Janus --

Welcome. Thanks for this contribution. Very thorough. It's interesting that the rule for abbreviation is so simple. I would think this could lead to some confusion. But perhaps Danes don't abbreviate ordinals that often.

Who else? What about French, Italian, Spanish, Dutch, and German?

-- K.

Spanish has variant forms in the written out numbers (e.g., decimoprimero and undécimo for eleventh), and after about 20 people don't really remember how to say them anyways like in English, but here goes. Most numbers change as follows:

Segundo (masculine singular)
Segunda (feminine singular)
Segundos (masculine plural)
Segundas (feminine plural)

All numbers ending in -ero (1st and 3rd) also have a variant form used when placed in front of masculine singular nouns:

Primer (masculine singular in front of noun)
Primero (masculine singular elsewhere)

The abbreviations are formed by using the number in digit form plus a period and a raised lowercase ending (underlined in the examples). In practice the period can be placed directly beneath the letter or omitted. Note in my examples I'm not superscripting the letters because I can't on here. Unicode has º and ª, but not the er, os, and as variants.

1 uno/una = primer/o/a/os/as = 1.er 1.o 1.a 1.os 1.as
2 dos = segundo/a/os/as = 2.o 2.a 2.os 2.as
3 tres = tercero/a/os/as = 3.er 3.o 3.a 3.os 3.as
4 cuatro = cuarto/a/os/as = 4.o 4.a 4.os 4.as
5 cinco = quinto/a/os/as = 5.o 5.a 5.os 5.as
6 seis = sexto/a/os/as = 6.o 6.a 6.os 6.as
7 siete = séptimo/a -or- sétimo/a/os/a = 7.o 7.a 7.os 7.as
8 ocho = octavo/a/os/as = 8.o 8.a 8.os 8.as
9 nueve = noveno/a/os/as -or- nono/a/os/as = 9.o 9.a 9.os 9.as

10 diez = décimo/a/os/as = 10.o 10.a 10.os 10.as
11 once = undécimo/a/os/as -or- decimoprimero/a/os/as -or-onceno/a/os/as = 11.er 11.o 11.a 11.os 11.as
12 doce = duodécimo/a/os/as -or- decimosegundo/a/os/as -ordoceno/a/os/as = 12.o 12.a 12.os 12.as
13 trece = decimotercero/a/os/as -or- decimotercio/a/os/as = 13.er 13.o 13.a 13.os 13.as
14 catorce = decimocuarto/a/os/as = 14.o 14.a 14.os 14.as
15 quince = decimoquinto/a/os/as = 15.o 15.a 15.os 15.as
16 dieciséis = decimosexto/a/os/as = 16.o 16.a 16.os 16.as
17 diecisiete = decimoséptimo/a/os/as = 17.o 17.a 17.os 17.as
18 dieciocho = decimoctavo/a/os/as = 18.o 18.a 18.os 18.as
19 diecinueve = decimonoveno/a/os/as -or- decimonono/a/os/as = 19.o 19.a 19.os 19.as

20 veinte = vigésimo/a/os/as = 20.o 20.a 20.os 20.as
21 veintiuno = vigésimo primero / vigésimo primera/os/as = 21.o 21.a 21.os 21.as
30 treinta = trigésimo/a/os/as = 30.o 30.a 30.os 30.as
40 cuarenta = cuadragésimo/a/os/as = 40.o 40.a 40.os 40.as
50 cincuenta = quincuagésimo/a/os/as = 50.o 50.a 50.os 50.as
60 sesenta = sexagésimo/a/os/as = 60.o 60.a 60.os 60.as
70 setenta = septuagésimo/a/os/as = 70.o 70.a 70.os 70.as
80 ochenta = octogésimo/a/os/as = 80.o 80.a 80.os 80.as
90 noventa = nonagésimo/a/os/as = 90.o 90.a 90.os 90.as
100 cien/ciento = centésimo/a/os/as = 100.o 100.a 100.os 100.as
123 ciento treinta y tres = centésimo vigésimo tercer/o/a/os/as = 123.er 123.o 123.a 123.os 123.as

And so on. i don't feel like tying out the rest, but basically you have

[number] + [period] + [raised er/o/a/os/as]

With the er only being used for 1 and 3.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

