Archivist Alex Pezzati of the Penn Museum was on the verge of discarding a “curious collection” in the fall of 2016, when I invited him to present at a workshop I was then co-organizing with other members of the Penn Humanities Forum. “Translation beyond the Human” was our chosen theme, and I was hoping he could divert us with anecdotes about the history of early computing in anthropology.

He delivered more than anecdotes. Among other materials, Pezzati brought out a box of letter-sized Termatrex punchcards from a project devoted to the Neolithic settlement at Erbaba in southwestern Turkey. Jacques and Victoria Bordaz had directed the project during the 1960s and 70s, with consulting help from the Faunal Research Group in the Department of Anthropology at Columbia University. In addition to offering insights on the material and social structure of the prehistoric community that had lived at Erbaba, the group piloted a new approach to semi-automated information retrieval on the basis of the archaeological data collected there. The now forgotten Termatrex machine held out distinct advantages in their view. It was easy to use; it secured a place for expert theorizing through “feedback”; it had a “virtually unlimited memory”; it blended seamlessly with other computing platforms; and last, but not least, it was cheap.

These are examples of a complementary notched card indexing system showing a preference to mix manual and automatic approaches. Object illustrations were paired with textual descriptions and edge holes were notched to note relevant characteristics. Typically, one would skewer a stack of cards according to a query on a given topic, allowing relevant cards to fall out of the stack. Part of the appeal of the Termatrex system was that it could be easily integrated with such “conventional” approaches. It seems that members of the Erbaba projects used these cards as “item records” in analyses using the Termatrex machine. Penn Museum: Jacques Bordaz Papers: Archaeological Site Survey: Termatrex Key Card System — Samples and Instructions (1966, 1976).

Engineers at Jonker Business Machines, Inc. developed the Termatrex machine over the course of the 1950s, and it was released with great anticipation in 1960. As opposed to “conventional” data-processing systems, the Termatrex (like its offshoot, the Minimatrex) featured an “inverted” approach to information retrieval. That is, rather than being devoted to a single entity, its attributes assigned via punched holes (“item records”), each card was devoted to a single attribute, with punches pointing to those records in which they appeared (“characteristic records”). If conventional systems were to be preferred for record keeping and accounting, the inverted Termatrex was considered ideal for directed search purposes– in a word, science.

Illustration of the “basic principle” of the Termatrex systems. Jonker, “The New ‘Termatrex’ Line of I.R. Systems—The ‘Minimatrex’ Line of I. R. Systems,” American Documentarian 14 (1963): 277.

This functionality certainly appealed to members of the Bordaz team, who subjected the Termatrex to a “critical examination” in 1966. As they explained to readers of American Antiquity, searches could be performed by superimposing punched cards over a desktop light source. This allowed for structured queries: layering multiple attribute cards, light spots revealed “hits,” which could be traced back to an individual item (e.g. a person, location, or artifact) possessing all the attributes in question. While “unspecialized clerical help” was enlisted to transfer field data to the cards, trained archaeologists decided which attributes came in for analysis according to a well-honed theory. Evincing a philosophy somewhat different from that of Big Data today, these archaeologists were decidedly unimpressed by mere “fishing” expeditions.

That said, size did matter. Even though other semi-automatic systems were adaptable to the principles just described, Jacques and Victoria Bordaz preferred the Termatrex machine for its highly extensible memory. The size of the card alone dictated how many items might be considered at once, and with space to accommodate 10,000 items, a card measuring just 9.625” x 11.5” boasted “virtually unlimited” capacity. If and when a research project did reach this limit, cards were miniaturized to yield further storage space.

All in all, the Termatrex machine offered the Erbaba researchers a way to cope with a perceived data deluge without sacrificing the “art” of associative taxonomy in archaeology. As might be expected of those with deep commitments to the human sciences, the Bordaz couple and their interlocutors were particularly keen advocates of human expertise during this pivotal moment in the history of electronic computers. This archaeological example confirms recent histories which emphasize the highly specific ways in which various disciplines, including the human sciences, entered the age of computing.



Jon Agar, “What Difference did Computers Make?,” Social Studies of Science 36 (2006): 869-907.

Marcia Ascher and Robert Ascher, “Chronological Ordering by Computer,” American Anthropologist 65 (1963): 1045-52.

Jacques Bordaz and Victoria Bordaz, “A Critical Examination of Data Processing in Archaeology, with an Evaluation of a New Inverted Data System,” American Antiquity 31 (1966): 494-501.

Nathan Ensmenger, “The Digital Construction of Technology: Rethinking the History of Computers in Society,” Technology and Culture 53 (2012): 753-776.

Jean-Claude Gardin, “Cartes Perforées et Ordinateurs au Service de l’Archéologie,” La Nature 3331 (1962): 449-57.

Frederick Jonker, “The New ‘Termatrex’ Line of I.R. Systems—The ‘Minimatrex’ Line of I. R. Systems,” American Documentarian 14 (1963): 276-282.

University of Pennsylvania Museum Archive. Jacques Bordaz Papers; Erbaba Project; Material-Fauna-Correspondence (1969-1971).

University of Pennsylvania Museum Archive. Jacques Bordaz Papers: Archaeological Site Survey: Termatrex Key Card System — Samples and Instructions (1966, 1976).

Judith Kaplan: contributions / website /