Künstliche Intelligenz zwischen Spezialisierung und Universalisierung

Cologne Media Conversations Künstliche Intelligenz zwischen Spezialisierung und Universalisierung

Auf welche Zukunft in welchen neuen Umgebungen muss sich die Medien- und Kulturtechnikforschung angesichts der rezenten Konjunktur von Künstlicher Intelligenz ausrichten?

Der Vortrag kommt hierzu auf eine grundsätzliche Frage von Marcel Mauss zur Konstitution von Techniken zwischen speziellen und allgemeinen Zwecken zurück. Er bietet, erstens, eine kurze Lagebeschreibung zur gegenwärtigen Entwicklung des maschinellen Lernens und neuerer KI zwischen Spezialisierung und Universalisierung.

Kern meiner Argumentation ist, zweitens, eine natur- und kulturtechnische Genealogie. Sie fragt nach der historischen Wiederkehr konnektionistischer Praktiken, Algorithmen, Denk- und Handlungsstile als Bedingung der aktuellen KI-Technologien. Wie konnten aus biologischen Neuronen artifizielle neuronale Netze werden, die als infrastrukturelle Grundierung Künstlicher Intelligenz fungieren?

Auf dem Spiel stehen dabei, drittens, die Medien selbst. Sie werden aufgrund der Naturalisierung und Habitualisierung von konnektionistischer KI zu statistisch operationalisierten, datenintensiven Netzwerken.

Künstliche Intelligenz zwischen Spezialisierung und Universalisierung:
Eine Wette mit Marcel Mauss
Cologne Media Conversations no. 55

Dienstag, 21.01.2025, 12–13.30 Uhr
Universität zu Köln, Hauptgebäude, Hörsaal VIII

Small Forms in Circulation

I am happy to give a keynote at this conference of the Research Training Group “The Literary and Epistemic History of Small Forms”. It is called Infrastructures and/as Environments: Practices and Ecologies of Circulation. Feel free to join us and think collectively about “Small Forms in Circulation” at Humboldt University on November 28-30, 2024.

Conference poster Small Forms in Circulation

Testing the Social Network

Blogging gets a bit harder throughout the start of the academic term. This is why I return to this series of image from The Connectivity of Things just now. Chapter 6 of the book tells a visual history of the network diagram. It follows the early history of graph theory in mathematics and chemistry. But the payoff here is a true crime history of sociology. While the name of Jacob Levy Moreno might sound familiar, the name of social psychologist Helen Hall Jennings is largely unknown even to scholars of social networks. Yet, it was Jennings who provided Moreno with a set of innovative methods to gather data on social relations. The sociometric test is what allowed for a visual formalization of a given network’s data. While Jennings still had a chapter of her own in the 1934 first edition of Moreno’s book Who Shall Survive, her work was hidden in a list of (numerous) references in the second edition of 1953. This is why I have started to reference both versions as written by Moreno and Jennings.

In the 1953 second edition, Moreno also made strong claims regarding “his” visual method of network visualization. A lot of the sociometric diagrams have been drawn by women though. Just consider an image like The Lady Bountiful, which was drawn by student assistant Pauline Lee in 1938. But in this case, Moreno’s copyright assertion for visualizing social networks has been more than mere rhetoric. His 1920s diagrams of stage movements in improvised theater (Das Stegreiftheater, 1924) can indeed be regarded as a crucial pathway towards visualizing dynamic social networks.  But we should still be careful whom to give credit to when it comes to the methods of testing and visualizing social networks.

Jacob Levy Moreno (?), Psychological Geography Map IV. A Reduction Sociogram, 1934. A reduction of the overall map on which plus and minus operators provide the scale. Every plus codes ten attractions and every minus ten rejections in one cottage. Double points on a sign halve the points. Every red and black line between the cottages stands for five attractions or rejections. The total can be calculated by multiplying by five. This technique reduces the quantity of lines so that “large populations” can also be mapped.
Jacob Levy Moreno (?), Psychological Geography Map IV. A Reduction Sociogram, 1934. A reduction of the overall map on which plus and minus operators provide the scale. Every plus codes ten attractions and every minus ten rejections in one cottage. Double points on a sign halve the points. Every red and black line between the cottages stands for five attractions or rejections. The total can be calculated by multiplying by five. This technique reduces the quantity of lines so that “large populations” can also be mapped.

