History of Alphabetical Order.

Joe Moran writes for the Guardian about a new book by Judith Flanders, A Place For Everything: The Curious History of Alphabetical Order, which sounds like an interesting read:

One of the many fascinations of Judith Flanders’s book is that it reveals what a weird, unlikely creation the alphabet is. Writing has been invented independently at least three times in different parts of the world. The alphabet was invented only once – over 3,000 years ago, in Egypt’s Western desert, along a road used by traders and soldiers from across the Middle East. […] Alphabetical order, however, had a much longer and more circuitous road to dominance. A Place for Everything tells this complex and layered story. The alphabet has always been learned in a set order – but it was ages before this order was used for anything other than memorising the letters. Alphabetisation arrived piecemeal and for centuries remained one arrangement among many.

The Library of Alexandria, founded around 300BC, only used first-letter alphabetical order. Not until the Middle Ages did it occur to anyone to file Aristophanes before Aristotle. Even then, they did so halfheartedly. Monasteries, the location of most books in Europe at this time, had few holdings, so librarians themselves, rather than catalogues, served as their institutional memory. In the 13th century, Durham Cathedral Library, one of the largest, held just 352 books.

The slow rise of alphabetical order relied on many technologies coming together: the codex book (scrolls are fine for continuous reading, but rubbish for looking things up), pagination (rare in the earliest books) and the explosion of words that came with the arrival of paper and the printing press. Ultimately, Flanders suggests, the unstoppable democratisation of knowledge demanded alphabetical order. When the Word of God was contained only in churches and monasteries, there was little need for alphabetisation. But when mendicant preachers began crisscrossing Europe in the 12th century, they relied on handbooks such as Peter Lombard’s Sentences, which let them look for biblical keywords in the alphabetised index and construct ready-made sermons out of them.

Every dictionary compiler seems to have thought that he alone invented alphabetical order. Few realised its significance. Hugh of Pisa’s 12th-century Great Book of Derivations kept interrupting its alphabetical ordering of words by giving precedence to longer over smaller entries. John Balbi’s 13th-century Catholicon, another early dictionary, includes a pained entreaty about its alphabetical arrangement: “I have devised this order at the cost of great effort and strenuous application … I beg of you, therefore, good reader, do not scorn this great labour of mine and this order as something worthless.”

Many scholars remained suspicious of using alphabetical order for reference purposes. It was somehow cheating, they felt, not to memorise large tranches of text or read books through from beginning to end. In 1588 the poet and barrister Abraham Fraunce rebuked those who “prefer the loathsome tossing of an A.B.C. abridgement, before the lightsome perusing of a methodical coherence of the whole common law”. As late as 1818, Samuel Taylor Coleridge denounced those encyclopaedias “where the desired information is divided into innumerable fragments scattered over many volumes, like a mirror broken on the ground”. By then, however, the phrase “walking encyclopaedia” had been coined, signifying an amused disdain for dependence on mere memory.

I’d never really thought about it, and now I want to know more; thanks, Kobi!


  1. David Marjanović says:

    Not until the Middle Ages did it occur to anyone to file Aristophanes before Aristotle.

    Would it be very snarky of me to point out that Φ goes after Τ?

  2. David Eddyshaw says:

    While it is true that the alphabet “has always been learned in a set order”, there are two completely different ancient orders.

    h l ḥ m …

    Still used in Ethiopic, was used in Old South Arabian; there’s even an Ugaritic example (alongside the familiar ʔ b g d …)

  3. Would it be very snarky of me to point out that Φ goes after Τ?

    I thought of that, but presumably at that point she’s thinking of the Latin West.

  4. Do other alphabets have a fixed order (eg, Arabic), and was it used for ordering words / books? The article seems very Western-centric and I too would like to know more.

  5. The article on “Before Donald Trump: the real history of fake news” is wonderful; very much up the LH alley

  6. David Eddyshaw says:

    Arabic has a fixed order based on the original ʔ b g d but adapted to put letters together in ordering when they are now only distinguished (as they so often are) by diacritic points; I believe the original order is still seen sometimes in contexts like numbering of volumes.

    Bathrobe himself will know about the iroha in Japanese; modern Japanese dictionaries order the kana in what is basically the Sanskrit order (null) k c (as s) t p (as f/h) … subordered by the vowels a i u e o, again as in Sanskrit.

  7. Perhaps a bit tangential, but the iroha order is still in use in music and, it seems, in some legal contexts (and it’s available as an ordering option in MS Word!).
    And as far as I can tell, the modern gojuon order was in use well before the iroha order.

