Before an important conversation, a person reaches for their phone but doesn't dial. Several possible phrases already exist inside, but none has been chosen. This state, when options exist but none has won, the authors call 'rogue variables.'
To describe this mathematically, they build a graph: nodes are images, words, memories, and the links between them show how close they are. Over time, the graph changes, just as the internal landscape shifts. Then all possible graph states are placed into one large space, as if all drafts of the conversation were in a single folder. A special operation in this space highlights directions with the most uncertainty. These are the 'rogue factors.'
Next, the authors add a 'Rosetta Stone'—a layer that translates one person's private coordinates into a common system. This allows comparing unresolved states across different people, even if their graphs don't share nodes.
Eventually, the uncertainty 'collapses': the person clarifies what they meant, and from several meanings, one remains. In the model, this looks like the fading of extra options upon interaction — for instance, when the conversation partner asks for clarification.
This approach could be useful not only in psychology but also in any system where decisions mature gradually: from chatbots to consumer behavior analysis.
The term 'Rosetta Stone' in the layer's name refers to the actual stone that helped decipher Egyptian hieroglyphs: it bore the same text written in three different scripts.
The idea that unresolved thoughts exist simultaneously as multiple options echoes stories about branching realities, where a hero sees all possible paths before choosing one.