Missed Opportunity for Ligatures

In 1916 Legros & Grant published their classic printers handbook Typographical Printing-Surfaces. In a small section in the middle of the book they advocate spelling reform for a different reason than most.

There have been many calls for spelling reform for the English language. The problem is that there are roughly 45 sounds in the language, but only 26 letters. The usual complaint is that causes spelling to be more difficult than necessary: traveler or traveller? embarass or embarrass? gauge or guage? untill or until? weird or wierd?

It is called a regular orthography when there is a 1:1 relationship between the sounds in a language and the letters to represent those sounds. Having more sounds in a language than letters to represent those sounds is called an irregular orthography. All alphabetic writing systems fall somewhere on this continuum. English is known for having an irregular orthography while Spanish is known for having a very regular orthography. Some researchers have proposed that dyslexia is more prevalent in countries that have irregular orthographies than countries that have regular orthographies. Reducing dyslexia seems an even better reason to call for spelling reform.

Legros & Grant provided a third reason. They pointed out that using two letters to represent a single sound is inefficient both for the amount of paper used and in the amount of time for compositors to create a page. Compositors are the people who manually creates pages of text to be printed by arranging all the metal slugs of lead that each carried a letter. L&G proposed creating two new ligatures: th and ng. They calculated that by creating these ligatures, which are associated with unique sounds in English, that they could save 3-4% in compositors time and in the amount of printing. They calculated the annual savings at 350,000£ (inflation adjusted that is over 6 million pounds).

Legros & Grant’s proposal for spelling reform is particularly elegant because it provides an easy transition path with relatively little re-learning. The ligatures are not completely new letters, but modifications of existing letters. Both th and the th ligature could comfortably exist during a transition period. Unfortunately we’ve missed the chance to save compositors’ time.

Cheers, Kevin Larson


Edit: Update Image Reference

Comments (17)

  1. Claw says:

    Why necessarily use ligatures for TH and NG, since they don’t look too distinguishable from existing letters?  For instance, the TH ligature looks too much like a captial H. and the ng ligature looks too much like a lowercase g.  Instead, how about using the thorn (Þ/þ) or theta (Θ/θ) for representing the TH sound and the engma (Ŋ/ŋ) for representing the NG sound?  The thorn character used to be used to represent the TH sound in Old English.  The theta and engma characters are commonly used to represent the TH and NG sounds, respectively, in phonology (there are actually two different TH sounds in English though, but I won’t get into that).

  2. Kevin Larson says:

    What caught my eye about the L&G proposal is the easy transition because of the clear connection between the old and new. Using theta for th requires a degree of memorization that the th ligature doesn’t. I absolutely agree with you about using engma for ng. It’s much simpler than what they proposed.

  3. John Hudson says:

    I agree with Claw, and while we’re about it why not throw in the ʃ from Pitman’s phonetic alphabet for the sh phoneme? For the two th sounds, we have ð and þ conveniently available in almost every font thanks to our Icelandic friends.

    The use of the terms regular orthography and irregular orthography seem prejudicial without clarification as to what aspect of the orthography is regular. They can be taken to favour phonetic correspondence above etymological correspondence. English spelling is really quite regular from an etymological perspective.

  4. Kevin LArson says:


    In this context people use regularity to mean something closer to consistency. Of course both irregular and inconsistent can be taken as negative descriptions, so it’s not clear what word would describe this feature that isn’t prejudicial. I think regular and irregular are perfectly accurate words for describing the closeness in relationship between spelling and sound.

    Though for many practical reasons I would vote for optimizing orthography for spelling to sound correspondence over etymological accuracy.

  5. John Hudson says:

    We can have a good phonetic vs etymological spelling fight over a bottle of wine at the next type conference. English has a complicated root system on account of the mongrel mix of Greek, Latin, French, German etc. word origins, but the value of root analysis in understanding the meaning of words remains. For example, the letter sequence ‘sig’ is pronounced differently in the words sign, signal, design and designate. Is there really much to be gained from changing the spelling of these words to match a ‘regular’ orthographic representation of how they are pronounced (bearing in mind that you will need to reform your spelling when the pronunciation changes, and that it will not actually represent the myriad ways in which these words are actually pronounced regionally), and hasn’t one lost rather more by obscuring their common root.

  6. Chris Kindschi says:

    The question in my mind is whether Legros & Grant would have wanted these ligatures used in words like "cathouse" and "vanguard"?

