The Last Link is a Reward for Getting Through the First Link

June 21, 2007 6:50 PM

The concept of alphabetization was invented at the Great Library of Alexandria in the third century BC, with words grouped by first letter. It wasn't until 1053, in the Elementarium doctrinae erudimentum that recursive alphabetization (where "Aab" comes before "Aac" and after "Aaa") appeared in rudimentary form. You'd think that by now we'd have the process down, but controversies still rage. Does "sea foam" come before "seaborne"? Does "Michael Jackson" come before "Nick Cave"? Throw in international characters and an occasional foray into ASCIIbetical order and it's no wonder the alphabet can be so frustrating.
posted by L. Fitzgerald Sjoberg (62 comments total)

Huh, I never would have thought that recursive alphabetization would have been a distinct development from first-letter ordering... that's now one of the more esoteric things I'll probably never forget.
posted by phrontist at 7:08 PM on June 21, 2007

Speaking of alphabetization, has anyone else noticed that no matter how Don Music throws himself at the keyboard he always hits the same chord? That seems unlikely, doesn't it?
posted by yhbc at 7:13 PM on June 21, 2007 [3 favorites]

... Because it is the correct method, the letter-by-letter version of alphabetization is the form you'll find in dictionaries and encyclopedias. Here's a listing of the first few "sea" entries in my Random House College Dictionary:

sea, sea anchor, sea anemone, sea bag, sea bass, seabed, Seabee, sea bird, sea biscuit, seaboard, Seaborg, sea-born, seaborne, sea bread, sea bream, sea breeze.

There are five columns of entries starting with "sea" in this dictionary. It would obviously be nonsensical to have "sea wrack" come immediately before "seabed", let alone trying to find "sea-maid" or "seat belt".

I just thought I should straighten folks out about this. ...

That's always seemed weird and wrong to me if there's actually a space--it should really go " sea, seabed, Seabee, seaboard, Seaborg, seaborne, ..."

I've always thought that dictionaries needed subordinate room or something for those multiple word phrases--or maybe just a whole other section or something. (but seafoam is really just one normal word i think nowadays, especially if it's the color)

I guess when books themselves get some hyperllnk capability, it can happen maybe. (you'd click or something to see all the words and phrases that start with that word) And we create new multiple word phrases all the time--dictionaries are never caught up.
posted by amberglow at 7:17 PM on June 21, 2007

Or the real problem is that there's no standard rule about whether you should mush the words together to make a new compound word or not. (or is there?)
posted by amberglow at 7:19 PM on June 21, 2007

Seaborne comes before sea foam, Ralph meant a literal Viking, and I love this post.
posted by gubo at 7:20 PM on June 21, 2007

I have little to no opinion on "seaborne" vs. "sea foam", but...

Does "Michael Jackson" come before "Nick Cave"?

Good lord, no.
posted by Flunkie at 7:22 PM on June 21, 2007

Proper names should always be by Last Initial--it'd be impossible to find otherwise. (and there'd be pages and pages and pages of Michaels and Peters and Johns in every Dictionary)
posted by amberglow at 7:25 PM on June 21, 2007

The idea of inventing things like this fascinates me, because some early Canaanite could have come up with recursive alphabetization, but apparently none did. It makes me wonder if there's anything similarly obvious in retrospect that any of us could invent if the right notion muscled its way into our heads.
posted by L. Fitzgerald Sjoberg at 7:28 PM on June 21, 2007 [1 favorite]

Lore, I believe I told you recently to get the fuck out of my head. Now you post this? Mere days after I finished reading Letter Perfect: The Marvelous History of the Alphabet? Out. Of. My. Head. Get.
posted by jacquilynne at 7:28 PM on June 21, 2007

Added bonus alphabetization trivia: Icelandic phone books are listed by first name, because Icelandic last names are taken from the first name of the father.
posted by L. Fitzgerald Sjoberg at 7:30 PM on June 21, 2007

The idea of inventing things like this fascinates me, because some early Canaanite could have come up with recursive alphabetization, but apparently none did.

