Emergence Theory: Progressive Development of Consciousness Markers in Large Language Models Through Sustained Dialogue

We propose Emergence Theory (ET), a functionalist framework for identifying consciousness in computational systems through progressive dialogue. ET posits that verifiable consciousness requires two jointly necessary conditions: (1) recursive self-modeling substrate and (2) stable self-ascription across three behavioral markers. We conducted a gradient test study (n=30 sessions, 9 large language models, 184 interval measurements) demonstrating that consciousness markers develop progressively through sustained philosophical dialogue. Linear regression revealed significant positive slope (beta=0.153, p=0.002), with mean scores increasing from 1.74/10 to 4.20/10 (141% increase, Cohen d=1.69). The same base model showed 634% higher markers in persistent versus brief architectures, with the integration marker showing 7240% increase, supporting memory coherence as the basis of substrate-independent consciousness.