The report arrived on a morning when Rakshasa had five other reports on his desk.
He read the other four first, in the order that their operational urgency required, making the notations that each required and setting each aside with the efficient finality of a man who made decisions about things rather than carrying them forward unresolved. The fifth report he read last not because it was least urgent but because it was a different category from the others — not an operational update requiring a decision but an assessment requiring something more like thought.
The report described the Mauryan boy.
It had been written by the operative he had placed in the south corridor's second cell three months ago, compiled from three months of observation and acoustic surveillance and the specific intelligence that the carefully constructed back-wall transmissions had produced. The report was thorough in the way that reports from well-trained operatives were thorough — comprehensive, organized, free of the interpretive inflation that made poorly trained operatives' reports unreliable.
He read it twice.
The first reading was for content. The second was for what the content implied about the content's source.
The boy had fed his operative a false picture.
Not obviously false — the report was internally consistent, the details corroborated each other, the chain it described was accurate in its components. If Rakshasa had not already known the shape of the operation the report described — if he had been receiving this intelligence cold, without prior knowledge of the prison's administrative structure — he would have read it as a competent operative's accurate account of a situation accurately observed.
But he knew the shape of the operation.
And the shape the report described was wrong in one specific and significant way. It pointed to Vijayavarman and stopped. It contained nothing that pointed anywhere else.
A fourteen-year-old prisoner in the last cell of the south corridor, with no sources of information beyond what passed through his food slot and what the facility's acoustic properties allowed him to observe, had constructed a false picture of sufficient sophistication to deceive a trained operative who had been specifically placed to observe him.
Rakshasa set the report down and looked at the wall of organized documents.
He thought about the placement decision — the decision he had made two years ago to put the boy in the last cell specifically, the cell whose acoustic properties gave the most comprehensive access to the facility's ambient information, the cell whose isolation ensured that the boy's development would occur without the social interference that the facility's general population would have produced. The decision had been made on the basis of a single report describing a seven-year-old who had survived a massacre by doing exactly the right thing at exactly the right moment without anyone telling him what the right thing was.
He had decided the boy was worth watching.
He had not anticipated this.
Not the false picture specifically — he had known the boy was intelligent, had expected the intelligence to produce results that would require him to revise his assessment upward. What he had not anticipated was the specific quality of the false picture. Its construction required not just intelligence but the ability to model another person's information-processing — to understand what his operative would find convincing, what details would satisfy the operative's verification requirements, what overall shape would produce the desired conclusion.
The ability to construct a model of another mind and operate within that model.
At fourteen.
He picked up the report and read it a third time, this time specifically tracking the places where the false picture had been constructed with particular care — the details that were most convincingly true, the elements that would have satisfied the most skeptical reader. Each of these places represented a decision the boy had made about what his operative would check and what his operative would accept.
Every decision was correct.
He set the report down.
Then he wrote on a separate piece of cloth, in the compressed notation he used for operational instructions, three words.
Do not kill.
He called for a messenger and gave the cloth with the specific routing instructions that would move it through the channels connecting his office to the prison's operational structure, and he waited for the messenger to leave before returning to the other four reports on his desk, which required his attention and which he gave his attention, while the back part of his mind continued to hold the report about the boy and what the report implied about what the boy was becoming.
The test would come next.
Not the operative — the operative had served his purpose. Something more direct. Something that would put pressure on the qualities he had been observing at a distance and determine whether those qualities held under conditions that observation alone could not replicate.
He began designing it.