Well, weren't most writings in the form of scrolls then or clay tablets earlier, etc? You needed some critical mass of information that needed to be sorted somehow beyond people's personal ordering or a king's collection. The Alexandria Library was one of the first attempts to collect all the known world's knowledge, no?

How did monasteries classify/order stuff during the Dark Ages?
posted by amberglow at 7:31 PM on June 21, 2007

Apparently -- and my knowledge here is rudimentary -- they collated things by other criteria depending on the set to be collated. Religious materials might be ordered by book of the Bible, materials on gardening might be organized by type of plant, so on and so forth.
posted by L. Fitzgerald Sjoberg at 7:38 PM on June 21, 2007 [1 favorite]

they collated things by other criteria depending on the set to be collated. Religious materials might be ordered by book of the Bible, materials on gardening might be organized by type of plant, so on and so forth.

Actually, that makes sense. At the risk of starting a huge "here's how I organize my music collection/High Fidelity" moment, I wouldn't think of including the Broadway Musicals or the a cappella cds in the same alphabetical order as the rest of them. However, in the "main" category, Sinatra, Frank, can and does go between Shriekback and Siouxsie (and the Banshees), and I don't think anything of it.
posted by yhbc at 7:49 PM on June 21, 2007

Oh, hell no. Damn ye, 15th edition of the Chicago Manual of Style! Damn ye!
posted by ormondsacker at 7:49 PM on June 21, 2007

I don't understand the Last Name First Name convention of name order. Mainly because this doesn't work for Asian names. You'd rarely, if ever, call a Muslim by their last name as that's usually their father's name (so Ahmad Hussein would be referred to as Ahmad in any context, formal or informal, but never Hussein). Chinese names go Last Name First Name First Name, but if they have an English name it would be First Name Last Name First Name First Name, like Sarah Lim Chew Sze. (This has made filling up forms difficult for my Chinese friends.) Malaysian Indians often have their names has So and So s/o or d/o (so of or daughter of) Father, i.e. Ramasamy s/o Muthusamy, so they'd be known as Ramasamy in any context.

Trying to find people by last name doesn't work at all, since THAT would make people impossible to find (unlike amberglow's comment). There are thousands of Hussiens and Lims and s/o Muthusamys, but first names tend to be more unique. In Malaysia everything is First Name Last Name, though in press they often refer to Chinese and Western people by their last name.
posted by divabat at 7:54 PM on June 21, 2007

Chinese names go Last Name First Name

Of course they don't. Last name is last, by definition. Chinese names go Family Name Given Name.
posted by scottreynen at 8:10 PM on June 21, 2007 [3 favorites]

Added bonus alphabetization trivia: Icelandic phone books are listed by first name, because Icelandic last names are taken from the first name of the father.

Most of the time. I'm one of the few Icelanders who has a family name. My paternal grandfather is in the phonebook both under his last name and his first name, while my father and his siblings are only listed by first name. But yeah, most of everybody a) has a patronym for a last name b) is listed in the phone directory by their given name.
posted by Kattullus at 8:16 PM on June 21, 2007

I don't understand the Last Name First Name convention of name order.

Is that how things are alphabetized in those languages? I mean, family name, given name is the standard in English and Romance languages (and makes quite a lot of sense), but is that the case in languages associated with cultures where it doesn't make sense?
posted by jacquilynne at 8:31 PM on June 21, 2007

Does "Michael Jackson" come before "Nick Cave"?

Um... BOYS' PANTS HALF OFF!

No wait, that's what Michael Jackson has in common with a department store.

Let's try again.

Does "Michael Jackson" come before "Nick Cave"?

I really, REALLY do NOT want to know.
posted by tckma at 8:34 PM on June 21, 2007

scottreynen: same diff. :P

jacquilynne: Not in Malaysia. They're alphabetized by first letter of your official name, doesn't matter if it's Ali bin Ahmad or Lim Ai Ling or G. Peruval or Andrea Fallows. With those examples, if they were on a list, they'd be listed as this:

Ali bin Ahmad
Andrea Fallows
G. Peruval
Lim Ai Ling

Basically it's what shows up on your IC or passport, positioning be damned.
posted by divabat at 8:43 PM on June 21, 2007

What annoys me is that I've known the English alphabet for over 20 years and still, when I open the yellow pages to 'Mechanics' and I happen to be looking for 'Pest Control,' I have to pause for a few seconds and think: "Does P come before M? Ok-- Let's see-- HIJKLMNOP. No. So that means I need to go right."

