If we want to understand the various forms of writing “in-the-loop,” then we must start with understanding the loops themselves. While discussions of accessibility, expertise, collaboration and more are at the forefront of the conversations here, if one wants to understand the impact that these topics have on their own loops, we must understand what a loop is in the first place, how it comes to be, and, most interestingly, how it might be changed if we find ourselves dissatisfied with the loops we participate in. Loops can be understood further as infrastructural rhetorical objects, specifically operational rhetorical objects. Studying their emergence through this lens forwards our understanding of all looped writing activity.
So, what is a loop? Traditional definitions are problematic, as loops in the present sense are already a hazy metaphor adapted from electrical engineering and flight (Safire, 2004). As such, Infrastructural Rhetoric may be of use to move our understanding beyond a simple definition (See: Adams, 2022 for further reading). In the context of Infrastructural Rhetoric, a loop is a repeated, definable, and definite process that exists as an operational infrastructure. Loops are not a physical object but rather an operational structure that is brought into being by human intent to support and maintain an ongoing rhetorical practice in time. This infrastructural understanding means that a loop begins in time when we say it does. Loops inherently require this intentional plurality to exist. One cannot “loop” singularly, even within the private process of writing, as loops, at a minimum, require two frames of mind to exist: One in the past, and one in the present, or one in the present and one in the future. Loops cannot exist until the activity is repeated through time, leading to a plurality. Loops are thus an infrastructural conflux of two rhetorical conversations made plural and repeatable through time. The loops that we create while writing or coding, as either Human-in-the-loop (HITL) or Machine-in-the-loop (MITL), emerge when our own rhetorical processes come to interact with those of others, our past selves, or AI (Knowles, 2024). It is not a singular event, then, but a multiplicity of loop-types that constitute human behaviors and human-machine interactions. A loop, in this sense, is something that can be understood and studied beyond metaphor; however, when a loop emerges specifically, and how one might interact with that emergence, are more complicated questions.
The emergence of loops is most easily seen in relation to their diachronic temporal existence. The first time something is written, it does not become a loop. Loops are created when a past activity or conversation reiterates within the present toward a future purpose. When loops are considered diachronically, however, we quickly realize what appears to be a confusing paradox: A loop requires repetition, yet, seemingly from the get-go, our writing process can be spoken about in singular loops, e.g., “I’m getting looped into this conversation.” This duality between loops as simultaneously repetitive and singular exists because operational infrastructural rhetoric may be singularly operational, but not singularly responsive. We write in a style, form, and practice that we have acquired through our teachers, coaches, and personal literacy. Though our writing is our own, it is not our own in the sense that the loop (or the process of forming our writing) is one that we learned and adapted through iteration and mimetic repetition of others. Infrastructure is always built on other infrastructure, and our writing is no different (Star, 1999). Even these letters that I write and transfer to you on this screen are not completely my own; they owe their original form to the Phoenicians in some sense, and it is only through thousands upon thousands of operations (teacher instructions, conversations, and written documents) that they now appear in this form to you. Like the letters, the words and thoughts are also infrastructural. In this lens, all writing is, to some extent, an operational infrastructure, meant to speed up and bring efficiency to communication through habituated manners of sharing information. It is loops within loops built upon loops—loops all the way down.
After years and years of writing and looping, we can begin a loop without acknowledging the conversations that established it, or even remembering them. Most of us do not remember how we learned to write, only that we now know how to. This compression happens with all of our writing throughout our lives. It all becomes infrastructurally looped in at some point, while the individual writing operations remain new in the same breath. This forgetting through ritualization and continued maintenance renders the main part of the loop invisible through a process of operational compression. Operational compression is both the focus and danger of the loop. As all loops are infrastructures, they have a strong desire to trend toward functional invisibility (Star & Ruhleder, 1996). As I have argued elsewhere, infrastructure is typically only noticed when it breaks. When our writing loops are functioning smoothly, they recede into the background, becoming the naturalized workflows that we grow accustomed to (Frith, 2020). Therefore, if you want to question, adjust, or even break your loops, it is this invisibility that we must map, make visible, and unpack. A careful rhetorician resists complete operational compression.
To change an operational infrastructure, one must change the repeatable and ritualistic maintenance that holds that operation in place: To change a loop, one can’t change it at the point of looping, but rather the point of habitation and compression for the loop. Instead of finding the point where the loop is most embedded into processes, we must ask the questions that originate at the points of compression. Allow me an example:
Consider the academic integrity section of a syllabus. The first time such a passage is created, it is crafted with care and a deep philosophical grounding from the instructor. Such passages are often drafted through committees, adding additional layers of care. However, these careful linguistic structures are also designed to be operationalized. Through years of rushed course planning and busy summer research, the policy paragraph becomes pasted from one document to another. It is a loop—a piece of ritualized maintenance that we upkeep while maintaining efficiency—this is operational compression in its purest form. The profound rhetorical act of ensuring honesty and integrity has been flattened into a clerical task. We only re-enter the loop and consider the policy for the rhetorical infrastructure that it is when a student challenges the wording, or a new technology like LLMs renders the invisible boilerplate visible (Salvo & Sherrill, 2025). We are forced to re-engage with it at this point, but it is already too late. The other infrastructures that depend on this piece of operation require it to go forward in its current loop. To truly change this document, we would need to change it at the point of compression at the beginning of the term, or even at the committee’s determination of the guiding ethical principles. Any other mid-loop change only breaks the infrastructure further, rather than shifting the discourse. This example is a minor one, all things considered, but it does give us a path forward for our own looped writing in more serious contexts.
As we move forward to read from the other authors here, I ask us to consider: What parts of our own writing loops have become so operationalized that they are now invisible to us? When we decide to share our rhetorical load with a machine, are we truly sharing a creative burden, or are we surrendering our process to the compressions of an algorithm? If so, those compressions may never be made visible until the algorithm breaks, or we do. The danger is a loss of our own infrastructural rhetorical knowledge. Ultimately, if we treat our loops as static objects, we will also lose our ability to steer them. But if we view loops as malleable infrastructures with several layers of nested compositions within them, we can begin to successfully find the possibilities for change.
Works Cited
Adams, J. (2022). A Theory of Infrastructural Rhetoric. Communication Design Quarterly, 10(3), 46–55. https://doi.org/10.1145/3507870.3507876
Frith, J. (2020). Technical Standards and a Theory of Writing as Infrastructure. Written Communication, 37(3), 401–427. https://doi.org/10.1177/0741088320916553
Knowles, A. M. (2024). Machine-in-the-loop writing: Optimizing the rhetorical load. Computers and Composition, 71, 102826. https://doi.org/10.1016/j.compcom.2024.102826
Safire, W. (2004, April 18). The Way We Live Now: On Language. The New York Times. https://www.nytimes.com/2004/04/18/magazine/the-way-we-live-now-4-18-04-on-language-hair-on-fire.html
Salvo, M., & Sherrill, J. (2025). Artificial Infrastructures (Online) [Digital]. WAC Clearinghouse. https://doi.org/10.37514/PRA-B.2024.2654
Star, S. (1999). The Ethnography of Infrastructure. American Behavioral Scientist, 43(3), 377–391.
Star, S., & Ruhleder, K. (1996). Steps Toward an Ecology of Infrastructure: Design and Access for Large InformationSpaces. Information Systems Research, 7(1), 111–134.