Dutch:

een (or één) = eerste = 1e
twee = tweede = 2e
drie = derde = 3e
vier = vierde = 4e
vijf = vijfde = 5e
zes = zesde = 6e
zeven = zevende = 7e
acht = achtste = 8e
negen = negende = 9e
tien = tiende = 10e

elf = elfde = 11e
twaalf = twaalfde = 12e
dertien = dertiende = 13e
veertien = veertiende = 14e
and so forth - 15e, 16e, 17e

twintig = twintigste = 20e
eenentwintig = eenentwintigste = 21e
tweeëntwintig = tweeëntwintigste = 22e
drieëntwintig = drieëntwintigste = 23e
vierentwintig = vierentwintigste = 24e
and so forth - 25e, 26e, 27e

honderd = honderste = 100e
and so forth

I believe officially numbers to twenty are written out full in Dutch; één, twee, drie .. elf, zestien, etc. (1, 2, 3 ... 11, 16). The factors of ten aswell; twintig, dertig, veertig (20, 30, 40). And also the rounded numbers above that; honderd, driehonderd, duizend, vijftienhonderd, achttienhonderd, etc. (100, 300, 1000, 1500, 1800).
Additionally for 1e, 2e and 3e, etc. also in use is/was: 1ste, 2de, 3de, etc.
The difference between 'een' and 'één', I think, is usually made if it stands next to the word 'een' in a text, which means 'a' or 'an' (a house) in Dutch. Or if you want to emphasize that it is really only one of something.

Other Dutch speakers, please correct any mistakes I may have made. :)

It would be helpful to identify the cases where the ordinal forms of the language may cause ambiguities when old-style numerals are used.

In Greek, I think there is a way of forming ordinals by attaching an omicron at the end of the arabic numerals. Then you don't want to have a font where it is difficult to distinguish the zero and a small omicron. In Georgia (the font you are probably reading this in), for example, 19ο (nineteenth) and 190 (one hundred ninety) are not the easiest to tell apart, though to its credit it does its best to make the troublesome letters distinguishable on screen.

Thanks all, so far. This is fascinating. English seems to have the most "exceptions" with "st" "nd" and "rd" abbreviations.

Maxim -- Declining ordinals? Yikes. Are the abbreviations ever superscripted, as sometimes occurs in English?

As Matthew (guifa) notes, the ordinal abbreviations in Spanish are typically superscripted. The English ones, only sometimes (much debate in some circles). Presumably, the languages that use only periods wouldn't superscript them.

I believe the French are fond of superscripting in certain circumstances. True? What about Dutch ordinal abbreviations?

Scalfin, Hebrew is not a particular focus of mine personally, but I'm sure it would add to this compilation if you wanted to explore this.

-- K.

Hey Jelmar,

you haven’t included my favourite Dutch word (apart from ‘sinaasappelsap’, that is): ‘achtentachtigste’ [the 88th]. ;°) Try pronouncing that, as a Non-Dutchman!

F

I adore the word sinaasappelsap, as well. And now achtentachtigste, too. :-)

It’s interesting that the rule for abbreviation is so simple. I would think this could lead to some confusion. But perhaps Danes don’t abbreviate ordinals that often.

Actually, no, it very rarely leads to any problems or misunderstandings. The only ambiguity is when the ordinal is last in a sentence, since it is then indistinguishable from a cardinal; but even in such situations, context is usually enough to differentiate them quite easily.

Ordinals of more complex numbers are usually avoided in speech, though. Instead of saying seks hundrede og toogtres tusind fire hundrede og seksogfirsindstyvende (662,486th), most people would paraphrase and say nummer seks hundrede og toogtres tusind fire hundrede og seksogfirs (number 662,486) instead. This also goes (even more so, actually) for numbers that end in 100, 1,000, 1,000,000, 1,000,000,000, &c., since those have such awkward ordinal forms.

Maxim — Declining ordinals? Yikes.

You should try Finnish, then. No genders, thankfully, but all numerals (cardinals and ordinals alike) are declined in all 15 or so cases both in the singular and the plural (yes, numerals in Finnish have both singular and plural forms! I’ll just ignore them here, though). And for complex numbers, all constituents of the numeral have to be declined, so ‘three hundred and fifty-seventh’ in the allative case, for example, would literally be ‘to-third to-hundredth to-fiftieth to-seventh’. And they’re written as one big long word, too (but that goes for all numerals).

