One hundred and fifty millimetres. That is how far the stone treads of the Geometric Staircase at St Paul’s Cathedral reach into the wall, and it is nowhere near enough to hold them. Eighty eight steps rise from the floor of the south west tower to the triforium with no central column, no stringer and no visible support of any kind, and they have done so since 1710. The staircase is almost always described as cantilevered. It is not, quite, and the gap between what floating stairs appear to do and what they actually do is the oldest unresolved question in stair design.

> The floating stair is not a contemporary invention. It is a Renaissance structural argument that modern fabrication finally made affordable.

Palladio’s Void and the First Unsupported Stair

The idea arrives with Andrea Palladio. Around 1570, at the Convento della Carità in Venice, he built what is generally recorded as the first cantilevered stone staircase: solid treads stacked one on another, each secured at one end into the wall, with the centre of the stairwell left empty.

In I quattro libri dell’architettura he explained why. Staircases void in the middle succeed very well, he wrote, because they take light from above, and because anyone already climbing them can be seen. Both arguments are spatial rather than structural, and both are still the reasons a client asks for a floating stair four and a half centuries later. Daylight and sightlines. Nothing about the brief has changed.

1616: What Nicholas Stone Added at Greenwich

Inigo Jones saw the Carità stair, sketched it, and brought the idea to England. Work on the Queen’s House at Greenwich began in 1616 and the building was completed around 1635. The Tulip Stairs are generally recognised as the first centrally unsupported geometric staircase in Britain.

The decisive contribution was not the architect’s. The mason Nicholas Stone introduced a rebate, a grooved interlock worked along the underside of each tread so that steps engaged one another rather than relying on simple overlap. That detail converted a stack of stones into a system. It spread stresses through the whole flight instead of concentrating them at each wall pocket, and it is why Christopher Wren could attempt eighty eight treads at St Paul’s seventy five years later with the same 150 mm of embedment.

> The rebate did not make each tread stronger. It made the flight behave as one object.

The method travelled quickly. By the eighteenth and nineteenth centuries it was common enough in Scotland to have its own name, the pencheck stair, used in Edinburgh townhouses and tenements and in country houses by Robert Adam.

Why the Cantilever Was Never Really a Cantilever

A cantilever is a beam anchored at one end and free at the other. Applied honestly to a stone tread with 150 mm in the wall, the arithmetic fails. The embedment is too shallow and the stone too brittle in tension to resist the bending moment generated by a person standing at the outer edge.

What actually happens is that each tread also bears on the one beneath it. Load and torsion are passed down the flight and delivered to the floor, so the staircase is a partially supported assembly behaving as a single stiffened element. Historians of the technique have long noted that the cantilever label is inaccurate.

The ambiguity never went away. Contrary to the way these staircases are marketed, a significant share of what the contemporary market sells as cantilevered are hybrids: treads supported at the wall and stabilised by a second, concealed load path. The visual claim and the structural claim have been diverging for 450 years, which matters because clients buy the visual claim and engineers inherit the structural one.

How the Modern Floating Stair Puts Steel Where the Stone Was

Contemporary practice resolves the same problem with four support strategies rather than one. A wall anchored cantilever runs steel arms, flat bars or plates deep into a structural wall. A single central spine carries treads sideways from one concealed beam. Twin stringers hide inside the depth of the treads. Suspension hangs the flight from the structure overhead.

Only the first reproduces the historic condition, and it carries the historic constraint with it. The wall has to resist a bending moment at every tread, which in practice means reinforced concrete, solid masonry, or a steel frame built into the partition. A conventional stud wall cannot, and the loads continue downward, so the foundation beneath has to accept the additional load and the overturning that comes with it. This is why the support strategy, the wall capacity and the embedment depth are decided before anything visible is drawn, and why the way a floating stair is supported and detailed determines far more of the result than the choice of tread material.

Material follows structure, and it changes the dead load more than the appearance suggests. Hardwood cladding on a steel core is light. Reinforced concrete treads of 80 to 110 mm can approach 150 kg each. Cast glass is specified from 38 to 100 mm because its span capacity comes from section rather than reinforcement. Two geometrically identical stairs in different materials can require entirely different walls.

What Floating Stairs Are Judged On, and No Code Measures

Codes settle the geometry. The 2021 International Residential Code permits open risers while requiring that no opening more than 30 inches above the floor below pass a 4 inch sphere. The International Building Code begins from the opposite premise, requiring solid risers and then listing exceptions. Treads are designed for 100 psf or a 300 pound concentrated load, guards for 50 pounds per linear foot and a 200 pound point load.

None of that addresses the thing occupants actually judge. Concentrating the load path into a few concealed connections makes these stairs far more sensitive to deflection and vibration than a flight sitting on two visible stringers, and a staircase that moves underfoot reads as unsafe long before it is anywhere near its structural limit. Designers answer with deeper sections, stiffer geometry, composite treads, concealed bracing, and occasionally dampers or neoprene bearings at the connections.

> Nicholas Stone solved stiffness with a groove. The problem has not changed. Only the vocabulary has.

The rebate at Greenwich was a stiffness detail invented four centuries before anyone could calculate it. Its modern descendants are an embedment depth and a deflection limit, written into a drawing rather than cut into stone. The floating stair has always been an argument that structure can disappear without being absent, and every generation has had to make that argument again with the tools available to it.