Kán yu andastánd wot aim seiing?

Whatever you think of the complexities and ambiguities of English spelling, reforming it is not a realistic prospect this side of an independently-governed moon colony.

Or, to put it another way: Woteva yu think ov dhi kompleksitíz ánd ámbigyúitíz ov Inglish speling, rifōming it iz not a rialistik prospekt dhis said ov an indipendentli-gavand mún koloni.

Because however unattainable reform might be in the real world, everything is possible in the imagination. Only very boring people permit their actions to be governed by the question, “Is it practical” when they could be asking, “Is it fun?”. So here’s my question: Suppose you were the governor of that moon colony. How would you propose a more-or-less phonemic English could be spelt?

Bikoz haoeva anateinabul rifōm mait bi in dhi rial wǎld, evríthing iz posibul in dhi imájineishon. Ounli veri bōring pípul pǎmit dhe ákshonz tu bi gavand bai dhi kweschon, “Iz it práktikul” wen dhei kud bi āsking, “Iz it fan?”. Sou hiaz mai kweschon: Sapouz yu wǎ dhi gavana ov dhát mún koloni. Hao wud yu propouz a mo-o-les fonemik Inglish kud bi spelt?

In this blog post I’ll present a system of my own, and for comparison also refer to a quite different system I designed a number of years ago.

In dhis blog poust ail prizent a sistem ov mai oun, ánd fo kompárison ōlsou rifǎ tu a kwait difrent sistem ai dizaind a namba ov yiaz agou.

— Where to Start —

To begin, you need to make a number of decisions. These include:

Tu bigin, yu níd tu meik a namba ov disizhonz. Dhíz inklúd:

  • Which dialects is your system intended for? Perhaps all of them (good luck with that), perhaps only your specific accent, or perhaps something in between. The system herein aspires to work for most non-rhotic dialects of English.
  • Wich daialekts iz yo sistem intended fo? Paháps ōl ov dhem (gud lak widh dhat), paháps ounli yo spesifik aksent, o paháps samthing in bitwín. Dhi sistem hiarin aspaiaz tu wǎk fo moust non-rotik daialekts ov Inglish.
  • Do you wish to exploit familiar spellings to keep your system easy to learn, or do you want to give English spelling a clean start by building a consistent, sensible system from the ground up? Herein, I’ve gone for the familiarity approach, up to a point.
  • Du yu wish tu eksploit familya spelingz tu kíp yo sistem ízi tu lǎn, o du yu wont to giv Inglish speling a klín stāt bai bilding a konsistent, sensibul sistem from dhi graond ap? Hiarin, aiv goon fo dhi familíáriti aprouch, ap tu a point.
  • Will there be a symbol reserved for schwa, and if so, which? You could re-use an existing letter to ensure it flows easily under the pen, but at the cost of making the remaining letters work that much harder to fill the gap. If not, people will go on misspelling separate as they always have. In this case I’ve chosen not to represent schwa.
  • Wil dhe bi a simbol rizǎvd fo shwā, ánd if sou, wich? Yu kud ríyúz an egzisting leta tu ensho it flouz ízili anda dhi pen, bat át dhi kost ov meiking dhi rimeining letaz wǎk dhat much hāda tu fil dhi gáp. If not, pípul wil gou on misspeling separeit az dhei ōlweiz háv. In dhis keis aiv chousen not tu reprizent shwā.
  • Will you have exactly one spelling per pronunciation, or will you build in some redundancy? Also, perhaps you’d like to use spelling to indicate which syllable is stressed. Here I’ll keep things simple for the most part, but discuss possible extensions at the end.
  • Wil yu háv egzáktli wan speling pǎ pronansíeishon, o wil yu bild in sam ridandansi? Ōlsou, paháps yúd laik tu yúz speling tu indikeit wich silabul iz strest. Hia ail kíp thingz simpul fo dhi moust pāt, bat diskas posibul ekstenshonz át dhi end.

My previous system was tailor-made for my specific dialect, was considerably more radical at the expense of being hard to remember, reserved the letter i for schwa, and included a rather convoluted system for marking stressed syllables.

