A few months ago in the Lounge we took a tour through the Visual Thesaurus based on the number seven. We thank the subscribers who pointed out Facts Previously Unknown to us, concerning two other languages in which the words for seven and week were etymologically related - i.e., Italian and Romanian. This month, we are taking a somewhat more specialized look at three. It's a number of nearly mystical powers, as a brief look at its wordmap suggests: even without considering its many appearances in other words, the words denoting a group of three (trio, troika, triplet, threesome, tercet, triad, etc.) suggest that it is a number with a mission.

Our specific interest in three, for which the VT will provide a few object lessons, stems from an observation made by the Danish linguist Otto Jespersen. It has come to be known as the "three letter rule" and it goes thus: content words in English (as distinct from function words) must have three or more letters. To put it another way, English is happy to spell function words (prepositions, pronouns, articles, conjunctions, exclamations and the like) with one or two letters, but words with meaning that is largely independent of context (that is, nouns, verbs, adjectives, and adverbs) cannot slip by with only one or two symbols to represent them in writing.

The corollary of this rule is that content words with pronunciations identical to function words are distinguished from them in writing by the addition of a (superfluous) consonant. To put it another way, and in the words of an author we'll mention in the links below, "a spoken content word that would logically correspond to two letters is 'padded out' with an extra consonant to avoid confusion with two-letter function words." To illustrate, all of the following content words are homophones of function words that are spelled with one or two letters in English: aye, bee, bye, inn, know, oar, sew, two, wee, and for some speakers, dew.

If you're like us, the first thing you think when you hear a new rule is whether or not there are exceptions, and in the case of language, the answer is nearly always yes. But over many centuries, English has settled down in its orthography remarkably well, such that the exceptions are few, and they tend to confirm the general application of the three letter rule.

At first blush, the commonest exceptions to the three letter rule are the verbs go and do. A look at their wordmaps (and you may want to don your eyeshades or turn off the other parts of speech before you look) reveal that these words are both true workhorses in English. In fact, you could almost argue that they have little meaning independent of context since they are both so polysemous. Dictionaries inform us that both do and go were not always spelled with two letters - they weighed in with three in both Middle English and Old English - and interestingly, all of their inflections are spelled with three or more letters. Along with being a denotative verb, do is also a function word in English, being required to form some tenses and various negations and questions.

Three other wee content fellows that come rushing to mind are ox, ax, and ex. There are all genuine exceptions to the three letter rule, though they exhibit a common pattern: a logical writing system (which English lacks to some degree) would spell them with three (or more) letters, since they each contain three phonemes: along the lines, for example, of German ochs, the ancestor of ox. It is only by virtue of x's habit of representing two phonemes (/ks/) that these words squeak by with two letters. Ax, by the way, is a spelling approved only in American English; the Brits spell it axe. Ex has become acceptable shorthand for one's former spouse or squeeze presumably because of the frequency with which these individuals appear in the modern world. (We note, however, that this usage dates back to the earlier part of the 19th century, which from this vantage doesn't seem all that modern!)

Other exceptions to the three letter rule are less compelling and tend to confirm the rule, because in the back of our minds we're thinking: these things denote rather ephemeral and airy entities anyway, and so perhaps don't deserve the full three-letter monty of a genuine content word. There are, for example, the syllables of the solfa scale, six of which are spelled with two letters (do, re, mi, fa, la, and ti). There is our mathematical friend pi, usefully displayed to many decimal places in the VT, and along with it, the names of a few other Greek letters (mu, nu, xi). Another example is id, which has probably squeaked into English with only two letters because of the direct borrowing from Latin, and from the fact that despite everyone's allegedly having one, no one has ever actually seen one. The printers' measures en and em also fit the general pattern of wee, abstract things that hardly merit additional letters. In the typography trade, these have longer nicknames anyway (nut and mut or mutton) since their pronunciations are so similar. Finally, we should not neglect el: anyone who has stood under one in Chicago would not say it was ephemeral and airy at all! But then it is only an abbreviation for elevated railway.

Evidence of the three letter rule among denotative words is plentiful. A cartload of words in English consist of two phonemes only - a vowel and a consonant, or in some cases a diphthong and a consonant - yet these are spelled with three or more letters, of which one, usually a consonant, is often phonetically superfluous. Examples include add, ebb, egg, odd, and err. A few other two-phoneme content words are spelled with three (or more) letters because English lacks a single symbol for their consonant sound, such as ash. The settled convention of indicating a long vowel by using two vowels to spell it or by suffixing a silent e at the end of a word accounts for a whole slew of three (or more) letter English concoctions which represent only two phonemes: take, for example, ache, eke, oak, ace, and ice.

Many of the quirks, methods, and peculiar logic of the way English is written can be studied in The English Writing System by Vivian Cook: he is the author of the quote we used above. His book is published by Oxford University Press:

http://www.oup.com/us/catalog/general/subject/Linguistics/TheEnglishLanguage/?view=usa&ci=9780340808641

Vivian Cook also has a website with several curiosities of interest to explorers of English:

http://homepage.ntlworld.com/vivian.c/index.htm

A theory we're tossing around in the Lounge at the moment is that nicknames in English exhibit a kinship in spirit with the three letter rule: English speakers like to reduce a full given name, regardless of length or complexity, to two or three phonemes whenever possible. When nicknames consist of two syllables, we also note, the final syllable is nearly always long e, though variously spelled. We haven't actually done the math, but here's a fairly complete list of name-to-nickname correspondences:

http://deron.meranda.us/data/nicknames.txt