Continuing reflections from the Connecticut College AI & Liberal Arts Symposium
If the first post was about waking up, this one is about finding your bearings.
At the symposium, the mood shifted. The panic over AI—the moral fog, the productivity hype—gave way to something quieter and braver: curiosity. Once we accepted that AI is no longer a visitor but a roommate, the real questions emerged:
How do we live and learn beside it?
How do we create with a system that accelerates answers but doesn’t guarantee understanding?
Those questions led me back to a word that keeps earning its keep the more I use it: polygnosis—many ways of knowing.
Where the Term Came From (and Why That Matters)
Confession: I thought I coined polygnosis. It arrived during a late-night exchange with a generative model while I was trying to name the frontier beyond “inter-” and “trans-.” A neat ego moment—until I realized the idea has been with us for millennia in different clothing. The point isn’t originality; it’s precision. Polygnosis names what I couldn’t quite put my finger on: not a new discipline, but a way of composing knowledge across differences—human and non-human—without flattening them.
Polygnosis isn’t a theory; it’s a temperament for learning beside machines.
From Disciplines to Directions
Interdisciplinary work still starts with the disciplines—a chemist and a poet at the same table, each speaking their dialect.
Transdisciplinary work stretches the table, turning it from a fixed surface into a living workspace—one where ideas, methods, and even machines can join the dialogue.
Polygnosis begins there. It sets a stage where the conditions of learning can play out—less about blending fields, more about cultivating the stance that lets multiple knowledges coexist and collaborate. In a world where learning happens with other intelligences, that stance isn’t a luxury; it’s survival literacy.
And yes, that’s a liberal-arts value at its core: the capacity to survive—and thrive—through changing times by reading context, holding paradox, and practicing judgment.
The Mirror Problem (and Kranzberg’s Reminder)
A recent Harvard study asks a piercing question: Which humans are our models built on? The answer won’t shock you—mostly Western, English-speaking, educated, highly online adults. That narrow slice of humanity quietly became the template for what many models assume is “normal.”
We can’t call that neutrality. We can only call it inheritance.
Melvin Kranzberg’s First Law lands squarely here: technology is neither good nor bad; nor is it neutral. Models aren’t villains or saints; they’re mirrors tuned to the cultures that polished them. The risk isn’t that they’re “biased”—it’s pretending they’re not.
So the task isn’t to scrub away particularity; it’s to expand the story the machine can tell and we can interpret. That is polygnosis in practice.
From Bias to Balance
Every prompt is a vote for what counts as knowledge. If we want better answers, we need better questions:
Whose perspective might be missing here?
How would this read through another cultural lens?
What assumptions am I reinforcing by treating this response as universal?
This isn’t performance or politeness. It’s epistemic honesty—being clear about where knowledge stands when it speaks. We’re not optimizing for optics; we’re optimizing for accuracy with context.
The Liberal Arts as a Living Laboratory
The liberal arts have always practiced polygnosis—even if they never used the word. They train the muscles we need now: interpretation, comparison, translation, discernment.
Keep it simple and grounded:
In a composition class, a student uses an LLM to draft an intro. The work isn’t to grade the draft; it’s to ask how the model thinks and where the student’s voice should diverge.
In biology, image models help visualize ecosystems. The lesson isn’t “cool pictures”; it’s What does the model miss about living systems, and why?
In history, we compare a human summary and an AI summary of the same source, side-by-side, then mark what each foregrounds and erases.
Down-to-earth, doable, honest. Not spectacle—pedagogy.
Polygnosis as Ethos (Not Coursework)
To theorize polygnosis is to pin the butterfly. I’d rather see it fly.
Polygnosis is an ethos—a discipline of attention. It’s curiosity with a spine, method with humility, design as dialogue. In classrooms, studios, and labs, it looks like:
- Co-creating a syllabus with an AI, then annotating its blind spots in the margins.
- Asking a model to produce three interpretations from different cultural frames—and having students choose, remix, or reject them with reasons.
- Building default prompts that say: “Answer without assuming age, ethnicity, or region; if an assumption is necessary, make it explicit.” Not for show. For truth.
What We Should Do Next (Practical and Small)
Set defaults that widen the lens. “Unless specified, respond for a general audience; flag any cultural or demographic assumption you made.”
Teach self-audit. Before accepting an answer, ask the model (and the student): What perspective did this come from? What would challenge it?
Diversify inputs. If we feed narrow corpora, we’ll get narrow mirrors. Bring in sources—texts, datasets, voices—that a general model is likely to underrepresent.
Reward interpretation. Grade the reasoning about outputs, not just the outputs themselves. We’re cultivating readers of intelligence, not just users.
None of this requires a new program or a moonshot grant. It requires habits, modeled consistently.
Polygnosis in Practice: Building Tools That Build Us
Theory becomes real when we put it to work. I’m thrilled to be teaching a class in the spring, Crafting Digital Identity, which aims help learners create professional web presence. This is where polygnosis isn’t an abstraction—it’s practice.
Students learn to prompt and code custom tools that help them craft portfolio narratives, articulate their professional voice, and position themselves strategically for graduate programs or the workforce. They build the tools, then use those tools to build their presence—websites, social media strategies, personal statements that sound like them, not like a template.
But here’s where polygnosis shifts from concept to practice: we don’t accept outputs at face value. We interrogate them using frameworks like Dakan and Feller’s 4Ds (Discover, Discern, Design, Deploy) and Mike Caulfield’s SIFT (Stop, Investigate the source, Find better coverage, Trace claims back to the original context). These aren’t just media literacy buzzwords—they’re diagnostic lenses. They help students ask: Where did this voice come from? What cultural assumptions are baked into this recommendation? How do I preserve my authenticity while leveraging algorithmic assistance?
This is polygnosis applied: using AI to amplify human agency while maintaining critical distance. Students aren’t consumers of AI-generated content; they’re collaborators who understand the system well enough to bend it toward their actual needs. They learn to see the seams, question the defaults, and design with intention.
The result? Graduates who can code a custom GPT to help draft a cover letter, then edit it with the discernment of a liberal arts thinker. Students who understand that personal branding isn’t about projecting an image—it’s about translating their complex, multifaceted selves into narratives that resonate across contexts. Professionals who can speak fluently to both recruiters and academics because they’ve learned to toggle between knowledge systems without losing coherence.
That’s the frontier we’re walking: not just teaching about AI, but teaching with and through AI in ways that make students more capable, more critical, and more themselves.
Finding Our Bearings
We’ve moved past the question of whether AI belongs in the academy. It’s already in the room, auditing everything. The better question is how we keep our humanity expansive, not defensive.
Polygnosis gives us a compass, not a map. It doesn’t dissolve disciplines; it stretches the table so more kinds of knowing can join the work. It asks us to prefer coherence over consensus, dialogue over default, and context over speed.
We don’t need a new field. We need a new fidelity—to curiosity, to complexity, to the courage of unknowing.
The frontier isn’t out there anymore. It’s between us. Within us. And, increasingly, beside us.
Next time: How the 4Ds and SIFT frameworks anchor practical AI pedagogy—and why critical making matters more than critical thinking alone.