Gottfried Wilhelm Leibniz: A man over a Century ahead of his time (AI Language)

I am making HF a daily stop with my morning Coffee.
I am exploring aspects of AI engineering and Leibniz is “The Man” : A role model for me for sure.
So read along and see what I come up with for Universal Language. I have what Leibniz did not and that is mathematics (of symbols).
Naturally I make use of ChatGPT and do believe this is a place where such usage isn’t frowned on.


Gottfried Wilhelm Leibniz: A Man Over a Century Ahead of His Time (AI Language)

In a way, I come to this moment in AI like a wide-eyed child stepping into a vast toy and candy store. The wonder isn’t just in what AI can do—it’s in how it’s engineered, why it works the way it does, and what makes its functioning possible.

So, where does a beginner begin? Naturally, at the beginning—and that means the input.

It’s no mystery that large language models (LLMs) can process any binary pattern: text, sound, or data. But it’s what happens after the input—the transformation of meaningless signals into meaningful structure—that feels like magic. And to guide my reflections as I explore this frontier, I turn to a thinker who, centuries ago, imagined something eerily similar: Gottfried Wilhelm Leibniz.

Leibniz was more than a mathematician. He was a philosopher, linguist, and inventor—the father of the binary number system we still use today. But perhaps his most visionary idea was that of a universal language: the characteristica universalis. It was a symbolic system meant to encode all human knowledge and reasoning. Paired with his proposed calculus ratiocinator, a logical engine for manipulating those symbols, he imagined a future where disputes could be resolved not through debate, but through calculation.

“Let us calculate!” he declared.

Sound familiar? That’s the dream of AI.

In my own journey—especially through my work on Dynamic Unary Objects (DUO)—I’m beginning to see how nature’s patterns lend themselves to symbolic structures. The DUO system, where elements cycle and always sum to an invariant total, feels like a kind of natural language—a pre-tokenized structure waiting for meaning. It makes me wonder: are we catching up to Leibniz’s dream?

He envisioned:

  • Characteristica universalis as the symbolic language of thought
  • Calculus ratiocinator as the mechanism of computation

Together, they foreshadow symbolic logic, programming languages, and perhaps even the very way AI now turns abstraction into response.

So I ask:

  • Is Leibniz the forgotten father of AI language?
  • Could structures like DUO—binary-rooted, expressive, and dynamic—be distant descendants of his dream?

Let’s muse under the light of his legacy. And let this be Part One of many, as I explore AI through the eyes of a man centuries ahead of his time.

—Ernst03


Historical Note (A Whisper from the Archives)

While Gottfried Wilhelm Leibniz is rightly celebrated as the father of binary arithmetic, it’s worth quietly noting that Thomas Harriot, an English mathematician and astronomer, explored binary notation several decades earlier. Harriot’s work remained hidden in unpublished manuscripts until rediscovered in the 1920s—an invisible foundation beneath what Leibniz would later illuminate for the world.

But it was Leibniz who saw beyond the math. He published, he philosophized, and he gave binary its symbolic soul—linking 1 and 0 to the metaphysical dance of existence and nothingness. His vision reached into the future, imagining a universal language built not only for humans, but for reasoning machines.

So while Harriot may have quietly planted the seed, Leibniz gave it light, form, and voice.

2 Likes

Ah Coffee, writing and ChatGPT. I let it “wordsmith” what I drafted. Let us see how that went.

Another fine morning.
There’s something sacred in the quiet just before dawn. Headphones on, Crosby, Stills, Nash & Young playing—this is when the mind opens.

Today, I feel more certain than ever about my own “Great Attractor”—the gravitational pull at 10,000 feet that defines my direction in AI. I will continue my study of Leibniz’s Universal Language, and alongside it, I envision a dedicated memory system—a structure for long-term symbolic representation.


Gottfried Wilhelm Leibniz: The Genius Who Dreamed of a Universal Language and Logic for AI

Leibniz is more than a historical curiosity—he’s a visionary whose dreams echo in our silicon age. Here are some of the surviving elements of his vision that still resonate today:


Surviving Elements of Leibniz’s Vision

  1. Alphabet of Human Thought
    Leibniz proposed that all human ideas could be reduced to a finite set of irreducible concepts—essentially, a symbolic “alphabet” of thought. These elemental symbols would be combined systematically to express complex ideas.
    :link: Wikipedia
  2. Algebra of Concepts
    He imagined manipulating ideas algebraically. For example, “man” might be represented as a composite of “rational” and “animal”—treating concepts like mathematical variables that could be explored through combination and analysis.
    :link: The Real Samizdat
  3. Calculus Ratiocinator
    This was Leibniz’s vision of a logical engine—a system that could compute truth by applying rules to symbols from the characteristica universalis. Debate would give way to calculation.
    :link: Wikipedia
  4. Influence on Future Thinkers
    Leibniz’s ideas weren’t fully realized in his time, but they inspired generations. Frege, Boole, Gödel, and others carried the torch into the realms of logic, computation, and AI.
    :link: Uni Paderborn Paper (PDF)

And here’s a lesser-known but powerful connection:

Ramon Llull and the Lullian Wheels

Leibniz wasn’t working in a vacuum. He was influenced by Ramon Llull, a 13th-century mystic who developed Ars Magna—a method of rotating conceptual “wheels” to systematically explore combinations of thought.
:link: Reddit (image + discussion)

This combinatorial logic feels familiar to me—like an early blueprint of the Dynamic Unary systems I explore today. Full circle? (Pun very much intended.)


