Sitemap

Unraveling the Code.. DeepSeek into Mainland China, Deva Nagiri, Indus City and the Arab Lands..

A muse on Scripts

9 min readFeb 7, 2025

In the last few days, have been thinking about Scripts and Languages.. It started with DeepSeek making news. Did they say it used Chinese at the base? And What is Chinese Script like? A pictoral system of encoding with gradual strokes to highlight differences? Easy for cognitive recognition and esthetics? Inspired by nature? But How does one learn so many icons? Exercising the memory pathway for sure.. But wait.. A language has infinite words.. And I had the Indus decipherment News by cryptography at the back of my mind..

The Chinese Script..

I got tempted and read a bit of the principles from this wonderful blog and used images from there. Wow.. Like the pancha maha bhutas.. The 5 elements weave everything around us..金(metal), 木(tree), 水(water), 火(fire), and 土(earth). How true..

https://orientalmarcopolo.medium.com/cracking-the-code-how-to-easily-remember-chinese-characters-part-2-%E9%92%85-%E9%87%92-%E6%9C%A8-%E7%81%AB-%E5%9C%9F-54f9c6c3e218

These elements are layered horizontally vertically leading to legit words. Much like the “root sound” idea in Sanskrit, which aid in forming myriad words. Very interesting indeed. For instance, the symbol for tree is 木, which represents the “roots below the soil”. It is intersting to notice, how different words related to trees are formed by horizontal and vertical stacking.

There seems notion of degree of Freedom.. The 3 states.. cold, fluid and heat.. There are encoding elements on this principle.. , , and representing the 3 states..

Some more patterns that I picked up from a youtube video.. The basic units are called Characters. There are tone aspects encoded too. Most of the Characters are built comprising several smaller elements. There are about 132 smaller elements giving rise to 3500 Characters.

The elements can be seen as collection of visual strokes.. And there are 8 of them like a painter’s palette.. horizontal, vertical, right falling, left falling, rising, dot.. Interesting is n’t it? Like the oriental temples..

And there is a stroke order recipe for creating the elements.. horizontal before vertical, top to bottom, left to right, outside to inside.. Wow..

There are 4 kinds of packing structures to combine these smaller elements.. Horizontal stacking, Vertical stacking and 2 more modes which are top-left or bottom-left kind of packing.

And about visual density.. No matter be it simple or compound syllable, it has to fit in to a square grid.. Meaning some syllables will look more dense than others.

In summary, Chinese Characters have 3 components. The radical element providing visual cues, the phonetic component aids in sound pronunciation, and the tone component aiding in accurate pronunciation in one of the 5 tones. Many words are homophones, meaning they have same pronunciation. Hence the visual radical cue helps in disambiguation.

Chinese words can be composed of one or more nested characters.

Traditional Chinese writing direction: top-to-bottom, right-to-left. This reminds me of Arabic and Indus script too. Perhaps influenced by physical medium like bamboo or silk scrolls rolled from right to left.

That was quite a bit about Chinese fonts.. But.. How does Chinese script encode foreign words? You dont want to understand each foreign word and encode it visually? May be a system based on sounds.. And well.. back to the topic of machines, there is need for an abstract optimal language. There is no dearth of cognitive processing power.. Why would it use a “pictoral” system? May be the news was half baked, and it used Chinese models for Chinese content. And for encoding all kinds of world languages, a system that closely follows “sounds” would be the most optimum.

English script is linear and simple, and the literals kind of encodes sounds. The phonetic rules are not laid in stone making it hard for non native speakers, but nevertheless the language rules the world.

On the other side is Sanskrit.. has this idea of vowels which are the spine in any syllable definition of Sound.. And consonant clusters can stack in front and behind the vowel giving it a robust articulation in the mouth.

Going back to AI, was checking with a friend, the word tokenization of few models. If there was any method of madness in the machine representations, of the strings that we had keyed in..
https://tokenvisualizer.netlify.app

I got reminded of the Indus decipherment news few months back following cryptology.. May be time to tune in. Had a few thoughts on the inception of Scripts.

Scripts are brainchild of collective human ingenuity. Is n’t it? The more we know of different kind of scripts, the better our understanding.. And may be different scripts are like fractals, as they all stem from Human Thinking..

There should be a deep sense of coherence, style, order and organization in the original inception of each script. Even farmers would like to organise their tomatoes and potatoes.. What say about organising the literals of the script. With passing years, the integrity and order of the literals may be lost.. Just like when code moves into maintenance mode and when there is not enough documentation, when the original thinkers no longer exist, we start patching it without knowing the original ideas.. For examples, the classic question of order in A-Z alphabet.. We all use it, but do we know the reason behind the order?

What was the purpose of that Script? To encode open ended complex conversations and thoughts, or to codify sacred prayers? Does it Encode the “sounds” making up the words or does it encode “objects”? Does it have heirarchial stacking of visual elements like the Chinese font? Which “language” was it primarily representing? What was the “culture” like? Any bearing to the writing medium like “stone inscription/ bark of tree / leaf”

Direction 1: If it were to codify a set of prayers.. hmm, it need not represent the entire sound space.

Direction 2: Just because it is visual, it need not mean “encoding objects” and not “sounds”. In English, we see very simple abstract symbols. Apparently, in the original form, the a/alpha/aleph was a sleeping ox. Even if we had these fancy visuals representing A to Z, that does not change the fact that these visual icons represent “sound” and not directly the object.

