# Data Spirits: On the Fabrication of Geist in the Humanities > [!tip] Reference > Talk held on 2024-11-14 for the [RUSTlab Lecture series](https://rustlab.ruhr-uni-bochum.de/lecture-by-fabian-pittroff-on-data-spirits-on-the-fabrication-of-geist-in-the-humanities/). > [!example] Resources > Comment on the talk by Mace Ojala on the [RUSTlab website](https://rustlab.ruhr-uni-bochum.de/comment-on-fabian-pittroffs-lecture-data-spirits-on-the-fabrication-of-geist-in-the-humanities/) ## Abstract 💁 Within the humanities, disciplines speak and work in different ways. There is a wide variety of methods and objectives, which makes collaboration difficult. Data is an integral part of this, and working with it is as essential as it is mundane. This infrastructural layer is not just another problem, but a starting point for tracing ways of doing research in the humanities. The talk draws on an ethnography embedded in a long-term research institution, and the notion of data spirits refers to the ways in which these researchers create and mobilise data through intelligent processes to generate what they are looking for: spirited ideas. ## Intro 📂 The talk is called *Data Spirits* because it is about *the use of data* in *humanities research*. Let me explain. I’m in the middle of doing an ethnography of a long-term research institution consisting of more than a dozen of subprojects, most of them from the humanities. I’m doing this with a sensibility from Science and Technology studies. Meaning, I’m interested in the way research is being done in practice and I focus on the role of technology and infrastructures in these practices. It turns out, one way to do this, it is useful to look at the use of data in this field. There are various methods and objectives within humanities research, but the handling of data is almost always part of it. I’m not talking about digital humanities or digital data only here. Data can take on many forms. Data can be a Word file, but also notes on a piece of paper. I will come back to this. First, Let me state my general interest here. On the one hand, data is an everyday part of humanities research. On the other hand, data is rarely a topic or a result of this type of research. This raises the question of what role data plays in the humanities. This question is not least useful because it helps to address a more general question. That is, how do the humanities work and what they are aiming for. Here, Data is a means to an end, but what end? ## Site & Method 🏢 Answers are provided by an ongoing ethnographic study of a research institution at a German university. At this site, around 50 people work together in 15 subprojects under a common program. Most of them are part of the humanities' tradition: Literature Studies, Media Studies, Art history, Medievalists. I hold a position in one of the subprojects of this institution. Doing an ethnography means I’m researching my colleagues. I work with them. I visit events or internal meetings. We as subprojects organize workshops ourselves. I’m part of collaborative publishing projects. I write ethnographic protocols. I did a few interviews. The 15 subproject are one of the strongest structural elements at the institution. Imagine a group of 4 – a PI (usually a professor), a postdoc, a doctoral researcher and a student assistant. This is the nuclear family of German academia. That said, every project is a little different, but it is typically not far from this constellation. In general, these groups share not only disciplinary interests and topics. Here, dependencies and authorities based on work contracts and academic qualifications create strong bonds. Of course, there are also structures to cross and connect the subprojects. Workshops organized by two or more subprojects. Moreover, there is a strong structure connecting the early career researchers across subprojects. So, this is the site: A research institution based in the humanities, around 50 people, with a vague common topic and program. ## Conceptual Framework 🧮 I draw on research at this site to address the question I mentioned before. What is the role of data in humanities research? More generally, I ask, how research is being done in the everyday work of the humanities and what are the objectives. It not about what the humanities should be, but about mapping out the very different forms of work (Martus/Spoerhase 2022: 13). There won’t be one answer to the questions above, but more like a catalogue of ways of doing humanities. To address this, I developed a conceptual framework revolving around three main concepts. Data: the everyday material of the research. Geist/spirit: the name I propose for the goal of humanities research. Intelligences: the procedures by with data is transformed into Geist. Here is a diagram of the relationship between data, intelligence and Geist. My idea for this talk is, to take a walk around this diagram. Elaborating on the three concepts along the way and giving you examples and insights from the field while we there. ![[geistkreis-eng.png]] *Fig. 1: Diagram of the relationship between data, intelligence and mind* ## Geister/Spirits 👻 Let's start with the weird one: Geist or spirit. I want to be frank, it