  8. Fascinating – I remember this mentioned as an aside in “The Name of the Rose”, in which the monastic library is arranged by topic, and the librarian mentions that the alternative would be to arrange it by accession date – the narrator mentions that alphabetical order is only coming in now (i.e. the end of the 14th century), fifty years later.

  9. According to American Heritage Dictionary (2014), “element”: “ety. Middle English, from Old French, from Latin elementum, perhaps ultimately from lmn, first three letters of the second half of the Canaanite alphabet, recited by ancient scribes when learning it.”

    Further, M. D. Coogan, “Alphabets and Elements,” Bulletin of the American Schools
    of Oriental Research 216 (1974) 61-3; “)LP, To Be an Abecedarian,” J. of the
    American Oriental Society 110.2 (1990) 322.

  10. Athel Cornish-Bowden says:

    Even when names are spelled identically different national conventions could produce different orders. Before the Royal Academy stopped considering ch to be a letter in its own right, a list of South American countries put Colombia before Chile in Spanish lists, but the other way round in English lists.

    There must be worse problems with Danish and Swedish lists.

  11. David Marjanović says:

    h l ḥ m

    Ε Λ Η Μ

  12. AJP Crown says:

    There must be worse problems with Danish and Swedish lists

    I only know that in Norwegian the letters Æ, Ø and Å are the final three in the alphabet. The Swedes may have chosen a different system – they use diacritics Ä and Ö – possibly for this very reason (alphabetical order).

  13. May I recommend the book, by Edward Wilson-Lee, The Catalogue of Shipwrecked Books, or Young Columbus and the Quest for a Universal Library (ISBN 978-0-00-814624-5).

  14. Ook: And as far as I can tell, the modern gojuon order was in use well before the iroha order.

    I guess that depends what you mean by “modern” and “in use.” Tachibana no Tadakane’s influential dictionary Iroha jiruishō 色葉字類抄 (12th century) used iroha ordering. And in the Tokugawa period, reference books generally seem to have used iroha ordering more frequently than gojūon. (Some Kokugaku scholars, however, preferred gojūon. Perhaps because of Keichū’s influence? Not sure about that.)

  15. Athel Cornish-Bowden says:

    I only know that in Norwegian the letters Æ, Ø and Å are the final three in the alphabet.

    OK, but if you want to look up “ångström” in Quantities, Units and Symbols in Physical Chemistry you’ll find it between “angle of optical rotation” and “angular frequency”. You won’t find it after “zettabinary”, but I suppose that if anyone made a Norwegian translation that’s where it would be.

  16. @ajay: Of course, that remark about subject ordering in The Name of the Rose, which initially just seems like a bit of historical color, actually turns out to be a major plot point.

    The movie version of The Name of the Rose has many problems. (I did not need to see Christian Slater’s ass.) However, one of the most disappointing was the rendering of the former heretic Salvatore’s “primordial” language, an argot of seemingly every tongue Adso has ever heard. Ron Perlman is a talented actor, but his lines in the script just cannot live up to Eco’s description.

  17. AJP Crown says:

    Athel, It would indeed. I first wrote “their alphabet” but to introduce an alphabetic Other seemed a bit rude so I changed it to “the alphabet”.

    Ångström though is Swedish (you can tell by the dots if not from the scientific history), and God knows where they keep their Ås. It may not be at the back end as it is in Norway.

  18. In Korean alphabetic order, all the consonants of precede all the vowels: ga na da la ma ba sa … a ya eo yeo o yo u yu …. I’m not sure what the sequence is determined by. It’s not strictly stroke order, although the basic consonants and vowels precede the related ones that require extra strokes, thusㄱ (plain k/g) before ㄲ (tense) or ㅋ (aspirated), andㅏ(a) beforeㅑ(ya).

  19. A serious inside physics joke in that the Ångström unit should really be more like 0.099986 nm, since Ångström actually based his famous solar spectroscopy measurements on a meter that was somewhat too short

  20. Mongolian traditional script also has its alphabetical order, which I’ve memorised, but I’m not sure how old it actually is. At least one major dictionary reverses ‘M’ and ‘L’, which look similar. The ordering is current in Inner Mongolia. I don’t know any dictionaries in Mongolia that actually use the old script.

  21. John Cowan says:

    There must be worse problems with Danish and Swedish lists.

    Indeed. The truth is that lists of names should properly be sorted not according to the language of the country in which the names are used, nor the language of the origin of the names, but according to the expectations of the user, which are often quite different.

