The distinctiveness principle in English spelling: “An eye for an I, a too for a two”

The mylk of human kindness

Why are the words two, to and too spelt differently, even though they sound the same?

  • What about meat and meet?
  • Or catarrh and Qatar?
  • Or OutKast and outcast?
  • And  why do greet and great not rhyme?
  • The answer lies in one of the central principles of English spelling, which is that different words should have different spellings, regardless of their pronunciation. Now this is so fundamental that it is almost not worth stating – obviously different words such as cat and dog should be spelt differently. But things get trickier with homophones, when different words sound the same, as in the case of two, too and to, as well as four and fore.
  • This post looks at how words end up sounding alike. Some cases arise due to accent changes, others arise when new words come into the language, and others are just made up from scratch, as can be seen in the picture above, where the spelling mylk is used for non-dairy milk. This technique is very common among names, especially first names, band names and brand names.
  • It should first be noted that we are only talking here about cases where two words sound alike but are spelt differently. The other logical possibilities will be discussed another day, namely:
    • same spelling, different sound (row, bow and lead, read, etc.);
    • same spelling, same sound (compare fingernail with nail on the head);
    • different spelling, different sound (cat, dog / right – roysh).
  • And finally – feel free to add more examples in the comments below.

Part I: It seems history is to blame

Etymological overlaps

  • The simplest case of distinctive spelling occurs when a word borrowed from another language happens to sound like a word that already exists. For example, French-derived quay is homophonous with the native word key. Other examples include fisher – fissure, deriving from Latin, and the unlikely pair of Qatar – catarrh, which derive from Arabic and Greek respectively. Some people pronounce the word Sheikh like chic and others like shake.
  • It is not unusual for British spelling to retain distinctive spellings, such as <cheque> and <tyre>, while American spelling has abandoned the distinction, using <check> and <tire> for all homophones. These cases exemplify a tension which defines English spelling more than anything else: the need to strike a balance between representing sound consistently and the need to attach a single spelling to each unit of meaning.

Vowel mergers

  • Many homophones have arisen due to changes in English accents over the centuries. For example, the spellings <meat> and <meet> reflect pronunciations that were different several centuries ago, but a systemic sound change has caused the two words to sound the same, resulting in a glut of homophonous pairs, as can be seen from the table below (examples taken from Kruse 2016, quoted in Ryan 2018):

  • Evidence of the old way of speaking comes from those conservative accents, heard in parts of rural Ireland, where meet and meat do not rhyme, sneak sounds like snake, and the Son of God is known as Jaysus.
  • Incidentally, a few words spelt with <ea> did not end up in the same group as the others, so that great does not sound like greet, and Yeats does not rhyme with Keats. Other examples are steak, Reagan and yea (as opposed to nay).
  • Phonological mergers account for a sizeable amount of the complexity of English spelling. Examples are legion and the following is a representative selection of homophones:

Ale – ail, plane – plain; wait – weight, slay – sleigh; rite – right; Dane – deign; no – know, groan – grown; loot – lute, do – dew (for Americans); birth – berth; for – four, horse – hoarse (not all speakers).

  • The number increases further for people who don’t pronounce /r/ in all environments (non-rhotic accents), rendering homophones such as:

spa – spar; Shaw – shore – sure and soya – Sawyer.

  • (Fuller lists, and further discussion, can be found in my PhD thesis (p245-256)

Consonant changes

  • Homophony also occurs due to consonant cluster reduction, so that knight and night sound alike, as do knob and nob. There was a time when these words contained a /k/ and that sound survives in cognate German words such as knopf, which can be heard here. Consonant-cluster reduction has also occurred in rap and wrap, as well as rite and write, etc, and it has occurred in some very high frequency <tw> and <sw> words, including two and sword (a word that would have been uttered a lot more in days of yore).  Another kind of change involves one consonant becoming another, as has happened in accents where the distinction between whine and wine has been lost (that’s most places nowadays), and in accents where three and free sound the same (a change that can be heard in many urban British accents).
  • Finally, there was a time when English had both long and short consonants, and this was the original use for double consonant letters. The <pp> of apple used to be a long /p:/ while hammer had a long /m:/. The difference can be still heard in the short /n/ of annoying versus the long /n:/ of unnecessary, but these sounds only occur nowadays in complex words such as misspelt and lamppost. Instead, double consonant letters have been repurposed to indicate the length of the preceding vowel, so that we can distinguish between hopping and hoping, bitter and biter, etc. (another day’s discussion).

Linguistic exaptation

  • The important point for now is that spelling patterns which have lost their original function (e.g. <ee> v <ea>, and <mm> v <m>) can be recycled for new uses.  Such orthographic repurposing is the lifeblood of creative spelling (c.f. Lass 1990 on ‘linguistic exaptation’) and it might be fair to say that spelling is only interesting in itself when new developments arise from within the writing system, in a manner that is independent of speech. The spelling of homophones is thus interesting because one spelling functions as a control group while the other(s) show the possible variation. The examples to follow are the result of changes occurring solely in spelling and not in speech. We shall begin by looking at some lexical examples and then move to the more glamorous world of names, or onomastics.

