Requests Roleplaying Video Games Entertainment & Media
Politics & World General Spirituality & Philosophy Worldbuilding
Creative Forum The Sports Center Science, Math, & Technology The Nostalgia Forum
Sexuality Hot Takes Complaints Awwww

Creative Forum

Poetry, short stories, song lyrics, what's your muse? What are you itching to share? ♥

Beyond lipograms and pangrams.

Posted Over 1 Year ago by chiarizio

A lipogram is a passage written without using a particular letter.
A pangram is a written passage that uses every letter.

It’s a fairly common exercise, at least among English authors, to write a series of 26 lipograms; each of which omits a single letter of English’s alphabet; and each long enough to use all of the other letters than the one designated to be elided from that particular lipogram.
If it makes a connected narrative that’s a real bonus!


I propose something beyond that here.
For each of the 325 unordered pairs of English letters, write a lipogram that doesn’t use either member of that unordered pair; but does use every bigram that actually occurs in English that doesn’t involve either of the two letters in the elided pair.
The passage written will be a sort of bigram-pangram.
The shortest possible one might have to be at least 576 words long; or it might be possible to shorten it to 1152 letters long.
I haven’t tried it yet so I don’t know; I haven’t even located an online list of all the legitimate bigrams that occur in English words.
Wikipedia seems to say they quote the most frequent 100 bigrams in a certain corpus and refer to a list of all the bigrams in some longer corpus.
In this post I’m referring to all the bigrams that occur in words in the latest edition of the OED, omitting bigrams that occur only in proper names or only in acronyms or only in foreign words or only in compound words.


One difficulty one novelist who wrote lipograms is that, if you follow the rule that all numbers have to be spelled out, there is a limited set of ages characters can be (assuming ages are always in years but you never have to write the word “years”) in any lipogram that elides the letter “e”. In his stated opinion this limited romantic scenes.

The allowed ages when omitting “e” are;
Thirty, thirty-two, thirty-four, thirty-six;
Forty, forty-two, forty-four, forty-six;
Fifty, fifty-two, fifty-four, fifty-six;
Sixty, sixty-two, sixty-four, sixty-six;
And the next number is:
Two thousand.

If one also omits “y” then no age older than six but younger than two thousand is available.

I suppose in that chapter there might be a six-year-old android who’d been cr*at*d as an adult and bootstrapp*d or IPL’d as an adult, who has a romance with a two-thousand-**ar-old vampir*, or something.


Anyone want to try?

There are 12 Replies

For each of the 325 unordered pairs of English letters, write a lipogram that doesn’t use either member of that unordered pair; but does use every bigram that actually occurs in English that doesn’t involve either of the two letters in the elided pair.

That's a ridiculously difficult challenge.

Challenge accepted.

I haven’t even located an online list of all the legitimate bigrams that occur in English words.

I'm installing a Wordnet database on the server as we speak. I'll be able to index it once done and there will be a searchable tool of bigrams --> words here. As well as a list of all of them. I need the database anyway for other projects.

The allowed ages when omitting “e” are;

You could get around it by prefixing "half-" and "quarter-" to numbers -- would give them a more interesting feel in the story. "He was half-forty, she was half-fifty".

Over 1 Year ago
Sky's the limit

I love that you accepted the challenge!


“Quarter” has in it both an e and a t.
“Month” doesn’t have an e or a y, but it has a t. “Moon” has no e no t and no y.
Fortnight has no e but has a t.
Summer and winter each have an e. Winter also has a t. Season has e.
Day has no e and no t but it has a y.

I think any chapter that follows Bechdel’s rule could possibly just skip romance. (Or heterosexual romance, anyway.)
Besides romance doesn’t have to mention characters’ ages, any more than non-romance has to.

Anyway I think people aged thirty to sixty-six are perfectly capable of being in romances; at least if they’re fictional characters in fictional romances.
Unless one of them is me. Once I passed sixty I think I was past even fictional romance; and once I was past fifty I think my romantic prospects were purely fictional. But that’s just me.

Over 1 Year ago

Are you going in the order ab ac bc ad bd cd .... tz uz vz wz xz yz
Or in the order ab ac ad ae af ag .... wx wy wz xy xz yz ?
Or in an order determined by in-story chronology?

I’m talking about the order of bilipogram chapters in the story, of course; not about the order of the words or the bigrams in each chapter.
Alphabetizing the bigrams in each quasi-pan-bigram chapter is hardly likely to make sense!

In the first case the chapter (if that’s the word) omitting the unordered pair {a, z} will be 25th; in the second case it will be 301st.

But you could do it in order by difficulty from most difficult to easiest.
In order of frequency of use in English the most frequent 14 letters are something like etaoinshrdlucm ..
You could omit e and t first; then e and a; then t and a; then e and o; then t and o; then a and o; and so on.
(An alternative order is etaonirsh .... And Jeopardy’s order is something like etnrsl ... )
And Scrabble makes the last five be ... kjzxq or something.
I don’t think xx is a bigram in any English word that’s not a brand name. The same is probably true of qq and qx, though xq occurs in eg exquisite.