started out as a joke. Humanities, in German, is Geisteswissenschaften, right? So I thought, Geist might be the thing, that it is all about. Like Nature for the Natural sciences, Geist is the thing the humanities would like to know and talk about. But seriously! I’m stating this not as a statement of fact, but as a heuristic strategy. I will not and cannot strictly define Geist. Its function is a heuristic one to point my research in a direction. The assumption is, there must be objectives within the humanities and because these are probably heterogeneous and multiple, it might be smart not to pinpoint these in advance. That’s why I choose this open and naive term. As such, Geist designates, very generally, what humanities research is interested in and values, both as a subject and a product of its work. But, the term is no empty signifier without any meaning. According to Niklas Luhmann > \[the term Geist is] only meaningful if one believes one has reasons to avoid the distinction between psychological and social system reference (Luhmann 1992: 44, fn. 47) This is meant critically and as an argument to avoid it. I think I have such reasons not to differentiate between what is thought and what is communication. Because it is conducive to the openness of my inquiry to leave it undefined, for now, to what extent Geist happens in heads, in the air, on paper or on the screens. Geist then refers to forms within the medium of meaning (*Sinn*). In this context, I might add: If Geist is understood as humanities’ medium – with the help of Annina Klappert (2020) – Geist could also be the name for the virtuality the humanities tap into. But more importantly for now: Geist (in the singular) understood as a medium means there are multiple forms possible in this medium. So there isn’t just one Geist, but many Geister or spirits. See e.g. Friedrich Kittlers volume from 1980. “The banishment of spirit from the humanities” or “Austreibung des Geistes aus den Geisteswissenschaften”. But again, to wrap this up: The function of the term Geist is a heuristic one. A tool for empirical analysis to circumscribe that which humanities endeavors to produce and for which it generates and processes data. ## Everyday Data 📓 Which brings us to the next part: *everyday research data*. What is research data in a humanities setting? I follow a praxeographic approach here, going back to Annemarie Mol (2002). Which means, data always goes back to data practices. I consider data practices to be processes through which… > “traces [...] are fixed in a suitable medium and [...] further processed in this permanent form" (Rheinberger 2021: 29). This results in digital or non-digital objects that can be moved through “space and time without losing their identity” (Rheinberger 2021: 32). Data practices produce objects that are both mobile and unchangeable. What Bruno Latour calls *immutable mobiles* (Latour 1990: 26). Data does not require the involvement of digital technologies. Handwritten notes and scribbles on paper can be data too insofar they are used in data practices. Talking about data means referring to practices and describing their genesis and use. This approach seems especially useful when dealing with the ambiguous and broad use of data in humanities research (Borgman 2009). Generally, data is used to prove phenomena (Borgman 2015: 29). Let's summarize: - Data only exist in relation to data practices (Mol 2002) - These practices fix traces in a medium (Rheinberger 2021) - … in order to obtain stable but mobile objects (Latour 1990) - … which serve as manifestations of a phenomenon (Borgman 2015) This general description is necessary to recognize data and their practices, even if they are not labelled as such in the field. It is equally important to listen to the definitions of the actors in the field. What constitutes a data practice in the humanities can be identified by those who conduct humanities research. This way, I can compile a list of data practices from my field. Necessarily incomplete, nevertheless, the list gives an impression of the multiplicity of data practices. ``` - Editing and exchanging shared texts - Annotation of texts, images and other objects - Preparation of notes and excerpts for personal use - Preparation of protocols for common use - Organisation of local or remote file systems - Use of software for literature management - Collection and arrangement in mapping processes - Recording of sensor data - Management of audio-visual media and software objects ``` This list shows, foremost, how data practices in the humanities are interested in the traces of thoughts and communication, forms in the medium of meaning, and traces of Geist. Speaking more practically, there are two groups of practices that are particularly widespread. The *creation and manipulation of text* and *the organization and annotation of literature*. In other words, there are data practices that belong to the field of writing and those that belong to the field of reading. As part of an editorial team, I have done experimental study on data practices of writing (Pittroff 2024). Working with text always involves data practices. When alternative data formats and infrastructural detours