I enjoyed this so much I shared it with my daughter, who is a freshman at UNC working through their many gen Ed requirements. She called me (really?!? called Dad about a random link I sent?) and described some good and bad--the bad is the dull part of known knowns you describe. The good was a gen Ed class where the professor had them going through a trove of material in the university archives that "no one had ever even read before" to relate to modern politics. I don't know if that's true but it did make her feel that she was rediscovering the "unknown knowns". One in the knowns knowns: a different professor getting them to relate 1930s European sources to personal experience and current events in small group discussion. I was overjoyed both that my daughter is finding interest in these courses and that the instructors are encouraging connections to what is personal and current.
This is a very illuminating framework. Moving forward, it would be worth investigating in greater depth how the four areas overlap and influence one another. If, for example, the delivery of known-knowns is a prerequisite for participation in the other three quadrants, but this is increasingly being delegated to AI, how can we ensure that AI handles knowledge transfer in a way that actually prepares students to produce knowledge? Also, how can implicit knowledge buried in the unknown-knowns be processed in such a way that it becomes productive for the knowledge-generation process within the unknown-unknowns quadrant?
An additional thought: what's missing from the four-field matrix is the system's environment and how it informs the knowledge-building process. The idea that hits you while walking in the woods, chatting with a neighbor from a totally different profession, and so on. These ‚external-world factors‘ can reveal tacit knowledge or contribute to innovation in ways that cannot be automated, remaining beyond the reach of AI. The challenge lies in how higher education institutions can help cultivate these factors, for instance, by sending students "out into the world" with specific assignments.
Unknown knowns are in some ways the most interesting. I'd think that things like open secrets and "emperor has no clothes" situations would be examples.
There is likely much more research going on into the various unknowns in the non institutional sector than there ever is at university.
As an individual I'm currently honing my new new take on Greek myth and fable, refining narrow vehicle dynamics, involved with a team involved in aether research, as well as following cold fusion work that is totally ignored by universities as it is unpatentable, not to mention studies into the very fundamentals English Law that are not covered by university courses.
I don't consider myself particularly intelligent, or special, just interested in where I see a lack of knowledge and I'm sure there are thousands like me across the planet.
Such an interesting framing, Hollis. Would you characterize Bell Labs or Xerox Parc as organizations that map to your fourth quadrant? Hard to imagine an undergrad working in such a place, but perhaps that’s the goal to aim for.
I will confess that I am not as enlightened by the RM as you, Hollis. We do not know what we do not know, which for me means its an empty category for me, and it's not at all clear what to make of the academy in general giving it much attention. Much less, for students.
People who do not yet know the known known must attend to learning it, just in case they would like to push on the margins of knowing more by spending time and effort on the known unknowns, which is the category known as research in the academy. People who have not got out of the known knowns category are not adequate for entering the known unknowns category.
The category unknown knowns is also meaningless for me. If it's knowledge that no one knows, because somehow it is misplaced or unperceived, why would I waste my time on it. How important could knowns in this category be?
Either I'm missing much of the point here, or this matrix is just not that fascinating for me.
I actually agree with way more of this than I thought I would. The mistake universities are making is not that they care about knowledge. It’s that they’re overvaluing delivery and undervaluing creation.
AI didn’t show up to destroy learning. It showed up to expose where the value really is.
If your institution’s main flex is moving known information from a syllabus into a student’s head, that game is already over. Not because learning doesn’t matter, but because the distribution cost of knowledge just went to near zero. You don’t get prestige for owning the DVD store anymore when Netflix exists.
What this matrix nails is something most leaders are avoiding because it’s uncomfortable. Known knowns are safe. They’re legible to legislators. They’re easy to audit. They feel serious. But they are no longer scarce.
Scarcity has moved to judgment, synthesis, verification, taste, and the ability to sit with uncertainty without freezing. That’s the work in the unknown quadrants.
Here’s the opportunity universities should be excited about instead of defensive about. AI lets you offload the boring part so you can finally double down on the part you claim makes you special.
Unknown knowns are a gold mine. Every university is sitting on insane intellectual assets that are fragmented, poorly indexed, or locked behind institutional habits that made sense in 1997. If you want legitimacy in the AI era, become the best organization in the world at surfacing, validating, and contextualizing what is already known but unusable. That’s real stewardship. That’s defensibility.