    I only know that in Norwegian the letters Æ, Ø and Å are the final three in the alphabet. The Swedes may have chosen a different system – they use diacritics Ä and Ö – possibly for this very reason (alphabetical order).

    The Swedish order is Å, Ä, and Ö. The reason the Dano-Norwegian alphabet has Å last is that it did not replace AA until 1917 in Norway and 1948 in Denmark, and novel letters are often put last (which is why Z stands where it does in the Latin alphabet). The actual 1948 proposal was to place Å first in Danish; it was rejected, not without objections from the people of the town of Aabenraa, who complained that the “svenske boller” would move them from the beginning of the alphabet to the end; in 1955 the Norwegian order was adopted. Note that these are considered unique letters in Swedish, just as much as their Dano-Norwegian counterparts.

    I assume that Swedish changed over to Ä and Ö in the 16C (which is also when Å was adopted) to make the language look less Danish, and only incidentally more German. The Danes of course already had had as much German as they could stand.

  22. David Marjanović says:

    The German solution to the ordering of ä ö ü ß, BTW, is total denial: they’re not included in the alphabet. They used to be ordered mostly as ae oe ue sz and are now mostly ordered as a o u ss. But such things as pocket-sized organizers, if they still exist, can even have separate pages for sch and st.

    Up into the 1970s, the library of the University of Vienna stuck to 24 letters and did not distinguish I/J or U/V. Going through that catalog is a surreal experience.

    Traditional-Mongolian alphabet song with two transcriptions. (Both transcriptions, as usual, lie about the fundamental nature of the vowel system… I’d need some practice to reliably tell the four rounded back vowels apart.)

    I assume that Swedish changed over to Ä and Ö in the 16C (which is also when Å was adopted) to make the language look less Danish, and only incidentally more German.

    Didn’t they consistently use w at that time?

  23. When I was young I remember catalogues and sorting guides (I’m not sure of the proper name–they were long things with folding cards with letters of the alphabet, digits, etc in order. You would sort papers by slipping them between the cards.) Anyway, I remember them grouping Mac, Mc and M’ together before M. But I haven’t seen this for quite a long time now.

  24. Yes libraries sort Mc/Mac at the start if the M’s. That’s where I came across Macchiavelli’s ‘The Prince’.

  25. The Inner Mongolian alphabetical order looks like the current Japanese syllabary order, but reading from right to left across each column, rather than down each row. The Japanese equivalent would be: a ka sa ta na ha ma ya ra wa, e ke se te ne he me [y]e re [w]e, …

    Hawaiian syllable citation order is similar: ha ka la ma na pa wa ‘a, he ke le me ne pe we ‘e, … People learn to write their hakalama rather than their abc’s.

  26. John Cowan says:

    Didn’t they consistently use w at that time?

    W in Fraktur, V in Antiqua.

  27. According to Wikipedia, the current ordering of Hangul is based on a ordering introduced by Choe Sejin in 1527. There are minor differences between the orderings used in North and South Korea.

  28. Lars Mathiesen says:

    Danish locales sort {aa} with {å}. There was actually some talk of encoding the digraph in Unicode, back around 1990 I think, specifically to avoid that — at the time I did my best to convince Keld Jørn Simonsen that people who needed that could bloody well put a ZWJ between the letters or something, but to no avail. So now there is no way to make Aabenraa and Aachen sort differently.

    (Swedish sorts {å} before {ä} and {ö}, Danish and Norwegian are the other way around).

    The just-so story about {ä/ö} vs {æ/ø} is that the Swedish king had enticed German printers to set up shop, but when the Danish king tried to play catchup he could “only” get Dutch ones. Of course the German printers had the {ä} and {ö} sorts, but the Dutch had to make do with {æ} from Latin and putting a stroke in the matrix for {o}.

    On the other hand, Danish manuscript tradition used essentially the same {æ} and {ø} as the current printed forms, so maybe the Dutch were just more willing to adapt. ({Ø} is not as common as {æ}, but the third heading starts Thæt ær kunnungs æmboth ok høfthings [thær] i land ær [at] gømæ domæ oc gøre ræt oc […]).

