The Ishango bone was unearthed in 1950 by Jean de Heinzelin, a geologist of the Belgian Institute of Natural Sciences, on the Congolese shore of Lake Edward. It is a fibula roughly 10 cm long, topped with a fragment of quartz, and it is nearly 20,000 years old. Data
Three columns of notches run along it. The museum describes the groupings as three and six, four and eight, and five and ten — that is, 2×3, 2×4 and 2×5.
That is the evidence. Everything after it is interpretation.
The doubling is on the bone. What the doubling is has been argued for seventy years without resolution. It has been read as a deliberate arithmetic table; as a lunar count; as a tally of something perishable; as a pattern that emerged from use rather than from intention. Each reading has serious people behind it and none has closed the question.
This chapter takes no side and has no standing to. Its interest is one step to the left of the dispute:
A dispute over what marks on an object mean is settled, when it is settled at all, by people handling the object — under magnification, alongside comparable material, with the excavation context to hand. For seventy years that has required a journey to Brussels. The scholars with the deepest contextual knowledge of the Upper Semliki are the least likely to be able to make it. Custody is not a neutral fact about where a thing is kept. It is a distribution of who gets to argue.
Chapter 1 separated two facts that look identical in the public record: where a commodity went, and what a producer is no longer free to sell. Heritage has the same pair, and the same gap.
H is approximately known and widely quoted. The 2018 Sarr–Savoy report, commissioned by the French presidency, estimated that roughly 90% of sub-Saharan Africa's material cultural heritage sits in Western collections; it singled out the Musée du quai Branly, holding some 70,000 African objects, for the quality of its provenance archive. Data
I and R are published nowhere. Museums catalogue what they hold. They do not aggregate the legal instrument under which they hold it, and no registry does it for them. That is the same shape as Chapter 1: location is visible, and the term structure that decides reversibility is not.
Section 5 of the British Museum Act 1963 sets out when trustees may dispose of an object. The grounds are: that it is a duplicate; that it is printed matter made after 1850 of which a photographic copy is held; or that in the trustees' opinion it is “unfit to be retained in the collections of the Museum and can be disposed of without detriment to the interests of students”. A further provision covers objects rendered useless by damage, deterioration or infestation.
Return to a place of origin is not among the grounds. A contested object is not a duplicate, is not post-1850 printed matter, and is emphatically not unfit to be retained. The holder's option to return has been legislated away, and the statute names no term at which it revives. In Chapter 1's vocabulary that is an encumbrance with T unbounded — the limit case of the instrument that chapter describes.
Belgium did something different, and it deserves to be stated precisely rather than gestured at. The law of 30 June 2022 carved an exception out of a doctrine that had made state property inalienable, imprescriptible and unseizable, allowing colonial-era assets to be moved into a restitutable category and title transferred free of charge to a state of origin.
Its scope is bounded. The asset must originate in the territory of the origin state; it must have been acquired between the Act of the Berlin Conference (26 February 1885) and that state's independence; it must belong to a Belgian federal institution; and human remains and archives are excluded. The route runs in three steps: examination by a Joint Scientific Commission for acquisition under duress or violence, a bilateral treaty between Belgium and the origin state defining how that examination is conducted, and only then a transfer of ownership — with physical return happening at the origin state's request.
The United Kingdom's instrument forecloses. Belgium's opens a door and puts a treaty in the doorway. Neither is the absence of an instrument, and a custody ledger that recorded only where things are would score the two identically.
The bone was excavated in 1950. Congolese independence came in 1960. The Berlin Conference Act is 1885. On dates alone the object falls inside the 2022 law's window, and it is held by a Belgian federal scientific institution.
Whether a palaeolithic object excavated by a geologist, held in a natural sciences institution, falls within the scope of a statute framed around colonial cultural collections is not settled by anything in this chapter, and the author is not qualified to settle it. It is a question for Belgian public law and for the Joint Scientific Commission, and it is precisely the kind of question that has no published answer because nobody has assembled the instrument-level ledger that would make the pattern visible. A second precondition is equally concrete: the route runs through a bilateral treaty, and a treaty that has not been concluded is a route that does not yet exist. Open
Stating the gap is the contribution. This chapter does not claim the bone should be somewhere else; it claims that the question of what governs its location has an answer nobody has written down, and that the answer is knowable.
Chapter 1 argued that a long-horizon supply commitment is the sale of an option — the credible ability to transact with somebody else — and that no accounting standard records the sale. Custody is the same relation with the parties fixed by history rather than by signature.
A holding institution acquires a bundle of options: to keep, to lend, to conserve, to display, to permit study, to refuse it. The place of origin loses the corresponding bundle. The asymmetry that matters is not in the objects but in the documentation of the options: the holder's are enumerated in statute, with procedures and appeal routes, while the originating community's are enumerated nowhere and must be asserted diplomatically, one object at a time, against a party whose position is written into law.
Foreclosure compounds here too. Each generation of scholarship conducted where the objects are produces expertise, apparatus and comparative collections in the same place, which is then a further reason the objects should stay. The argument from research capacity is not dishonest. It is self-fulfilling, and the ledger would make that visible.
It was researched from Newark, New Jersey, from published sources, without one conversation with anyone at the holding institutions or in the country the object came out of. Ernesto Sirolli’s volunteers grew magnificent tomatoes in the Zambezi valley and lost the crop to two hundred hippos that every local farmer knew about; asked why nobody had mentioned it, the answer was “you never asked”. Being careful about what one does not know is not the same as having asked. This chapter should be corrected by people who were not consulted about it, and the door for that is open.