Among the reactions to my book A New Vision for Islamic Pasts and Futures, one that stands out especially concerns ChatGPT. The large language model software application has been much in the news lately. A few months after the book’s publication, someone told me that the chatbot comes up with a plan that closely matches the paradigm that my book critiques and is designed to replace when asked to devise a syllabus for “Islamic history.” The artificially intelligent application arranges history along an arrow of time. The story begins with the life of the Prophet Muhammad and moves teleologically, producing a chain of names of dynastic empires and regional states. The narrative structure of the chatbot syllabus compels students to apprehend “Islam” hypostatically: it makes Islam into a bounded entity, a concretized unit that progresses along time’s vector in a purposeful way.
The ChatGPT response reproduces my object of critique because of how artificial intelligence modules (re)produce knowledge. A chatbot differs from a straightforward informational source (such as an encyclopedia) in that it apprehends, processes, and reproduces not just content but also form. The machine’s “intelligence” exceeds copying and pasting. Its algorithm compares multiple examples to arrive at the conditions that make knowledge recognizable as such. Subsequently, it utilizes the results of comparative processing to synthesize its answer to the query. The limitations of this way of reproducing knowledge are especially clear in AI “hallucinations.” When a chatbot cannot find adequate resources to algorithmically produce answers, it manufactures references and facts that it anticipates questioners will find convincing. The credibility a chatbot projects while hallucinating reflects its capacity to mimic form.
When prompted to create a syllabus for Islamic history, however, ChatGPT has no need to hallucinate. The algorithm regurgitates what the large language model digests from the vast corpus of scholarly and other published materials on Islam that have been produced in English since the nineteenth century. The chatbot restates authoritatively the singular form it can find replicated endlessly in its source material.
A New Vision for Islamic Pasts and Futures disrupts the hegemonic, self-authenticating form of Islamic history that the ChatGPT response distills. It does so by putting new frames around the strange and familiar materials of the Islamic past. The book’s form — a custom-designed digital interface — is its argument. It aims to disestablish patterns of scholarship that would lead ChatGPT to create its formulaic syllabus of Islamic history. My intention is to offer new ways to process and represent historical data pertaining to Islam at both micro and macro levels. Like ChatGPT, my book addresses the relationship between knowledge and digital form. However, my work activates this connection in a way opposite to artificial intelligence: it invites the reader to navigate not a mimicry of existing knowledge but a new form and its possibilities.
One of my ultimate objectives in A New Vision for Islamic Pasts and Futures is to move beyond monolithic representations of Islam. Doing this requires that scholars dismantle presumptions about Islam as an object of historical investigation and invent alternative conceptual frameworks. Scholarship remains beholden to limiting presumptions especially when scholars claim that their work is a straightforward empirical relaying of what is found in “primary” sources. Whenever we cite, our decisions about what to include or exclude are conditioned by the paradigms from which we derive our knowledge and to which we aim to contribute. Our citations adhere to existing forms. This is precisely the epistemology that allows ChatGPT to produce its bland and mediocre syllabus. In the field of Islamic history, the underlying presumption is that historical research should make Islam’s presence in time cohere into a “tradition” or a “civilization” that moves along a temporal continuum. The calendrical sequencing of the tradition runs parallel to a spatiotemporal emplotment that makes it possible to produce dichotomies: center and periphery, core and margins, orthodoxy and heterodoxy, and so on. This diachronic historical representation of Islam also underpins synchronic accounts that posit matters such as a core set of beliefs, around which hierarchically arranged understandings of practices cohere. At the heart of monolithic representations of Islam and Muslims lies a singular history for Islam.
By contrast, A New Vision for Islamic Pasts and Futures has a visual table of contents that invites readers to see the book as an interconnected web made of nodes that represent the book’s various sections. The web represents an understanding of Islamic history as a perpetually expanding and transforming network that has no predetermined boundaries. History, imagined as a web rather than a line, can account for patterning, connectivity, parallelism, causality, and so on, without confining the story to a singular timeline. This networked understanding privileges the idea that historical time is a human construct that can vary greatly between differing actors and situations. Time is made by whoever makes a claim about it. When thinking historically, we must account for the agency of the person or entity that posits relationships between events to create time. While the timeline version of history buries this agency beneath a fixed form, the web model allows, and even encourages, multi-temporal understandings that provide access to the ever-changing human worlds that we wish to represent in our work.
In the conceptualization I offer, the web of history is three-dimensional, pluriform, and malleable. It accommodates overlaps, contestations, leaps, and continual invention. It compels us to appreciate the Islamic past as a conglomeration of densely interconnected stories that need not be tied to a bounded tradition, civilization, or singular narrative. There are many Islamic pasts that are contingent on who makes claims about them — how, when, and from where. We can see these multiple pasts in the multitude of materials available for interpretation. Our own accounts can also vary greatly depending on our commitments and preferences. In the future, new Islamic pasts will be created that we cannot and need not anticipate, since these will respond to material conditions not yet known to us. The pasts we document contain our apprehensions about the future. This diversity of pasts and futures reveals how time functions as a precondition as well as a product of human endeavor.
The interface of A New Vision for Islamic Pasts and Futures is designed to encourage jumping between chapters and sections based on immediate instinct rather than authorial guidance. Readers who respond to this encouragement often comment that doing so can be confusing and even disconcerting. Sticking with it, however, makes it possible to see patterns, especially upon arriving at familiar points via different pathways. Encountering the same thing placed in a different sequence makes intuitive how the same evidence can be put to very different uses in the creation of historical narratives. This reading process builds a web of ideas in the reader’s mind that is derived from the web projected by the book, but with nodes rearranged according to the pathways the reader opted to follow.
Every reader who spends time with the book experiences it in ways I could not have anticipated. My current experience of the book attests to this effect. Now that significant time has passed since the book was completed, many of the issues that were top of mind when I composed it have receded. When I read it again, the web of history that forms in my mind differs from what I thought was starkly immanent during the writing process.
The possible and likely disjunction between authorial intention and reader experience is a desired effect of the design choices I made in A New Vision for Islamic Pasts and Futures. I understand my own representations of Islamic pasts as parts of the larger web of history. The way I tell stories derives from my positionality. It is perfectly understandable — and indeed a central point of the book — that these stories will change if and when they are taken up or transmitted by others. As the author, I join the primary sources I cite and each reader as a node in the web of Islamic history, which contains us all as agents and as recipients of stories. Through this (re)conceptualization of Islamic history, the book aspires to engender new epistemological, aesthetic, and ethical relationships between us as readers and writers and the evidence pertaining to Islam in the past, present, and future. The book’s experimental digital form assumes that new horizons for creating and operationalizing knowledge about Islam and Muslims will keep emerging in the future. There is no standard form. The book shows that the materials from the past that are available to us today are already chock-full of evidence to make this case.
Am not able to grasp the nuances indicated here … need to spend time with the book more diligently