Part II: Split the difference

  • English spelling was not always as stable as it is today and sometimes two variant spellings become associated with two separate words. For example, the spellings <flower> and <flour> were both listed as variants of the same entry in Dr Johnson’s (1755) English dictionary. Since then, one spelling has become associated with the colourful plant and the other with the bland powder. Similarly, the spellings <mettle> and <metal> are now associated with two different words, as are <discreet> and <discrete> etc. Carney (1994: 413-6) provides a few dozen such examples, although not all of them have established fully. For example, <disc> is apparently used in British spelling except when referring to a computer disk (OED 53642). It would appear that this spelling pair is undergoing the very slow process of disentanglement by consensus. At first, both spellings might be used indiscriminately but, over time, one spelling becomes associated with each meaning.

The so-called three-letter rule

  • Distinctive spelling is particularly common among very short words, partly because it has to be (shorter words are more likely to be similar to one another), but also because a the so-called ‘three-letter rule’ of spelling, whereby normal words must have at least three letters. Only function words, such as I, me, he, we, it, no, of, etc. can be have fewer than three letters, and this results in several homophonous pairs having distinct spellings, including in – inn, by – bye, be – bee and to – too. The only exceptions are abbreviations, such as ad, homophonous with add, and foreign borrowings such as pi, homophonous with pie. A recent paper by Martin Evertz (2017) explains how the rule is a little more complex than words needing to have three letters, and he explains why gnu and chi still look foreign while blue and shoe look Englishy.

The spelling of surnames

  • Another relic of the erstwhile variation in English spelling survives in the spelling of surnames such as Browne, Webb and Payne, where the names are pronounced the same as the lexical words brown, web and pain. Carney (1994: 449-458) observes that there is a very limited number of changes in occurrence, and this reflects areas of English spelling which were never fully systematised, including double consonants (Kitt, Scragg, Foxx), the addition of final <e> (Wilde, Bowe, Toye), and variation between <i> and <y> (Wynn, Smellie). It is not fully clear to what extent these distinct spellings arise from historical accident, and to what extent they were introduced as deliberate archaisms, although there is a certain amount of value in distinguishing oneself from a common noun, especially for persons named Bottome, Fagge or Freake.

Part III: Made-up spellings

The spelling of first names

  • The spelling of first names is a lot more variable and this is because parents can actually choose the spelling of their child’s name – for good or ill. This is in contrast to surnames which aren’t really yours to change, unless you are an emigrant alone in a new country looking to shed the outward marks of your heritage.
  • As a result we see much more variation in the spelling of first names, especially girls’ names like Kristen, Kristin, Christen and Christin, or Vicky, Vikky, Vikki, Vykki and even Vycci. While some of these spellings might seem odd, they are not chosen willy-nilly and they adhere to the following rule:

You can spell you name any way you want as long as it does not suggest a different pronunciation.

  • There are several ways to interchange <c> and <k>, but you can’t change <v> because there no other letter can be used to represent /v/ (with the exception of the word of). This is why we don’t have <Ficky> or <Phikki>. Creative spelling is therefore constrained by existing variation within English spelling and hence there is a finite set of possible changes. The full range of possibilities is perhaps best seen in the spelling of band names.