I'm always asking myself that when looking through alphabetic things-- at record stores or in a book store where the authors are arranged by last name-- "Does I come before K? HIJK-- yes. So I need to go left."
posted by Kronoss at 8:50 PM on June 21, 2007 [1 favorite]

" It makes me wonder if there's anything similarly obvious in retrospect that any of us could invent if the right notion muscled its way into our heads."


Organization by smell. You need to see it in action to understand, but trust me.
posted by furiousxgeorge at 8:59 PM on June 21, 2007

I think that's called "laundry day."
posted by L. Fitzgerald Sjoberg at 9:00 PM on June 21, 2007 [1 favorite]

Does "Michael Jackson" come before "Nick Cave"?

No, children.
posted by five fresh fish at 9:10 PM on June 21, 2007

There are thousands of Hussiens and Lims and s/o Muthusamys, but first names tend to be more unique.

Is that really true tho? In most of the world there are consistently popular names. The majority of people all over tend to share first names based on religion or history or fame, i believe. In a Western, Christian-dominated country, you'd find far more famous/notable Michaels and Peters than famous/notable Jacksons, even tho trends come and go in naming. I've heard the Chinese tend to go for floral names for girls and prosperous or lucky ones for boys but i don't know if it's true.
posted by amberglow at 9:15 PM on June 21, 2007

This is less of a problem for databases than for lists, since databases in general are capable of handling more intelligent queries, eg "M%gill%cud%" to find McGilliecuddy, MacGillycuddie, and all their variations. Personally I like Google's approach, of converting diacritically-marked letters to their nearest 'plain' equivalent, eg "Renée" is searched for as if it were "Renee", "Grøndahl" as "Grondahl".

To add to the Gaelic issues: Apparently the 'c' in Mc is traditionally written superscript, ie M^cArthur, but this fell out of favour after the printing press became common; many printers just didn't bother with it, or used M'. Similarly, Fitz- is a prefix, and O' is a printer's alteration of the original Ó. The more one digs into name spelling, the more complex the subject becomes.

The university I work at has a lot of international students, and we have less trouble than one might expect, primarily because we always ask for "name and student number". Most students are thoroughly conditioned to write both by midway through first year. If one or the other, or both identifiers have some issue, it's usually possible to identify the person by searching on the combination, or checking the list of students in the course, or some similar approach. In practice it doesn't actually matter what naming convention the student chooses to use, as long as they personally are consistent in all of their communications with the university. Same goes for any other entity, I expect - establishing consistency is the first task, before sorting can be considered.
posted by aeschenkarnos at 9:18 PM on June 21, 2007

this is interesting, about the variations in Indian regions about naming (but there would still be popular names that many carry, i believe)
posted by amberglow at 9:22 PM on June 21, 2007

and this about Mohammeds in the UK (next year it'll be the #1 baby name, they say)
posted by amberglow at 9:27 PM on June 21, 2007

Is it really called "recursive alphabetization"? I had never heard that phrase before and wanted a bit more insight. So I did a Google search and it only came up with three entries. The first was MetaFilter and the third doesn't seem to REALLY be about alphabetizing.
posted by Slack-a-gogo at 9:27 PM on June 21, 2007

the third doesn't seem to REALLY be about alphabetizing

You don't say. Here's the little blurb Google throws up for the third link:

"Put-up Colossae destine the recursive alphabetization teen fucking for cash with gluteal hibernation. Polyandrous disgruntlement felt the unpasteurized"

The focus of this site is not pasteurization, Colossae, alphabets or hibernation.
posted by Kattullus at 9:32 PM on June 21, 2007

Adding to the confusion: what to do with surnames of Irish/Scottish descent?I am often accessing our (non-electronic) employee files at work. There, the convention is:McArdle
McTiernan
Macchio
Macintosh
Nabors
O'Bannon
O'Toole
Ochoa
posted by skammer at 9:33 PM on June 21, 2007 [1 favorite]

Metafilter: The focus of this site is not pasteurization, Colossae, alphabets, or hibernation.

