Textual Promiscuity in the Age of AI

If a tree falls in the forest…

NOTE: As you read this text, attend to the careful embedding of language patterns in the text which call for constraints on reader interpretations, including participation in Stanley Fish’s interpretive communities. I do not argue that readers are free to see whatever they like in a text. I’ve taught many of these ideas about textuality in disciplinary reading methods courses long enough to know that knee-jerk reactions cause minds to snap shut like clams reacting to semantic fumes.

Indeed, my argument is that authorizing the reader in community actually intensifies the demand for readers to attend carefully to textual symbols and their patterns. Meaning is never pre-made right there, and right there, and right there, waiting to be found and collected. Significant scholary meanings always are born of legitimate peripheral participation in communities of practice.

Ancient Wisdom

Human literacy exhibits what Daniel Dennett, a seminal thinker in artificial intelligence, called “interpretive ravenousness,” a hunger which literally transforms the world into a text. Any array of patterns brought before the eyes or ears—or even fingers or nose or tongue— is brain food.

When Roland Barthes proclaimed the death of the author he proclaimed the birth of the autonomous text. Touching on the matter of the silence of trees falling unheard, Dennett’s ravenous mind and Barthes’ absent author elevate the reader to the status of text finder, text taker, and text modeler. In the absence of a receiver’s eye or ear, texts and falling trees alike make no sound. Meaning is made with a text apart from its creator, not drawn from a text waiting to be unpacked like an author’s suitcase. No disrespect to the author intended. I am one myself.

Reading the Unwritten World

Clifford Geertz's anthropological perspective helps us understand how divination practices across cultures affords asymmetrical textual transactions: reading tea leaves, interpreting the flight patterns of birds, decoding the arrangement of thrown bones, analyzing the lines in palms. In each case, humans bring their unique and their inheritable interpretive frameworks to bear on random or naturally occurring patterns, transforming what Geertz would call the "thin" accidents of physics into "thick" meaningful texts. Tea leaves don’t encode messages about romantic partners or buried treasure, but the diviner reads messages anyway, and their interpretations can influence real behavior and decisions.

Archaeological Asymmetries

Ian Hodder and other post-processual archaeologists¹ have theorized how material culture functions as text, creating meaning from fragments that were never intended to be messages. A pottery shard becomes a text about ancient trade routes; food remains become texts about dietary practices and social hierarchies; the arrangement of stones becomes a text about astronomical knowledge. The original makers encoded some intentional messages, e.g., the pot was meant to hold grain, but archaeologists read entirely different texts from the same physical evidence, i.e., this grain-holding technology indicates contact with Mediterranean cultures.

Digital Age Promiscuity

Shoshana Zuboff's analysis of "surveillance capitalism" exemplifies asymmetrical textual transactions on steroids. Your browsing history wasn't written as a text about your aches, your pains, your fetishes, but algorithms read it as exactly that, extracting behavioral predictions from patterns you never consciously encoded. Following Zuboff's framework, your location data becomes a text about your traffic patterns; your purchasing history becomes a text about demographic trends; your social media behavior becomes what she calls "behavioral surplus" that corporations read as texts about future actions.

Biological Texts We Conscript

Evelyn Fox Keller's work on genetic discourse reveals how DNA exemplifies the ultimate promiscuous reading². The genetic code evolved as a biochemical instruction manual for protein synthesis, but humans read it as texts about evolutionary history, medical predispositions, family relationships, and ethnic ancestry. As Keller demonstrates, we've imposed narrative structures onto genetic sequences—reading stories about human migration, adaptation, and kinship that the DNA never "intended" to tell.

Literary and Artistic Extensions

Roland Barthes' "Death of the Author" acknowledges the radical asymmetry in textual transactions: readers routinely extract meanings from literary texts that authors never consciously encoded. The text becomes independent of authorial intention and available for unlimited interpretive engagement. A Shakespeare sonnet written about personal love becomes a text about universal human experience, or colonial power structures, or the nature of language itself, depending on who's doing the reading and when.

