To describe the way a fabric feels in the hand — its stiffness, its springiness, its silkiness — we have always leaned on subjective words. In the textile trade, those words carry immense weight, but they drift between cultures, individuals, and even moods. The Kawabata Evaluation System (KES) was built to arrest that drift, to anchor fabric hand in numbers that anyone, anywhere, could reproduce. It does this by measuring a fabric’s low-stress mechanical and surface properties through four specialized instruments, then translating that data into standardized Japanese hand descriptors via linear regression equations. I have been pulling apart that translation layer — the primary hand equations, their 16 input parameters, and the sensory logic that binds a bending moment to *koshi* — and what follows is the anatomy I have reconstructed.
The KES-F instrument suite captures a full mechanical portrait of a fabric through 16 parameters, each obtained under minute deformations that mimic a hand’s light touch. Tensile properties are captured by the linearity of the load-extension curve (LT), the tensile energy (WT), and tensile resilience (RT). Bending is described by bending rigidity (B) and the hysteresis of the bending moment (2HB), which separates elastic stiffness from energy lost when a fold is reversed. Shear behaviour comes in three numbers: shear stiffness (G), hysteresis at 0.5° of shear angle (2HG), and hysteresis at 5° (2HG5), the latter intended to reflect the frictional component during larger shear displacements. Compression yields linearity (LC), compression energy (WC), and resilience (RC), while surface character is split into the coefficient of friction (MIU), its mean deviation (MMD, which captures variation in friction), and the geometrical roughness (SMD). Completing the set are fabric weight (W) and thickness (T) under a standard pressure. This is not an arbitrary list; each number corresponds to a particular way a fabric yields, resists, or slides against skin and fingers.
The primary hand equations sit at the system’s heart. They are multiple linear regressions, with each hand value — *koshi* (stiffness), *numeri* (smoothness), *fukurami* (fullness and softness), *shari* (crispness), and several others — expressed as a linear combination of the logarithms of the 16 KES parameters. The logarithmic transform accounts for the roughly Weber-Fechner-like way human perception compresses mechanical magnitudes. A typical equation for *koshi* in men’s suiting fabrics looks like:
Koshi = β₀ + β₁ log B + β₂ log 2HB + β₃ log G + β₄ log 2HG + … + β₁₆ log T
The coefficients are determined separately for each hand descriptor and each fabric category (men’s suiting, women’s thin dress, etc.) by regressing expert panel ratings against the mechanical measurements. These coefficients are not universal constants; they encode the collective sensory judgment of Japanese textile specialists who first defined the hand terms. This is both the system’s strength — it embeds a reproducible, culturally grounded standard — and its limitation: apply it outside that cultural and fabric category context, and the predictions can stray.
What makes this more than a statistical trick is the traceable sensory logic woven into the coefficients. Bending rigidity B and its hysteresis 2HB dominate the equations for *koshi*, giving weight to the resistance a fabric offers when folded between the fingers. Shear stiffness G and shear hysteresis terms contribute strongly to *koshi* and *shari*, capturing the crisp, papery feel of a fabric that resists skewing and recovers quickly. Compression energy WC and its linearity LC feed directly into *fukurami*, because fullness is perceived largely through the volume of air trapped in a fabric and the springiness of that trapped space. Surface friction MIU and its variation MMD steer *numeri*: high smoothness comes from low, uniform friction, while roughness SMD opposes that sensation. These aren’t one-to-one mappings — every hand value draws on several parameters — but the dominant pathways align well with everyday tactile intuition.
In my own hand-taxonomy project, I treat the KES primary hand equations as a candidate “sensory mapping” that links physical measurements to a structured vocabulary. The 16 parameters form an objective fabric fingerprint; the equations translate that fingerprint into a profile of perceived qualities. By reversing the logic — asking which mechanical changes would raise *fukurami* without stiffening *koshi*, or how to design a finishing process that shifts *shari* — I can begin to predict how a textile will feel from its construction and finishing specs. The equations are not the destination, but they are a rare, quantifiable anchor in a field that otherwise drifts on subjective language.
For Vivina, where we build fabric-first intelligence for independent designers, this kind of anchoring matters. A sourcing manager who can say “I need a jersey with *koshi* above 6 and *numeri* below 3” is speaking a language that can be connected directly to KES measurements, mill specs, and objective quality control. The Kawabata system taught us that the hand of a fabric is not a mystery; it’s a fingerprint we can learn to read.
Comments
No comments yet — be the first.