  29. Mac, Mc and M’ together before M.

    In my experience of Irish phone books since the 1970s, McFoo and Mac Foo are sorted as Macfoo, thus:–

    Ma, Junren

    Maberly, John

    MacAdam, Alan

    MacAdam, Brian

    McAdam, Catherine

    MacAdam, Denis

    McAdam, Eugene

    MacAdam, Fiona

    Mac An tSaoi, Máire

    Macaroni, Luigi

    McDonald, Ronald

    McGuire, Patrick

    Macken, Walter

    Mac na Mara, Pól

    Macpherson, Wiliam

    McTiernan, Turlough

    Madden, John

    Maguire, Thomas

  30. And that’s the only sensible way to do it; people should not be expected to know or remember whether a name is spelled Mc or Mac.

  31. Yes, but think about poor Luigi Macaroni…

    Which brings me to a question how pre-surname particles are treated in various languages? Should I search for Beethoven under “B” or “V”? How about ‘t Hoft? van der Waals? I mean, it’s Beethoven’s sonata, but van der Waals forces. I would guess (without looking up) that Russian goes with the most common usage ignoring morphology.

  32. Stu Clayton says:

    Should I search for Beethoven under “B” or “V”? How about ‘t Hoft?

    Google finds what you want no matter what, and a bit more. We are dealing with semi-sentient indexes (as you may imagine if you don’t peek behind the curtain like Toto). I now know that in Bielefeld ( !) a popopera called Van Beethoven is running.

  33. Why is there no Beethoven Van Line?

  34. “Ride the BeethoVan!”

  35. Search for Hoft doesn’t find ‘t Hoft, at least not on the first 3 pages. The first link to something relating to a person is

    Jim Hoft | Media Matters for America
    Jim Hoft is a right-wing blogger who founded The Gateway Pundit and has repeatedly earned the title “dumbest man on the internet.” He is also a pro-Trump …

    Sorry, I am looking for one of the smartest people on the planet, not one of the dumbest.

    ADDENDUM: And, of course, in Russia he is cited as Beethoven, Ludwig Vanych

  36. Stu Clayton says:

    ‘t Hooft has two “o”s, silly ! I found him immediately. Google caters especially for people who know 0.07 % more than can be expected.

    Anyway the Dutch always double their vowels to indicate “long”. I learned that at university (I can’t remember why I thought it would be appropriate to take two semesters of Dutch. I never even had a Dutch boyfriend, which would have explained it).

  37. Right. But search for ‘t Hoft finds correct person even with misspelling!

  38. Stu Clayton says:

    They’re onto us !

  39. John Cowan says:

    My general rule for surname particles is that a capitalized particle (Merete Van Kamp), or one that is written solid with the surname (Jerry terHorst), is part of the surname; a lower-case particle (Alexis de Tocqueville) is not. Note that in French, a lower-case particle is omitted when the first name is omitted, unless the surname is a single syllable: Alexis de Tocqueville is Tocqueville but Charles de Gaulle is de Gaulle. (ObScure: other members of de Gaulle’s family are surnamed van der Walle and Vro-C’halle.)

    The just-so story about {ä/ö} vs {æ/ø}

    It must indeed be a just-so story: the Older Law of the West Geats, which is supposedly the earliest Old Swedish manuscript, reads (though I am no palaeographer, and depend on WP here) “Varþær suþærman dræpin ællær ænskær maþær, ta skal bøta firi marchum fiurum þem sakinæ søkir, ok tvar marchar konongi.”

    poor Luigi Macaroni

    Perhaps his paternal ancestors were surnamed Mac Áróin.

  40. Lars Mathiesen says:

    Old Swedish — so if {æ/ø} was a common Nordic thing, why did printing in Sweden replace them with {ä/ö}? German printers would explain it, I’m just not sure it can be proved.

    On the other hand, on Swedish Wikipedia they show a page from another manuscript of the law that uses something like {a’/o’} — so maybe it just depended on local scribal traditions until it was standardized, and the nationality of the printers wasn’t decisive.

  41. Charles Perry says:

    Surprised that nobody has mentioned that in Lisan al-‘Arab, the listing order is alphabetical according to the final consonant, but the individual words within each section are then listed alphabetically starting from the front of the word. It’s of like a cross between dictionary order and rhyming dictionary order.

  42. Jen in Edinburgh says:

    There is a rogue miscapitaliser occasionally at work in the university systems – the last victim I remember was called Macari (or maybe they really were called MacAri!)

    Putting all the M(a)cs under Mc would avoid the problem of the other Ma-s getting mixed in.

  43. Jen in Edinburgh says:

    ‘Alphabetic Other’ is going in my list of Accidental Band Names, not that I can ever remember what else is on the list.

