[seqfan] How to speak long numbers -- now I'm really curious (-:

Robert Munafo mrob27 at gmail.com
Mon Sep 26 23:10:29 CEST 2011

I have found some folks over on math-fun who have encountered "and" when
spelling out (or I guess speaking) names of numbers over 1000. Also a lot of
disagreement to my claim that commas should be written in the same places
where the comma would appear in the digits version.

But no-one is giving specific rules (yet :-) and I've spent a while looking
but haven't found any rules myself. (Just isolated examples without

So here are some questions for our colleagues who speak the word "and" as
part of a number name -- and regarding the comma, consider how you would
write it out in words:

  * Do you ever or always use "and" after a power of 1000 if something else
follows? ("one million and thirty thousand")

  * Do you ever or always use "and" in places where there would be a comma
in the digits? ("five thousand three hundred forty seven" but "twelve
thousand and three hundred forty seven")

  * Or instead of "and" would you put a comma in the words? ("twelve
thousand, three hundred forty seven")

  * Do you use "and" wherever there are 0s between digits? ("one hundred and

  * Do you do two or more of the above? ("five thousand and one hundred and
one" or maybe "five thousand, one hundred and one")

  * Do you put "and" anywhere else? ("three hundred and forty seven")

  * Does it matter how many syllables the part after "and" has? ("one
thousand and twelve" but "one thousand seventy-seven")

  * Other rules?

  * If you consider it proper usage in only part of the English speaking
world (like the UK, Australia, Canada, U.S. etc.) let me know that too.

It would be nice to formalize this, because it is used a lot, and almost
everything else in OEIS never uses "and", and is mute on commas.

I found 7 OEIS sequences where it is clear if they use "and" or not, there
might be more...

A005589 includes a PARI program which gives A005589(1000)=11
and A005589(1001)=14.
A052360 has the same program.
A052363 agrees with A005589 and A052360 because it includes 1103 but not
A058230 specifically claims "not to put the word 'and' in the names of
A092320 agrees because it contains 1005.
A134629 uses Noll's program. [1]

*but* A126259 uses "and" in its spelling of 108

- Robert Munafo

[1] http://isthe.com/cgi-bin/number.cgi

On Sun, Sep 25, 2011 at 13:33, Victor Miller <victorsmiller at gmail.com>wrote:

> On the program cartalk last week, there was a puzzle of the following form:
>    A list of numbers was given, and one was asked what they had in common
> (I give the actual puzzler at the end).  The answer was that each of
> these numbers was divisible by the number of letters (excluding
> spaces) in the standard spelling out of the number in words.  This got
> me to thinking of the following modification:
> [...] (there followed speculation about A092320 vs. A126259)
  Robert Munafo  --  mrob.com
  Follow me at: gplus.to/mrob - fb.com/mrob27 - twitter.com/mrob_27 -
mrob27.wordpress.com - youtube.com/user/mrob143 - rilybot.blogspot.com

More information about the SeqFan mailing list