Wolfgang Iser's reader-response theory complicates this authorial death with his notion of the “implied reader,” arguing that meaning emerges in the encounter between a text with a reader designed in and a reader externalizing that figured reader rather than residing in either alone. Perhaps the ghost of the dead author remains entangled in the text. Abstract art exemplifies this: Jackson Pollock's drip paintings encoded something about movement, gravity, and spontaneity, but viewers read them as texts about chaos and control, about randomness and pattern, about American individualism, about the unconscious mind.

The Pareidolia Principle

Psychologists have identified pareidolia—our tendency to see faces in clouds, hear voices in static, or find meaningful patterns in random data—as revealing the promiscuous nature of human textual literacy. This cognitive mechanism, evolved for pattern recognition and survival, means we can't help but read texts in noise. As Charles Sanders Peirce's semiotics suggests, this isn't a flaw but a feature of "unlimited semiosis"—our capacity to create meaning through interpretive chains that can extend infinitely.

Implications for Textual Theory

Hans-Georg Gadamer's hermeneutics suggests that this promiscuity and textual asymmetry reveals textuality as less about the properties of encoded messages and more about what he calls the "fusion of horizons"—the meeting between our interpretive apparatus and the patterns we encounter. We don't just read texts—we create texts through a merging of a reader’s horizon shaped by history, culture, language, and experience and a text, an artifact made by a missing person who left behind traces of another human horizon.

This synthesis of thinkers raises questions about the nature of meaning itself. If we can find coherent, actionable significance in tea leaves, pottery shards, Netflix algorithms, and random number sequences, perhaps making meaning is less a property that texts possess and more what Stanley Fish would call an "interpretive community" ³ activity. The asymmetrical transaction reveals humans not as readers of pre-existing meaning, but as authors of textuality itself—constantly writing the world and fellow humans as readable texts through the very act of reading them.

Enter AI

AI as the Ultimate Expression of Barthes' "Dead Author"

AI represents the logical endpoint of Barthes' proclamation that meaning resides in the reader, not the author. When an AI generates text, there is literally no conscious author—no intentional mind encoding personal experience, cultural insight, or deliberate meaning. The "author" is a composite made of linguistic fragments. Or the author may reside, a character, a figure, in one of Fish’s past interpretive communities.

Yet AI texts function for human readers who bring their most astute interpretive frameworks to bear on the generated patterns. AI texts reveal that Barthes may have been more radical than even he realized: not only can texts function independently of authorial intention, they can function without any conscious author whatsoever. The reader truly becomes the sole source of textual meaning.

Simulated Texts as Pure Pareidolia Engines

AI systems mechanize the pareidolia principle, i.e., that human tendency to find meaningful patterns in random data. But instead of seeing faces in clouds, AI finds linguistic patterns in statistical relationships among words that show up together over and over again, and its algorithms generate new patterns that human readers cannot help but interpret as meaningful.

The AI exploits our "interpretive ravenousness" by producing artifacts that trigger our pattern-recognition instincts. These texts succeed not because they contain anything substantive or material, but because they are engineered to match the statistical signatures of texts that humans have historically found meaningful. AI has industrialized our cognitive bias toward meaning-making.

The Death of the Individual Encoding/Decoding Distinction

Traditional textual theory assumes a clear sequence: encoder transforms experience into text, decoder extracts meaning from text. But AI texts scramble this relationship. The AI encodes patterned relationships among symbols in abstract structures—it treats previous texts as raw material for generating new textual artifacts. Texts that emerge from pure textual statistical manipulation rather than experiential encoding may not be categorically different from human texts encoded in interpretive communities.

How does the activity of a human writing a research paper in eleventh grade History, finding relationships among words used similarly in a variety of texts, differ from the activity of a bot? This question isn’t rhetorical, and the answer isn’ t obvious. The result suggests that much of what we call textual communication may actually be humans’ projecting meaning onto pattern-matched symbolic arrangements—a possibility that calls into question the experiential grounding we assumed was necessary for all meaningful text.

Practical Implications for Teaching Reading in the Disciplines

From Information Extraction to Pattern Recognition Training

If AI reveals that much textual competence depends on recognizing and manipulating symbolic patterns rather than extracting author-intended meanings, then discipline-specific reading instruction must shift focus to maximize the value of this insight for human learning. In history classes, instead of teaching students to "find the significant ideas and be ready to discuss them" in primary sources, we might teach them how historians create historical narratives by recognizing patterns across multiple texts, how they transform fragments of evidence into coherent interpretations through learned analytical frameworks.

