Condition Reports and What They Leave Out

The houses have already told you what a report is worth

The strongest statements about the limits of a condition report are in the conditions of sale, written by the houses themselves. Sotheby's provides that all lots are offered for sale in the condition they are in at the time of the sale, that many are of an age and type such that they are not in perfect condition, and that catalogue descriptions and condition reports may refer to imperfections while the lots may have other faults not expressly referred to.

The next clause is blunter. Information provided regarding any lot, including estimates and condition reports, is a matter of opinion only and not a representation of fact. The disclaimer of warranties then states that all lots are offered for sale AS IS, without any guarantee, representation or warranty, and that no statement anywhere, whether oral or written, will be deemed such a warranty or assumption of liability. The buyers' guide adds the clause that does the most work: the absence of reference to the condition of a lot in the catalogue description does not imply that the lot is in good condition, working order, free from faults, imperfections, restoration or repair.

Christie's language runs parallel and adds an admission about staffing. Its descriptions and condition reports are its opinion and not to be relied on as a statement of fact; reports are provided free of charge as a convenience and are for guidance only; and they may not refer to all faults, inherent defects, restoration, alteration or adaptation because Christie's staff are not professional restorers or conservators. The same section states that Christie's does not carry out in-depth research of the sort carried out by professional historians and scholars.

Lining is routine, and it is irreversible

The conventional vocabulary is understated, and the conservation literature supplies what the catalogue phrasing omits. The Getty Conservation Institute's Conserving Canvas glossary defines lining as adhering a fabric or solid support to the reverse of a painting where the original support has degraded, and relining as the lining of a canvas that has been lined before, involving removal of the old lining canvas and adhesive.

Stephen Hackney, writing for Tate, states the consequence that a condition report will describe in one word. We can take off an old lining, he writes, but it is unlikely that we can repair flattened impasto or emphasised canvas weave. Of glue lining specifically: impasto was squashed, the canvas texture reinforced in the paint. He adds the general principle that a truly reversible adhesive or consolidant is by definition not possible, and dates the change in professional opinion to the Greenwich Comparative Lining Conference of 1974.

The Getty glossary has a term for the visible result, moating: any raised portion of paint, such as impasto, can be pressed down level with the rest of the surface, creating a small hollow all around it. A picture described as lined has undergone a permanent alteration to its surface, and the word carries no indication of how much was lost.

Cradling was standard practice and it caused the damage it prevented

For panels the equivalent term is cradling: fixing a lattice of wooden slats to the reverse to control warping. A study published in Image Processing On Line in 2017 records that until about 1950 cradling was common practice for the restoration and preservation of paintings on wooden panel, and describes the procedure as permanently fixing hardwood slats or lattices to the back of a thinned panel.

The National Gallery in London has published what that meant for one object. Treating a three-board poplar altarpiece by Giovanni Martini da Udine in the nineteenth century involved thinning the panel from four or five centimetres to just five millimetres and flattening it before the cradle was glued on. The museum's assessment is unequivocal: while the cradle did structurally support the altarpiece, it also restricted the natural movements and contractions of the wooden panel which, over time, led to splits, distortion and cracks. The combined treatment of thinning the panel and attaching an inflexible cradle was probably the most damaging structural intervention this altarpiece has undergone.

A catalogue entry recording that a panel is cradled has therefore recorded that roughly nine tenths of the original support was planed away and that the remedy is a known cause of the condition it was meant to prevent. That is a great deal of information compressed into one adjective.

The crack vocabulary is precise, and catalogues use it loosely

Conservation bodies define the terms narrowly. The Smithsonian's Museum Conservation Institute gives craquelure as a pattern of cracks that develops on the surface of a painting as a result of the natural drying and ageing of the paint film, and distinguishes inpainting, which is paint applied over losses only, from retouching, the broader work done to replace areas of loss or damage.

