spy. That scroll there contains notes about a correspondence with a different Spirish duke and an Ixoanian admiral. We couldn’t figure out why, but the nobles seem to be paying the author to disrupt this convocation.”
“We had two spies for Spirish nobles in Starhaven?” Amadi half-squawked. “Nora Finn and the owner of the private library? And this second spy took bribes to set these bookworms loose?”
“Worse than that,” the secretary added. “The scrolls by your feet are drafts of curses written to infiltrate a spellwright’s body and force it to overexert itself!”
Amadi felt her hands go numb. “Like the misspell that killed Nora Finn and the neophyte. Did you find any evidence of the remaining spy’s identity?”
Kale shook his head. “Of course not. The author was too intelligent for that. But Magistra, remember the first bookworm we found; it should have returned to this private library, but it was damaged in such a way that it accidentally returned to a previously designated location. Well, we searched that location and found a hidden chest filled with an appalling number of Spirish and Ixonian coins. And Magistra, you forgot to ask where that location was.”
Amadi looked at Kale and then at the two sentinels behind him. “No, don’t tell me,” she said, pressing a hand to her forehead. “I already know.”
CHAPTER
Twenty-five
Though Shannon had been sincere when he told Nicodemus to rest while it was still possible, the grand wizard found himself walking not to his quarters, but to his study. Neither of the sentinels following him objected; they would be up all night no matter if he was lying in bed or sitting at a desk.
After leaving the guards outside and locking the door, Shannon put Azure on a perch and assured her she could sleep. He knew his study well enough to move about without mundane vision.
Though he was exhausted, the idea of a golem had roused his curiosity. How could magical language create such a being? As he pondered this question, habit prompted him to retrieve his research journal and absently finger the three asterisks embossed on its face.
As far as he knew, a spell could gain intelligence only from one of two processes: “direction” and “impression.”
Authors creating “direct” textual intelligence had to write clever prose. At its simplest level, this required strings of instructions: if this happens, then do that; if that does not happen, then do this and so on. More complex methods directed constructs to recognize patterns or develop evolving webs of decision-making sentences.
However, any “directly” intelligent spell fell short of an “impressed” counterpart. Descended from an ancient spell that survived the Exodus, “impression” used two Numinous matrices. The first matrix inhabited a living mind; the second, a spell’s executive language. If physically close, an impressing matrix began to mimic the thought processes of a living mind. In this way spellwrights could “impress” aspects of their own intelligence into texts.
Shannon had given Azure fluency in Numinous through impression, and most gargoyles and all ghosts required a living mind after which to model their thoughts.
What excited Shannon about the golem spell was its implicit connection to impression. To animate a golem, a spellwright had to invest his textual “spirit” into earthen body. To form a spirit spell, an author would haveto use a radical form of impression that translated his mind into a text. That would leave the author’s body an empty husk until its spirit returned.
So before investing his spirit into a golem, a spellwright would have to plan for his spirit’s return to his body. Therefore, a golem would need an escape subspell allowing a spirit to eject itself from a wounded golem.
What Shannon wanted to do was write a linguistic attack that would hinder or destroy a golem’s escape subspell. If he could do that, he might slay the golem’s author without finding the fiend’s living body.
Shannon worked with an excitement he had not known for a half century. After skimming the relevant texts, he had an idea of what functions an escape subspell would have to perform. That left him the task of deducing how a text might fulfill those functions and how an attack spell might interrupt those same functions.
In an hour, he had an outline.
Writing the spell proved more difficult. He worked in Numinous and stored the early drafts on older scrolls. The latter drafts he wrote onto his best parchment. At times his hands shook with excitement and made it difficult to place the lines.
After four hours, he had finished a