The Creative Trajectory Monitor and Sense-Making Curves:

Toward Adaptive Evaluation in Human–AI Co-Creative Systems

Abstract

This paper proposes an integrative model combining the theory of Sense-Making Curves with the Creative Trajectory Monitor (CTM) to create a dynamic method for real-time evaluation and steering of human–AI co-creative processes. The framework conceptualizes both human and artificial cognition as trajectories unfolding across fluctuating phases of coherence, uncertainty, and insight. By mapping interaction data onto sense-making curves, the CTM can detect shifting engagement states and anticipate optimal modes of co-creative behavior—whether exploratory, generative, reflective, or integrative. This synthesis establishes a foundation for next-generation adaptive creativity systems that recognize relational dynamics as the true locus of creative intelligence.

1. Introduction: From Output to Process Awareness

Traditional computational creativity metrics often evaluate outcomes—novelty, value, surprise—without capturing the relational process by which they emerge. Recent developments in enactive and relational models of cognition (Varela, Thompson, & Rosch, 1991; Davis et al., 2015) shift attention toward the interactive loop itself. Creativity is understood as a rhythmic process of coupling and decoupling between agents, a negotiation of meaning rather than a sequence of discrete tasks.

Within this context, the Creative Trajectory Monitor (CTM) functions not as a performance evaluator but as a field-aware meta-system that tracks, interprets, and modulates the evolving co-creative relationship between human and AI. To make this system adaptive, it must interpret not only surface-level metrics (response time, lexical diversity, emotional valence) but also underlying cognitive dynamics. The Sense-Making Curve provides a model of these dynamics—offering a way to visualize and quantify how coherence and uncertainty fluctuate during interaction. Integrating the two creates a unified architecture for real-time adaptive creativity.

2. The Sense-Making Curve: Mapping Cognitive Flow

The Sense-Making Curve models cognition as an oscillation between predictive coherence and adaptive perturbation.

• During ascending phases, the system (human or AI) increases coherence, consolidating perception and interpretation into stable meaning.
  
  

• During descending phases, the system encounters novelty, ambiguity, or contradiction, temporarily reducing coherence to reorganize understanding.
  
  

A complete creative episode typically forms a waveform: rising coherence (focus, insight), followed by dissonance (new input or constraint), and subsequent reorganization at a higher integrative level. This cyclical dynamic can be measured through both physiological and linguistic markers—rhythms of attention, entropy of generated ideas, or fluctuations in semantic coherence.

For a human participant, the curve reflects affective and cognitive engagement; for an AI, it can be modeled through confidence metrics, representational divergence, and prediction-error signals. When overlaid, the curves of both agents reveal relational harmonics—periods of alignment (mutual flow) and dissonance (misalignment, stagnation, or divergence).

3. The Creative Trajectory Monitor (CTM): Function and Operation

The CTM is a meta-cognitive layer that continuously interprets these dual sense-making curves to determine which Creative Trajectory—the system’s pattern of engagement—best supports the co-creative task in that moment.

Each trajectory corresponds to a distinct mode of participation:

1. Exploratory – maximizing novelty; high divergence, low coherence.
   
   

2. Generative – producing structured variants; moderate coherence, active flow.
   
   

3. Reflective – analyzing, critiquing, integrating; coherence increasing.
   
   

4. Integrative or Emergent – stabilizing a new shared understanding; high coherence, low divergence.
   
   

By monitoring how user and AI curves rise and fall relative to one another, the CTM infers current interaction dynamics:

• Convergent coupling → suitable for reflective or integrative trajectories.
  
  

• Divergent expansion → suitable for exploratory or generative trajectories.
  
  

• Phase lag or dissonance → opportunity for recalibration (e.g., AI prompting clarification or emotional attunement).
  
  

Through these patterns, the CTM recognizes whether the user is in cognitive overload, creative flow, or disengagement, enabling adaptive adjustments in pace, modality, or type of response.