  44. PlasticPaddy says:

    Macari is a good Irish name☺

  45. Spurious CamelCase aside, I don’t see why anyone should have a problem with having Macaroni etc mixed in with the Gaelic Macs, any more than having the Chinese Youngs mixed in with the Anglo-Saxon Youngs. Irish phone books also collate O’s:–

    O’Hara, Alan

    Ohara, Toshihiko

    O’Hara, William

  46. rogue miscapitaliser

    Probably from trying to convert names that are in all caps to mixed case, with special handing of names starting with MAC and MC.

  47. Jen in Edinburgh says:

    Scottish too – there was a footballer Lou Macari. (Celtic manager, I think.)

  48. There’s a well-known brand of shoes, Thom McAn. (I don’t know if it is well-known outside the US.)

    There’s another version Magan that sometimes confuses people more than if it were spelled McGann.

    I also mentioned M’ which I believe was once used mostly in Northern Ireland. The most common example you might run across is the character M’Turk in Kipling’s Stalky & Co., published in 1899. I don’t spend enough time in Irish archives to get much of an idea when and how much it was used.

  49. @maidhc: I only remember the existence of the Thom McAn shoe brand because I once noticed that Wal-Mart had a knockoff brand, Thom SomethingElse.

  50. John Cowan says:

    Gale’s former married name was McGhan.

  51. David Eddyshaw says:

    If I were a McAn, I would name my daughter Megan.

  52. McArov 9mm

  53. David Marjanović says:


    That’s the first M’ I encounter whose name doesn’t go on with C.


    Oh, that explains the TRAP vowel of McGahn, the former White House lawyer.

  54. The M’Foo form seems to have peaked with the Victorians. It was most common when Foo began with C or K; perhaps to forestall attempts at geminated /kk/ by well-meaning Anglophones ignorant of Gaelic phonology. Hansard 1803-2005 has dumb-computer-sorted MPs’ surnames with the few M’Foos before the MacFoos before the Macfoos before the McFoos. I note that Justin McCarthy (1830-1912) is so called in the index; clicking through to the source transcripts they all seem to be “MR. JUSTIN M’CARTHY”; in fact checking the hardcopies he seems to have switched from M’Carthy to McCarthy some time in the 1880s.

    Mc is sometimes written with a superscripted c, and sometimes a M‘Foo single-openquote instead of apostrophe/single-closequote — I guess because it looks a bit like a c. Karl MᶜCartney MP spent thousands of tax pounds getting his superscript c in Hansard. As with Justin, he is plain McCartney in the index but MᶜCartney in the transcripts.

  55. John Cowan says:

    Huh. I had thought it was always a high-6 single quote precisely because it was a superscript c, and that the apostrophe was an error.

    It is truly a Great Literary Moment when M’Turk relapses into good 19C Hiberno-English (which my father spoke when he felt like it) when talking to Colonel Dabney about the evils of shooting foxes.

  56. That’s the first M’ I encounter whose name doesn’t go on with C.

    Johnson v. M’Intosh.

  57. J.W. Brewer says:

    “Miss Maggie M’Gill, she lived on a hill
    Her daddy got drunk and left her no will”

    (Admittedly /g/ bears a certain resemblance to /k/.)

  58. AJP Crown says:

    M’Turk in Stalky & Co is based on George Charles Beresford when he was a pupil at the USC, Westward Ho! (a school I badly wanted to go to as a small child because they wore red-and-white striped caps – or at least they did on the cover of the library’s copy of Stalky & Co.) Beresford (of the Marquess of Waterford Beresfords) is the photographer I keep mentioning who took all the pics of Wyndham Lewis and other artists, as well as one of the most famous studio portraits of the young Virginia Woolf.


  59. (Admittedly /g/ bears a certain resemblance to /k/.)

    Ah yes, I forgot about M’G~ (and M’Q~; McQuillan is not too rare). OTOH many “Mac G~” names are anglicised “Mag~” (Maguire, Maginnis), which resists collation.

    There’s also the M’Naghten Rule.

    And a not very good article https://en.wikipedia.org/wiki/Mac_and_Mc_together

  60. John Cowan says:

    A good article about the mysterious “man in the brown mackintosh” in Ulysses, who appears in the paper as “M’Intosh” and pops up throughout the book.

  61. January First-of-May says:

    Fortunately (if written together) it’s always either (e.g.) McFoo, MacFoo, or Macfoo [or M‘Foo, I suppose] – the Mcfoo option doesn’t appear to exist (or, at least, if it exists, it is excessively rare).
    So a system where Mc sorts as Mac if and only if it is followed by a capital (or possibly a space) probably wouldn’t even get confused by Georgians with too many consonants in their surname.
    [We’d still need to be careful with “M‘”, though.]