are introduced, researchers tend to avoid these to keep their data routines familiar and stable. This avoidance refers to the role of data in the acquisition of Geist. Data is meant to disappear. Let’s have a look at data practices of reading. Which includes the handling of literature data. Academic literature in the form of books, printouts, or PDF files is one of the essential data types in the humanities research. Literary objects are data as soon as they are integrated in data practices. That is, when they are treated as traces of meaning. Let's say this less abstract, this refers to a broad spectrum of practices. Annotation on paper or on screens. Organizing literature in digital or non-digital folders, through software or on shelves. Even downloading or deleting a PDF file. These are everyday humanities research data practices. ## After Data 🐀 Let’s talk about the fabrication of Geist from data. Together with my colleague Leman Çelik, I held a group discussion regarding research data with a group of humanities researchers from the institution in question. One question we asked the participants was, what they do with folders on their computers. Even though the group had already discussed different data types, they responded to the question about the use of the file system by talking about literature files. This data type seems especially important to them. At the same time, participants report very different ways of organizing literature files. Stored unsorted in a single folder named literature. Distributed to thematic sub-folders and sub-sub-folders. Stored locally or in remote folders. Or simply kept in the system's download folder. Practices for organizing these items involve the use of literature management software, or via bibliographies of own publications, or by cataloguing in text files. ``` I02: What do you do with folders on your computer? P05 (36:59): Desperation. (*laughter*) I always say to myself, yes, now I will have a very highly organized, curated PDF system or something. And then it's mostly chapter by chapter, big bulk. And I, whenever I use it, then they sometimes get better names where I can find them, but sometimes they are a random string of numbers, which they come from the download, and then I'm like, I'm pretty sure there were three sevens in this one. Yes. But sometimes I have sporadic periods where it works out, but then I get overwhelmed by the mess, I think. I'm not sure anyone can relate. ``` This is just one of the statements from the group discussion, but most of them consider their data practice to be inadequate. The question is, in terms of which goal, the researchers consider their data use insufficient. What are the next steps that the data should enable? What should be the outcome of working with it? I collected a few of these goals. Some researchers want to find without “knowing what I am looking for”. “Because everything is connected” they need “overview”, “bringing things together” and “finding out” things. And, of course, it's about writing: Sorting the data “chapter by chapter”. Finding what is “relevant to what I want to write”. “I do research”, reports one participant, “when I write something new by looking up what I read in old articles that I have written”. ## Intelligences 🪸 As stated in the diagram above, the key here is to find procedures that fabricate Geist from data. I refer to the processes of this transition as *intelligences*. I have written elsewhere about my idea of a sociology of intelligences (Pittroff 2024). For today, let me just summarize the key points. My concept of intelligences draws mostly from John Dewey and Niklas Luhmann. With their help, I’m able to describe intelligences by four traits: First, Intelligence is *collective*. Intelligence doesn’t happen in your head only, but it is also a social thing. Second, Intelligence is *multiple*. There is more than one kind of Intelligence. Third, Intelligence is a *medium* (Luhmann 2017). On the level of semantics, Intelligences is the fundamental possibility of observing observations as observations and proposing alternative distinctions. You can ask if something is true or false, but you can also replace the distinction between true/false by another distinction. Fourth, Intelligence is a *method* (Dewey 2001). On a practical level, intelligence is the ability to connect active action and passive experience. Like in an experiment, intelligence, in its basic form, is about the link between means and ends. | | Medium | Method | | --- | --- | --- | | Collectivity | Semantics | Practices | | Multiplicity | Alternative distinctions | Action-experience ensembles | *Table 1: Elements and relations of a sociology of intelligences* The goal of this framework is to make it possible to distinguish and compare different types of intelligence. For this, you have to ask two questions. What are the distinctions that a way of doing intelligence is using? And wow, does it aim to connect action and experience, means and ends? ## Read-write access 💾 Now, tentatively, let’s look at one of these intelligent procedures that are supposed to fabricate Geist from data. It is the seemingly but truly complex set of semantics and practices that is