Mai prívios sistem woz teila-meid fo mai spesifik daialekt, woz konsidarabli mo rádikul át dhi ekspens ov bíing hād tu rimemba, rizǎvd dhi leta i for shwā, ánd inklúded a rādha konvolúted sistem fo māking strest silabulz.

[The self-translations will cease at this point. They’ve been proofread a few times, but errors may remain.]

— Consonants —

In English, n becomes a velar nasal when followed by g or k (as in anger, angle, ankle, anchor, etc), and in those cases it makes sense to spell the combination ng or nk as we normally do (since the velar nasal can be regarded as an allophone of n). But we also use the spelling ng to represent a velar nasal on its own, leading to the ambiguity whereby the g in anger is pronounced but that in hanger is not.

In my old system, I decided that a velar nasal not followed by g or k would be spelt yn, resolving the ambiguity and taking advantage of the fact that the letter y (always a consonant) cannot occur immediately before another consonant. This time, however, I’ve decided that the anger/hanger ambiguity is a tolerable one, and to simply spell the velar nasal ng as English speakers are used to.

Similarly, in English the spelling th sometimes denotes the unvoiced fricative of thieves, sometimes the voiced fricative of these, and sometimes simply the sequence of sounds represented by the letters t and h (the famous-to-the-point-of-being-a-cliche example being pothole).

In my old system I used the spellings hs and hz for unvoiced and voiced th respectively, taking advantage of the fact that the h sound never occurs immediately before another consonant, and having the second half of the digraph be something that normally represents a sound of the some phonological category (unvoiced and voiced fricatives respectively). This time I’ve decided to spell the unvoiced th simply as such, and (borrowing from some other languages) to use the spelling dh for its voiced equivalent.

In this system, the letter j has the pronunciation that English speakers would expect, and the same goes for the digraphs ch, sh and zh. The letter c never occurs on its own (as its two main pronunciations are spelt s and k), but only as part of the digraph ch. One can envisage a later reform of the reform in which the surplus h is dispensed with, but for the sake of familiarity I’ve decided to leave ch alone. The letters q and x do not exist in my alphabet at all.

— Vowels —

That’s enough about consonants. The real fun is with the vowels, which I’ll describe with reference to John Wells’s lexical sets.

In my dialect, the PALM/BATH/START sets are merged, as are the LOT/CLOTH sets, and THOUGHT/NORTH/FORCE sets. If your dialect has slightly different mergers (e.g. BATH merged with TRAP instead of PALM), you might have to modify the system slightly and use alternative spellings for some words, but we’d still be able to understand each other. If you dialect has significantly different mergers, then my system may not be suitable for your dialect. I expect that nearly all non-rhotic dialects of English would — with minor adjustments — find it workable.

Here’s my list of vowel spellings and the corresponding lexical sets:

á … TRAP
ǎ … NURSE [was ü in earlier drafts]
i … KIT
u … FOOT
ai … PRICE
ao … MOUTH
ei … FACE
ia … NEAR
ou … GOAT

In my old system. I chose spellings for vowels and diphthongs that closely mirror their pronunciations in Australian English (e.g. reserving certain symbols for front, central and back vowels respectively). In this new system, I’ve stuck more closely to traditional and familiar values that are not tied to a specific accent. That said, I’ve referred to accents I’m familiar with in order to make sure diacritics change vowel quality in reasonably consistent ways.

There is no symbol for schwa. To represent schwa, one should choose the vowel best suited to the role in the context (taking into account the pronunciation used when the word is overarticulated or sung, the pronunciation used in more conservative accents, the traditional spelling, the etymology, and so on). I do have one rule, though: to keep things simple, vowels that are frequently reduced to schwa should normally be represented with one of the basic five symbols a, e, i, o, u — no diacritics, no digraphs. For example, the re in reform contains the FLEECE vowel (í) when clearly articulated, but because it is often reduced to schwa, we compromise and use the KIT vowel symbol (i) instead.