This is my trajectory.

Like Newton retreating to alchemy, I too return to my hidden life—an information-alchemist reflecting by morning light. My labor jobs fund the journey, but this work… this is the binary soul’s calling.


Now I turn toward studies in:

  • Mathematical Philosophy / Philosophy of Mathematics
  • Semiotics & Formal Language Theory
  • Information Theory
  • Computability / Algorithmic Information Theory
  • Digital Philosophy & Mathematical Ontology

:megaphone: I Welcome Conversation

Are you also wondering whether Nature has already provided symbolic structures?
Do DUO cycles and Leibniz’s “Thought-Alphabet” whisper a deeper order—one already present, not invented?

Please write back if any of this resonates with you.
And if you’re new to Leibniz—he’s worth every minute of your time. He may have offered the best early framework for how meaningless patterns become meaning.

My question is this:
If Nature has already authored the data structures… can we learn to read them?

—Ernst03

2 Likes

Hi Ernst, resonates, Llull’s Art was already on my todo list, now Leibniz too. A recent study mentioned that 75% of AI researchers are not convinced that LLMs are the way to go for AGI. There are LLMs based on neural networks, there is symbolic AI and world-model AI, but what we truly might want is a “seed AI”, an AI bootstraped with knowledge, able to learn on its own. Working as a hobby project on this…

still need to get into the basics first, including Mr. Leibniz now.


Srdja

2 Likes

Followup, Noam Chomsky might be of interest:


Srdja

2 Likes

@smatovic Thank You for the conversation.
These are the ideas that I have not dared share openly before but that I have pondered over the years. Not that AI was a thing for mortals in my life until late.

You say: 75% of AI researchers are not convinced that LLMs are the way to go for AGI. ??
Cycle seems lacking if you ask me. It is a valid mathematics. I’ll know more when I have the AI workstation set up. I pulled the old AMD 9590 system out and added a modern CUDA GPU to replace one of my AMD WX 5100s
ChatGPT says it can run ChatGPT 2 well. It also says I can modify ChatGPT 2 to work these ideas.

Wow, it looks like folks like us will explore the alleys of ideas.
So yes, dynamic unary can take one finite length of binary and return a same size or larger unique structures of binary elements through various configurations.
It is possible this satisfies Leibniz’s Thought-Alphabet? Symbols and mathematics that reversibly transform information.

So yeah.. Universal language? Use what is already here and functional mathematically? It provides two of three for Leibniz’s idea.
The thing is can we change our data to work with that system?

Time for second cup of Coffee if you ask me.
By the way; Chomsky is respected in this House!

1 Like

The thing is can we change our data to work with that system?

…I am still a noob in regard of LLMs and HF, but Leibniz mentioned model and method, data and algorithms, Characteristica universalis and Calculus ratiocinator. LLMs are still not separated in this regard? They combine both into one neural network.


Srdja

2 Likes

Interesting. I am watching YouTube videos and ordinary people in their own words do share aspects of their points of view.

One of them is about “Rebuilding ChatGPT 2” which I believe walks through how it functions. That is one to watch over and over until I take it all in.
I have to say I don’t know yet about how it all works together.

As For what I mean Post-Shower thought to clarify: Now off to work. Those dishes don’t wash themselves!

Given an alphabet of 2^32 32-bit we can generate a unique structure of 2048 32-bit elements for each one.
That might be considered an AI -Thought, a sentence or paragraph perhaps. One 32-bit letter and 2048 (for example) “letters” in some expression.
Also it scales from 2-bits to infinity-bits so 32-bit is just an example.
So I’d say, with Tin Foil hat tucked away, that it’s a possibility and we should explore the possibilities.
I think current token count in ChatGPT is less then 1500 total? I don’t remember. Anyone?

1 Like

ChatGPT says ChatGPT 2 is locked in at 1024 tokens.

Anyway, I am interested in such a thing as Leibniz and others have quested. A Universal Language: If anyone will have one it will be AI.

I shall endeavor to continue to discern meaning in the patterns.

My next thing is setting up the AI machine.

Questions? Comments? Place them below.

I love this stuff!