apparently well known as *writing*. More precisely, the writing of scientific publications. This dazzling process is certainly more than just one practice. And surely, it is more than a mental occurrence. The condensing and reflection in the process of writing can be described as a form of intelligence in accordance with the elements introduced above. *Semantically*, writing is about the development, proposal, and application of distinctions. Often as a superior alternative to other distinctions in the form of concepts or systematics, e.g. let's not talk about real and possible but about actual and virtual. *Practically*, writing is a sophisticated method of combining action and experience because it produces results in meaning, through the writing in front of you that can immediately be experienced and tested through reading. The author Rainald Goetz describes this as follows: > “The writer looks at the words he has written and reads them. The writer constantly reads what he has written: what does it say? What does it mean? Is what is written what is meant? Is it what he wanted to say?" (Goetz 2024: 49) Additionally, humanities researchers Martus and Spoerhase (who published a “praxeology of the humanities”, which is especially interesting as a practice theory from the field) find writing “is not explained by the fact that the participants simply open their eyes and document in writing what they observe” (2022: 50), but a procedure in which “thoughts arise in the course of dealing with materials and goals \[…] emerge in the course of the work” (2022: 250). ## Ends 🌈 This brings me to the end of the talk. Let's reiterate my question one more time. *What are the intelligent procedures by which Geist is fabricated from data in humanities research?* The detour via the data practices, we took today, shows how reading, writing and their alteration are connected and interrupted by data practices. If you ask about files, you learn about literature and the lack of an overview. When alternative data practices are introduced, the writing process tries to deviate into the familiar routines. Reading and writing are the procedures to make data disappear and fabricate Geist instead. If you ask humanities scholars about their most recent contact with data, you may be met with a discussion about the meaning of the term. This not only points to the interests in the humanities, but also refers to the under-representation of data in the field. In the search for the processes that enable the intelligent leap from data to Geist the focus on the dazzling set of activities around reading and writing can only be the beginning. To me, it’s promising though to try to better understand how data practices of reading (ordering, marking, excerpting) and data practices of writing (texting, note-taking, drafting) are synthesized into spirits and thus forgotten. ## References 📚 - Borgman, Christine L (2009): »The Digital Future is Now: A Call to Action for the Humanities«, in: Digital Humanities Quarterly, 3/4. - Borgman, Christine L (2015): Big data, little data, no data. Scholarship in the networked world, Cambridge (Mass.): The MIT press. - Dewey, John (2001): Die Suche nach Gewißheit. Eine Untersuchung des Verhältnisses von Er-kenntnis und Handeln, Frankfurt a. M.: Suhrkamp. - Fleck, Ludwik (1980): Entstehung und Entwicklung einer wissenschaftlichen Tatsache: Ein-führung in die Lehre vom Denkstil und Denkkollektiv, Frankfurt a. M.: Suhrkamp. - Fleck, Ludwik (1983): »Über die wissenschaftliche Beobachtung und die Wahrnehmung im all-gemeinen«, In: Erfahrung und Tatsache: gesammelte Aufsätze, Frankfurt a. M.: Suhrkamp. - Goetz, Rainald (2024): Wrong: Textaktionen, Berlin: Suhrkamp. - Latour, Bruno (1990): »Drawing things together«, In: Lynch, Michael/Woolgar, Steve (Hrsg.): Representation in scientific practice, Cambridge, Mass: MIT Press. - Luhmann, Niklas (1992): Die Wissenschaft der Gesellschaft, Frankfurt a. M.: Suhrkamp. - Luhmann, Niklas (2017): »Gibt es ein System der Intelligenz?«, In: Die Kontrolle von In-transparenz, Berlin: Suhrkamp, S. 30–45. - Martus, Steffen/Spoerhase, Carlos (2022): Geistesarbeit. Eine Praxeologie der Geisteswissenschaften, Berlin: Suhrkamp. - Mol, Annemarie (2002): The body multiple. Ontology in medical practice, Durham, N.C: Duke Univ. Pr. - Pittroff, Fabian (2024): »Text, plain«, In: Early Career Forum des SFB 1567 (Hrsg.): Voka-bular des Virtuellen. Ein situiertes Lexikon, Bielefeld: transcript. - Pittroff, Fabian (2024): »Künstliche und künstlerische Intelligenz. Zum ästhetischen Umgang mit algorithmischer Technik«, In: Krenn, Karoline/Kropf, Jonathan/Laser, Stefan/Ochs, Carsten (Hrsg.): Varianten digitaler Bewertung. Wiesbaden: Springer VS. (Im Erscheinen) - Rheinberger, Hans-Jörg (2021): Spalt und Fuge. Eine Phänomenologie des Experiments, Berlin: Suhrkamp. - RUSTlab/Amelang, Katrin/Asai, Ryoko/Çelik, Leman/Eggel, Ruth/Galanova, Olga/Laser, Stefan/Ojala, Mace/Pittroff, Fabian/Sørensen, Estid/Werner, Lynn (2024): »Please Go Away… We’re Reading«, in: On\_Culture, S. No 16 (2024): Ways of Reading. - Serres, Michel (1987): Der Parasit, Frankfurt a. M.: Suhrkamp. - Tarde, Gabriel de (2009): Monadologie und Soziologie, Frankfurt a. M.: Suhrkamp.