A simple example, since the entire system is too complex to even hint at here, using 357th to go by:

Formula (see below for abbreviations and explanation):
[Num]. [C nom] [C part] [C gen] [C ess] // [O nom] [O part] [O gen] [O ess]

3. kolme kolmea kolmen kolmena // kolmas kolmatta kolmannen kolmantena
100. sata sataa sadan satana // sadas sadatta sadannen sadantena
5. viisi viittä viiden viitena // viides viidettä viidennen viidentenä
10. kymmenen kymmentä kymmenen kymmenenä // kymmenes kymmenettä kymmenennen kymmenentenä
7. seitsemän seitsemää seitsemän seitsemänä // seitsemäs seitsemättä seitsemännen seitsemäntenä

Putting those together to form 357 is quite simple:

C nom: kolmesataaviisikymmentäseitsemän
C part: kolmeasataaviittäkymmentäseitsemää
C gen: kolmensadanviidenkymmenenseitsemän
C ess: kolmenasatanaviitenäkymmenenäseitsemänä

O nom: kolmassadasviideskymmenesseitsemäs
O part: kolmattasadattaviidettäkymmenettäseitsemätta
O gen: kolmannensadannenviidennenkymmenennenseitsemännen *
O ess: kolmantenasadantenaviidentenäkymmenentenäseitsemäntenä

Abbreviations:
C = cardinal
O = ordinal
nom = nominative
part = partitive
gen = genitive
ess = essive

The singular stem (and the plural stem in many words, but thankfully not in the ordinals) is used to form all cases except the nominative and partitive, which often have stems of their own. It exists in two grades, strong and weak. The genitive singular always gives the weak grade, while the essive singular always gives the strong grade (this isn’t true for all declensions, but it’s true for the numerals). It is thus necessary to know both the nominative and partitive for both the singular and plural, as well as the genitive and the essive singular, for any word in order to be able to decline it correctly.

As numerals, the Finnish do the same as the Danes and Germans: add a simple period after the last number. To be honest, I’m not sure how they handle the cases. With cardinals, they simply add the (vowelless) case ending to the number, separated (as is common in Finnish) by a colon, so ‘in 357 houses’ would be written as 357:ssä talossa (or kolmessasadassaviidessäkymmenessäseitsemässä talossa). But it doesn’t seem likely that ‘in the 357th house’ would be written as 357.:ssä talossa (or kolmannessasadannessaviidennessäkymmenennessäseitsemännessä talossa).

* That word took about half an hour to type. Cripes.

All right, moving on to something slightly easier: Chinese.

In Chinese, to make a cardinal into an ordinal, you simply prefix it with the character 第 (dì). It’s always written out as a character, and it’s spoken where and as it’s written, whether the numeral itself is written as Chinese numeric characters or as Arabic numerals.

Thus:

1 or 一 (yī) => 第1 or 第一 (dì yī)
36 or 三十六 (sānshíliù) => 第36 or 第三十六 (dì sānshíliù)

A lot simpler. ;-)

I’d give you some info on Irish ordinals, as well, but I can barely wrap my head around half of those myself, so I should do some more research before trying my hand at it. Irish numerals are notoriously complex and tricky.

> It’s interesting that the rule for abbreviation is so simple

Turkish is the same (e.g ordinal that used in titles '7.' = yedinci, seventh; streets -- '12. Sokak' = 12th Street)

Ordinals in Polish

Ordinals declines like other adjectives in Polish, it means the ending of the word changes depending on
1) a number (single, plurar)
2) a gender (masculine/feminine/neuter gender)
3) a declension (8 forms: Rzeczownik, Mianownik, Dopełniacz, Celownik, Biernik, Narzędnik, Miejscownik, Wołacz)

It gives such a big amount of possible combinations I'd write examples only for single/masculine/rzeczownik forms.

Here in Poland we got three ways of writing orinals: full name; only ending after the numeral and dash; dot after the numeral.
However the best solution is to write whole words, other possibilities apply if you are short of space for a whole words, they can be long. It means you'd rather don't meet a 1-szy or 1. in a book, but you may spot it in mags, letters, etc. Grammar rules allow all three forms. So here it goes:

num:[basic form/declined form [1],[2],[3]]
1: jeden/(1st) pierwszy,1-szy,1.
2: dwa/(2nd) drugi, 2-gi, 2.
3: trzy/(...)trzeci, 3-ci, 3.
4: cztery/czwarty, 4-ty, 4.
5: pięć/piąty, 5-ty, 5.
6: sześć/szósty, 6-ty, 6.
7: siedem/siódmy, 7-my, 7.
8: osiem/ósmy, 8-my, 8.
9: dziewięć/dziewiąty, 9-ty, 9.
10: dzesięć/dziesiąty, 10-ty, 10.