Known unknowns are where credibility still compounds. AI can help you move faster, but it cannot replace the hard work of experimentation, failure, replication, and proof. If students are not inside that process, not watching how fragile knowledge actually is, then you are graduating consumers, not builders.
And the unknown unknowns part is where universities can still be magic. Not because you have answers, but because you can protect questions before the market kills them. That’s underrated. That’s a real moat. The ability to notice anomalies, keep them alive, and give them time before demanding ROI is something almost no other institution can do at scale.
Here’s the mindset shift I’d push. Stop framing AI as a threat to authority. Authority was never the point. Trust is. Speed is. Credibility is. If AI can handle knowledge transfer cheaper, great. Take the savings and put it into environments where students are learning how knowledge is made, challenged, documented, and defended.
The universities that win are not going to be the ones that argue hardest for why lectures still matter. They’re going to be the ones that say we used AI to free ourselves to do the work only humans and institutions can do together.
This is not about abandoning known knowns. It’s about admitting they’re the floor, not the ceiling.
If higher education wants to keep its seat at the table, it has to stop confusing familiarity with value and start organizing itself around producing people who are comfortable living in uncertainty and turning it into something real.
Very quickly: Rumsfeld famously did not seek to account for what you're calling the unknown knowns, the things that you don't know you know. Zizek aligns the latter with the Freudian unconscious and thus darkens Rumsfeld's grid. Give me a beat to see how I might leverage this in your analysis. But for now, this might be interesting:
Zizek:
In March 2003, Rumsfeld engaged in a little bit of amateur philosophizing about the relationship between the known and the unknown: "There are known knowns. These are things we know that we know. There are known unknowns. That is to say, there are things that we know we don't know. But there are also unknown unknowns. There are things we don't know we don't know." What he forgot to add was the crucial fourth term: the "unknown knowns," the things we don't know that we know-which is precisely, the Freudian unconscious, the "knowledge which doesn't know itself," as Lacan used to say.
If Rumsfeld thinks that the main dangers in the confrontation with Iraq were the "unknown unknowns," that is, the threats from Saddam whose nature we cannot even suspect, then the Abu Ghraib scandal shows that the main dangers lie in the "unknown knowns" - the disavowed beliefs, suppositions and obscene practices we pretend not to know about, even though they form the background of our public values.
Thus, Bush was wrong. What we get when we see the photos of humiliated Iraqi prisoners is precisely a direct insight into "American values," into the core of an obscene enjoyment that sustains the American way of life.
Ah, now I know why I jumped in before reading more carefully: when you introduce the Rumsfeld matrix you give him a little too much credit. He only ever names three of the four terms. It’s telling, and in retrospect almost a psychoanalytic joke, that he misses or, um, represses the “unknown knowns.”
More substantively, I’m interested in what it would mean to take that fourth term seriously for the university itself. I like your account of unknown knowns as buried infrastructure, uncatalogued expertise, and underused institutional memory. But the psychoanalytic version would have to be sharper and stranger, no? Institutions sometimes need not to know something they already know in order to function. Disavowal is both enabling and an absolute constraint.
I wonder then whether some of the moral panic around AI (not universally, but in its most anxious forms) stems from an uncomfortable recognition. LLMs may be too close to what the contemporary university already is: a system optimized for recombination, credentialed paraphrase, managed novelty, and the circulation of second-order knowledge. What appears as an external threat would then be disturbing not because it is alien, but because it is familiar/uncanny in a way the institution cannot comfortably acknowledge. Again, uncanny as thought by Freud: too close and too true and threatening.
This would be closer to Žižek’s reading of Abu Ghraib: the scandal does not reveal a deviation from American values, but their obscene supplement. In that sense, AI may be forcing into view not just inefficiencies or missed opportunities, but the university’s own disavowed self-knowledge—what it knows about itself but cannot officially say.
Yes yes yes please write an essay-long version of this! "I wonder then whether some of the moral panic around AI (not universally, but in its most anxious forms) stems from an uncomfortable recognition. LLMs may be too close to what the contemporary university already is: a system optimized for recombination, credentialed paraphrase, managed novelty, and the circulation of second-order knowledge. What appears as an external threat would then be disturbing not because it is alien, but because it is familiar/uncanny in a way the institution cannot comfortably acknowledge. Again, uncanny as thought by Freud: too close and too true and threatening."
