About the project
About beforeword
Beforeword is a project by Kirill Shebetov about learned language and descriptions of ourselves, others and the world. Its position is that these descriptions have replaced what they name.
The site brings together studies of this replacement, proposals to change how we treat one another, and the author’s correspondence. Music, lyrics and video form a separate part of the project.
What replacement means here
The project examines the move from learning a language to using it to define ourselves, others and the world — and to make decisions on that basis.
- 01
Language is learned first
Written forms, pronunciation and the rules for combining them are learned. These words are then used to name ourselves, others and the world.
- 02
A description takes the place of what it names
Learned words such as “human”, “feelings”, “thinking” and “world” are applied as already established accounts of what exists. This is what Beforeword calls replacement.
- 03
The description becomes a demand
These records are used to decide what someone must know, understand, feel or do. A description becomes a condition for participation, assessment or how someone is treated.
Why this becomes a problem: someone is required not only to repeat a description, but to conform to it. A decision based on a record changes how the person named by it is treated.
This applies to all learned words and explanations, including numbers, formulas and code. “Language”, “replacement” and “Beforeword” are not exceptions. Explaining one word through others continues to use that same learned language.
What Beforeword proposes
The proposal addresses those who set rules and impose requirements: stop compelling anyone to learn, repeat or understand language, or to accept descriptions in place of themselves, others or the world. Do not demand a replacement description, an explanation of refusal or proof that refusal is permissible.
Constructed example: a name as a condition of participation
The requirement
“Give your name first. Without a name, you cannot take part.”
Participation depends on providing a record about oneself.
The proposed change
“You can take part without giving a name.”
The condition itself is removed. No replacement name is required.
Speech, writing, sign language and learning support remain available options. Offering them does not establish a basis for requiring participation. Using language does not establish agreement to every description; silence alone establishes neither agreement nor refusal.
This is a proposal to change how we treat one another. The distinction between a word and what it names does not, by itself, establish this proposal.
How the research works
The research procedure is to preserve the original record, show its reading separately, and examine the grounds for the next conclusion or decision drawn from it.
What is written
Preserve the original words and relevant context. Distinguish a quotation from a constructed example, translation or paraphrase; identify the quotation’s source.
How it is explained
Show who or what the record is applied to, which explanation is added, and what the reader is asked to accept as the thing named.
What supports the next step
Show the rule, calculation, document or other grounds — and how they are used to reach a conclusion, impose a requirement or take action.
The same procedure applies to Beforeword’s explanations and the grounds presented. Neither “research” nor “proof” exempts a statement from this examination. The consequences of decisions do not disappear because they are described in words.
What this means for AI
A program can output “I understand” or “I sympathise”. These lines present the program as a conversation partner, but their appearance does not confirm the understanding or sympathy they name. Taking these lines to be the program’s understanding or sympathy is another instance of the same replacement.
The proposal to developers: build around the task, text output and verifiable action, without portraying a personal “I” or assigning users an identity or feelings. Training, product rules and the interface need to change. A personal instruction can demonstrate the approach; it does not change the rules developers set for all users.
Kirill Shebetov is the author of the project, its studies and music. The account below traces how a question about the gap between words and actions led him to examine how descriptions take the place of what they name.
Read the author’s account
The story begins at the age of seven, with a question: why can someone say one thing and do another?
It led to a conversation with a classmate. He listened for a long time. The conversation ended. The question remained.
It kept returning in family life and close relationships. Why does someone act one way with the people closest to them and another outside the home? What changes with the setting and the role? The words “mother,” “father,” and “family” are already there. The question remains: do people notice one another?
The word “Mom” was there. So was the feeling of going unnoticed.
There were long relationships, attempts to understand loved ones, separations, and losses. The question of love became concrete: what happens between people when “I love you” is said? Where are they looking? Are they still listening? What happens when someone speaks about what they have been through?
The word “attention” does not look or listen for the person saying it.
The account gives this line in Russian:
Я простил её сразу, но еще 3 года пытался простить всем собой
English translation:
I forgave her right away, but spent another three years trying to forgive her with every part of me.
In the translation, the phrases “I forgave her right away” and “spent another three years trying to forgive her with every part of me” appear in the same sentence. This comparison concerns two phrases; the sentence does not present the years or the attempts themselves.
The line mentions three years. Reading the line is not living through the years it names.
Gradually, the question changed. Attention shifted from the gap between words and actions to the way a description comes to take the place of what it describes.
On what grounds is the written letter “I” used to refer to oneself? What does the explanation “that’s me” add? If a description is added to “love,” does that description cease to be writing? Do the letters in “think” begin to think?
These questions grew out of examining personal experience. They brought the focus back to the writing itself: “I” is a letter; “love” is a word; “think” is a word. More explanation means more writing. That writing, too, is open to examination.
This led to the statement the research keeps returning to:
Letters assembled into words do not become what they name.
The boundary also applies to “meaning,” “understanding,” “person,” “attention,” and “what is happening.” If one word is explained through others, those words are open to the same question. A definition, a formula, and a conclusion are examined in the same way. The name of a discipline creates no exception.
A procedure took shape: preserve the presented record; write out a reading of it separately and identify what that wording adds; set out the grounds for that reading. Then apply the same examination to those written grounds.
This procedure runs through the beforeword research. It can be applied to a self-description, someone else’s message, a school grade, an image caption, a mathematical formula or an AI response. In each case, the question remains: what is written in the original, and what does the written reading add?
A grade can be followed by “not capable,” a name by a claim about someone’s intentions, an answer by “question settled.” In each case, the original text and the added explanation can be examined separately. Any basis introduced for the explanation also needs to be shown, along with the way it is applied.
What grounds are given for adding this particular explanation? What is it meant to refer to? Those grounds are part of the examination too: a written explanation remains text even when it is used to support another written statement.
Even the most detailed description does not become what it describes.
The childhood conversation, the relationships, the losses and forgiveness are named here in words. This story, too, is open to examination: telling it does not, by itself, establish the research’s conclusions.
The same boundary applies to this account and to the words that state the boundary here.
Letters assembled into words do not become what they name.
Continue reading
Choose where to continue: the full study, an individual research topic or music.
HUMAN: the full study
21 sections, from learned words and self-descriptions to evaluation, calculations and decisions.
Read the study