k: /’INglIS &z Si Iz spEld/

I should start by saying that this week's column has more to do with linguistics than with wordplay. I'm assuming readers have no formal background in linguistics; linguists in the audience can skim ahead.

First off, a definition: the field of linguistics has to do with study of the structure of language in general, not with learning to speak particular languages. Linguists are often asked, "So you're a linguist? How many languages do you know?", prompting one linguist to retort, "So you're a polyglot? How many linguistic theories do you know?"

One of the areas that linguists study is the structure of the sounds of language, a branch of linguistics known as phonology. The sounds of speech can be broken down into component parts; the smallest such parts are called phonemes. The International Phonetics Association (IPA), established in 1886, maintains a standard system for written representation of all phonemes used in all languages: the International Phonetic Alphabet (also known as the IPA). The IPA(lphabet) was most recently updated in 1996, after fairly extensive revision in 1993.

Unfortunately for those of us using computers, most of the IPA consists of characters which aren't in the ASCII character set. (I'm aware that there are other systems for representing the characters of various languages on a computer: Unicode, for instance. For the moment, however, ASCII still thrives as the standard representational system for text.) So Evan Kirshenbaum, with the help of other folks from the alt.usage.english and sci.lang newsgroups, created a very sensible ASCII version of IPA. The Kirshenbaum system includes some complex and detailed ways to precisely represent particular phonetic features; formal phonetic transcription can be very tricky. Fortunately you can ignore most of the complexities if all you want is broad transcription of English, to show how a word or phrase sounds without giving a precise phonetic specification for it.

A broadly transcribed word or phrase is written between slashes. /kip/, for instance, transcribes the way I say "keep": the /i/ represents the vowel sound spelled as "ee." Examples (in my pronunciation) of the way other vowel sounds are transcribed: "mop" is /mAp/; "hat" is /h&t/; "hit" is /hIt/; "buy" is /baI/; "bet" is /bEt/; "bait" is /beIt/; "put" is /pUt/; "now" is /nAU/; "bone" is /boUn/; "saw" is /sO/; "soy" is /sOi/; "tune" is /tun/; "bun" is /b@n/. Pretty easy once you get used to it. An apostrophe before a syllable indicates that syllable is emphasized ("receives primary stress").

I've occasionally bemoaned the fact that English isn't written in IPA or some other phonetic form; I like to know how words are pronounced, and it would certainly make spelling easier. Many others have complained about lack of regularity in English spelling over the past few centuries, perhaps best summed up by George Bernard Shaw's well-known comment that one could spell "fish" as "ghoti"—"gh" as in "rough," "o" as in "women," "ti" as in "motion." The problem has broadened as English has become more and more the lingua franca (as it were) of the world, requiring more non-native speakers to struggle with the vagaries of English spelling. The Simplified Spelling Society has a variety of plans toward the goal of spelling reform; some people (such as Ben Discoe) have even gone so far as to create Web pages using a phonetic spelling system.

But I'm less enamored of switching over to a phonetic spelling system than I once was, largely due to Steven Pinker's book The Language Instinct. Pinker (an MIT linguist) provides several reasons that phonetic spelling might not be such a great idea:

  • Phonetic spelling removes some connections between words, connections that are obvious given current spellings. "Sign" and "signature" are spelled similarly, so a reader can identify the two words as having the same root; but in IPA there's no indication that they're remotely related.
  • Different speakers would spell differently. Pinker points out that some people pronounce "career" as /k@ 'ri R/ and "Korea" as /k@ 'ri @/, while others switch those pronunciations.
  • Straightforward rules govern the spelling of about 84% of English words, so learning about the exceptions to the rules is not quite as daunting a task as spelling reformers suggest.

I'm still not entirely convinced phonetic spelling is a bad idea. It could be argued that we can usually understand people who speak with accents; phonetic spelling could actually help us understand where people are coming from. On the other hand, it could also be argued that standardized spelling lets us communicate with people whose accents might be unintelligible to us...

For now, I'll be satisfied with more people getting a rough idea of how to use the IPA (and its ASCII equivalent), so that we can at least communicate fairly well in print when we discuss pronunciation.


  • Disinterested plug: The alt.usage.english FAQ provides about 250K of fascinating reading about English usage, history, spelling, and so on.
  • Undisinterested plug: Those interested in the IPA might enjoy Magnetic Phonetics, in which the IPA is printed on a set of magnetic tiles. The set includes tiles suitable for playing IPA Crossword, a game in which players take turns placing criss-crossing word transcriptions in the playing area. I said "undisinterested" because I helped create the game and I know the publisher; but as I have no financial interest in the set, I hope I may be allowed to mention it here without the taint of commercialism.

Join the Conversation