11: jedenaście/jedenasty, 11-ty, 11.
12: dwanaście/dwunasty, 12-ty, 12.
13: trzynaście/trzynasty, 13-ty 13.
14: czternaście/czternasty, ...
15: piętnaście/piętnasty
16: szesnaście/szesnasty
17: siedemnaście/siedemnasty
18: osiemnaśice/osiemnasty
19: dziewiętnaście/dziewiętnasty

20: dwadzieścia/dwudziesty,20-ty, 20.
21-29: dwadzieścia + [basic form 1 to 9]/dwudziesty + [declinated form 1 to 9]

30: trzydzieści/trzydziesty
31-39: trzydzieści + [basic form 1 to 9]/trzydziesty + [declinated form 1 to 9]

40: czterdzieści/czterdziesty
41-49: czterdzieści + [basic form 1 to 9]/czterdziesty + [declinated form 1 to 9]

50: pięćdziesiąt/pięćdziesiąty
51-59: pięćdziesiąt + [basic form 1 to 9]/pięćdziesiąty + [declinated form 1 to 9]

60: sześćdziesiąt/sześćdziesiąty
61-69: sześćdziesiąt + [basic form 1 to 9]/sześćdziesiąty + [declinated form 1 to 9]

70: siedemdziesiąt/siedemdziesiąty
71-79: siedemdziesiąt + [basic form 1 to 9]/siedemdziesiąty + [declinated form 1 to 9]

80: osiemdziesiąt/osiemdziesiąty
81-89: osiemdziesiąt + [basic form 1 to 9]/osiemdziesiąty + [declinated form 1 to 9]

90: dziewięćdziesiąt/dziewięćdziesiąty
91-99: dziewięćdziesiąt + [basic form 1 to 9]/dziewięćdziesiąty + [declinated form 1 to 9]

100: sto/setny, 100-ny, 100.
101: sto jeden/ sto pierwszy, 101-szy, 101.

Counting from 100 you don't decline hundreds only singles and decimals UNLESS it is the EXACT number, so you decline eg. 100th word (setne słowo).

Cheers,
Antoni

Ah yes, I once took a Finnish class. I had to stop because of tongue and cheek cramps, not to mention lack of oxygen. Kissa on sovala or something about a cat and a sofa is all I remember.

Haha! Yes, Kissa on sohvalla.

My favourite words in Finnish have to be jalkapaloilijoille (‘to the football/soccer players’) and the somewhat artificial but highly amusing riiuuyöaieuutiset (‘news about intentions for a dating night’), which begins with no less than eleven vowels. :-)

Oh well. Nothing new in Portuguese, for it is very much similar to Spanish:
Primeir|o|a|os|as = 1º, 1ª, 1ºs, 1ªs
...
and similarly until you get to
centésim|o|a|os|as = 100º, 100ª, 100ºs, 100ªs
because then it goes like this:
centésim|o|a|os|as primeir|o|a|os|as = 101º, 101ª, 101ºs, 101ªs
...
until the mind boggles ;-)
or you reach aleph-null.

Ronald Kyrmse
ronald.info.ms

So, today is still the most seldom, still most useful day of the year, so let's give an example of Hungarian.

In Hungarian you write ordinal numbers as in Danish, German, Polish (and several other European traditions): with a period following the Arabic numeral: 1., 2., 3., etc. so that's simple. The first ten are spelled out (cardinal ~ ordinal)

1 egy ~ első 1.
2 kettő ~ második 2.
3 három ~ harmadik etc.
négy ~ negyedik
öt ~ ötödik
hat ~ hatodik
hét ~ hetedik
nyolc ~ nyolcadik
tíz ~ tizedik

Basically the rule is (except for the first three; they always seem to be special cf. one ~ first, two ~ second... uno ~ primero) to add -Vdik where V stands for a vowel harmonising with the vowels of the word (vowel harmony is a basic feature of Hungarian) [and some minor vowel length changes in the stem might take place].