(The indefinite article a would be spelt ei when there is a reason to emphasise it, but otherwise I favour a rather than e for several reasons. These include familiarity, maintaining the similarity between a & an, and avoiding confusion with the word air. Even the most consistent languages in the world have exceptions for some of their most common words. The word and maintains its diacritic because, although the vowel is often reduced in a sentence, it is always articulated when the word is spoken in isolation.)

Because the DRESS, KIT, LOT, CLOTH and FOOT vowels (also TRAP, but that’s unimportant here) never occur word-finally in English, you may omit the diacritics on ē, í, ō, ú when they occur as the final letter of a word (i.e. spell them e, i, o, u respectively).

A comment on my use of diacritics. Although I’ve tried to keep the spellings of phonemes reasonably familiar, I think this can be taken too far. A system based entirely on the most common correspondences between spelling and sound in English (say, komyoonikayshun for communication) risks being perceived as juvenile. I think diacritics are widely perceived as sophisticated, so may help counteract that effect. However, they do place an extra burden on handwriting and risk being confused with commas from the line above, so it pays not to overdo them. That’s why I’ve outlawed diacritics on reduced vowels, advised omitting most of them from word-final vowels, and made a point not to include them in digraphs.

— Extensions —

As an optional extra, here’s a suggestion for how my system might be extended to include stress marking:

  • Add a h or y between the nucleus and coda of the stressed syllable (aesthetically, I generally favour h after ao, u, and y after e, i). For example, communication without stress marking is komyúnikeishon. With stress marking, it becomes komyúhnikeiyshon.
  • To mark stress on a syllable that lacks a coda, represent the coda with an apostrophe. Without stress marking, ambiguity would be ámbigyúiti. With: ámbigyúh’iti.

This could be used routinely, or only when potential for confusion exists. Redundant stress marking might be used to differentiate between homophones.

(Note: I do not use apostrophes in other contexts, e.g. no apostrophes for common contractions or possession. Apostrophes and diacritics seem a little too fly-specky in conjunction.)

We might use etymological spellings as well as redundant stress marking to differentiate between some homophones. Perhaps dhe for there, dhēy for their, and dheia for they’re, for example. If a system such as this were actually used, the community would soon develop some conventions.

— Postscript —

I’d like to end by reiterating what I said at the beginning: that this is not a serious proposal for an English spelling reform, but is intended as entertainment. Please feel free to use the comments section appropriately as a playground.


7 Responses to “Kán yu andastánd wot aim seiing?”

  1. dainichi Says:

    I thought it would be interesting to create a system which keeps some of the quasi-regularities of English orthography. I’ll try to render some of the text that you wrote with your system – with my system. I apologize in advance for mistakes, I’m sure there are some.

    Wutévver u think uv the kumpléksiteze and ambigúiteze uv Inglish spelling, rifórming it iz not a realístik prospekt this side uv an indipénduntly govvernd moone kolluny. Bikúz howévver unutánubl rifórm mite be in the reul werld, evrything iz possibl in the imagináshun. Onely verry boring peple permít thare akshunz to be guvvernd by the kweschun “iz it praktikl” wen thay kood be asking “iz it fun”. So here’z my kweschun: supóze u wer the guvverner uv that moone kolluny. How wood u prupóze a mor-or-les funémik Inglish kood be spelt? In this blog poste y’l prizént a sistem uv my one, and for kumpárrizun olso rifér too a kwite difrunt sistem y dizíned a number uv yerez agó.

    I’m not a native English speaker, but the dialect I’m assuming is a cot-caught merged rhotic dialect. I’ve far from completely formalized the rules, but below is a draft of some of the most important ones.

    1. Stressed “long” vowels are followed by a single consonant, then a vowel (but see below) – or a vowel.
    2. Stressed “short” vowels are followed by 2 or more consonants or a word-final consonant. If the vowel is only followed by one consonant in the pronunciation, double it.
    3. For unstressed long vowels and stressed long vowels followed by 2 or more consonants in pronunciation, insert a silent “e”. This would usually come after the first following consonant, but certain 2-consonant sequences (st, sk, pl etc.) are not split. In case of these sequences, the silent e comes after the 2-consonant sequence.
    4. e’s are pronounced as in “he” at the end of one-syllable words.
    5. y’s are pronounced as in “by” when stressed at the end of a word, otherwise as in “fully”.
    6. Stress is marked with an accent unless it comes on the first syllable of a word.
    7. Vowels are pronounced (and spelled) differently at the end of words. I won’t list all the cases.

