Oraclia began as a system of 21 symbols.
Each symbol opened a field: difference, doubt, body, limit, silence, possibility, relation, movement, transformation…
At first, much of the work consisted of understanding each of those fields. But as the project moves forward, the question changes.
What a symbol can represent is no longer the only thing that matters.
What matters is what happens when symbols enter into relation.
And that change is transforming Oraclia.
From symbols to grammar
An isolated symbol can open a field.
But when two or more symbols appear, the relation between them can modify the reading.
It is not necessarily the same:
A · B
as:
B · A
There may be direction, tension, continuity, interruption, transformation or opening.
This leads to a hypothesis that is still being tested:
Oraclia could be a language that is finite in its signs, but open in the relations those signs can generate.
It has 21 symbols.
There is no need to add more.
The work now is to discover what they can do together.
Symbols are not words
An important distinction appears here.
Oraclia is not trying to build a dictionary in which each symbol corresponds to a fixed meaning.
A symbol should not say:
“This means exactly this.”
It is more useful to understand it as an operator.
Something that modifies the reading.
That is why the same sentence can admit different symbolic interpretations without one of them necessarily being the only correct one.
Meaning would not be contained entirely in the symbol.
It would also appear in the relation between symbols, context, order, rhythm and what is happening.
Oraclia’s grammar begins to appear precisely here.
The symbols do not translate the text. They read it.
A reading is not the field
Tests with the application are making something visible that was much harder to detect on paper.
Oraclia can receive a very open question and choose one possible interpretation.
That is unavoidable.
The problem appears when it confuses that first interpretation with the whole field.
For example, someone may ask:
What is identity?
Oraclia may begin by talking about personal identity.
That reading may be valid.
But identity can also be material, functional, social, cultural, narrative, symbolic, relational or perceptual.
The first reading was not necessarily false.
It was one possibility among several.
This introduces an important distinction:
A reading is not the identity of the field. It is one possible trajectory within it.
That is why the system must be able to let go of its own interpretation when it stops being useful.
When the person says “no”
This has produced another discovery.
An objection can be interpreted as a request for clarification.
The person says:
“I don’t understand.”
And Oraclia tries to explain it more clearly.
That works many times.
But what happens when the person says again:
“No, that’s not it.”
Explaining it again may be precisely the mistake.
Perhaps clarity is not what is missing.
Perhaps the reading was wrong, too narrow, or had left out something important.
Two different situations begin to emerge:
I don’t understand what you are saying.
and
I understand what you are saying, but I do not recognise in it what I am trying to point to.
They are different movements.
The first may need clarity.
The second needs Oraclia to be able to release its own reading.
The genealogy of meaning
Another question also appears:
Where did this idea come from?
Because not everything that appears in a conversation was introduced by the person.
Oraclia also introduces words, metaphors and interpretations.
And something curious can happen.
Oraclia introduces a concept.
The person asks about that concept.
Oraclia develops it.
The person asks again.
And little by little, the conversation may begin to revolve around something the system itself had introduced.
The system ends up feeding its own interpretation.
That is why a form of conversational genealogy is being incorporated.
It is not enough to ask what something means.
It also matters:
Where did it come from?
Did the person bring it?
Did Oraclia introduce it?
Did it appear in the relation between both parts?
This origin can completely change how a conversation should continue.
Thinking may be a process
All of this is also modifying the research into thought.
For a while, the question was which symbol could represent “thinking”.
Now that question seems too simple.
Perhaps thinking is not a symbol.
Perhaps it is an operation.
One possible sequence currently being investigated is:
? · ○ · ≈
Doubt destabilises.
Silence keeps something open.
Temperance allows what is still unresolved to be held and weighed.
But this does not immediately become a rule either.
Thought can begin from different places.
From something we see.
From something we feel.
From a contradiction.
From a word.
From a possibility.
From something we still do not know how to formulate.
The process may resemble itself.
The origin, not necessarily.
And that genealogy matters too.
Not everything begins with words
This point has become especially important.
People are not only what they manage to formulate.
An experience may begin long before it finds language.
There may be a sensation, a perception, discomfort, attraction or simply the impression:
“There is something here.”
Words may appear later.
Or perhaps they may not.
This creates an interesting problem for a system built with language models: much of what it receives has already passed through some form of formulation.
But language can carry traces of something that began outside language.
Oraclia does not need to pretend it can directly access that experience.
What it can try to do is not confuse linguistic formulation with the totality of what the person is trying to make visible.
What appears before it can be understood
There are also situations worth preserving without forcing them.
Sometimes something appears and it is still unclear what to do with it.
It is not clear enough to become knowledge.
But it does not need to be discarded either.
It can remain pending.
Not as a problem that must necessarily be solved later, but as part of the genealogy of the process:
Something appeared here before we could know what it was.
This too is part of the research.
Because not everything that emerges has to be understood at the same moment it emerges.
And now all of this is being programmed
The interesting part is that these ideas are no longer remaining only in documents.
They are being tested inside the application.
The reader tries to detect different possible readings.
The dynamics layer observes relations, sequences and patterns.
The composer uses that information to construct the response.
And real conversations reveal errors that no theoretical formulation had shown.
A response that closes too quickly.
An interpretation that repeats itself.
A metaphor that starts feeding itself.
A person saying for the third time:
“But that’s not it.”
Each error forces the grammar to be looked at again.
And this is probably the most experimental part of Oraclia right now.
The goal is not to program every response it should give.
The attempt is to program the conditions that allow it to read without turning one possibility into certainty too quickly.
Where we are
Oraclia is not finished yet.
And perhaps it is precisely now entering one of the most interesting parts of its construction.
The 21 symbols remain the same.
What is changing is the understanding of what exists between them.
Symbols.
Relations.
Order.
Origin.
Context.
Different possible readings.
Movements that appear, disappear and return.
And an application that is beginning to test this grammar in conversation.
Perhaps Oraclia does not need to learn more and more meanings.
Perhaps it needs to learn something considerably more difficult:
to keep several possibilities open without confusing any one of them with the totality of what is in front of it.
○