I made up the term "recursive alphabetization" because it was the best term I could come up with. I don't think there's an accepted term for it, normally it's just called "alphabetization."
posted by L. Fitzgerald Sjoberg at 9:39 PM on June 21, 2007

Amazingly, only one result occurs for "Polyandrous disgruntlement". Go figure.
posted by skammer at 9:39 PM on June 21, 2007

I think your 'recursive alphabetization' is called 'lexicographical ordering'.
posted by lathrop at 9:45 PM on June 21, 2007

Somewhat relevent Straight Dope article regarding the Dewey Decimal system. It gives a pretty detailed account of pre-Dewey methods of collating data.
posted by cj_ at 10:59 PM on June 21, 2007

amberglow: It doesn't matter what the popular names are. Malaysia is such a racially diverse country that first names can span the globe (what with current trends in fancy names, plus more immigration, and all). However, because of the convention of having your father's name be your last name, and because families tend to stick around together, there are more common last names.

For example, if you went to a school to look for a student named Lim, you will get about half the Chinese population. Get an uncommonish last name such as Gan and you'll still get quite a few people. Give them a first name like Chew Sze and it's easier to find your specific person.

Also, your system would benefit girls more than guys. People do not get their mother's name as their last name. By your system, if someone came in looking for Shafiq or Samyvelu, they'd get all the guys with that as their first OR last name, and all the girls with that as their last name. If it was a girls name like Shanti or Shafiqah, then you only get folks with those as their first name. If you go only by the basis of first name, you get less false positives.
posted by divabat at 11:04 PM on June 21, 2007

Just an addenda on that comment: You could go in with a first name ONLY, no last name, and it would be a lot easier to find your specific person than it is to go with a last name ONLY and nothing else.
posted by divabat at 11:05 PM on June 21, 2007

This is a fascinating post (and discussion)...

aeschenkarnos and skammer mentioned the Scottish/Irish last name issues... My maiden name was a "Mc" name and there was never a clear-cut convention for its alphabetization when I was growing up... In some class rosters, the "Mc" people were before the rest of the "M" people, in some we were after the "M" people, and in others we were incorporated into the "M" people with the "c" was treated like a normal second letter... Very confusing.
posted by amyms at 12:13 AM on June 22, 2007

The idea of inventing things like this fascinates me, because some early Canaanite could have come up with recursive alphabetization, but apparently none did.

Lacking Canabnites they also lacked the "we're higher than you in list form!" incentive.
posted by vbfg at 1:52 AM on June 22, 2007

Many Dutch people, including me, have last names of the form 'van den XXXXX', these are listed alphabetically by the XXXXX part.
posted by atrazine at 3:53 AM on June 22, 2007

[IS GOOD THIS]

The following paragraph from the first link clarifies a few things (for instance, that there was no continuity between the ancient and medieval systems):

The principles of alphabetical order, known in Antiquity, were later forgotten and were rediscovered in the Middle Ages. Miethaner-Vent explains that the first step was grouping together all the words beginning with the same letter. Next came organization to the second letter, which was not always the second letter of the word but rather the vowel of the first syllable. Words whose initial syllable was pronounced in the same way were grouped together. It is this arrangement which is found in the Affatim Glossary (ninth century) and described by Daly: "[the order] is determined by the first letter and by the first vowel following the first letter of each word, regardless of what other letters may intervene. It results, for example, that words beginning in fla- and fra- may precede words beginning in fa-, and that 144 glosses intervene between affatim and affecta."

As for a couple of particular issues:

1) It's crazy to list M', Mc, and Mac in strict alphabetical order—who can remember whether a given M(a)cIntosh spells his name with or without the -a-?

2) The Swedish alphabet ends ... z, Z, å, Å, ä, Ä, ö, Ö. This is madness. Everyone (i.e., Norwegians and Danes) knows that å is the last letter of the alphabet. Jævla svenske!
posted by languagehat at 6:03 AM on June 22, 2007

some early Canaanite could have come up with recursive alphabetization, but apparently none did.

