Base-4 fractions in Telugu | Hacker News Reader

Base-4 fractions in Telugu

blog.plover.com

179 pointspavel_lishin39 comments

> the digits for 3 have either three horizonal strokes ౾ or three vertical strokes ౻, and the others similarly. I have an idea that the alternating vertical-horizontal system might have served as an error-detection mechanism

The Chinese Suzhou numerals use alternating 〡〢〣 and 一二三 but for a different reason:

    "21" is written as "〢一" instead of "〢〡" which can be confused with "3" (〣)

from https://en.wikipedia.org/wiki/Suzhou_numerals

Telugu's a spectacular language - it's ancient, it's spoken by tens of millions of people, it has its own cinema industry and an enormous literary corpus reaching back to before English even existed... and yet almost nobody outside the subcontinent even knows it exists.

Same with Tamil, Kannada, Bengali, Malayalam etc.,

Tamil is a classical language like greek, hebrew, chinese etc. Its much older than telugu and the other Indian languages. Malayalam and kanada is derived from tamil

Malayalam diverged from Tamil. But, Kannada is not a derivative of tamil.

Of course Kannada evolved from ancient Tamil. If you know both the languages, you can easily find root words for most current day Kannada words in ancient Tamil sangam poetry (like ooru=>oor referring to town, kaaveri=>kaaviri, etc.) https://en.wikipedia.org/wiki/Dravidian_languages

Kannada, Telugu, Malayalam all have higher degrees of impact due to Sanskrit but they are all dravidian languages, derived from ancient Tamil.

Your Wikipedia link disagrees that all of the Dravidian languages are derived from ancient Tamil. It describes them as common descendents of Proto-Dravidian. It cites the attested literature of Tamil as being 2,000 years old.

Calling Tamil much older than Telugu is probably not appropriate. They both developed during the same time.

The Tamil language currently used is not same as the one when the other Dravidian languages separated, So calling other languages 'derived' from the Tamil is not correct. All Dravidian languages have same root, so calling it 'Dravidian Language' and Tamil, Kannada, Malayalam & Telugu etc. derived from that language is more accurate and avoids the confusion

Modern-day Greek, Hebrew, and Chinese are all very different from their "classical" versions.

> an enormous literary corpus reaching back to before English even existed

Certainly true of Telegu, Tamil, Kannada, and Malayalam. But because Bengali is, like English, an Indo-European language, not true of it. Also, not sure what you mean by "etc".

Sanskrit is also Indo-European, and definitely had a large literary corpus before English became separate from Old Norse.

Let the regional language wars begin.

Aren't most wars about regional languages, when you get down to the nuts and bolts?

English is less than 1,000 years old in its identifiably modern form.

Off-topic: The plover.com domain caught my eye. He is the author of 'Higher Order Perl'[1][2], which is to this date, one of my most favorite programming language books I've read and my first (and imho, best) introduction to functional programming. The book is free to download!

[1] http://hop.perl.plover.com/ [2] https://en.wikipedia.org/wiki/Higher-Order_Perl

Thank you both for your kind words!

HOP is awesome!

As someone who speaks Telugu as their first language, I am ashamed to admit I don't know how to write numbers in Telugu, let alone the fractions. Though I had learnt how to read and write Telugu, even in school, one is never taught how to write numerals in Telugu. But I am glad people are still working on it. It's an amazing language where most of the words end in vowles and I feel it doesn't get enough interest from the Telugu community.

As a native speaker, taking a class is mind-blowing. Sadly I had to move out of town after 3 classes, but I had no idea how sophisticated and well structured the formal language was. The colloquial form we speak at home doesn't do it any justice.

>The colloquial form we speak at home doesn't do it any justice.

In a certain sense the spoken form of a language is its true form. The written form are rules that get frozen at an artificially defined moment in time, but the spoken language keeps evolving. Another way to look at this is that writing systems are a relatively new phenomena for humans. In terms of time scale, if you consider that spoken language has been around for 24 hours, writing has only existed since about 11ish pm.

The written form is also heavily sanskritized. The spoken form of Telugu is closer to other south Indian Languages.