The most frequent bigram is th and the second most frequent is he.
Any lipogram that omits e or t or h is going to have to bypass the definite article “the”. And if it omits t or h it will also have to skip demonstratives such as this and that. Those chapters (? still not sure that’s the right word?) will be interesting.


Did you yet find a list of all bigrams that occur in English words (in for instance the most recent edition of the OED) that aren’t
  • acronyms
  • compound words
  • foreign loan words
  • proper nouns

  • Over 1 Year ago

    Do you still accept this challenge?
    It’s starting to look like you’ll never have time to work on it!

    Over 1 Year ago

    What can I do to help?

    Over 1 Year ago

    Oh yeah I have no idea, I'm not even sure if the database is on the server or not -- will check that later.

    Over 1 Year ago
    Sky's the limit

    Thanks! That’s encouraging!

    Over 1 Year ago

    @chiarizio: I imported a kind of words dictionary and made a bigram searching tool:

    11 Months ago
    Sky's the limit

    I also discovered that the following bigrams have no english words that contain them:

  • bq /qb

  • dx /xd

  • fq /qf

  • fv /vf

  • fz /zf

  • jq /qj

  • jv /vj

  • jx /xj

  • kq /qk

  • pq /qp

  • qt /tq

  • qv /vq

  • qw /wq

  • vw /wv

    Because of this the challenge becomes:

    For each of the 311 possible English unordered bigrams,

    Create a passage that contains every single other bigram but does not contain this bigram.

    There are also a bunch of single-word bigrams:

  • novgorod

  • earthquake

  • daugavpils

  • velazquez

  • maxzide

  • fjord

  • chongqing

  • foxglove

  • merckx

  • dojc

  • muzjik

    It looks like every single passage will be exploring earthquakes in Chongqing, Novgorad or Fjords.

    Incidentally, the database I'm using here omits acronyms, anything with a dash (like "forty-six"), however it does contain proper names of places or whatever. This seemed like a good starting point for my other dictionary-based projects.

  • 11 Months ago
    Sky's the limit

    My main question, beyond the more open-ended question of "what is English really", is "is this necessarily about the English letters or could a similar thing be done with the phonetic inventory of English". It would be astronomically more complicated to collate the appropriate data from a sizeable corpus of text, given how English works.

    While the most common English letter is "e", this is because that symbol is capable of taking responsibility for a bunch of phonemes like /i/, /ɪ/ and /ə/ (which are the three most common vowel phonemes in English).

    11 Months ago

    I always thought English's use of /ʒ/ and /ð/ was interesting given their lack of distinct graphemes. "zh" and "dh" make logical sense, but instead they end up becoming something like ghost phonemes. Meanwhile, the "dzh" sound is so common it has its own letter (J), and yet is absent in a lot of common languages.

    11 Months ago
    Sky's the limit

    @galbraith: Thank you for your interest!

    There are some pan-phoneme passages for some ‘lects of English, including the “comma gets a cure” story which is designed to bring out the differences between dialects.
    I don’t think a lipo-phoneme passage is totally out of the question.
    But a series of phoneme-eliding passages each of which contained every other phoneme than the one elided, would be quadratically longer. If there are 44 phonemes, it would contain 44/26 times as many chapters, and each would be 44/26 times as long.

    I think a pan-biphoneme passage would be enough more difficult that consideration of such a project might better be postponed until the pan-bigraph project is nearing completion.


    @Xhin: This is a major advance! Is there any way you can make it available to the rest of us?
    [edit]yes, there is, you’ve included a link![/edit]

    If I remember right my initial proposal omitted those bigrams that occur only in foreign words or proper names or compounds or acronyms or trademarks or service-marks. There may be other categories of lexemes I (or we) also might want to omit, or should omit; I can’t think of any of them at the moment, but maybe I thought of one or two earlier but didn’t write them down.

    For most of the bigrams that you’ve found occur in only one word, that word is a proper noun and/or a foreign word and/or a trademark and/or an acronym.
    I’m not sure words like “foxglove” and “earthquake” count as compound words. If not maybe we should leave them out anyway by creating a new category.

    Or, maybe, we should just declare a minimum number of English words that a bigram needs to occur in before we’ll need to include it in every passage; say maybe >=25 . Or maybe >=311 ?
    And maybe we should have just one passage that omits every bigram that occurs in only some low number of English words —— say, <=1 or <=10 or <=24 or something —— but includes every bigram that occurs in more than that many English words.

    Whatever. For all I know a 311-chapter story of which 310 were about earthquakes and 310 were about foxgloves might be fun to read!
    There was a novel written by some SF author in which he never failed to mention in the first third and the last third that one of the main antagonists was a primate. But he didn’t mention that at all in the middle third. Sometime in the last third he has one character ask another what a hypothetical reader’s reaction would be to a hypothetical such novel.
    It might be fun for a reader to figure out which bigram(s) was(were) elided from which chapters. Kind of like a reverse “Easter egg” hunt.


    This is glee-inspiring data, to me! Thank you very much!

    11 Months ago

    This thread is archived