Oof! You’re very kind. And: I wish I had the bandwidth for that project—or the fortitude to survive the smash-and-grab energies of the panicked, or the intelligence to understand the tech better than I do. And I wish Bruno Latour were still alive. He would help me (a) avoid theoretical inflation (stay away from claims that presume some universal like Thought, Seriousness, Creativity, or Humanity), which I can do, but also (b) insist on following the actors, human and non-human, and the weak or strong, fragile or durable connections that bind them.
That latter is the difficult bit; it’s slow and patient work, and frankly beyond me, not only at the moment, but forever, ‘cause not my discipline. But perhaps it’s not strictly necessary to register what’s happening. One can already see a gap opening between what is said about AI—often in tones of panic or idealization—and what people are actually doing with it: tentative experiments, misuse, quiet avoidance, pragmatic incorporation into existing work and play. Yeah, and it’s being done to us as much as we’re doing it. But that’s a topic, not a reason to abandon thoughtful engagement. I’m better with dramas that circulates discursively; the practices are messier, more provisional, and far less apocalyptic.
Reading Latour has taught me that institutions get into trouble when they mistake their own talk for the thing itself. My worry is that some of the most consequential conversations about AI and the future of the university are happening at exactly that level of abstraction—where “AI,” “rigor,” or “work” start to behave like actors in their own right, rather than shorthand for heterogeneous practices that deserve to be followed rather than judged in advance.
Last week in class (ok, a class called Beauty Problems) I affirmed what I called “slow pretty,” that is, sitting with pleasures that might be light but neither simple nor throw-away… all motivated by, um, Heated Rivalry. Is there something like a slower AI? Can we just slow down, just a bit? In our use, sure. But also to ask whether the panic isn’t telling us less about AI than about the university’s uncertainty about itself. And whether there might be sites (Visual Studies is one I know best) where iterative making, interpretive judgment, and genuinely dialogic back-and-forth with tools—old and new—are already happening without the need for grand claims. If that’s right, then the question isn’t how to save the university from AI, but how to notice, protect, and extend forms of work that have quietly been preparing students for this strange present all along.
I enjoyed this so much I shared it with my daughter, who is a freshman at UNC working through their many gen Ed requirements. She called me (really?!? called Dad about a random link I sent?) and described some good and bad--the bad is the dull part of known knowns you describe. The good was a gen Ed class where the professor had them going through a trove of material in the university archives that "no one had ever even read before" to relate to modern politics. I don't know if that's true but it did make her feel that she was rediscovering the "unknown knowns". One in the knowns knowns: a different professor getting them to relate 1930s European sources to personal experience and current events in small group discussion. I was overjoyed both that my daughter is finding interest in these courses and that the instructors are encouraging connections to what is personal and current.
Thank you and yay!
This is a very illuminating framework. Moving forward, it would be worth investigating in greater depth how the four areas overlap and influence one another. If, for example, the delivery of known-knowns is a prerequisite for participation in the other three quadrants, but this is increasingly being delegated to AI, how can we ensure that AI handles knowledge transfer in a way that actually prepares students to produce knowledge? Also, how can implicit knowledge buried in the unknown-knowns be processed in such a way that it becomes productive for the knowledge-generation process within the unknown-unknowns quadrant?
Yes! This is the conversation I wanted to start, exactly...
An additional thought: what's missing from the four-field matrix is the system's environment and how it informs the knowledge-building process. The idea that hits you while walking in the woods, chatting with a neighbor from a totally different profession, and so on. These ‚external-world factors‘ can reveal tacit knowledge or contribute to innovation in ways that cannot be automated, remaining beyond the reach of AI. The challenge lies in how higher education institutions can help cultivate these factors, for instance, by sending students "out into the world" with specific assignments.
Unknown knowns are in some ways the most interesting. I'd think that things like open secrets and "emperor has no clothes" situations would be examples.
There is likely much more research going on into the various unknowns in the non institutional sector than there ever is at university.
As an individual I'm currently honing my new new take on Greek myth and fable, refining narrow vehicle dynamics, involved with a team involved in aether research, as well as following cold fusion work that is totally ignored by universities as it is unpatentable, not to mention studies into the very fundamentals English Law that are not covered by university courses.
