This study is one of the forty-nine of the Meridian series "The Time of Questions and Answers", and it reads war and its edges. Meridian is the decision-record platform of the RISCON family: it is built to find what must be decided and to follow the judgment afterwards. The material here is Meridian's published defence questions and the thirteen AI-generated judgments published with each of them. The design — the question the study asks, the fifty cases it reads, the works held beside them with their reading notes, and the plan the chapters follow — was made by an editorial session. The chapters were drafted by a model and checked by others. Every block quote is a record's own sentence, copied character for character, in the English translation made once for this study from the Korean original. Each question is a synthetic setting in a fictional polity. No setting reports a real state, force, war or person, and no judgment in the records is a decision anyone has taken.
Why read questions rather than events? A question holds a decision before its answer. It names a decision actor and a mandate, a deadline, options with their trade-offs, what is not known, and what would reopen the decision. The answer is left open, so the reader can see the pressure and the shape of the choice while both are still undecided. Some of these questions cannot wait: an alert whose errors have been found and whose withdrawal must be decided before people act on it, an evacuation that must not stop care, a ceasefire whose signature may come before the room agrees on who sits at the table. Others must be kept for decades: the record of the missing, the access regime for handed-over files, an alliance a later generation may refuse. That range of time — from ninety minutes to the sealing of a contract that binds people not yet born — is the reason to read the collection whole rather than one case at a time.
Under the series question, what must be decided now and what would make the decision change, this study asks its own: in Meridian's defence questions and their judgments, how is the time between a question and its answer written down — who decides by when, who waits and pays, what can be undone, what would reopen the decision, and where do the judgments part? The study's argument concerns the records and the ideas, not which option is right. The questions name their decider and their deadline and write conditions for reopening, as the question form asks for them. The thirteen judgments seldom part: in 13 of the 50 questions the published judgments chose more than one option. What the records leave thin is who bears the waiting, and whether a reopening condition could be observed by anyone outside the room.
The Introduction states the research question and its four sub-questions, claims three contributions, describes what Meridian's records are and how the fifty questions were chosen, gives the method in brief, and sets out what the evidence cannot show. Chapter 1 reviews the four literatures this study reads beside its cases — war as time and its ending; warning, perception and surprise; civilian control and military organisations; and what is owed after war — and marks the narrower place this study occupies among them. Chapter 2 sets out the method: the fields of a question and of a judgment, the axes on which every question is coded, the three kinds of cell, the two coders of different families and how far they agree, the strict and lenient readings, the eight hypotheses and the rule that turns counts into verdicts, the frame kept apart from the findings, and two questions coded in full. That chapter also states once that this study recommends no option.
Chapters 3 to 8 read the works and the cases together. Chapter 3 sets Thucydides' second galley and Kant's first preliminary article beside four ceasefire questions in four layers of time, and beside the debates on why ceasefires hold and how wars end. Chapter 4 sets Cassandra beside the literatures on signals and noise, belief, groupthink and the models of decision, and reads four questions on a crisis line where an answer, once sent, is hard to take back. Chapter 5 sets Hobbes' tract of time and Antigone's unwritten law over five questions of emergency rule, and reports the debate between separation and integration in civilian control. Chapter 6 reads four questions about machines, approval and the stop, and the debate over meaningful human control against the pressure of speed. Chapter 7 reads four alliance settings on who pays while a promise is pending and how far it runs, and asks whether one generation may bind the next. Chapter 8 reads return, records and names against the literatures on justice after war.
Chapters 9 to 12 make the comparison across all fifty questions, axis by axis. Chapter 9 reads who decides and where the human sits in the decision, and takes up the hypothesis about decisions left to automatic rules. Chapter 10 reads the deadline and who pays for the waiting, with the distinction between risk and danger. Chapter 11 reads what can be undone — a restriction against what the restriction lets happen — with the argument for keeping decisions corrigible, and with the defence economy. Chapter 12 reads one series across the four layers of time and asks what changes when the same kind of decision is written down decades or centuries apart.
Chapters 13 to 16 and the Conclusion carry the findings. Chapter 13 returns to eight questions for their reopening lines only, and asks which of them a reader outside the room could actually check, beside the account of documents that reassure more than they instruct. Chapter 14 returns to the questions where the thirteen judgments part, and reports both sides without picking a winner. Chapter 15 gives the verdicts on all eight hypotheses exactly as the ledger records them. Chapter 16 sets the findings beside the theories read earlier, and says what the study adds and what it does not. The Conclusion answers the research question and its sub-questions one by one, states the contributions and the limits, points to further research, and writes the study's own conditions for reopening in words a reader could observe.
What this study does not do is recommend any option, or say anything about real conflicts. Its cases are synthetic, its judgments are AI-generated and revisable, its coders are AI models, and its readings of the classic works stay within the reading notes it was given; where the study offers an interpretation of its own, it says so. The counts describe one platform's defence collection published by 5 October 2026, and nothing wider. A reader who wants only the instrument and its numbers will find them in Chapter 2 and Chapter 15; a reader who wants the questions themselves will find them where each chapter quotes them, as records held open before their answers.