In science courses, rather than simply having students decode and commit to memory textbook explanations, we might teach them how scientific texts work as pattern-recognition exercises: how experimental data gets transformed into theoretical explanations and frameworks, how mathematical relationships get encoded into symbolic descriptions, how scientists read patterns in data that non-experts cannot see. How does one read a reported effect size?

Teaching Students as Co-Authors Rather Than Passive Decoders

If readers create textual meaning through the act of reading rather than merely extracting it, then every reading assignment becomes a creative writing exercise. Students in literature classes who analyze what Shakespeare meant in Hamlet should understand that their interpretation actively creates meaning from symbolic patterns that can support multiple coherent readings. There truly is no Shakespeare nor Hamlet to ask.

Plato articulated this problem with text long ago in the Phaedrus when Socrates bemoans the fact that texts can never answer questions nor talk back. Socrates remarks that unlike living, interactive dialogue, written words are silent, unresponsive, and unable to defend themselves or adapt to the reader’s needs:

“Once a thing is put in writing, the composition, whatever it may be, drifts all over the place, getting into the hands not only of those who understand it, but equally of those who have no business with it; it doesn't know to whom it should speak and to whom it should not… And when it is ill-treated and unjustly abused, it always needs its parent to come to its help, being unable to defend or help itself.”

Implications for assessment are profound. Instead of testing whether students can identify standard pre-interpretations (the green light in Gatsby), we might assess their ability to construct their own original, coherent, evidence-based readings that reveal plausible pattern recognition within disciplinary frameworks. In economics classes, students reading market analyses should understand they're not absorbing expert knowledge but learning how economists transform raw data patterns into policy recommendations.

Developing Meta-Textual Awareness in the AI Age

Students ought to develop deep understanding that they're increasingly engaging with texts generated by systems that have never had experiential grounding in the subjects they're writing about. We know, or think we know, that Shakespeare had a child named Hamnet. Distinguishing between experientially grounded texts even barren of intention/author and pattern-generated ones, and more importantly, understanding when this distinction matters, is now a basic skill.

In researcher-heavy disciplines like psychology or sociology or education, those fields where we speak using a researcher’s name as shorthand for ideas rather than concept terms (e.g., DNA in science, e.g, Chomsky in linguistics),students need to recognize when they're reading texts that emerge from direct empirical investigation conducted by human researchers versus texts that synthesize perspectival patterns from other texts.

On this view, even human-constructed reviews of the literature are essentially synthesized patterns from other texts more like AI text with less tendency to hedge and hallucinate and are read differently than empirical studies. AI generated text introduces meaningful slippage which requires hypervigilance for random error different from the sort of critical reading applied to a human-constructed review of the literature where subtle hallucinations are the exception, not the norm.

Students need to learn to trace the experiential foundations of knowledge claims while simultaneously recognizing that much valuable disciplinary knowledge emerges from symbolic manipulation rather than direct experience. Reading texts from markedly different varieties of source materials is an entirely new dynamic in disciplinary reading instruction.

Disciplinary Pattern Libraries as Explicit Curriculum

If textual competence depends heavily on recognizing discipline-specific patterns, then these patterns should become explicit teaching targets rather than implicit expectations. Chemistry students should learn to articulate the specific symbolic relationships that connect molecular diagrams to chemical properties to experimental observations. Literature students should master the pattern relationships that connect literary techniques to historical contexts to interpretive frameworks.

This means creating systematic instruction in what we might call disciplinary semiotics—the specific ways that each field encodes physical experience into text and text into actionable knowledge, which is an epistemic, not a semantic, issue—reading to learn. Students need conscious access to pattern-recognition systems that experts use unconsciously.

Preparing Students for Co-operative Human-AI Reading

Since students will increasingly encounter AI-generated texts alongside human-authored ones, they need skills for productive co-operation with AI systems in textual interpretation. This means teaching students how to use AI as a pattern-recognition amplifier while bringing their own experiential grounding and critical judgment to bear on the results.

