Oraclia continues to change.
Not because we are adding more answers, more symbols or more rules, but because we are becoming better at distinguishing what each part of the system should do.
Over the past months, the prototype has gradually moved away from a very simple idea: that a sentence can be reduced to a symbol or to one specific interpretation.
The current direction is different.
Oraclia is trying to read relations, forms and transformations.
Symbols are not words
Oraclia works with an alphabet of 21 symbols.
But these symbols do not function as a dictionary.
∞ does not always mean the same thing. Neither does ?. ∆, ≈ or ○ do not have fixed translations that can be mechanically applied to any conversation.
They are better understood as operators.
A symbol may modify what becomes visible, which difference appears, which relation gains relevance, or how a situation is reorganised.
This is why one of the central ideas of the project remains:
Symbols do not contain meaning. They modify the field where meaning emerges.
From symbols to configurations
A conversation may activate several symbolic fields at the same time.
What matters is not only which ones appear, but what happens between them.
They may coexist, enter into tension, reinforce one another, separate, reorganise or modify the whole configuration.
This has gradually produced a clearer architecture within Oraclia.
First comes a symbolic reading. Relations are then observed. From those relations, a minimal form is built. When another conversational turn arrives, that form can be compared with the previous one.
Oraclia therefore does not only look at what is there.
It can also begin to observe what continues, what disappears, what appears and what changes relation.
An absence may still remain active
This change has made one particularly interesting question visible.
If something stops appearing explicitly in a conversation, this does not necessarily mean that it has stopped having an effect.
An absent element may remain related to elements that are still present.
But Oraclia should not automatically conclude that this absence “weighs”, “blocks” or “determines” anything.
It can only preserve the previous relation and observe whether later language supports the idea that the absent element is still active.
The distinction may seem small, but it matters:
not present does not necessarily mean not active.
What if there is not enough structure to interpret?
This has led to one of the most recent developments in the system: the null symbolic field.
A null field does not mean that there is nothing there.
Nor does it mean that there is no weight, relation, emotion or movement.
It means something much more specific:
Oraclia does not yet have enough grounding to build a reliable symbolic configuration without inventing one.
A system may be tempted to fill this gap with a likely interpretation.
Oraclia is trying to do the opposite.
If it does not know, it can preserve that not-knowing.
And if the conversation needs to continue, it can ask for a concrete distinction.
For example, someone might say:
“I have a meeting tomorrow and I don’t want to go.”
The position is clear, but the reason is not.
It could be disinterest, tiredness, discomfort, disagreement, or something else entirely.
Instead of selecting one explanation, Oraclia can ask.
The question is not an interpretation.
It is a way of introducing a distinction where there is not yet enough differentiation.
Null is not infinity
This distinction has also required separating two things that initially looked similar.
A null field is not the symbol ∞.
Null describes the state of the reading: there is not yet enough configuration.
∞, on the other hand, may function as an operation of the system: preserving possibility while asking for differentiation.
This prevents memory from storing something the person never expressed.
If three consecutive turns do not contain enough configuration, the trace should not become:
∞ → ∞ → ∞
It is instead a persistence of null symbolic field.
This does not mean “openness”, “blockage” or any psychological state.
It only means that across several consecutive moments, there was not enough structure to support a reading.
Memory should not turn trajectories into identities
As Oraclia becomes able to preserve forms and compare them over time, another risk appears.
Recognising a recurring trajectory is not the same as defining a person.
If a configuration has appeared several times, the system may say:
this trajectory has appeared before.
It cannot conclude:
this person is like this.
Memory can preserve possible continuities, but every new situation must be able to contradict them.
The person remains prior to the model Oraclia constructs.
A reading must also be correctable
This principle is now entering the system’s governance.
When a person corrects an Oraclia interpretation, the correction is not treated as an objection the system needs to overcome.
It is new information.
If Oraclia reads tension and the person says there is no tension, that interpretation loses priority.
The system should not defend it, reformulate it or immediately replace it with another theory.
It should revise its model.
The reading is provisional.
Describe before interpreting
Another rule that has become increasingly important sounds simple:
first describe what is observable; then interpret only if there is enough grounding.
A decision is not necessarily rigidity.
Withdrawal is not necessarily fear.
A boundary is not necessarily protection.
Change is not necessarily openness.
Any of these may be true in a particular case.
But the conversation has to support it.
The pipeline is changing too
Internally, Oraclia is gradually becoming organised into more distinct layers.
The reader detects and proposes. The relations layer organises connections. Form represents the configuration. Dynamics compares different moments. Governance decides what the system may do with the reading. The composer then translates the result back into human language.
This separation serves an important purpose:
the final language should not have to carry all of the system’s intelligence.
The answer can remain simple because much of the work has already happened before it reaches language.
Some questions remain open
Oraclia is still a prototype.
We are still exploring how to represent forms that the current grammar cannot describe, how to govern different ways of deciding without imposing one, and how to build memory without turning recurrence into identity.
There is also more work to do in separating internal layers and progressively simplifying the composer.
But perhaps the most important change is not technical.
It is introducing a possibility that is often difficult for both systems and people:
being able not to know yet.
Not filling the gap immediately.
Not turning possibility into certainty.
Not confusing a reading with reality.
And allowing a difference to appear before giving it form.
○