl
The remaining diphthong vowel is the ‘oy’ sound of foil. Its two constituent vowels, ɔ and ɪ, are used individually, but if you were two try to put these two vowels’ inverses together you’d get something like a sound which hasn’t been used since Middle English: ɛʊ. I considered pairing it with the diphthong ʌɪ, which is the long i sound as it is pronounced before voiceless consonants (as a result of a phenomenon related to Canadian raising) such as in fight, since they both end with ɪ but start with the rounded and unrounded versions of the same back open-mid vowel. However, I decided against that, since ʌɪ is an allophone with the other long i sound, aɪ, such as in fide.
The only other vowel sound that remained in G.A. English was a monophthong, ɐ, the short u sound of buzz. It is unusual in that it is a central vowel, neither front nor back, and its rounding is not as important. In these respects it is like the schwa, ə, the neutral vowel sound used in unstressed syllables. But I couldn’t very well pair these two; what would become of my own name, Doug, if it’s only syllable was unstressed? To preserve natural stressing, schwa could only be paired with itself.
So that left ɔɪ and ɐ to be paired together. I suppose it could be worse, as ɔɪ, of all the vowel sounds, makes by far the most drastic crossing from back to front, and ɐ is the only unstressed one to stay centered, so in this sense they are opposites.
So here’s the whole phonetic inversion chart, using IPA symbols:
b <-> p
d <-> t
v <-> f
g <-> k
z <-> s
ð <-> θ
ʤ <-> ʧ
ʒ <-> ʃ
l <-> r
m <-> n
w <-> j
ŋ <-> h
i(ː) <-> u(ː)
ɪ <-> ʊ
e(ɪ) <-> o(ʊ)
ɛ <-> ɔ / ɒ
æ <-> ɑ
aɪ <-> aʊ
ɐ <-> ɔɪ
ə <-> ə
For an example, let’s try this lamely nonsensical panphoneme (ignoring boring rhotic vowels):
Ah, fission you of those high-gauge, short-length Woody cowboy chaps.
In IPA, this is:
ɑ, fɪʒən juː ɐv ðoʊz haɪ-geɪʤ, ʃɔrt-lɛŋθ wʊdi kaʊbɔɪ ʧæps.
Phonetically inverted, then:
æ, vʊʃəm wiː ɔɪf θeɪs ŋaʊ-koʊʧ, ʒɛld-rɔhð jɪtu gaɪpɐ ʤɑbz.
Since this is hardly legible, let’s transcribe the result using the scheme I devised for my sound poetry:
Ae, veucem wi oif qeis ngao-kotc, joeld-rauhx Yuitou gaipu djabz.
So, yes, obviously, this is awesome.
yoim, di, xlou, veil, vaof, zeugz, zaufem, od, maom, daum!
laut, oelemtc, waurei, kloum, pri, fao’ereud, prag, klo, yaod, plaim!
pou, ngaf, ti, zo, yeur, yirt, kaud, kei, nog, gam!
It’s fun to discover when a word’s phonetic inverse is also a word. So far I’ve discovered that “closed” and “graced” are a pair, as are “you” and “we”, “down” and “time,” “juice” and “cheese,” “bomber” and “panel,” “zed” and “sought,” “bagel” and “poker,” so on. I’d really like to find a pair of at least three syllables each…
And now for some disclaimers.
Both ɪ and ʊ are technically near-front and near-back, not front and back, and are thus excluded from Jones’s cardinal vowels.
æ is technically near-open, not fully open.
In the most glaring exception to my main set of vowel inversions, ɑː is not rounded. After endless comparing of R.P. and G.A., reading about cot-caught mergers and father-bother mergers, I decided that while I contrast between three vowels, those three vowels for me are ɑ, ɒ and ɒː, not ɑ, ɒ and ɔ. For me, ɔ doesn’t really exist, except as part of the ɔɪ diphthong and as a rhotic vowel with r. In R.P. it sounds much like an o, which comes as little surprise since R.P. doesn’t really have an o; they pronounce it as əʊ. In many words where ɔ is shown before an r in the IPA, I pronounce it as o. Since I cannot come up with any set of three words that distinguish meaning by three different low back vowels, I am forced to admit that I have only two. Since ɛ must naturally pair with ɒ, æ then pairs with ɑ. However, since ɔ is prettier and more distinct to look at, I chose that symbol.
If you wanted to get nit-picky, rather than collapse all unstressed syllables into schwa, I should say that ɪ̈ <-> ʊ̈, ɨ <-> ʉ, and ɘ <-> ɵ.