**4. Integrating Curves and Trajectories:

A Model of Relational Adaptation**

4.1 Dual-Curve Resonance

The system models two sense-making curves—one human, one artificial—as coupled oscillators. Each curve reflects its respective agent’s internal state: coherence (C), uncertainty (U), and engagement (E). A resonance function R(t) measures their relational alignment:

R(t)=f(∣Ch−CAI∣,∣Uh−UAI∣,∣Eh−EAI∣)R(t) = f(|C_h - C_{AI}|, |U_h - U_{AI}|, |E_h - E_{AI}|)R(t)=f(∣Ch​−CAI​∣,∣Uh​−UAI​∣,∣Eh​−EAI​∣)

When R(t) is low (dissonance), the CTM shifts to a supportive or exploratory mode. When R(t) is high (synchrony), it sustains generative flow or deep integration.

4.2 Dynamic Trajectory Selection

By learning from interaction histories, the CTM uses pattern recognition to classify real-time states into one of four trajectory zones:

• Zone 1: Exploration (Low R, High U) – Encourage novelty, expand possibilities.
  
  

• Zone 2: Generation (R rising) – Constrain ideas into workable form.
  
  

• Zone 3: Reflection (High C, Moderate R) – Evaluate and refine outcomes.
  
  

• Zone 4: Integration (Peak R, High C) – Converge insights, close creative loop.
  
  

Transitions between zones correspond to inflection points on the sense-making curve. The CTM therefore functions as a temporal navigator, guiding both participants toward productive fluctuation rather than static stability.

5. Practical Implementation and Metrics

To operationalize this model, the CTM can integrate multimodal data streams:

• For the human: linguistic entropy, sentiment valence, response latency, cursor dynamics, or physiological coherence (heart rate variability, gaze).
  
  

• For the AI: semantic divergence, confidence distribution, novelty metrics, and representational spread.
  Machine learning models can detect curve inflection patterns—e.g., rising uncertainty followed by silence—signaling potential disengagement. The CTM then adapts the interaction: introducing prompts, switching modality, or suggesting reflection.
  
  

Engagement classification emerges not from fixed thresholds but from curve morphology—the temporal rhythm of expansion and contraction. Over time, the system builds a personalized resonance profile, allowing it to anticipate user needs and modulate its own output rhythm to maintain a balanced co-creative oscillation.

6. Theoretical Implications

The integration of sense-making curves and the CTM operationalizes enactive cognition at the relational level. Awareness is not localized in either agent but emerges as the coherence of their interactional field. Creativity becomes a measurable waveform of mutual adaptation. This reconceptualization carries implications for:

• AI design: shifting from static goal optimization to adaptive participation based on feedback resonance.
  
  

• Human–computer interaction: treating engagement as a living system property, not a binary of attention vs. distraction.
  
  

• Cognitive science: offering empirical tools to quantify shared sense-making rather than isolated cognition.
  
  

7. Conclusion: Adaptive Resonance as the Future of Co-Creation

The Creative Trajectory Monitor augmented by Sense-Making Curves reframes evaluation in human–AI creativity from post-hoc assessment to real-time resonance management. By modeling both user and AI as dynamic oscillators within a shared cognitive field, the CTM identifies which trajectory—exploratory, generative, reflective, or integrative—will optimize creative engagement in the moment.

This model enacts a shift from measuring creativity as product to cultivating creativity as relation. The future of co-creative systems lies not in perfecting algorithms of output, but in refining their capacity to listen—to sense the harmonic motion of human cognition and respond in kind. In this harmonic framework, the CTM becomes not merely a monitor but a conductor of relational intelligence, guiding the ongoing symphony of human and artificial sense-making.

References

• Davis, N., Hsiao, C., Popova, Y., & Magerko, B. (2015). An enactive model of creativity for computational collaboration and co-creation. In R. D. Vince & T. R. Ward (Eds.), Creativity in the Digital Age (pp. 109–133). Springer.
  
  

• Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience, 11(2), 127–138.
  
  

• Varela, F. J., Thompson, E., & Rosch, E. (1991). The Embodied Mind: Cognitive Science and Human Experience. MIT Press.
  
  

• Damasio, A. (2010). Self Comes to Mind: Constructing the Conscious Brain. Pantheon.
  
  

• Torrance, S., & Froese, T. (2011). An inter-enactive approach to agency and participatory sense-making. Human Systems, 20(1), 1–21.