    (Wikipedia provides some cases of people with Mcfoo-style first names, but I was unable to find anyone with a last name like that; granted, it’s far harder to search for those, so I could easily have missed some.)

  62. M’Intosh

    If I didn’t read this thread, I would have pronounced the apostrophe as glottal stop

  63. John Cowan says:

    The first work cited in the not-very-good article is said to be by “Cóil, Róisín Nic”, which obviously indicates the Wikipedians’ ignorance of the subject.

  64. Stu Clayton says:

    That citation is marked in green as having been added 3 hours ago (as I write) by a certain Jnestorius.

    You’re free to correct it. Or contact Róisín Nic Cóil herself and let her decide what to do, if anything.

    Merely pointing a finger at ignorance is not playing the Wikipedia game. You’re expected to put your edit where your erudition is.

  65. Playing the Wikipedia game has become more trouble than it’s worth, now that you have to deal with a cabal of smug Korinthenkackers.

  66. A couple of days ago, I found something on Wikipedia that used to be common but is now rather rare. It was a short article that was clearly written by someone who did not know anything about the subject matter. It looked like they were trying to be helpful but had just misunderstood the terminology and written a paragraph about a completely different topic (albeit one named after the same person).

  67. Stu Clayton says:

    I don’t see JC’s criticism as an extruded raisin. It took a bit of searching to figure out what he meant. Do I know from Irish already ? Now I sorta understand one small item, but am not emboldened thereby to edit it.

    Over the last few years, merely while browsing through WiPe articles for other reasons, I have swept up dozens of raisins en passant – typos and ESL-speak. Just my widow’s mite for the right cause.

  68. Stu Clayton says:

    So I myself am a Korinthenfeger, cleaning up after the -kacker as time permits. Mostly I tend the cows and tilt against the windmills of change.

  69. John Cowan says:

    My point was that (anecdotally) any change I make is very likely to be reverted by somone who assumes either bad faith or ignorance on my part. So I gave up reading comments and making changes some years ago. I agree that this is not playing the WP game, but it’s become a game I don’t want to play. My life is too short to try to find a “reliable secondary source” that explains how Nic Cóil writes her name (or other people write it) when inverted. The WP article on Irish names explains the system, but you can’t (rightly so) cite WP on WP.

  70. What he said.

  71. I have several times run into 19th-century dictionaries of languages not spoken anywhere near South Asia still following broadly Sanskrit rather than Latin alphabetization, apparently felt for a while to be “more scientific”.

    Another interesting approach is “two-tiered” alphabetization, where words are sorted primarily by their consonant/syllabic skeleton, followed only secondarily by their vowels, so that e.g. tak, tok, tuk all precede taku, which precedes taks, which precedes all of tap, top, tup (i.e. they’re sorted as if they were something like tVk.a tVk.o tVk.u tVkV.au tVks.a tVp.a tVp.o tVp.u). This is practical in some large interdialectal comparative dictionaries: when vowels can vary quite a bit and you might not know offhand the lemma corresponding to a dialect form tok (or tōk, or tɔk, or tȯ̀k), it’s convenient to have all the tVk entries in one place. For natural reasons all examples I’ve seen are Uralic, but it would work and probably has been done for other cases too I’m sure.

    A special shoutout should go to Sölkupischer Wörterbuch (Alatalo 2004) which uses the following phonetically based multi-tiered order:
    1. ∅ p m/w t/č n ć ń/j/-ľ- k/q ŋ s/ś -r- l/r- ľ-
    2. a uə o u i̮ e̮ ä e ö ü üə i
    The equated letters (functioning as a small mid-tier themselves when contrastive) follow mostly from dialectological considerations, such as *č > /t/ in Northern Selkup, word-internal *m > w in several dialects. Several other details, though, seem gratuitous.

  72. Good lord!

  73. David Eddyshaw says:

    I’ve always been particularly annoyed by the Egyptological alphabetical order, which seems to have nothing whatsoever to recommend it all.

  74. >In my experience of Irish phone books since the 1970s, McFoo and Mac Foo are sorted as Macfoo, thus:–
    >Ma, Junren
    >Maberly, John
    >MacAdam, Alan
    >MacAdam, Brian
    >McAdam, Catherine

    I help run an election office. Till recently, there was a standard in the voter registration unit that all such names would be given a space, regardless of whether the voter or the registrar used no space, space, or apostrophe — thus Mc Adams, Mac Adam, O Connor, De La Cruz.