The Getty glossary separates the failure modes. Drying cracks, described as vehicular, are caused by the failure of the film to withstand its own contraction during drying or by the artist's incorrect use of paint. Cupping is islands of aged paint bounded by cracks with upward curving edges forming saucer shapes, which often draw a less stiff canvas support with them. Flaking is the breaking away or detachment of paint and ground layers from the support. Cleavage is separation between paint layers, or between paint and ground, or between ground and support.

Read against those definitions, catalogue phrasing is doing considerable work. Craquelure consistent with age is normal and expected. Cupping and flaking describe a picture that is actively losing its surface, and they describe processes rather than states. No published guidance quantifies the phrase 'some restoration', which can cover a few retouched losses or an area substantially in modern paint; the nearest thing to an official acknowledgement of the gap is Christie's own admission that its staff are not conservators.

Ultraviolet shows a great deal, and conceals more than is assumed

Under ultraviolet light traditional natural resin varnishes such as mastic and dammar give off a characteristic greenish yellow. The National Gallery explains the diagnostic effect: retouchings applied over a layer of varnish show up clearly because they prevent the ultraviolet from reaching the varnish, appearing as dark patches in an otherwise fluorescent area. A few minutes with a torch will map the distribution of restoration faster than several paragraphs of prose, which is why viewing in person still matters.

The National Park Service's conservation guidance sets out the failure modes, and they are numerous. Areas of retouching may occur over early varnish layers and under later ones, in which case they may carry a fluorescent haze over them and read as original. A lack of fluorescence does not necessarily indicate the absence of a varnish. Superficial dirt and grime may obscure fluorescence significantly. Many synthetic resins do not fluoresce at all. The companion guidance adds the warning that matters most here: avoid drawing conclusions about an object based on ultraviolet-induced visible fluorescence alone, and this caveat is especially true with regard to issues of authenticity.

The practice is also less formal than outsiders expect. In Thwaytes v Sotheby's the court recorded that ultraviolet examination is used particularly to reveal retouchings or restorations made in the last eighty to a hundred years, that these show up as purple patches, that specialists do it as a matter of course using a little torch, and that it is not something that usually generates an image. The National Gallery describes retouchings as reading dark and the judgment describes them as purple, which is a fair indication of how much the appearance depends on the varnish, the retouching medium and the lamp.

What condition costs, and where it fuses with attribution

Loss of value has been quantified in litigation, though not standardised. When Picasso's Le Marin was torn at Christie's in New York in 2018, appraisers appointed by the house concluded the painting was worth USD 100 million and had suffered a loss of value of 20 per cent, or USD 20 million, based on the extent of the physical damage and the accompanying reputational damage. The insurer settled at USD 18.74 million and Christie's paid conservation costs of USD 487,625. For Le Reve, damaged by its owner in 2006, restoration cost about USD 90,000 while the owner submitted a sworn statement of a USD 54 million loss of value against a cancelled USD 139 million sale.

The chartered surveyor who set out those cases states the position plainly: there remains no objective, uniform and definitive guide for measuring the loss of value factor. She adds two rules of thumb that are worth more than most catalogue prose: the first instance of damage always results in the greater loss of value, and earlier methods of lining or relining a canvas customarily produced a loss of impasto.

At the top of the market condition stops being a separate question from attribution. Dianne Dwyer Modestini's own published account of the Salvator Mundi records that much of the original background was scraped off including the upper part of the hair, that the area of the chin, neck and breast was harshly cleaned with strong solvents, that the panel had separated, and that little of the original design survived in the knotwork, which is largely later repaint. Her opening sentence is that the restoration of a badly damaged painting can never be satisfactory. Vincent Delieuvin's summary of the specialists the Louvre consulted shows where that leads: they pointed to passages worthy of Leonardo alongside surprisingly weak details where they had difficulty distinguishing between repainting during restoration and the awkward intervention of an assistant. Condition had become inseparable from the attribution question on the most expensive picture ever sold.