The German number system has the peculiarity that the ones are read out before the tens. For example:

634542 = Sechshundertvierunddreißigtausendfünfhundertvzweiundvierzig
       = "six hundred four-and-thirty thousand five hundred two-and-fourty"

All other languages I am familiar with read out numbers by starting at the highest digit (with respect to the underlying number system) and then proceding linearly to the lowest one. I think it's the other way round in Sanskrit (i. e. from least to most significant digit) but I'm not aware of any language where the digits get jumbled up in a fashion similar to German.

Are there examples of other languages where numbers are not necessarily read out in a linear fashion?

(Of course, a lot of languages use composite words for certain (especially two-digit) numbers like "thirteen" or "duodecim", where the order of the composits defies the otherwise strict order, but I'm looking for languages where the rules themselves of forming arbitrary numbers from the fixed expressions proceed nonlinearly.)

  • 1
    Check this question at German.SE and its excellent answers. Commented May 16, 2016 at 1:47
  • In Czech, we use both the straight as well as the German reverted names for numbers like 21. In particular, the word "jednadvacet" (einundzwanzig in this order) is a popular word that is said to "last exactly one second", a timer of a sort. That's why it may be more frequent than "dvacet jedna", i.e. "twenty one". It's hard to figure out which is more frequent but the German order is usually cooler and used especially for years in the past. For example, the 1968 occupation took place in "osmašedesátém roce" - "in dem achtundsechzigen Jahr" or whatever is the right German form. Commented May 16, 2016 at 11:00
  • Forms like "five and twenty" were common in English up to the 19th Century. I used to know somebody only a few years older than myself (he'd be in his sixties now) who routinely said "five and twenty" when telling the time, though I don't think he used it in other contexts.
    – Colin Fine
    Commented May 16, 2016 at 23:11
  • Numbers in Afrikaans are also the same as Dutch & German.
    – user13264
    Commented May 24, 2016 at 11:19
  • The reading of numbers is a property of locale/variant, not language per se. fr-ch reads numbers differently than fr-fr, for example. English, Arabic etc have varied over time and place on this front. Commented May 25, 2016 at 13:21

4 Answers 4


Arabic: 1984 could be read as "one thousand and nine hundred and four and eighty" (the modern reading) or "four and eighty and nine hundred and one thousand" (traditional). I have heard the latter on the radio.

and you could probably get away with "four and eighty and one thousand and nine hundred".

P.S. On a related note: it is often claimed, even by "experts" (like the Unicode folks) that Arabic numbers are written left to right, and since the language is written RTL, Arabic (and other RTL languages) are "bidirectional". This is total hogwash. You can read and read numerals in whatever order you like. The semantic rule is simply "base 10, least significant digit first", in contrast to (most) LTR languages, which go most significant digit first. The myth of bidirectionality is 100% fabricated by the tech industry.

  • There are also email addresses, ISBNs, brand names etc that are LTR even in Hebrew and Arabic. Commented May 15 at 4:00

In Irish numbers are famously messy, putting the three counting systems aside to focus on the system for non-people/inanimate objects to address the question, numbers can be read in two ways, base ten and base twenty.

If you are counting things you have to name the thing you are counting, or at least use a placeholder like "ceann" (lit. Head, means thing in this case). In the base ten example twenty three things would be "trí chinn is fiche" "Three things and twenty" and fifty three things would be "trí chinn is caoga" "Three things and fifty". Above 100 the counting switches to a MSD to LSD system as in English, "dhá céad fiche is a trí chinn" "two hundred twenty and three things"

There is also the base twenty system. In this system, which is going out of fashion quite rapidly twenty three is still "trí chinn is fiche" but fifty three is "trí chinn déag is dá fichead" or "three things teen and two twenty".


Not surprisingly, numbers in Dutch are organized the same way as in German. For example, 634542 written out becomes

zeshonderd vierendertig duizend vijfhonderd tweeënveertig

Which translates to:

six-hundred four-and-thirty thousand five-hundred two-and-forty

  • 3
    can you please translate that (literally) into English?
    – mobileink
    Commented May 22, 2016 at 19:27
  • 1
    @mobileink: looks like "six-hundred four-and-thirty thousand five-hundred two-and-forty." It's not really all that mixed up--no more than English "sixteen" for "16" (which would more "logically" be "ten-six"). Commented Aug 22, 2016 at 19:17

Please be aware that this peculiarity in German (and other languages mentioned here) affects only the combinations of small order numbers, i.e. tens and units and their iterations when these are reused again due to the lack of higher order numbers (tens of thousands, tens of millions - if we had a separate word for this akin to Greek "khillia", most likely this would not happen).

As your question mentions, there are lots of compound words for smaller numerals, particularly the ones between 10-20 - these operate on exactly the same principle but may be further on the univerbisation scale. "fourteen" also gives the order in reverse to the typical one, the I believe applies in German and definitely in Slavic languages (Czech "třináct" comes from "tři-na-deset" or "three on ten", not so dissimilar from "three-and-twenty").

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.