  7. Kevin Larson says:

    That’s a great question Chris. I have no idea what L&G would have said. Clearly using the ng and th ligatures for all instances including lighthouse and vanguard would save compositors time, but they might have chosen these two ligatures in part because they usually also lead to an increase in orthographic regularity. If so, they would need to be more careful in specifying when to apply the ligature. For these two cases it is a straight-forward rule to apply them when the letters appear within the same syllable, and don’t apply if they cross a syllable boundary.

  8. Nick B. says:

    I recall back in HS in the mid 70s reading an anthology of SF shorts wherein the author, relating the story "from 50 years in the future" discussed how the spelling had been altered "in the last 50 years" — as he introduced each change he incorporated it into the story. By the end of the story, the spelling was far more logical (‘c’s became the appropriate s or k sounds, for example) and it was still completely readable by anyone.

    The only real reason for this not being done is social inertia.

  9. Francis says:

    Why just th and ng? If you are going to change the language, why not go whole hog?

    What this post neglects to mention is the advantage English has in having no odd characters or accents (unless you count j,k,w,x,y.) It might have saved time typesetting to introduce these ligatures, but it would definitely have added time for people on non-English keyboards wanting to type English text or names correctly.

    Ask a non-Icelander to type an Icelandic name correctly on his own keyboard, and see how long he takes. Or a non-German to write "Straße."

  10. Tobias says:

    And it seems that Office 2007 missed that opportunity again…

  11. Kjartan Þór Kjartansson says:

    Francis, as an Icelander I must say in my case getting people to say my first name correctly over the phone is more of a problem, but I know completely what you mean but we Icelanders tend to ask people to use th and ht for þ and ð respectively. But correct me if I’m wrong but didn’t English have at least on if not both Þ and Ð a few hundred years ago? Wikipedia seems to think so http://en.wikipedia.org/wiki/Eth and http://en.wikipedia.org/wiki/Thorn_%28letter%29 so picking them up again should not be a very difficult task.

    But there is one problem for adding characters to the language that has not been raised here and that is that for every character you add you need to move some tokens about for instance in Icelandic keyboards the Þ is where americans have / and ? and Ð is where [ and { are if I remember correctly on US keyboards, add to that the fact that US keyboards have 101 keys while most of the rest of the world uses at least 102 and you will have an escalating problem.

    But I must say seeing English written using Þ and Ð would be very pleasant for Icelanders

  12. SC says:

    And all this should be used to give more space to ads??? Terrible

  13. Amit Agarwal says:


    Could you please point to the fonts that contain the "Th" ligatures? I see the Adobe Minion Pro & Adobe Garamond Pro containing them in the private use area but these don’t seem to be very common in other fonts.

  14. I’ve been researching the thorn TH (in capital) used in the Stratford Shakespeare monument "STAY PASSENGER, WHY GOEST THOV BY SO FAST/ where the TH in THOV is an odd character that welds the T to the H by eliminating the left hand side asender I and connecting the T directly to the right side asender of the H, I. Has anyone seen this "ligature" or Digraph ever before? Is it a unique capital thorn, th,? Anything on this would be much appreciated. I have a possible issue of a "diagraph" in an alleged cryptograh. Thanks a lot. (I’m afraid one might need to see the text of the Shakespeare monument to see exactly what I mean)

  15. Nathan Pizar says:

    Although some might suggest the use of thorn and theta in place of these ligatures, I believe that approach would defeat the largest advantage of the ligatures proposed; the transition (at least in reading if not quite as smooth for the writer) is rather negligible. One can scan the example text with ease because of the very fact that the forms are already familiar and very similar to the existing letters.

    The article is not a call for sweeping reform or adoption of entirely new (to native English readers) glyphs for each phoneme as some have proposed in the past. Simply an example of how some of the most common combinations could be easily implemented using immediately recognizable forms. This would mitigating the typical transition period during which the eye would ordinarily be slowed by pausing to phonetically process input that is not yet an imprinted reflex, especially for those who sight read beyond the speed of sub-vocalization.

  16. Barbara Henry says:

    The character described by Elwood Miller exists in Serbian Cryilic, both in upper and lower case though it stands for a soft ‘tch’ sound.

  17. ecs says:

    Seems there was a missed opportunity for apostrophes, too.