The same can be said of Malayalam. There's a number system, but nobody uses it anymore.

" even in school, one is never taught how to write numerals in Telugu"

Would if have had any practical uses? At the end of the day, everyone uses the arabic symbols.

There are several changes I’d make to the number system if I could. Chief among them:

(1) Try to mostly use a positional system in speech as well as writing. This saves a lot of time because "three five nine" is a lot easier to say than "three hundred fifty-nine", removes a huge amount of confusing irregularity, and overall makes life much nicer for children just learning. While we’re at it, scrap percentages.

(2) Allow signed digits, and give the negative versions their own unambiguous names. It’s amazingly convenient to be able to have a way to directly express e.g. 200 – 3 without needing to call it 197. Students should learn to convert between an all-positive-digit form of a number and a round-via-truncation form of the number, and generally use the latter in most practical circumstances.

For more on this point see http://ethw.org/Ancient_Computers

(3) – a pipe dream – general use of base twelve, and in particular use of binary logarithms (“doublings”, “bits”, “octaves”) written using duodecimal fractions.

Using alternate bases can be nice, and binary divisions are a lot more useful than division by 5 – especially for e.g. measuring circular arcs where binary divisions are much easier to compute because they only require square roots https://en.wikipedia.org/wiki/Binary_scaling#Binary_angles – but these Telugu symbols are too complicated to write for a base 4 system to use for general numeration. If we want divisions by two, hexadecimal (with appropriately redesigned glyphs) is better.

"Three five nine" is easier to say than "three hundred and fifty-nine", but the second option lets me have an order-of-magnitude number from the start. Take "one seven nine five four two zero eight three". Are you keeping track of how many digits you're hearing? "one hundred seventy nine million, <...don't need to pay attention to the rest...>".

But you did say "mostly". Are you thinking "one seven nine million, five four two thousands, zero eight three" or something like that?

If it were up to me, all the digit names would be one syllable, and the order of magnitude would get its own syllable (or a few, for very large/small numbers) up front. Basically a verbalization of scientific notation, but with the exponent leading.

This could be safely dropped for numbers of less than a few digits. "Two six" or "eight one four" is not going to be ambiguous/confusing.

So, more or less: Japanese.

https://en.wikipedia.org/wiki/Japanese_numerals

No, not at all like Japanese. You’d have a single indication of the order of magnitude of the largest digit, up front, followed by a string of positional digits, where you just say the syllable for 0–9 for each place, including zeros.

The Japanese version is basically like other languages (including English), just with slightly fewer special-cased names.

Oh.. wouldn't things like 1,000,001 be harder to say, then?

Yes, instead of “one million and one” (5 syllables) it would be 8 syllables (or maybe 9).

On the other hand, 1,762,354 would be the same 8 syllables compared to 18 syllables in the current system (in English).

It’s possible some shortcuts could be added for several zeros in a row if that ever became useful. Or if the speaker were willing to break it up into the sum of two numbers this number could be expressed using 2 syllables for the 1e6 part, then the word “plus”, and another 1 syllable for the 1 part, so 4 syllables overall.

This reminds me of the only Unicode character (as of 10.0) which numeral value is negative: U+0F33 TIBETAN DIGIT HALF ZERO has a value of -1/2. Andrew West has written an interesting article [1] about this and other "half" Tibetan numerals.

[1] https://babelstone.blogspot.com/2007/04/numbers-that-dont-ad...

Trivia: The first dictionary in Telugu which is still in use is written by an Englishman by name Charles P. Brown.

https://en.wikipedia.org/wiki/Charles_Phillip_Brown

Another one.

First Malayalam-English( another south Indian Language) dictionary was compiled by a German linguist Hermann Gundert who is the grandfather of German novelist and Nobel laureate Hermann Hesse.

https://en.wikipedia.org/wiki/Hermann_Gundert

And so is the first Kannada-English dictionary by a German, Ferdinand Kittel, in 1894.

https://en.wikipedia.org/wiki/Ferdinand_Kittel

As a person whose native language is Telugu, I am forever indebted to this man! His contributions to Telugu go beyond dictionary.

Shocked to see my language mentioned here

Reply on news.ycombinator.com