I don't consider myself particularly intelligent, or special, just interested in where I see a lack of knowledge and I'm sure there are thousands like me across the planet.
Such an interesting framing, Hollis. Would you characterize Bell Labs or Xerox Parc as organizations that map to your fourth quadrant? Hard to imagine an undergrad working in such a place, but perhaps that’s the goal to aim for.
I have been thinking about how to make a campus Bell Labs yes. This is the conversation I want to be having!
I will confess that I am not as enlightened by the RM as you, Hollis. We do not know what we do not know, which for me means its an empty category for me, and it's not at all clear what to make of the academy in general giving it much attention. Much less, for students.
People who do not yet know the known known must attend to learning it, just in case they would like to push on the margins of knowing more by spending time and effort on the known unknowns, which is the category known as research in the academy. People who have not got out of the known knowns category are not adequate for entering the known unknowns category.
The category unknown knowns is also meaningless for me. If it's knowledge that no one knows, because somehow it is misplaced or unperceived, why would I waste my time on it. How important could knowns in this category be?
Either I'm missing much of the point here, or this matrix is just not that fascinating for me.
This was about the only thing Rumsfeld said that made any sense. Personally I’m counting on having a lot of unknown knowns…
I actually agree with way more of this than I thought I would. The mistake universities are making is not that they care about knowledge. It’s that they’re overvaluing delivery and undervaluing creation.
AI didn’t show up to destroy learning. It showed up to expose where the value really is.
If your institution’s main flex is moving known information from a syllabus into a student’s head, that game is already over. Not because learning doesn’t matter, but because the distribution cost of knowledge just went to near zero. You don’t get prestige for owning the DVD store anymore when Netflix exists.
What this matrix nails is something most leaders are avoiding because it’s uncomfortable. Known knowns are safe. They’re legible to legislators. They’re easy to audit. They feel serious. But they are no longer scarce.
Scarcity has moved to judgment, synthesis, verification, taste, and the ability to sit with uncertainty without freezing. That’s the work in the unknown quadrants.
Here’s the opportunity universities should be excited about instead of defensive about. AI lets you offload the boring part so you can finally double down on the part you claim makes you special.
Unknown knowns are a gold mine. Every university is sitting on insane intellectual assets that are fragmented, poorly indexed, or locked behind institutional habits that made sense in 1997. If you want legitimacy in the AI era, become the best organization in the world at surfacing, validating, and contextualizing what is already known but unusable. That’s real stewardship. That’s defensibility.
Known unknowns are where credibility still compounds. AI can help you move faster, but it cannot replace the hard work of experimentation, failure, replication, and proof. If students are not inside that process, not watching how fragile knowledge actually is, then you are graduating consumers, not builders.
And the unknown unknowns part is where universities can still be magic. Not because you have answers, but because you can protect questions before the market kills them. That’s underrated. That’s a real moat. The ability to notice anomalies, keep them alive, and give them time before demanding ROI is something almost no other institution can do at scale.
Here’s the mindset shift I’d push. Stop framing AI as a threat to authority. Authority was never the point. Trust is. Speed is. Credibility is. If AI can handle knowledge transfer cheaper, great. Take the savings and put it into environments where students are learning how knowledge is made, challenged, documented, and defended.
The universities that win are not going to be the ones that argue hardest for why lectures still matter. They’re going to be the ones that say we used AI to free ourselves to do the work only humans and institutions can do together.
This is not about abandoning known knowns. It’s about admitting they’re the floor, not the ceiling.
If higher education wants to keep its seat at the table, it has to stop confusing familiarity with value and start organizing itself around producing people who are comfortable living in uncertainty and turning it into something real.
That’s not a crisis. That’s an opportunity.
Very quickly: Rumsfeld famously did not seek to account for what you're calling the unknown knowns, the things that you don't know you know. Zizek aligns the latter with the Freudian unconscious and thus darkens Rumsfeld's grid. Give me a beat to see how I might leverage this in your analysis. But for now, this might be interesting:
Zizek:
In March 2003, Rumsfeld engaged in a little bit of amateur philosophizing about the relationship between the known and the unknown: "There are known knowns. These are things we know that we know. There are known unknowns. That is to say, there are things that we know we don't know. But there are also unknown unknowns. There are things we don't know we don't know." What he forgot to add was the crucial fourth term: the "unknown knowns," the things we don't know that we know-which is precisely, the Freudian unconscious, the "knowledge which doesn't know itself," as Lacan used to say.