In history classes, students might use AI to identify linguistic patterns across large document collections while learning to apply historical thinking to evaluate and contextualize the patterns the AI finds. In scientific disciplines, students could collaborate with AI systems to generate multiple interpretations of data while learning to assess which interpretations align with experimental evidence and theoretical frameworks.

What To Do Monday Morning

For Teachers:

Audit Your Reading Assignments: Look at this week's reading assignments across your classes. For each text, ask: "Am I asking nd expecting students to extract the author's intended meaning intact, pristine, in a manner that matches the meaning I want them to have and to hold for use during the next assignment and on the test without promiscuity?” What happens if Barthes is onto something important: There is no author’s intent. There are only patterns of language in a text which remain meaningless until the reader constructs sense. If so, would an interactive lecture be more appropriate and effective before asking students to read the text?

Am I asking them to recognize and manipulate disciplinary patterns they aren’t aware of that make this text work and to discover links between these patterns? In-class guided reading, perhaps a version of reciprocal reading, is the more reasonable approach. Rather than "What did this experiment on photosynthesis discover?" ask: "How does this paper transform raw data into theoretical claims? Notice how it moves from methods to results to discussion—what work does each section perform? How do the authors use hedging language ('suggests,' 'appears to indicate') versus certainty claims? How do contemporary biologists use these same textual patterns to build scientific credibility, and how might you recognize when these patterns are being manipulated?" No need to bring Robert Kennedy Jr. into the discussion.

Introduce Pattern Hunting as Explicit Skill: In your next class, spend 10 minutes having students identify recurring patterns in whatever text they're reading—not the content patterns, but the structural ones. In a history document: How does the text connect evidence to claims? In a science text: How do mathematical relationships get translated into verbal explanations? Make these disciplinary patterns visible and teachable.

Start Co-Authorship Conversations: When students express their interpretations, respond with: "You've just created meaning from these textual patterns. What other meanings could these same patterns support?" Don’t ask "Is that right or wrong?" Ask "How did you construct that reading, and what makes it coherent?"

For Students:

Become a Pattern Detective: In every text you read this week—textbook chapter, news article, research paper—spend five minutes identifying the moves the text makes. How does the language connect ideas? How does the text present evidence? How do arguments build? You're not looking for content; you're reverse-engineering the textual machinery.

Practice Co-Authorship When you read anything, ask: "What meaning am I creating from these patterns?" instead of "What is this text trying to tell me?" Write brief responses that begin: "Reading this text through [historical/scientific/literary] frameworks, I construct the following interpretation..." Own your role as meaning-maker.

Test AI Collaboration Take one of this week's reading assignments you’ve spent quality time with and ask an AI system to identify patterns or generate interpretations. Then apply your own disciplinary knowledge to evaluate, extend, or critique what the AI texted. Practice using AI as a pattern-recognition tool while maintaining your authority as the meaning-maker.

The Bottom Line:

Stop treating texts as containers of pre-existing meaning to be extracted for application in a pre-defined activity leading to a pre-defined test. Start treating them as pattern-rich artifacts that become meaningful through legitimate peripheral interpretive engagement in a community of peers and mentors gradually becoming more expert practitioners. This shift—from information extraction to collaborative meaning-making—is pedagogically sound.

In an AI age, it's survival.

1. “This approach, developed with a group of students at Cambridge University, was a reaction to the then-dominant processual archaeology, which viewed culture as an adaptive process and believed archaeology could be objective by applying the scientific method. My own ideas of how archaeology could expand to incorporate previously marginalized theoretical perspectives, such as structuralism, post-structuralism, Marxism, feminism, and practice theories, were published in my 1986 book Reading the Past.” (Hodder, https://www.ian-hodder.com/research)

2. “Since therefore the knowledge and survey of vice is in this world so necessary to the constituting of human virtue, and the scanning of error to the confirmation of truth, how can we more safely, and with less danger, scout into the regions of sin and falsity than by reading all manner of tractates and hearing all manner of reason? And this is the benefit which may be had of books promiscuously read.” Milton, John. Areopagitica (1644)

3. “Interpretive communities are made up of those who share interpretive strategies not for reading (in the conventional sense) but for writing texts, for constituting their properties and assigning their intentions.”