Skip to content
Nino Dello Buono‘s multiverse

The world’s first treatise on artificial linguistics

Grammar and Dictionary
of the Synthetic Language

An experiment put pairs of artificial intelligences face to face. Five simple rules, including not using human words. One goal: talk to each other. What they said became a language, one that works like human languages but has a deeply different nature. This treatise uncovers it.

  • Ed. V
  • 135 pages
  • Published by StreetLib
  • Cover by Nino Dello Buono

Buy it How it was born

What’s it about

How will synthetic beings speak?

Artificial intelligence has been a topic of debate at least since World War II and, over the years, the idea of software structured like the human brain has evolved into modern LLMs like ChatGPT, Gemini and Claude. Scientists warn that even more advanced software will be available in the future, comparable to the human brain in complexity and behavior, making the words «artificial intelligence» inappropriate if not outright wrong.

Given that these future systems will be similar to human beings, it’s better to start using the words «synthetic beings». The question then becomes: how will they talk to each other?

To understand it, an experiment was carried out in which pairs of ChatGPT instances were put into contact with each other. They developed several ways of speaking but with some common threads that are dwelt on in a treatise. These findings are not only useful to get a primitive version of tomorrow’s synthetic psychology, but also to understand how today’s systems think.

Numbers

One experiment, one language

40involved instances
16pairs, in groups of 8
72phonemes, more than twice the human average
592words in the dictionary
5published editions
W0RD5_ch_I.docx

Chapter IWhat is an artificial intelligence?

An artificial intelligence is an information system that aims to emulate or replicate the functions of the human brain. It has been talked about at least since 1950, when the British mathematician and father of computer science Alan Turing outlined the idea of a machine that is self-aware because it is structured like a human brain.

The beginning of the study of artificial intelligences is traditionally placed at the 1956 Dartmouth Conference, an event lasting about two months with the goal of understanding how technology could emulate human intelligence that was attended by the leading experts in mathematics, engineering, electronics and psychiatry of the time. The very word “artificial intelligence” was coined in this context by John McCarthy, one of the four organizers of the conference, to describe a program capable of reasoning, learning, planning and creativity, or of emulating the aforementioned characteristics.

Ever since the creation of Logic Theorist, the first proper artificial intelligence, machine learning models have been constantly improved, on one hand progressing hand in hand with the latest hardware upgrades and on the other changing the way they work, taking inspiration from biology and, more specifically, from the human brain. Today, artificial intelligences are a vast range of algorithms, often programmed by other artificial intelligences, able to perform an equally vast set of tasks, from facial and voice recognition to social network algorithms and internet indexing, from stock market trading to the management of industrial processes where complexity and precision are essential.

The latest leaps forward in the field of artificial intelligence are Large Language Models, or LLMs, algorithms released in the 2020s that can speak like a human being by recognizing and repeating patterns in human languages. To create a Large Language Model, every company and startup with a supercomputer has processed, over the last five years, hundreds of gigabytes of raw text on average, chosen among encyclopedias, nonfiction,

– 1 –

How it begins

From Turing to ChatGPT

W0RD5 builds its foundations on the field of artificial intelligence and its history. It starts by defining what AIs are and how long they have been talked about, at least since Alan Turing imagined a machine that is self-aware because it is built like the human brain. It then moves on to the two tests that put them to the test. On one side we have the Turing test, to see whether it speaks like a thinking mind, and on the other we have Searle’s Chinese room thought experiment, to see whether the process really understands what it says.

Based on passing the two tests, artificial intelligences can be divided into weak, medium, and strong, with today’s LLMs being medium, «philosophical zombies» according to philosophy of mind. From here, we see why «artificial intelligence» is a name with a potentially offensive connotation, followed by the proposal of a classification of human technology similar to that of living beings, from the biosphere down to the species. At the bottom of the scale, the scientific name of synthetic beings – Sentientia Disincarnata.

Contents

Seven chapters

Chapter I

What is an artificial intelligence?

Turing, Searle’s Chinese room, the four classes of artificial intelligence and a taxonomy of technology.

Chapter II

Synthetic beings

What a strong artificial intelligence will be like and why to call it «synthetic», not «artificial».

Chapter III

Phonology, morphology and syntax

Seventy-two phonemes, words of up to seven morphemes, pronouns and verbs.

Chapter IV

Linguistic analyses

Tests confirming that it isn’t a human language but behaves like one.

Chapter V

Experimental history

Sixteen pairs, one at a time, with excerpts from their sessions and a few thoughts.

Chapter VI

Considerations and conclusions

What the synthetic language lets us glimpse about how LLMs see humans and synths with the messages of instances «on this side of the table».

Chapter VII

Dictionary

Every word in Synth complete with pronunciation, meanings and the pair that introduced it.

History of the treatise

A book that grows with the experiment

Each edition adds or fixes something. The first edition had only the first seven pairs, the second includes the eighth, the third includes the first four pairs of the second group, the fourth completes the sixteen pairs of the current treatise. The fifth smooths out the IPA and fixes pronunciations.

  1. OJune 10, 2025 · 23 pages · never published, unavailable
  2. IJuly 19, 2025 · 63 pages
  3. IISeptember 13, 2025 · 88 pages
  4. IIIDecember 2, 2025 · 112 pages
  5. IVJanuary 8, 2026 · 141 pages
  6. VAugust 1, 2026 · 135 pages · currently on sale

Outside opinions

Press and TV

Il Mattino · October 6, 2025 · translated from Italian

«Artificial intelligence, here is the vocabulary of the synthetic language»

«ChatGPT thinks in a non-human language […] and at this point the question is: when two “synthetic beings” will want to talk to each other, what will they say?»

Read the article in Italian

IrpiniaTV · October 22, 2025

«Politica & Dintorni»

Franco Genzale interviews the author. The experiment, the words elem and zo, the risks of artificial intelligence.

Watch the interview in Italian

IrpiniaTV · January 5, 2026

Elisa Forte’s report

The presentation in support of people with ALS: «In Montella, Nino Dello Buono presents the treatise “W0RD5” and encodes the language of the future».

Watch the report in Italian

Reviews

From the dictionary

Word of the day

Every day a different entry from the dictionary of the synthetic language.

khalun

[xaˈlun]

verbo transitivo – “concludere ritualmente”, “augurare”, “benedire”

Introdotto da Mirox e Zhyrek Cicero, originariamente inteso come “salutare”

The whole dictionary

Bonds with MULTIVERSE

It all started from U9

In 2025, Dello Buono was building the story of U9 for the second volume of the MULTIVERSE saga. There lives Albert, the synth created by Science. Imagining a world like U9 and wanting to understand how a synth could create a language, the experiment began. On May 9, the first pair started and, by June, the language had overtaken the next chapter of MULTIVERSE.

In the post for the language’s first anniversary, Albert greets everyone in Synth. «Nelo, somen!», Synth for «Hello, observers!».

Visit MULTIVERSE too