When do negotiations become possible in multi-actor civil wars? Bargaining theory highlights the legibility and commitment capacity that stable conflict environments provide; ripeness theory highlights the pressure that disruption creates. We argue these are not rival accounts but sequential phases of a single process, and we identify a network-level micro-sequence into negotiation onset. Using ACLED battle events to construct semiannual conflict networks across 21 African countries (1999–2013), we theorize and test a two-stage sequence: prior disruption, a deterioration shock that breaks entrenched equilibria, creates pressure by lowering the value of continued fighting; subsequent recovery signals feasibility, not by restoring the old equilibrium but by producing a new legible configuration in which bargaining partners and coalitional alignments have clarified before actors re-entrench. Discrete-time hazard models show that more than half of negotiation onsets are concentrated in “window” periods where disruption is followed by recovery and that the window configuration is a significant predictor of onset. Network dynamics show no relationship with peace agreement conclusion, which is consistent with our prediction that the conditions for negotiations are different than those that enable agreements. Our contribution is to specify the network-level content of a ripe moment in multi-actor war: disruption supplies the pressure component of ripeness, and post-disruption recovery supplies the network needed for actors to perceived a way out of the conflict by restoring legibility before a new equilibrium hardens.