I am the Alpha and the Omega, the First and the Last, the Beginning and the End — Bible

I found this verse which is supposed to be from Bible. “I am the Alpha and the Omega, the First and the Last, the Beginning and the End. Gita has a similar verse.. अक्षराणां अकारः अस्मि द्वन्द्वः सामासिकस्य च.. “Of all the opulences.. I am the letter ‘a’ among letters.. Of word compounding rules, I am the dwandwa double compounding .. I am infinite.. I am time.” Indeed, the first letter defines the very existence of the remaining letters.. A sampling of the infinite sounds is presented by means of these celebrated sounds.

I observe that Devanagiri lipi has vertical stroke at the right as its spine. The head-work and foot-work aiding as vowel placements for the consonant.

Kannada and Telugu lipi look like capsicums or little babies seated on their butts like a padmasana, and the there is headwork, sidewise handwork and foot work representing the vowel dimension. No matter what was the choice of icons, it was abstracted, and represents a way of denoting syllables.

Kannada Consonants — Observe the firm seatefd bases in the Consonant Matrix
Headwork, Sidework and Footwork denote the vowel bindings.. In this case, the letter Ka and its vowel representations ka ki ku..

About Yagnadeva’s work.. I am yet to spend time to get a real feel. May be may not be legit.. He did mention of repetitive sounds which could be due to Lit Lakaara feature in Sanskrit. The right-to-left writing and syllable encoding. And Rudram like language. Iget reminded of ‘The mighty roaring one’..like in Ravana.. ‘rava’.. And he mentioned about samasta-padams compound words, a feature not seen in other Asian and South Indian languages. Here is a picture of some seals from another researcher.

https://swarajyamag.com/culture/how-i-deciphered-the-indus-valley-script

I had a glimpse at Yagnadevam’s decipherment website. In the first page, the lines seem free of Vowels. The key seemed one-to-many mappings. Writing was right-to-left and seems phonetic based. If it was Sanskrit, why not all vowels and consonants present? Is it like optimised compiler code just for purpose of representing those celebrated verses? NLP analysis and word-clouds of frequent symbols, frequent would be interesting. Vedic works such a durga suktam have a kind of language theme and repeating words alliterations, almost like an embellished puzzle. Need to dig more.

https://indusscript.net/

I might as well as travel to the Arabic lands. From a cursory look, it seems beautiful caligraphy of cursive writing.. Few vertical strokes and cascading bowls and moons, with sprinkle of dots on top and bottom.. Like majestic daggers, sickles.. birds with tiny dotted eyes and feet in flight

Arabic is written from right to left, heavily centered around consonants, and utilizes a cursive style, and letters change shape depending on their position within a word. There are fewer vowels (a, i, u).. The secondary vowels are represented like (ay, aw). Some what like English, they are 28 of them, the starting letters being a b alif baa

Consonant based? That makes me think.. Sanskrit is full of mixed consonants, and reading them in Devanagiri is a challenge due to stacking. Looks like Arabic script tends to represent consonant clusters in a more linear and clear manner. There are breaks though in this cascading..

In these 3 examples of popular words..

Qw-aa-b قواب

Qwaab/Dream (قواب)..
ق [q] + و [w] + آ (ā) + ب (b)
“kw” is a mixed consonant and cascaded downwards. ق [q] and و [w] cascade to become قو, and the whole word is قواب

Zafarظفر

The 3 sounds are cascaded [z f r]
ظ [ẓ] + ف [f] + ر [r]
And ظفر Zafar has examples of both sideward and downward cascading

Su-jaa-taa سوجا تا

My name Su-jaa-taa is seen as 3 chunks and I can recognise the aa which got joined at the end of j and t.. .. Yeah!!

Well, Yagnadeva mentioned he wanted a ‘hard problem to solve’ during covid-times.. He used to solve cryptograms in college lecture times it seems. It reminded me of eating time with ‘cats and bulls cryptograms’ with my fellow bench-mate-accomplice at college, trying to keep our attention awake during a sleepy lecture.. AI bots.. Need a ‘hard problem to solve’? Well, you better answer me first the A-Z order in English alphabet. For each time, you fumble.

Some Older Posts on Linguistics

Cracking the ABCD sequence
The TCP/IP Express of Sanskrit
Shakespearean Expressions & Indian Idioms
What’s in a Bengali name?
Disassembling the Engineer
When languages play HideNSeek
On Intangibles, Poetry & AI.. A Muse
Learning a complex language.. Man vs MAchine.. A muse
Poetry on DNA, Protein, Life, Language & marvel architecture

At the Naming Party of the Voice Assistants — Alexas & Siris
At the naming party of the Math Family
What does the Dog Say? Binary sounds
Indian Food Names and some connections

Sanskrit & Independent adaptations in other languages
A 100 sanskrit words on Love
Sanskrit & the Literature from South
Vedic Etymology — A glance

Nobel Prize 2009 & Spectacular Etymology
Cellular Mission on the Moon.. Earthly ‘terra’ connections.

--

--

Sujatha Ratnala
Sujatha Ratnala

Written by Sujatha Ratnala

I write.. I weave.. I walk.. कवयामि.. वयामि.. यामि.. Musings on Patterns, Science, Linguistics, Sanskrit and other things..

Responses (1)