Download csv file of dataset, last updated July 17, 2023.
A data-focused project brings with it the expectation of objective, clear, numeric data, but the hard facts remain elusive in this project, as they do in most data-driven endeavors. Humanistic data is subject to change over time, interpretation of compilers and researchers in the past and present. However, this does not mean our project is void of information, just that we must be flexible in our interpretation and willing to make adjustments to the data and our interpretations as new information becomes known to us. If you have information to share or alternative interpretation(s), please contact us.
Sources: Our data derives from Dutch East and West India Company archives and documents, which account for the movement and value of goods as they related to the Dutch economy, and which were produced from a Dutch perspective that justified territorial expansion and the enslavement of human beings in the name of trade, profit, and exploration. These documents are often dry and abstract, but it is imperative to us as scholars and as humans to provide reminders that these documents and this trade were and are not inert, neutral accounting. Transparency about the choices we have made in aggregating these data into digital form is a critically important responsibility for anyone working with colonial archives such as these.
Textile names: This project contains 300+ textile names, many of which are currently obsolete, which were recorded in cargo lists and ship invoices (Dutch: factuur). Early modern spelling irregularities, translation then and now across multiple languages and alphabets have produced many variations for each term. Deciding on authoritative spellings required choices, which we discuss in the individual entries for each textile in the glossary, and as a general rule, we have standardized to feature the spellings we see most often, with an aim to be legible to an Anglo-Dutch readership. Each textile glossary entry also includes variant terms and related terms. The textile names are specific to time and place, and it is our aim that when all the data is compiled, changes/patterns will become visible.
Textile measurements: The most common standard measurement of a textile in this data is ‘piece,’ meaning the size of the cloth when it is removed from the loom. Piece sizes depend on the width of the loom and the length of cloth that the loom can accommodate (which depend on local practices and equipment). While piece size should be relatively consistent regionally, unfortunately this information is inconsistent between and even within textile types. European merchants and consumers measured textiles by ell (Dutch el), a length of roughly 27 inches or 69 cm, which also varies regionally. A piece might measure 20 els by 1.5 els, or a textile can be sold in cut lengths.
Currencies and value: Our dataset indicates value using the currency of the Dutch records, the tripartite system of Dutch guilders, stuivers, and pennigen. Some VOC accounts use Indian (Indisch) guilders, which are approximately 7/10ths of a Dutch guilder; these have been normalized to Dutch guilders in the dataset. In colonial-occupied regions, value was determined by the relationship between commodities, rather than in currency—often goods were traded directly for other goods, or, when money was exchanged, it was not always in the same currencies and denominations indicated in the documents.
Geographies: In this dataset, we are considering the movement of a shipment of textiles overseas from one port to another. The textiles were manufactured, traded, and consumed inland from the ports, which is not (yet) reflected here. Each point of trade is ideally indicated by a specific port and its historically relevant region, but in many cases only the region was indicated by the source, either because the ship made several points of call along a coast line, or because the clerk saw no reason to record this information. In the downloaded dataset, you will find the port/region as indicated by the original document, but the data visualizations draw from names of ports/regions we’ve standardized for the sake of clarity and comparability. Our intention was to name places that are legible to readers today, but are historically relevant. This was a tricky undertaking and we had many difficult decisions to make, considering that from the 17th century to the present, these places have changed names, borders, and sovereignty. These changes are in large part due to European trade and colonization intervening in global exchange, and the legacies of colonialism still reverberate through these geographies. Within this messiness, a standard, however imperfect, allows our data visualization apps to make comparisons between regions and ports. It was also necessary to create regions that would be visible on the Textile Geographies app, which lose some historical accuracy in favor of visibility. Ports, like Jakarta (former Batavia) and the Cape of Good Hope, which were trading centers where goods were warehoused and transshipped, have been treated as entities separate from their broader regions, so that trade from Jakarta to elsewhere on Java can be seen. These decisions are not perfect, and we welcome any feedback from users about how we can best balance user needs (clarity of terms and visuals), historical accuracy, and machine-legibility of the data.
Dates: Dates in this dataset derive ultimately from the source documents, and have been organized into YYYY and YYYY-MM-DD format (two columns of data). At minimum, each exchange of cargo has a year of departure or arrival, and ideally a full date for departure and arrival. You can expect dates to vary by a few days depending on whether the document reflects the movement of the ship or the writing of the document. In the dataset, we have indicated the source of the date (archive, document, other dataset, our estimation, or NA for a ship that never arrived). The data visualizations draw from the calendar year as we understand it today (Jan.–Dec.), while bookkeeping practices measured the year differently (for instance, the Batavia ‘book year’ was Sept. 1–Aug. 31). The 1582 shift from Gregorian to Julian calendar (10 days; https://en.wikipedia.org/wiki/Gregorian_calendar) was adopted by different provinces of the Dutch Republic in 1582/1583 and 1700/1701. These 10 days do not meaningfully impact the data presented here.