The link that a death makes

Ninety-seven links in this archive move an object because an owner died

Ninety-seven of the 2,033 ownership links recorded across the 291 objects in this archive transfer the object because its owner died. We reached that figure by counting every link whose event is an inheritance, a bequest or a descent, and by leaving out sales made by an estate, which are ordinary purchases that happen to have a dead man's name on the receipt. The 97 links are spread across 70 separate objects, so this is not a quirk of one long aristocratic chain. It is roughly one object in four.

A word about what that number is and what it isn't. The event vocabulary is ours: we assign it when we build a record, and a chain written by a museum registrar in 1890 does not arrive labelled. Where a source says only that an object passed to a son, we grade it as a descent and note the wording. So the count is a count of how this archive reads its own evidence, and a different editor would move a handful of links across the boundary. What we are confident about is the shape, because the shape is large and it repeats.

The interesting property of these links is not that they exist. Everyone knows objects are inherited. It is that they behave differently from every other kind of link in a provenance line, in two opposite directions at once, and the difference is measurable here.

A death-transfer holds an object more than three times as long as a purchase

A link created by a death holds the object for a median of 29 years, against 8 years for a link created by a purchase. We measured this on every link in the archive that carries both a start year and an end year we could read as a number, discarding anything longer than six centuries as a sign that the two ends were not really the same custody. The gap between the two figures is the largest of any pair in the table, and it is not close.

The reason is mundane and worth stating plainly. A purchase is made by someone who has decided to own the object, and who may decide otherwise in a few years. An inheritance is received by someone who did not choose it, who often cannot sell it without breaking a settlement or a trust, and who will in the ordinary course hold it until they die in turn. The market moves objects at the speed of intention. Families move them at the speed of mortality.

The last row is there for scale rather than for comparison. A gap is not a kind of ownership, it is the absence of a record, and its median of 110 years is a measure of our ignorance rather than of anyone's behaviour. We include it because a reader checking a chain should know what the units look like at both ends.

The price record goes dark for as long as the family holds

None of the 97 death-links in this archive records a price, because in none of them was a price paid. That is obvious when you say it and easy to forget when you are reading a chain, because a provenance line prints a purchase and an inheritance in the same typeface and at the same size. The consequence is that an object entering a family disappears from the price record for the whole length of the holding, and the median length of that holding is 29 years.

Velazquez's portrait of his assistant Juan de Pareja shows the full arc. Painted in Rome in 1650, it reached the Earls of Radnor at Longford Castle early in the nineteenth century and remained in that family for about 140 years. Then, on 27 November 1970, it was sold at Christie's for 2,310,000 pounds to the dealer Alec Wildenstein, almost tripling the previous world auction record and becoming the first painting to pass a million pounds. The Metropolitan Museum of Art acquired it the following year, and its credit line still records the funding scramble that took: purchase, Fletcher and Rogers Funds, and a bequest applied by exchange, supplemented by gifts from friends of the Museum, 1971.

Read as a price series, that object has two data points separated by fourteen decades. Anyone valuing it in 1969 had nothing recent to anchor against, and the 1970 result is usually told as a story about Velazquez or about the arrival of serious money. It is at least as much a story about a picture that had not been priced within living memory.

The proof of an inheritance is written to be proved

At the moment the price record goes dark, the documentary record usually gets better. This is the part that surprises people who assume a family holding is a documentary void. A will is drafted by a lawyer in the expectation that somebody will contest it. A probate inventory is compiled because a tax authority requires it and will audit it. A settlement or an entail names the object precisely so that a later heir cannot quietly dispose of it. These are adversarial documents, and adversarial documents are the ones that survive and the ones that hold up.

Compare that with the evidence for an ordinary purchase in the same centuries: a dealer's note, a line in a stockbook that may be a code, a receipt that names neither the object nor the buyer with any precision. We have written elsewhere in these pages about what a stockbook can and cannot prove. The asymmetry runs the other way from what most people expect, and it has a practical consequence for anyone checking a chain. If a provenance line stalls at a family holding, the right next step is usually the probate registry rather than the auction archive.

The British Museum's record for the Lycurgus Cup shows the institutional version of this. Its acquisition note names Baron Lionel Nathan de Rothschild, who died in 1879, as the man who originally acquired the cup, and Victor, 3rd Lord Rothschild, as the man it was purchased from in 1958. The museum documents the two ends of a roughly eighty-year family holding and does not attempt the middle, because the middle is three deaths and no transactions. That is honest cataloguing. It is also a map of exactly where a researcher would have to go.

The longest holdings often end in the largest prices, and that is not the same as causing them

Several of the longest family holdings in this archive end in a sale that broke a record. The Codex Leicester is the extreme case: The Art Newspaper, writing in 1994, recorded that the manuscript had been in the possession of the Earl of Leicester for 260 years before Christie's sold it in London on 12 December 1980 for 2 million pounds excluding premium. Thomas Coke had acquired it in Italy around 1717 and had it copied in Florence, and the Museo Galileo's account has it in his London residence through the middle of the century, accessible to English scientists. Two and a half centuries of continuous ownership, one price at each end.

It is tempting to run that observation into a rule, and we want to be careful here, because the archive cannot support the strong version. Objects that stay in one family for two centuries are not a random sample. They are disproportionately the objects that were worth keeping, which is to say the good ones, and a good object would command a high price whatever its recent ownership history. The absence of comparables may well add something on the day. We cannot separate those two effects with 291 records, and we are not going to pretend otherwise.

What the archive does support is narrower and still useful. When an object emerges from a long family holding, the person valuing it is working without a recent comparable for that specific thing, and has to reason from the cohort instead. That is a different and harder estimation problem than pricing an object that changed hands eight years ago, and it is worth knowing which of the two you are looking at before you read an estimate.

What the death link does not tell you

A death-link is silent about condition, location and use, and those silences can run for decades. The object was somewhere for 29 median years, in some state of repair, possibly on a wall in a damp house and possibly in a strongroom. A purchase at least fixes a date on which somebody looked at the thing and formed a view about it. An inheritance fixes nothing except that the previous owner stopped being alive.

It is also the link where a chain most often quietly goes wrong. Families divide, names change on marriage, an object moves between houses that are all called by the family name, and a later cataloguer collapses three generations into a single line reading by descent. Those two words can conceal a sale to a cousin, a division under a will that separated a set, or a decade in a dealer's hands that nobody wrote down. When we grade a link as disputed in this archive, a fair share of the time the disputed thing is a descent that somebody has smoothed.

So the practical reading is not that inheritance is good evidence and purchase is bad. It is that they fail in opposite ways, and that a chain made mostly of one kind should be checked differently from a chain made mostly of the other. A run of purchases can be checked against sale records and stockbooks. A run of descents has to be checked against wills, and if nobody has been to the probate registry, the run has not really been checked at all.

Why this archive is free to read

We do this for everyone who loves art. The people who have studied it for years, and the people who just stopped to look. Iconotheca exists so that what we learn belongs to all of them.

Nobody pays us for this. No ads, no sponsors, no commercial interest. These works belong to all of us, and we would rather share them with as many people as we can than keep them to ourselves.

If it gave you something today, tell us to keep going. Follow us, leave a like, or write a positive comment. We read every one, and they are what keeps us going.