    As you notice, there’s not a grapheme-phoneme one-to-one correspondence, but the idea is that pronunciation-to-orthography and orthography-to-pronunciation translation should be (almost) deterministic and not too complicated. Also, I’ve made it an explicit goal to try to keep some resemblance to current English orthography.

  2. Adrian Morgan Says:

    Thanks for sharing that. I haven’t had time to examine it in detail, but here are some comments based on a cursory glance.

    First, you didn’t say but I’d guess you speak an American dialect of English, given that you sometimes use “u” and sometimes “o” for what I would consider the same vowel (“u” in “whatever”, “of” and “complexities”; “o” in “not”, “prospect” and “colony”). I suspect some of these are mistakes, but others probably show that the LOT and CLOTH vowels are not merged for you. I must say I have very little grasp of the distribution of these vowels in dialects where they are distinct. But elsewhere you use “u” for the STRUT vowel (e.g. in “unattainable”), which I can’t explain.

    The desire of speakers around the globe to remain mutually comprehensible in writing is, of course, one of the barriers to reform. No conceivable system would make sense everywhere, so reform would have to mean the final divorce between Englishes of different regions, putting trivial arguments over the “u” in “colour” firmly in their place. The world isn’t ready for that.

    One definite mistake I’ve spotted is the “o” in “governed”.

    Your system reminds me somewhat of Interspel, which is one of the better sincere reform proposals I’ve seen, although the Wikipedia page ain’t kidding when it warns that “This article’s tone or style may not reflect the encyclopedic tone used on Wikipedia” (it’s obviously based on a pro-Interspel source).

  3. dainichi Says:

    Hi Adrian, thanks for your comment on my comment!

    Yes, “govvernd” should definitely be “guvvernd”.

    As I said, I assume COT and CAUGHT are merged, which I think means LOT and CLOTH are too, i.e. I assume all vowels in this table are the same: http://en.wikipedia.org/wiki/Phonological_history_of_English_low_back_vowels#Table. But the vowel of OF and NOT are different for me, the former is an open-mid back unrounded vowel /ʌ/, while the latter is an open back vowel /ɑ-ɒ/ (roundedness allophonic). Apart from the one you found, I don’t see any errors in my use of “u” versus “o”.

    To explain, in my orthography, “u” corresponds to 3 sounds: Stressed, it’s either /ju/(“long”) or /ʌ/(“short”), depending on the following consonant(s) (as explained). Unstressed, it’s usually a schwa. I realize that can cause problems in cases like “unutánubl”, where you might say the first syllable is not a schwa, but an unstressed but unreduced /ʌ/, or even that it carries secondary stress. I only operate with one stress per word, and I personally find that pronouncing the first syllable the same as the second, as a schwa, sounds fine. If you insist it’s not a schwa, I guess that’s one case of indeterminism in my system, unstressed “u” is either schwa or /ʌ/ (which are pretty close in the dialect I assume, but I realize maybe not in yours).

    “o” is either /oʊ/(“long”) or /ɑ-ɒ/(“short”).

    “reform would have to mean the final divorce between Englishes of different regions”

    I only agree with this if you absolutely want a graphemic/phonemic system. But why let perfection be the enemy of improvement? There are many one-off spellings that make no sense to any dialect (AFAIK). I believe if you could get political agreement to remove e.g. the “i” in “friend”, the “s” in “island”, the “b” in “debt” etc, the current spelling would seem strange and archaic in a matter of decades.

  4. Adrian Morgan Says:

    I completely missed your statement about the COT-CAUGHT merge. I guess I scanned your comment by assuming each paragraph had one main point, and mentally labelled the third paragraph as being summarised in its second half (an introduction to the following list), which made the first half invisible to me.