Purely technically, Hungarian ordinals can also take one of the 20+ cases and be in singular as well as plural (cf. Finnish), but that's really not what you are looking for* ;-). In most such cases, you'd write it out. (Again, technically, one would write it with number, dot, dash, ending: 1.-ben (first[inessive case])). When writing dates (which are spoken with the ordinal in possessive form (Hungarian marks linguistically what is possessed as opposed to what is possessing in most IE languages (genetive case)) they have a slightly different writing rule, as in not writing/dropping the dot, just writing "május 1-én" e.g. (May 1-[possessive marker][superessive case] meaning "on the 1st of May")

_____
* You really don't want to go into that, as there are more than just the ca. 50 combinations of cases and singular/plurar, as Hungarian marks several other things, as. e.g. possession and the person and number of possessor on the item as well, and additionally, whether the possession is in singular or plurar... so that adds up. -- However as opposed to fully inflecting languages, it's comparably easy, you can imagine it as a building kit, just adding modules (so you don't have to remember 6x2x25x2x... endings, just 6+2+25+2+... endings which you can combine then in 6x2x25x... ways... it's soo simple!

This cannot comete with the "somewhat constructed" Finnish riiuuyöaieuutiset (‘news about intentions for a dating night’), but one the (acutally atypical) Hungarian vowel-word is: "fiaiéi" which means "his sons' somethings (that were referred to previously)". And merely formally, you could go on saying fiaiéiéiéiéiéiéi... forever, if those somethings together again possess several somethings possessing together several somethings... but THAT's constructed (and noone could track it, right?), while the first could happen in a real conversation...

Thomas --

Well, I happen to be working on Whitman superscript/ordinal letters right now, so that piqued my curiosity. But in addition to being a typophile, I also have a fascination with linguistics, so it's a little bit of both fueling my interest.

I don't really have a grand scheme in mind, certainly nothing as ambitious as a contextual {ordn} feature. I doubt that would be of any practical benefit.

But I thought the discussion might be of general interest to the type design community. And I am interested to see if there is any use out there that is unfamiliar and eye-opening.

I'm aware that, strictly speaking, the {ordn} feature really refers only to superscripted abbreviations -- which is a purely typographic convention, and really only a limited one, with seemingly no application to many languages.

So, really there are three semantic layers. There's the ordinal form itself (which is an essential semantic element); then there's the convention for abbreviation; and then there's potentially a typographic tradition of superscripting the abbreviation. I'm always interested in understanding the linguistic elements behind the typographic traditions and conventions.

But, practically speaking, I was pretty much planning to put a complete superscripted l.c. alphabet* in the {ordn} alignment -- different from the {sups} alignment. Just of good measure. I suppose to some extent, I am curious about the minimum letters required to handle the most common uses of superscripted ordinal abbreviations. I figure I should try to be aware of which of these would actually qualify as true "ordinals."

*(unaccented -- unless I can discover a convincing argument for an é or è in French.)

-- K.

Oisín, sorry, but I’m nitpicking here... It’s jalkapalloilijoille.

Whoops, that was a typo, hadn’t noticed that. Pallo has two l’s, of course. Unless one could make a case for jalkapaloilijoille meaning something like ‘to the people who cause foot fires’... perhaps not so likely. :-P

I’m very impressed with your grammatical knowledge of Finnish. As a native speaker I couldn’t have presented it better. I suppose I tend not to think of the structure so much, just get on with speaking/writing.

And as someone who’s taken one year of Finnish (and has, unfortunately, been unable to take the second year this year), I do the exact opposite: I still have to carefully think my way through declensions and conjugations before being able to say very much. Such a very different language to wrap one’s brain around.

"have you heard M.A. Numminen "

Mili, my daughter played some of his music for me on a road trip. My daughter has a way of finding things beyond the typoical on American radio.

ChrisL

Kent wrote:

I was pretty much planning to put a complete superscripted l.c. alphabet* in the {ordn} alignment — different from the {sups} alignment.

That's pretty much what we're doing at Adobe in new fonts, adding the egrave.

Cheers,

T

adding the egrave.

...and alternate g's and a's :-)

I guess Nick is referring to having both single- and double-storey alternatives for those letters, as Thomas has included them in his Hypatia Sans (see this PDF).

Very interesting reading all of you. Could I ask your thoughts on the use of capital letters for ordinals above 20 written in full in English?
e.g. Twenty-first or Twenty-First etc?
Many thanks!

I’d have to agree, unless it’s for the title of a song, book, movie, etc.; then I’d say ‘Twenty-First’. Not sure if that’s just me or an actual rule, though.

Or is this just type-designer vanity...

No, I meant that if a typeface has alternate versions of "a" and/or "g" in its lower case, it should also have alternates in its superscripts.