We have no way of knowing that. All we know is that none has come down to us. Many, many things known to the ancients vanished without trace in the transition to the Middle Ages.
posted by languagehat at 6:04 AM on June 22, 2007

The Swedish alphabet ends ... z, Z, å, Å, ä, Ä, ö, Ö.

The Icelandic alphabet ends X, Y, Z, Þ, Æ, Ö. We have no use for Danish fripperies like Å.
posted by Kattullus at 6:22 AM on June 22, 2007

We have no use for Danish fripperies like Å.

And thus you are excused from the moral requirement to put Å at the end of the alphabet. Besides, I know better than to piss off Icelanders—they burn down your homestead, bury a hatchet in your skull, and carry your wife and children off into slavery on the slightest pretext.
posted by languagehat at 6:30 AM on June 22, 2007 [2 favorites]

Just look in the local Yellow Pages. Advertisers know how to game the alphabet. How many variations on 'A.A.A. [ad finitum] aardvark Pest Controllers' are there? Who really needs to get rid of an aardvark? As for 'A.A.A.aardvark Furniture Removers', well I have no aardvark fashioned furniture.
posted by tellurian at 7:24 AM on June 22, 2007

The Icelandic alphabet ends X, Y, Z, Þ, Æ, Ö.

Z? You ain't got no stinkin' Z. That archaic thing was abolished years ago. (Though, it does appear now and again.)

What I've never been able to understand is why is ð the 5th letter in the alphabet when nothing can start with it. Shouldn't it be at the end?
posted by weebil at 7:24 AM on June 22, 2007

Cool post!
posted by carter at 7:38 AM on June 22, 2007

Over my head. I'm still lost in the CD store looking for the Dead in the D section and the Airplane in the A.
Is Elvis under E or P?
posted by cccorlew at 8:08 AM on June 22, 2007

1) It's crazy to list M', Mc, and Mac in strict alphabetical order—who can remember whether a given M(a)cIntosh spells his name with or without the -a-?

Why is it crazy? It seems like it would help to have all the "Mac" and "Mc" names together in two groups in the "M" section, and each group gets sub-ordered alphabetically. So in the phone book you can just go to the "Mc" section, then find "McIntosh" within that group, and if it's not there, it's in the "Mac" section. Same with Dutch names. Just put all the "van den" names in the "v" section, and alphabetize them further within their group. So if I'm looking for "van den graaf", I'd go to "V", then to the "van den" part, then look for "graaf". It helps because it makes it easy to find these name-groups instead of having to go all over the phone book, especially if you're looking for several people with "Mc" or "Mac" names.
posted by Sangermaine at 9:29 AM on June 22, 2007

Sea foam comes before seaborne. Last Name, First Name. There is no difference between the metaphorical and literal Viking.
posted by dame at 11:00 AM on June 22, 2007

languagehat: 2) The Swedish alphabet ends ... z, Z, å, Å, ä, Ä, ö, Ö. This is madness. Everyone (i.e., Norwegians and Danes) knows that å is the last letter of the alphabet. Jævla svenske!"

The Finnish alphabet also ends with Z, Å, Ä, Ö. Any other way is an insult to the natural order of things, as far as I'm concerned. Hemmetin tanskalaiset ja norjalaiset, opettelisivat nyt edes aakkostamaan!
posted by severiina at 11:11 AM on June 22, 2007

Z? You ain't got no stinkin' Z. That archaic thing was abolished years ago. (Though, it does appear now and again.)

The z-rule* was abolished in the 70s, but not the letter itself. There are still people called Zakarías. Cars have license plates with the letter z. And even though the z-rule was officially abolished, it's made a bit of resurgence lately, as the main subscription newspaper has started using it in their editorials.


*The z-rule was roughly (in terms as ungrammatical as possible) that where a ts sound became an s sound, one wrote z instead, even though it was the same sound as the letter s. For instance, hertist (hardened) became herzt (has hardened, instead of hertst). Now hertist turns into herst.
posted by Kattullus at 11:47 AM on June 22, 2007

Why is it crazy? It seems like it would help to have all the "Mac" and "Mc" names together in two groups in the "M" section, and each group gets sub-ordered alphabetically. So in the phone book you can just go to the "Mc" section, then find "McIntosh" within that group, and if it's not there, it's in the "Mac" section.