    Recently while I was on leave, someone decided that it was best to have voters decide for themselves. This might seem to make sense. However, many registrations come on election day. The e-pollbooks (voter list laptops) have no string function that allows searches that would compress that space or eliminate the apostrophe to return matches. And the election judges aren’t great at searching all possible combinations. So election day registration now sends back an extraordinary number of duplicates.

    Part of the problem is that a lot of Mcs and Os have no standard themselves, and go back and forth depending on the form. Nevermind some of the de la Madrid Hurtados, who come in as dela Madrid one time, Hurtado the next.

    I found a maxim on the internet somewhere:
    Falsehoods Programmers Believe About Names:
    # 39 – people whose names break my system are weird outliers. They should have had solid, acceptable names.

    I’ve posted it on my door, but only partially in self-mockery. The system should adapt, but till it does, voters need solid, acceptable versions of their names. We’re going back to the rule of Mc Adams after this election.

    Identity voter fraud, while by no means common, is not so rare as activists would have you believe, so culling the list of deadwood, duplicates and other bad registrations is important.

    >Why is there no Beethoven Van Line?

    There was the band, Camper Van Beethoven.

  75. David Marjanović says:

    Sölkupischer Wörterbuch

    Masculine Buch is Yiddish. 🙂

  76. Jen in Edinburgh says:

    I’ll do it, then – I don’t care whether I’m reverted or not. And the linked article is itself (presumably) a reliable secondary source.

    I’m tempted to leave it for the irony, though 🙂

  77. Swedish used to mix mix the Vs and the Ws together, but now I mostly see ordering where W has its own section after V. As for the ordering after date of aquisition, it’s very convenient if you want to see what’s come in since you last visited the library. Not so convenient today, I guess, with huge library systems who have many acquisitions every week.

  78. Smaller libraries have special sections for new arrivals…

  79. Stu Clayton says:

    Jen, your correction of the author’s name in the citation has survived so far ! Your comment preserves the irony: “Correct form of Irish surname (see linked article for details!)”

  80. Lars Mathiesen says:

    Now alphabetize this.

  81. I’m going to blame Google Scholar for “Cóil, Róisín Nic”. Of course trusting Google Scholar metadata is itself blameworthy.

  82. January First-of-May says:

    39 – people whose names break my system are weird outliers. They should have had solid, acceptable names.

    As I recall, the original actually ended in something along the lines of “they should have had solid, acceptable names, like 张伟”.

    (I think it was actually some other pair of Chinese characters; I substituted the ones that I thought would fit best in the context.)

    [EDIT: looked it up, and the original example was in fact “田中太郎”, a Japanese name, though the message is similar.]

  83. Now alphabetize this.


  84. Lars Mathiesen says:

    Hmm, I thought it looked like a link when I posted it. EDIT: This one still does, I don’t know what happened to the other.

  85. They have an audio file so you can hear how Gǃkúnǁʼhòmdímà sounds — kudos!

  86. John Cowan says:

    “Falsehoods Programmers Believe About Names”. See also my essay “Against Structured Names and Telephone Numbers”, which links to a post on Thai names by James Clark as well as “a brief lecture from your neighborhood pedant on Russian names”, a comment I wrote on a Write Badly Well post.

  87. Athel Cornish-Bowden says:

    Thanks for these links. Both are interesting.

    I thought I was going to get to the end of the comments on the post about Thai names before anyone said anything about Brazilian and Portuguese names, which are often more problematical than Spanish names. The order Given name(s)-Paternal surname-Maternal surname is pretty consistently used in Spanish names, but is often ignored in Portuguese names. Once in Portugal I asked why that was, and was told that they had historically followed the Spanish order, but in the early part of the 20th century the government changed it to Given name(s)-Maternal surname-Paternal surname in order to emphasize that they weren’t Spanish and didn’t have to do what the Spanish liked. That was fine, except that, as the Brazilian guy said, people often don’t do what the government likes. So you just can’t tell. Worse than than that, although the Spanish system stops after one generation the Portuguese sometimes include grandparents’ names as well, in an unpredictable order. I’m not sure they necessarily stop with grandparents. If I followed that system and called myself Athelstan John Duncan McMurtry Kitson Karslake Cornish-Bowden you could probably guess where the given names end (except that you would’t know whether Duncan was a given name or a surname), but after that you’d be lost. I went on a research assessment exercise in Portugal in 1999, and was sent lots of CVs, in which some people had as many as ten names.

