A baby is found alive after five days, still inside her babysitter's arms

When the ground moved under Port-au-Prince on the evening of 12 January 2010, a woman called Marie Ange was singing an infant named Jenny to sleep in a fifth floor bedroom on Rue Canape Vert. Jenny's mother Nadine was in the kitchen. The shaking threw Nadine down twice, and before she could reach the bedroom the floor opened and she went down with the building, buried and pinned. Marie Ange had wrapped herself around the child as the ceiling came in. She was killed, by a blow or a broken neck or the weight above her. Jenny took a fractured skull and several broken ribs and lived. Trapped in the dark and choking on concrete dust, Nadine sang a hymn to Erzilie, the spirit who protects mothers and children, because Haitians hold that a prayer sung counts twice. Rescuers reached the child five days later. She was still inside the dead woman's arms.

Human rights organisations inside and outside the Dominican Republic argued that stripping citizenship by circular contradicted the constitution itself. In 2010 the country adopted a new constitution which says that children of foreigners in transit or living illegally in the territory are not citizens by birth. The founding document was thereby brought into line with the migration law of 2004 and with what registry offices had already been doing. The same constitution created a Constitutional Court of thirteen judges charged with enforcing constitutional justice. Because the audit of the civil registry ran backwards, a birth recorded in 1983 could be suspended under a rule written in 2010, and the government also kept the Book of Foreigners for Dominicans of Haitian descent who could not prove their parents were legally resident. Generations of children born in the country are entered in it at birth.

Philip DeArmond Curtin died in 2009. Eltis and Richardson name his work as the foundation their atlas of the transatlantic slave trade rests on. Before Curtin, the scale of the traffic was handled with round figures that travelled by repetition, some running to twenty million and more, none of them tied to a document anyone could inspect. Curtin went to the shipping records and produced an estimate with its method attached, so that every scholar afterward could argue with his total by arguing with his sources. That is what made the field cumulative instead of rhetorical. The voyage database behind the atlas, with its record of individual ships, ports, owners and mortality, exists because he established that counting was possible and that a count could be corrected. I keep this on a Haitian timeline because Saint-Domingue took in hundreds of thousands of captives in its last fifty years, and the size of that traffic explains both why the colony was rich and why it went up.

Windward and Leeward are administrative terms European powers applied to the Lesser Antilles in the seventeenth century, and the boundary between them was drawn to suit shipping and colonial jurisdiction. Colonial usage puts the divide at Dominica. Many archaeologists put it at Guadeloupe and count northward from there. The disagreement is small and it exposes something larger. These names were never meant to describe how the islands were inhabited before Europeans arrived, and the societies living there were organised around canoe routes, marriage, exchange and language, none of which respect a line drawn for administrative convenience. Using the colonial vocabulary makes island groups look separate and fixed where the evidence shows movement and mixing along the whole chain. Keegan and Hofman argue for an archipelagic view instead, treating the sea as the thing that joined people rather than the thing that divided them. The map you are handed decides in advance what you can see.