La science et la vie

There’s much to be said about this title page of the French journal La science et la vie. It comes with all the ambivalent connotations of gender, telephony, labor and automation history. We can still see a shadow of „la demoiselle du téléphone“ in the center of the image. Yet, the telephone girl or „Fräulein vom Amt“ has become a non-human switching machine in 1927. Experts of telephone infrastructure are going to notice the rotary selector and a stepping switch at the bottom of the image. Plus, the switching network in the center shows a line of alphabetic letters.

This circuit with letters points to an unusual feature of Paris telephone switchboards of the time. The first three letters of a telephone code indicated the telephone exchange. They were followed by four digits representing the person being called, which correspond here to the four dials of the “brain.” The seven-digit number could be entered using dials on which the twenty-five letters were printed.

Besides this technical dimension of social addressing, this title page also invokes older, reticular representations of networks. Blue and red lines remind us of the color conventions in anatomy and blood circulation, and thus provide a rather corporeal link for the telephone call of a man and a woman. More on telephone networking can be found in chapter 5 of „The Connectivity of Things: Network Cultures since 1832.“

Title page of the journal La science et la vie. Color lithograph, May 1927.
Title page of the journal La science et la vie. Color lithograph, May 1927.

Charles Joseph Minard’s Statistical map of European railroad traffic

So I continue with another image that represents a significant entry in my network history The Connectivity of Things: Network Cultures since 1832. The book came out on October 15, by the way. Chapter 4 follows the twisted paths of Saint Simonianism. Yet, this week’s image is not as ambivalent as the rest of the chapter which deals with early socialism’s colonial ambitions. Polytechnicien Charles Joseph Minard’s 1862 statistical map of European railroad traffic rather underscores the Saint-Simonian ambitions for a mobilized industrial Europe. Minard’s other statistical maps and diagrams have become a classic in the history of visualizing information. But this image primarily expresses the ambitions of Saint-Simonianism to create a vibrating infrastructural network for the good of society.

Charles Joseph Minard, Statistical map of European railroad traffic of 1862. Lithograph, 1865. 97.5 cm × 76 cm.
Charles Joseph Minard, Statistical map of European railroad traffic of 1862. Lithograph, 1865. 97.5 cm × 76 cm.

Prosper Enfantin: Colonisation de l’Algérie

This week’s image from The Connectivity of Things is a colonialist map. It was published as a fold-out part of Prosper Enfantin’s 1843 book Colonisation de l’Algérie. In this treatise, the Saint-Simonian cult leader Enfantin was upfront about the strategic placement of settlements and settlers in Algeria. While the early socialist movement of Saint-Simonians movement had started with utopian dreams of a new networked unity of orient and occident in the 1830s, this map deliberately drew a reséau de soumission –– a network of colonial submission and police power.

Chapter 3 of the book is focused on theory and heuristics of writing network history in „An Archive of Networking.“ Yet, this map features prominently in chapter 4 of the book. It is called „Channels: The Politics of Networking around 1850,“ and it deals with the ambivalent history of network ideology and practice in early socialism. There’s not just a breathtaking colonial history to be told here –– which includes the Suez Canal –– but also one of early feminism, romanticism and social engineering. I’ll come back to the Saint-Simoniennes and Saint-Simoniens next week.

Prosper Enfantin, Colonisation de l’Algérie. Engraving by L. Bouffard, lithograph by Joseph Lemercier, 1843. 49.5 cm × 66.5 cm. Courtesy of Bibliothèque Nationale de France.
Prosper Enfantin, Colonisation de l’Algérie. Engraving by L. Bouffard, lithograph by Joseph Lemercier, 1843. 49.5 cm × 66.5 cm. Courtesy of Bibliothèque Nationale de France.

The Stele of Vultures

Chapter 2 of The Connectivity of Things: Network Cultures since 1832 is called „Six strata of network history: Genealogy of a cultural technique“. Actually, this one has been written almost exclusively from scratch for the English version of the book. It carves out two main themes of the book. Firstly, this chapter is about research on cultural techniques, and what this means for an Anglophone audience. And secondly, I am telling the longue durée story of nets and networks in a very condensed form.