The spelling of band names

  • Bands of musicians have a tendency to name themselves after existing words and phrases, and standard examples include The Kinks, The Rolling Stones and Nirvana. Like other proper nouns, capital letters are used to distinguish these names from their lexical counterparts, but many bands go further, respelling their name in a unique way. The simplest examples involve switching interchangeable letters, as in the band Filosofy or Phuture, which swap <f> for <ph> and vice-versa.
Music theory
  • A fuller range of possibilities can be seen by focussing solely on bands with the word ‘music’ in their name, including:
    • The Music
    • Musikman
    • Musiq Soulchild
    • µ-Ziq
    • Mu$ic
    • Musique
    • Salsa Musika
  • The Music simply use capital letters to distinguish themselves from the music, and once again we see some swapping of <c> and <k> (Musikman), with <q> being added to the list of options (Musiq Soulchild). This time <s> is replaced by <z>, but only in places where it indicates the sound /s/. Not all instances of <s> can be replaced by <z>, so we can’t have <Zalza Musika> as that would suggest a new pronunciation.
  • The names Musique and Salsa Musika are not examples of distinctive spelling because they indicate a different pronunciation to the word music. I have just included them to show how creativity begets creativity. Once it is permissible to change the spelling then it quickly becomes permissible to play around with the pronunciation, and from there it’s a short step to full-on punning.
Beat it
  • It is worth noting that very few bands have managed to use distinctive spelling as a way of adding extra meaning to their name, with The Beatles being a notable exception. Here they have swapped <ea> for the <ee> of beetle, and thus added the word ‘beat’ into their name. Another rare example is Creedence Clearwater revival, but it’s such a shite pun that it scarcely deserves a mentch. Even worse is the title of Michael Jackson’s retrospective album HIStory.
Micro developments
  • The IDM producer µ-Ziq has a much higher concept substitution, where the name of the symbol ‘µ’ replaces the first syllable of the word music, and this is then followed by the more normal substitution of <z> for <s> and <q> for <c>. Syllable-level changes are common with certain letters and numbers, giving us bands such as 2ManyDJs, Altern-8, INXS and NOFX.
Add ))) to O
  • Finally, the name Mu$ic involves a visual substitution of <$> for <s>, a change also seen in Ke$ha and P!nk. There is a very limited set of options here and the scarcity of examples such as 5ive! where the numeral <5> replaces the <F>, shows how unwilling people are to make changes that do not adequately reflect the original form. However, this does not stop bands from using punctuation as a kind of visual decoration, as we see in the bands MǀAǀRǀRǀS and Add N to (X). The band Sunn O))) claim to be pronounced like sun, but it is very tempting to stick an ‘o’ on the end, while a few bands, including Gr†ll Gr†ll (grill grill) and GL▲SS †33†H (glass teeth) use obscure symbols in order to make themselves hard to find on the internet.
  • Here is a link to my Masters Thesis on the spelling of band names, written in 2011.

Further uses

  • Distinctive spellings feature in several other domains of the language, most notably brand names and newspaper headlines, where they frequently occur in conjunction with other kinds of language play. The simplest kind of example is Google, which is a respelling of the large number googol, while a wittier example is Froogle, the original name for Google Shopping. Such wordplay often needs to be explained when decontextualized, although the explanations can be worth it: Optimeyes is an eye-testing centre, A Salt & Battery is a fish&chip shop. C’est Cheese is a cheese shop. In each case, the distinctive spelling is used to squeeze in extra information, something that seldom done in band names for example (The Beatles is a rare exception, being a put on the beat movement). Open a tabloid newspaper and you’ll find every page wedged with distinctive spellings which make puns relating to the story at hand (see what I did there ?-) Finally, distinctive spellings can be wedged into a mnemonic, something that is common among Academic Acronyms that are Shite (AAS). Here are some other examples of bad acronyms.


  • Distinctive spelling is a very useful but its occurrence among normal words is largely haphazard, being the result of phonological changes (wait and weight), etymological coincidence (fisher and fissure), and existing spelling variation (flower and flour, disc and disk). Distinctiveness is useful for separating function words from lexical words (by and bye) and a few modern examples have arisen from semantic developments of a existing words (hence dairy milk and vegan mylk). These spellings make life difficult for the writer because one has to remember which letters to associate with which word, but they make life easier for readers because they allow us to map directly from the spelling to the meaning of the word, rather than requiring context to understand the meaning.
  • However it is in creative spelling where distinctiveness flourishes, most notably in names, because an existing word can be put to new uses. This is why we have Muzikman and µ-Ziq, 2 Many DJs and INXS, etc. The reason these changes are possible are because they provide people with a rare chance to create their own spellings, and to express themselves in a medium which can be frustratingly inflexible. Normal words must be spelt by consensus and respelling words is simply not accepted unless you happen to write newspaper headlines or are the kind of person who invents new words on the reg. Giving your child’s name a special spelling is an rare act of freedom in a life of orthographic enslavement and, like making puns, it provides a chance for adornment, albeit in a tiny and heavily constrained way, just as a man will choose a colourful tie or pocket handkerchief in order to offset the shackles of wearing a dull suit.


Carney, E. (1994). A survey of English spelling. London: Routledge. (n.b. link is not the full text).

Evertz, M. (2017). Minimal graphematic words in English and German: Lexical evidence for a theory of graphematic feet. Written Language and Literacy, 19(2), 192-214.

Johnson, S. (1755). A Dictionary of the English Language. London: W. Strahan.

Kruse, J. (2016). Accent variation reflected in the standard writing system of English, (log in required to access text), in The Routledge Handbook of the English writing system’ (Cook and Ryan, eds.). Routledge pp. 175-187.

Lass, R. (1990). How to Do Things with Junk: Exaptation in Language Evolution. Journal of Linguistics, 26(1), 79-102.

Ryan, D. (2010). Kre-8-iv Spell!nk: Why constructed homophony is key to understanding patterns of orthographic change. (MSc Thesis), University of Edinburgh.

Ryan, D. (2018). Principles of English spelling formation. (PhD Thesis), Trinity College Dublin.

Leave a Reply

Your email address will not be published. Required fields are marked *