    Your handling of schwa threw me, partly because you have schwa share a letter (“u”) with other phonemes, and partly because your standard for whether to write a vowel as schwa is based on how it is pronounced in the context of continuous speech, rather than when articulated carefully. Both of these design decisions were unanticipated, particularly in conjunction.

    There are two ways to approach the identification of stressed syllables. One is to identify all stressed syllables and then identify which of these has the primary stress, and the other is to identify primary stress first and then identify which of the remaining syllables have secondary stress. I always follow the first approach, as it better reflects how I naturally think about stress in English. It sounds like you lean more to the second. I have no idea what studies have been made of their merits.

  5. dainichi Says:

    With regards to stress, I think my system agrees with Ladefoged’s analysis mentioned here: http://en.wikipedia.org/wiki/English_phonology#Lexical_stress, where there is only one degree of stress, but unstressed vowels are distinguished for reduction.

    I’d be interested in knowing what spellings make you think my system is based on continuous speech, not careful articulation. I generally tried to use http://www.dictionary.com‘s pronunciation guide when I wasn’t completely sure, but I’ll admit I didn’t look up all words. I’m wondering how much dialects differ in what vowels can remain unreduced. For examples, I consider an unreduced first vowel in “complexities” highly unnatural. The only reason you would pronounce it unreduced would be to explicitly give a hint as to how it’s spelled. Then again, I’m neither a linguist nor a native speaker, so I could be completely wrong.

    I did consider, however, allowing all vowel graphemes to stand for reduced vowels (e.g. first vowel in “kompléksiteze”) thereby making the system more morphemic. But firstly, this would require more non-pronunciation-related insight from the speller, thereby partially defeating the purpose of spelling reform, and secondly, the definition of what is and isn’t a morpheme is not clear-cut. For example, should the verb “object” be spelled “objékt” because it’s the same morpheme (sequence of morphemes?) as the noun “object”? Surely there’s not much semantic overlap between the two.

    Oh, another error is “ago”, which should be “ugó”. My system treats word-final “a” as schwa (whereas word-final “u” is /ju/, i.e. the cheese “feta” is “fetta”, the verb “argue” is “argu”), and I was playing with the idea of making word-initial “a”s schwa as well, hence the error. But then I realized I wouldn’t be able to distinguish “object” and “abject” (both second syllable stress), so I gave up the idea.

  6. Adrian Morgan Says:

    You’re right that the “o” in “complexity” is reduced in most circumstances, but in exceptional circumstances it might not be. Perhaps when sung, for example. There are many words in which the underlying phoneme is ambiguous, but “complexity” is not one of them.

    Come to think of it, going by the “sing” test, to me the third syllable of “complexity” is even less likely to be unreduced than the first.

  7. iastphonetic Says:

    One may visit this site.


    strat traep paam baeth staart nars dres skwer kaait fliis laat klawth
    thawt nawrth fawrs fut guus praais maeuth feis nir chawis got


    stra̩t trăp pām băth stārt na̩rs dres skwer kāit flīs lāt klŏth
    thŏt nŏrth fŏrs fut gūs prāis măuth feis nir chŏis got

    strət ˈtræp ˈpɑm ˈbæθ ˈstɑrt ˈnərs ˈdres ˈskwer ˈkaɪt ˈfliːs ˈlɑt ˈklɔθ
    ˈθɔt ˈnɔrθ ˈfɔrs ˈfʊt ˈguːs ˈpraɪs ˈmæʊθ ˈfeɪs ˈnɪr ˈtʃɔɪs ˈgoʊt

    સ્ટ્રટ્ ટ્રૅપ્ પામ્ બૅથ્ સ્ટાર્ટ્ નર્સ્ ડ્રેસ્ સ્ક્વેર્ કાઇટ્ ફ્લીસ્ લાટ્ ક્લૉથ્
    થૉટ્ નૉર્થ્ ફૉર્સ્ ફુટ્ ગૂસ્ પ્રાઇસ્ મૅઉથ્ ફેઇસ્ નિર્ ચૉઇસ્ ગોટ્

You are welcome to add your thoughts.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s