Page **4** of **5**

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Wed Sep 02, 2015 6:04 pm**

by **janKipo**

But it doesn't solve the problem of number names: telephone numbers, SSNs, PINs, addresses and so on.

On the other hand, the tricimal system is really hard for people to get right, for some reason, I suppose mainly because it does not connect with familiar systems regularly, e.g. x tu is not regularly even nor is x ala nor is x wan regularly odd. Nor does it help with large numbers by providing convenient round numbers (well, it does but they don't seem so significant).

Nothing but decimals will ultimately solve this problem, if it is a problem (maybe tp just shouldn't be used in the modern world -- which is sorta the idea).

(9/2/15 See a decimal scheme -- or several -- on Facebook Toki Pona)

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Thu Sep 03, 2015 3:44 am**

by **Cvi**

It is true that a decimal (or hexadecimal, octal or whatever) is better for large numbers, and for telephone numbers, SSNs, PINs, addresses we would need decimal numbers anyway. I shall try to find in facebook the proposals.

I proposed the trinary system as a solution for things like "I am 77 years old". I know that proposing changes while learning the language is rather ridiculous (as an esperantist I feel that very often). However, I really think that the existing system is very limiting and absurd.

I thought that within the existing words, and I understand that it is a central feature of toka pona - this could be the best compromise.

As for it not being a familiar system - much about toka pona is not familiar, but it does have its own logic. This could fit in very nicely. The round numbers - we could get used to them and they can be as significant as in any other system. Even and odd numbers are not so basic, certainly compared to the use of mute or mute mute for anything

If there is an acceptable decimal system - I am all for it.

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Thu Sep 03, 2015 10:30 am**

by **janKipo**

I personally am a fan of duodecimal numbers, which have the advantages of hex but deal well with thirds also, a practical advantage. Both of these (and vingesimal, tregesimal and sexagesimal, too) contain the decimal and so solve naming problem.

For the naming problem (and also the counting one, since, once the names are in place, they will be used to count), there are two paths. One creates arbitrary new words for the digits (the nine consonants expanded by vowels that don't duplicate existing words, say, as in the current suggestion. There have been dozens of these proposals, by the way, though I don't think they have all been collected in one place yet.) The other follows the pattern of the added 'luka' (and the ill-advised 'mute' and 'ali') by interpreting some existing words as number words, either arbitrarily (all the biota words, say, more or less alphabetically) or on the basis of some perceived (or imagined) relation between referent and some number ('pipi' and 6 because of the legs, say). In both of these the status of 'wan' and 'tu' is more or less left up for grabs. Most, once they get going, go on to suggest a variety of round numbers for E1, E2, E3n and then move to the fractional notions as well (so clearly not about naming any more).

It took me three minutes to work out in my head that 77/10 was 2212/3; not encouraging.

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Mon Sep 07, 2015 2:09 am**

by **Cvi**

I personally am a fan of duodecimal numbers, which have the advantages of hex but deal well with thirds also, a practical advantage. Both of these (and vingesimal, tregesimal and sexagesimal, too) contain the decimal and so solve naming problem.

I don't understand what you mean by "both of these". Duodecimal and hex don't solve the decimal naming problem, only the vingesimal, tregesimal and sexagesimal systems do.

There is another posibility - using the decimal system to write the numbers, but a kind of biquinary to name them in speach.

I met a kind of biquinary as the internal number representation in the IBM 650 computer, the first one I programmed in, but we didn't have access to it.

The advantage - we need to add just two new words, for 3 and 4, and the written numbers enable to use the

telephone numbers, SSNs, PINs, addresses and so on.

. It could be easy to learn and use, and reasonably short. Like "lukatu lukatu" for 77, "tu luka" for 25 and "wan ala ala" for 100

So I really would prefer this system - and go back to ternary only if there is an absolute tabu against new words, even two of them.

Thanks for your comments

Cvi

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Mon Sep 07, 2015 8:54 am**

by **janKipo**

Of course both base 12 and base 16 number systems solve the naming problem; the first ten digits, of 12 or 15, do fine (and even usually have the same names, though the rest vary wildly).

I never thought of biquinary systems, in spite of abacuses and chisanbop. It is an interesting compromise. tp actually had a 3 and 4 at some point ('tuli' and 'po' -- 'tuli' should probably be 'tuwi' for clarity -- or even just 'si' for uniformity).

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Mon Sep 07, 2015 11:51 am**

by **Cvi**

OK - I only meant that with biquinary we can just leave the decimal cifers as they are, and they have the accepted value.

If I read 100 as wan ala ala and use hexadecimal - I won't be quite sure if I meant 100 or 256.

As you see in my example, I would propose a slightly modified biquinary system. The first digit in each pair would be nothing or luka, not ala or wan.

People who don't like math could just learn that 6 is lukawan and 7 lukatu, etc., not specially dificult.

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Mon Sep 07, 2015 5:55 pm**

by **janKipo**

But how are base 20 and base 60 saved from the odd value problem? Presumably number names are contextually different from enumerations.

The modification makes sense, since the use of 'wan wan' etc. is ambiguous. I assume that 5 is 'luka ala' else we get that problem there as well (maybe abbreviated to 'lukala').

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Tue Sep 08, 2015 12:41 am**

by **Cvi**

About base 20 and 60 you are right - the problem remains.

About 5 - it would have to be either luka (in spite of the bequinary origin of the proposed system) or lukala. I would interpret luka ala as 50

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Tue Sep 08, 2015 9:19 am**

by **janKipo**

If 5 is 'luka' then we get the problem with 'luka ala' repeated for all of 6/51. 7/52, etc. I think that the drop bit has to be just 'lu' so 'wan, tu, si, po, lu ala, lu wan, lu tu, lu si, lu po, wan ala, ...'

### Re: nasin nanpa kepeken toki pona/How to count in toki pona

Posted: **Tue Sep 08, 2015 3:24 pm**

by **Cvi**

I like your proposal. I only think, that in real speach it would naturally contract to luala, luwan, lutu for example 1937 = "wan lupo tu lutu" actually much and nicer than "one thousand nine hundred thirty seven"