How a stone tool is cut and polished

A flat stone tool starts as a block and reaches you through four operations: sawing to a slab, shaping the outline, grinding through a sequence of progressively finer abrasives, and polishing. Every stage removes the marks of the one before it, and the object in your hand records which stages were done properly. This is why two tools of identical material can look and feel entirely different.

Sawing

The block is reduced to a slab of roughly final thickness with a diamond blade — a steel disc with diamond grit bonded to its rim, run wet. Diamond is used because it is harder than everything, so one tool cuts every material, and water carries away heat and swarf.

Slab thickness is decided here, and it is the most consequential decision anyone makes about the object’s strength. A plate’s resistance to bending rises steeply with thickness, so the difference between a slightly generous slab and a slightly mean one is much larger than it looks — thickness dominates breakage, and the saw sets it.

Sawing also determines orientation. In a material with a fabric — fibrous nephrite, banded rock, a cleavable mineral — the cutter chooses the relationship between the plane of the slab and the plane of the structure, and this decides whether an internal weakness runs across the object or along it.

Shaping the outline

The slab is cut and ground to its silhouette, by hand against a wheel, with a saw and a template, or in volume production by machine.

The outline determines where the object will fail. Any concave notch — the bay of a heart profile, the neck of a comb form, a scalloped edge — is a stress concentrator, a point where stress bunches up under load and a crack preferentially begins. A simple convex outline is mechanically much better behaved than an elaborate one in the same material.

Machine-produced pieces are consistent and often have crisper geometry; hand-shaped pieces vary and sometimes carry flats where the wheel was held too long in one place. Neither is a statement about the material.

Grinding: the sequence that matters

This is the stage where quality is made or lost, and it is a sequence rather than a step. The piece is worked through progressively finer abrasives — a coarse grit to remove saw marks and establish form, then a series of steps each finer than the last, each one removing the scratch pattern left by its predecessor.

The rule is that a stage can only remove scratches coarser than the ones it makes itself. Skip a step, or cut one short, and its predecessor’s scratches survive into the polish where they cannot be removed without going back. This is exactly why a parallel scratch field shows on a finished surface: it is a coarse stage that a fine stage never erased.

Skipping steps is the standard economy in cheap production, because each stage costs time and the defects it leaves are invisible except at a grazing angle. It is the single most common corner cut in the trade.

Polishing

The final operation is not abrasion in the same sense. A polishing compound on a soft wheel or leather lap produces a surface flat at a much finer scale, partly by very fine abrasion and partly by other surface effects. Different compounds suit different materials: cerium oxide is standard for quartz and glass, chromium oxide and diamond pastes for other stones, and matching compound to material is part of the craft.

An aggregate cannot be polished perfectly flat, because its grains abrade at different rates — the undercutting that shows as faint relief. This is a property of the material, not a failure of the polisher.

Tumbling, and what it conceals

Volume production often uses barrel tumbling: pieces rotated for days in a barrel with abrasive grit and water, progressing to finer grits and a polishing charge.

Tumbling is cheap per piece and produces a characteristic result — softly rounded edges, slightly uneven contours, and a good overall shine with no crisp lines anywhere. It also disguises things. Crude shaping, small chips and coarse grinding marks all soften into the general roundness, so a tumbled finish carries less information about what happened earlier than a hand-finished one does.

WHAT YOU CAN TELL — from the workmanship

  · Parallel scratches under the polish
      → a grinding stage skipped or shortened
  · Softly rounded, slightly uneven edges, no crisp
      lines → barrel tumbled
  · Flats on what should be a continuous curve
      → hand-ground against a wheel
  · Mould seam, no grinding marks anywhere
      → cast, not cut. Glass or resin
  · Very thin section at a narrow waist
      → the saw and outline decided its lifespan
  · Whether the material is what it is sold as
      → NOT DETERMINABLE from workmanship. Fine
        finishing tells you nothing about species

That last row is the one worth holding onto. Workmanship and material are independent variables, and the market conflates them constantly: a beautifully finished object reads as a good stone, and a plain one reads as a poor stone, when the two facts have nothing to do with each other. Glass takes a superb polish. Fine nephrite can be finished badly.

Why the economics produce what they produce

Cutting stone is time on a machine with a human attending it, and the time is a large share of the cost of an inexpensive object — often larger than the rock, which for serpentine or quartzite is close to worthless in small quantity.

Two consequences follow directly.

Softer materials are preferred. Soft stone cuts and polishes faster, so it costs less to work. Hardness is a production cost as well as a property, and it biases the low end of the market toward softer material regardless of what any listing says.

Moulding beats cutting. A cast glass or resin piece skips sawing, shaping and most grinding entirely. When the finished appearance is what sells, casting is not a marginal saving but a different order of cost, and that is why imitations exist.

The limit

Workmanship is readable and species is not. You can tell that a piece was tumbled rather than hand-finished, that a grinding stage was skipped, that it was cast rather than cut — and none of that identifies the material.

The one genuinely load-bearing thing production tells you is about geometry: thickness and outline were decided in the workshop, and between them they predict how the object will fail far better than the mineral does. Everything else about it — species, treatment, origin — remains a question for instruments.