# Is 千{せん} a “current” number construct?

The reason why I ask this question is because of the pattern I've seen with Japanese numbers. Once you have to repeat a number to describe itself, you get a new unit. For example, 十十 =　百,　百百 =　万, 万万 =　億. So it seems that 千 is a number that had some Western influence in its creation. Implying that it was a number that created in more recent times (ie. the last 200 to 300 years). Am I in the right direction with this pattern I'm seeing? Or is this just coincidence?

-
Interesting idea (no idea if it's grounded in truth). From a statistical standpoint, I would point out that, considering you conveniently remove the one number that does not fit the pattern, the "pattern" you are seeing is not very convincing, when compared to Ockham's preferred version of "there's a number for each multiple of 10, until it gets too big to really be a concern". – Dave Nov 14 '12 at 6:39
The reading セン for 千 is listed as a 呉音 reading. (I also see it listed as both 呉音 and 漢音 in another dictionary, but 呉音 is earlier and therefore more relevant here, I think.) According to Wikipedia's article on go-on, this means it was borrowed into Japanese during the 5th or 6th centuries. That makes it rather unlikely that the modern word せん was created in the last 200 to 300 years, I believe. – snailboat Nov 14 '12 at 13:06
@dotnetN00b Note that Old Japanese had ち to express 1000 even before せん replaced it in common use. (ち is still around, of course.) – snailboat Nov 14 '12 at 19:54
@dotnetN00b: that's where your non-statistician's mind fails you... ;-) In hypothesis-testing terms, the null hypothesis (what Ockham tends to prefer) is that there is no pattern. You are trying to invalidate that hypothesis by trying to show a pattern. However the amount of points fitting your pattern (3) is pretty low given the amount of outlier (1) created by your hypothesis, not to mention the fact that the null hypothesis offers the same fit (3 out of 4). Jokes aside 億 sounds a better candidate than 千 for being a recent creation... – Dave Nov 15 '12 at 8:01

The Japanese number system (which I believe is derived from the Chinese one) is pretty similar to the western one (just to call it something), except it breaks at every 4 digits instead of every 3 digits.

Japanese:

0-9999 兆 0-9999 億 0-9999 万 0-9999

English:

0-999 billion 0-999 million 0-999 thousand 0-999

(assuming you're from a part of the English speaking world which doesn't use milliards or "thousand million"s).

The confusing thing about the Japanese system, though, is that when written with digits, it still puts commas at every third digit. This is most likely from western influence.

Slightly off topic, but...

Your idea is interesting, since it would be theoretically possible to have a number system where digits break in a "binary" fashion:

0-99999999 億 0-9999 万 0-99 百 0-9 十 0-9

I don't know if any language breaks digits in this way, though.

-
Actually in ancient China, that kind of number system was theoretically considered. Wikipedia says: 漢代あたりから、上数（じょうすう）が記載され始めた。数詞が表す位の2乗が次の数詞となる。万万が億（10^8）であるのは今日と同じであるが、次は億億が兆（10^1‌​6）、兆兆が京（10^32）となる。実際に使われたことはないようであり、数学書では用いられていない。 – Gradius Nov 20 '12 at 15:59
@Gradius, interesting! Nice find. – dainichi Nov 20 '12 at 23:43

your theory breaks down at 億。　億億　!= 兆　 （兆　＝　万億） （億億　＝　京）

It would to be necessary to have every increment of power of 10 up until 10 000 under the Japanese system. Things would be akward without the 1000 unit as well.

I do think it looks weird though when I see a number like 56000 written as ５万６千

-