This week’s image resonates with that genealogy which uncovers six strata of network history: material culture, mythology and religion, reticular nets, infrastructural nets, social networks, and network science. (Connectionist artificial intelligence might develop into a seventh stratum, but this remains to be seen.) The so-called „Stele of Vultures“ resides in today’s Louvre. In the spatial logic of the Paris museum’s collection it represents the beginning of cultural history. In the Louvre you’ll not just find the fragments of a net of power that holds and punishes the enemies of the city-state of Lagash. The stele is surrounded by a multitude of smaller objects of Sumerian everyday culture. Nonetheless, it remains a stunning testimony of ancient civilization’s symbolization of the net as a network of binding, holding, and ruling power.

Eannatum catches the enemies of the city of Lagash with a net. Fragment of the Sumerian Stele of Vultures, circa 2440 BCE. Limestone. Reconstructed original size: 130 cm × 180 cm; depth: 11 cm. Louvre, Paris, inv. no. AO 50.
Eannatum catches the enemies of the city of Lagash with a net. Fragment of the Sumerian Stele of Vultures, circa 2440 BCE. Limestone. Reconstructed original size: 130 cm × 180 cm; depth: 11 cm. Louvre, Paris, inv. no. AO 50.

 

Paolo Veronese: L’industria

Since The Connectivity of Things: Network Cultures since 1832 is coming out on October 15 I might as well blog about some of the book’s visual elements. The book has eleven chapters, so this will be a weekly series of eleven images that convey and carry Connectivity’s narrative. These will be brief entries, since the images themselves are key protagonists of my network history. I do hope that they inspire deeper reading.

Let’s start with Paolo Veronese and his workshop who finished their allegory of „industry“ in 1577. This oil painting makes its appearance in chapter 1, which is called „Getting caught up“. It has already served as a frontispiece for the German original Die Verbundenheit der Dinge. From my point of view, it is the first heterogeneous network that represents and transcends the material net as web between nature and culture. Veronese’s network evokes a multitude of meanings, some of which I address in the first chapter. Up until today it is one of the most stunning artworks in the Palazzo Ducale of Venice.

Paolo Veronese, L’industria. Oil on canvas, 1575–1577. Palazzo Ducale, Sala del Collegio, Venice.
Paolo Veronese, L’industria. Oil on canvas, 1575–1577. Palazzo Ducale, Sala del Collegio, Venice.

The Connectivity of Things: Network Cultures since 1832

Nets hold, connect, and catch. They ensnare, bind, and entangle. Our social networks owe their name to a conceivably strange and ambivalent object. But how did the net get into the network? And how can it reasonably represent the connectedness of people, things, institutions, signs, infrastructures, and even nature? The Connectivity of Things by Sebastian Giessmann, the first media history that addresses the overwhelming diversity of networks, attempts to answer all these questions and more.

Reconstructing the decisive moments in which networking turned into a veritable cultural technique, Giessmann takes readers below the street to the Parisian sewers and to the Suez Canal, into the telephone exchanges of Northeast America, and on to the London Underground. His brilliant history explains why social networks were discovered late, how the rapid rise of mathematical network theory was able to take place, how improbable the invention of the internet was, and even what diagrams and conspiracy theories have to do with it all. A primer on networking as a cultural technique, this translated German classic explains everything one ever could wish to know about networks.

A media history of the material and infrastructural features of networking practices, a German classic translated for the first time into English.

The MIT Press, Infrastructures Series
October 15, 2024
444 pp., 6 x 9 in, 82 b&w illus.
Open Access version: https://doi.org/10.7551/mitpress/12547.001.0001

„The Connectivity of Things: Network Cultures since 1832“ weiterlesen

Lehrveranstaltungen im Wintersemester 2024/25

GNSS-Referenzpunkt Siegen

Im Wintersemester 2024/25 biete ich folgende Lehrveranstaltungen in der Siegener Medienwissenschaft an:

  • Medien- und Wissensgeschichte Künstlicher Intelligenz
    Unisono | Moodle
  • Ringvorlesung: Media Environments: Between Capture and Surveillance
    Unisono | Moodle
  • Internetgeschichte und Web History (Seminar Mediengeschichte)
    Unisono | Moodle
  • Susan Leigh Star und Bruno Latour
    Unisono | Moodle

Worum geht es in den Lehrveranstaltungen? „Lehrveranstaltungen im Wintersemester 2024/25“ weiterlesen