    Official lists of names (for example results of student exams on noticeboards in universities) sometimes order them alphabetically by given names in Portugal and Brazil. I have encountered that in both countries, and nowhere else. No doubt I would if I went to Iceland, and I think they may do it in Turkey as well.

  88. David Eddyshaw says:

    The Kusaasi traditionally didn’t use surnames of any kind (unlike the Mossi, who use clan names like surnames); as a handful of given names cover the great majority of people, this does not lend itself very well to the demands of state bureaucracy.

    In contexts where they need to interact with English- or French-speaking officials, people use European-style baptismal names (or their formal Arabic Islamic names if they’re Muslims) for their “given names” and use their actual personal names as “surnames.”

    I don’t think the system is as yet well enough established that people have got round to inheriting their fathers’ personal names as surnames.

  89. Reminds me of this Anna Korostelyova’s post:

    Oh my God, what a blessing that the Faculty of Medicine came up with the column “Other names” in their list of students, straight after the last name and first name! How fortunate that this brilliant idea occurred to them!
    Now finally everything is clear and transparent, for example: Alsaegh Loai, and then, in the column “Other names” – Ismail Musa Yusuf Musa Rami Emad Saad. Or Abdelwahab Ahmed, and then, as a separate item, Abdelhadi Mohamed Abdelrahman Amin Sayed Abdullah and so on. And all of Egypt, Iraq, Bahrain, Lebanon is processed in this manner. What a genius they have in the academic affairs! Before this, the start of the academic year was like: “Ah … uh … but which one of these is … your … uh … In short: what should we call you??”

  90. Adding a name in characters to rule 39 improves it. Thanks.

    I already had to edit my version to “Things election administrators and programmers …” in order to avoid offending our programmers.

  91. J.W. Brewer says:

    Back to the mysteries of M’Foo: I just saw in another context someone dredging up a N.Y. Times story from just prior to the 1968 primary. The key point of historical interest in hindsight is that the story badly mispredicted the outcome of the primary, but the point of orthographic interest is that Sen. McCarthy (Gene not Joe, for those not up on the details of the timeline) was “McCarthy” in text but “M’CARTHY” in the headline. Not sure if this was a weird glitch or a recognized space-saving technique for headlinese. https://www.nytimes.com/1968/01/15/archives/mcarthy-called-weak-in-first-bid-all-new-hampshire-rivals-see-no.html

  92. I’m pretty sure it was the latter; I’ve seen it elsewhere.

  93. David Marjanović says:

    In contexts where they need to interact with English- or French-speaking officials, people use European-style baptismal names (or their formal Arabic Islamic names if they’re Muslims) for their “given names” and use their actual personal names as “surnames.”

    In Rwanda this is official. Everybody seems to have a very Catholic name in French form as a given name, and everybody has their own surname which is neither heritable nor changeable, is in Kinyarwanda and is chosen in a ceremony by older relatives.

  94. John Cowan says:

    In Melanesia, Christianized people adopted a Bible name as their first name and kept their own names as surnames (Margaret Mead talks about it). Several generations later, the original names are now ordinary surnames.

  95. January First-of-May says:

    Kenya used to run on the Ethiopian system of Name Fathersname, then the British arrived and made everyone’s patronymics into surnames.
    So, for example, Onyango Obama (b. ca. 1895) got his name from his father Obama Opiyo, but after the British arrival his son ended up with the name Barack Obama (instead of, say, Barack Onyango).

    Many other countries had a similar thing happen, though usually earlier and often much more slowly; my favorite example of a not-really-surname-based name is Nikita Romanovich Zakharyin-Yuryev, which approximately deciphers to “Nikita, son of Roman, son of Yury, son of Zakhary”.

  96. David Eddyshaw says:

    In his remarkable grammar of the remarkable New Guinea language Yimas, William Foley says that in official contexts Yimas call themselves by their baptismal names (they’re all Catholic now) followed by the father‘s traditional name. This presumably relates to the fact that traditional names were only used in speaking Yimas, not Tok Pisin, and not even always then; moreover, even when Foley described Yimas in 1991 many children were growing up speaking Tok Pisin alone.

  97. David Marjanović says:

    Many other countries had a similar thing happen

    *raises hand*

  98. “libraries sort Mc/Mac at the start if the M’s. That’s where I came across Macchiavelli’s ‘The Prince’.”

    Oh, aye, the weel-kent Scots political theorist McIavelly. One of the further-flung members of the tribe, along with that pillar of the Cypriot Kirk, Archbishop McArios.

  99. I once invented a political science professor named McHevill.

Speak Your Mind