This diagram presents embroidery as a structured language system composed of three interdependent layers: Structure, Path, and Tension. Structure represents the visible form, Path defines the generative trajectory of stitching, and Tension acts as the governing mechanism. On the left, structural typologies are classified into Continuous, Interlaced, Radial, and Non-local systems based on Canvas studies. On the right, an evolutionary progression is illustrated, showing the transition from non-local tension systems to interlaced structures and finally to radial stabilization. The diagram establishes embroidery as a rule-based generative system rather than a purely manual craft.
This diagram presents embroidery as a structured language system composed of three interdependent layers: structure, path, and tension. Structural variations emerge from different tension configurations and path behaviors.
圖 2. 三種刺繡結構系統的比較。
Embroidery systems differ fundamentally in how tension is distributed and controlled.
This diagram compares three fundamental embroidery systems: non-local (Canvas 215), interlaced (Canvas 152), and radial (Canvas 129). The comparison is structured across three layers: structure, path, and tension. Each system demonstrates a distinct configuration of generative logic and force distribution, revealing that embroidery structures are governed by underlying tension systems rather than surface patterns alone.
Figure 3. Path–Tension Relationship in Embroidery Systems.
Different paths generate different tension fields, which determine structural outcomes.
This diagram illustrates how different path configurations generate distinct tension systems. In the non-local system (Canvas 215), long-distance jump paths produce non-local tension fields. In the interlaced system (Canvas 152), crossing paths distribute tension across local interactions. In the radial system (Canvas 129), continuous paths generate centralized tension. The comparison demonstrates that tension is not an independent factor, but is directly determined by path behavior.
introduction
Embroidery has traditionally been understood as a manual craft defined by stitch techniques and visual patterns. However, this study proposes a shift in perspective: embroidery is not merely a collection of stitches, but a structured language system. By analyzing a series of controlled Canvas studies, it becomes evident that embroidery structures emerge from underlying generative mechanisms rather than surface-level repetition. This research introduces a three-layer model—Structure, Path, and Tension—to redefine embroidery in computational and structural terms.
Theoretical Framework
The proposed framework consists of three interdependent layers:
(1) Structure: the visible geometric or surface configuration formed by stitching. (2) Path: the trajectory of the thread, representing the generative process. (3) Tension: the governing force system that determines how the structure is stabilized.
Among these, tension plays the most critical role. While structure is observable and path can be reconstructed, tension operates as a hidden mechanism that controls both formation and stability.
Structural Typology
Based on empirical analysis of Canvas-based embroidery systems, four structural typologies are identified:
- Continuous Systems: generated through uninterrupted path progression. Interlaced Systems: formed through crossing and interlocking paths. Radial Systems: characterized by centralized anchoring and outward distribution. - Non-local Systems: defined by long-distance tension jumps across the structure.
These typologies demonstrate that embroidery is governed by distinct generative logics rather than stylistic variation alone.
Evolution Model
The study further identifies a tension-based evolutionary sequence:
Non-local → Interlaced → Radial Non-local systems exhibit high degrees of freedom and discontinuity, where tension spans large distances. Interlaced systems introduce structural coherence through crossing paths. Radial systems achieve stabilization by centralizing tension and distributing it symmetrically.
This progression suggests that embroidery structures evolve toward increasing stability and predictability.
Conclusion
This research redefines embroidery as a structured and generative system governed by tension dynamics.
The proposed model bridges traditional textile practices with computational logic, opening new possibilities for AI-based analysis and generative design.
Rather than viewing embroidery as a static craft, it should be understood as a dynamic system of rules, forces, and transformations.