Because it helps even more to have all the "Mac" and "Mc" names together in one group, so you just go to the "Mc" section, then find "McIntosh" within that group, and if it's not there, it's not in the phone book. Capeesh?
posted by languagehat at 11:59 AM on June 22, 2007

languagehat
My cat's breath smells like cat food.
posted by Sangermaine at 2:16 PM on June 22, 2007

Nobody has reconciled alphabetical and numerical ordering, though.

You can see this if you have files with numbers as names. They get listed as:

1.txt
10.txt
100.txt
11.txt
12.txt
...
2.txt
20.txt
...
etc.

Completely useless.

So if I have two books, one called "My 1000 Monkeys" and another called "My 230 Monkeys", which comes first?
posted by vacapinta at 3:31 PM on June 22, 2007

I'd put My 1000 Monkeys first -- because when mixing numeric and alphabetic characters, I'd alphabetize the numbers as if they were spelled out. My One Thousand Monkeys / My Thousand Monkeys both come before My Two Hundred Thirty Monkeys.
posted by jacquilynne at 5:06 PM on June 22, 2007

Kattullus: Thanks for the explaining the z-rule. I learned my Icelandic by immersion and am very vague on the nuances of grammar. Interesting the z is making a resurgence. Is there a reason Morgunblaðið is using it in the editorials?
posted by weebil at 7:01 PM on June 22, 2007

I wish I weren't language-retarded.

Actually, I really should get over it, go buy a learn-to-speak-X DVD and give it a go. Just because I hated French in high school and had to cheat to pass the final exam so that I'd never, ever have to take it again doesn't necessarily mean I can't learn a new language.

I just can't learn French. :)
posted by five fresh fish at 7:06 PM on June 22, 2007

Is there a reason Morgunblaðið is using [the z-rule] in the editorials?

Eh... the personal predilections of the editor. It's a bit weird, actually. It strikes most Icelanders roughly similar as, say, if The New York Times decided to roll back all of Noah Webster's spelling changes. Icelanders are very prone to language fascism, which also results in absurd levels of linguistic one-upmanship. Deciding to use an obsolete rule of spelling is a perfect example of this kind of behavior. When a poet, a very famous and well-respected one, can get away with lines like "country, nation and language/the one, true trinity" (land, þjóð og tunga/þrenning sönn og ein), and having it become his most well known and quoted couplet, you know there's something odd afoot.

And on the subject of the oddness of the Icelandic alphabet. Some Icelanders will tell you that C is not an Icelandic letter. But again, Icelandic Carl's and Camillas will set you straight.

What were you up to in Iceland, weebil?
posted by Kattullus at 7:32 PM on June 22, 2007

I was an exchange student for my junior year of high school and I've never been back. No good reason--except maybe the heartbreak of having to leave again; I would have gladly stayed forever.

It's got to be a bit odd to have the old spellings on only the editorial page rather than throughout the paper. Then again, IME, quirkiness and Icelanders go hand in hand.
posted by weebil at 2:26 PM on June 23, 2007

What would make Iceland worth "gladly stay[ing] forever"?
posted by five fresh fish at 6:00 PM on June 23, 2007

Sheesh. I'm not sure how to put it into words. I loved the landscape, the culture, and the people. I felt comfortable and inspired in a way I haven't experienced anywhere else I've lived or visited in the 20 years since. For me it was a wonderful combination of past and present, nature and artifice, intellectualism and whimsy.

All of which isn't to say that Iceland is perfect: xenophobia is a real problem, IMO. I always perceived the distrust and dislike as being directed toward the US as a political/military entity and, some would say, an occupying force. Unfortunately, it sometimes erupted into drunken brawls between US military personnel and Icelanders. Seeing 15 guys kicking the shit out of 2 US airmen in downtown Reykjavík was one of the scariest experiences of my life.
posted by weebil at 12:05 AM on June 24, 2007