If Rumsfeld thinks that the main dangers in the confrontation with Iraq were the "unknown unknowns," that is, the threats from Saddam whose nature we cannot even suspect, then the Abu Ghraib scandal shows that the main dangers lie in the "unknown knowns" - the disavowed beliefs, suppositions and obscene practices we pretend not to know about, even though they form the background of our public values.
Thus, Bush was wrong. What we get when we see the photos of humiliated Iraqi prisoners is precisely a direct insight into "American values," into the core of an obscene enjoyment that sustains the American way of life.
Yes that's in a footnote! (Or at least I put a shorter version there.) So I agree this is necessary context in a texture...
Ack! Sorry, I missed that note! That's on me.
It’s all good! Now more people will notice it after they read your comment! Zizek amplification protocols!
Ah, now I know why I jumped in before reading more carefully: when you introduce the Rumsfeld matrix you give him a little too much credit. He only ever names three of the four terms. It’s telling, and in retrospect almost a psychoanalytic joke, that he misses or, um, represses the “unknown knowns.”
More substantively, I’m interested in what it would mean to take that fourth term seriously for the university itself. I like your account of unknown knowns as buried infrastructure, uncatalogued expertise, and underused institutional memory. But the psychoanalytic version would have to be sharper and stranger, no? Institutions sometimes need not to know something they already know in order to function. Disavowal is both enabling and an absolute constraint.
I wonder then whether some of the moral panic around AI (not universally, but in its most anxious forms) stems from an uncomfortable recognition. LLMs may be too close to what the contemporary university already is: a system optimized for recombination, credentialed paraphrase, managed novelty, and the circulation of second-order knowledge. What appears as an external threat would then be disturbing not because it is alien, but because it is familiar/uncanny in a way the institution cannot comfortably acknowledge. Again, uncanny as thought by Freud: too close and too true and threatening.
This would be closer to Žižek’s reading of Abu Ghraib: the scandal does not reveal a deviation from American values, but their obscene supplement. In that sense, AI may be forcing into view not just inefficiencies or missed opportunities, but the university’s own disavowed self-knowledge—what it knows about itself but cannot officially say.
Yes yes yes please write an essay-long version of this! "I wonder then whether some of the moral panic around AI (not universally, but in its most anxious forms) stems from an uncomfortable recognition. LLMs may be too close to what the contemporary university already is: a system optimized for recombination, credentialed paraphrase, managed novelty, and the circulation of second-order knowledge. What appears as an external threat would then be disturbing not because it is alien, but because it is familiar/uncanny in a way the institution cannot comfortably acknowledge. Again, uncanny as thought by Freud: too close and too true and threatening."
Oof! You’re very kind. And: I wish I had the bandwidth for that project—or the fortitude to survive the smash-and-grab energies of the panicked, or the intelligence to understand the tech better than I do. And I wish Bruno Latour were still alive. He would help me (a) avoid theoretical inflation (stay away from claims that presume some universal like Thought, Seriousness, Creativity, or Humanity), which I can do, but also (b) insist on following the actors, human and non-human, and the weak or strong, fragile or durable connections that bind them.
That latter is the difficult bit; it’s slow and patient work, and frankly beyond me, not only at the moment, but forever, ‘cause not my discipline. But perhaps it’s not strictly necessary to register what’s happening. One can already see a gap opening between what is said about AI—often in tones of panic or idealization—and what people are actually doing with it: tentative experiments, misuse, quiet avoidance, pragmatic incorporation into existing work and play. Yeah, and it’s being done to us as much as we’re doing it. But that’s a topic, not a reason to abandon thoughtful engagement. I’m better with dramas that circulates discursively; the practices are messier, more provisional, and far less apocalyptic.
Reading Latour has taught me that institutions get into trouble when they mistake their own talk for the thing itself. My worry is that some of the most consequential conversations about AI and the future of the university are happening at exactly that level of abstraction—where “AI,” “rigor,” or “work” start to behave like actors in their own right, rather than shorthand for heterogeneous practices that deserve to be followed rather than judged in advance.
Last week in class (ok, a class called Beauty Problems) I affirmed what I called “slow pretty,” that is, sitting with pleasures that might be light but neither simple nor throw-away… all motivated by, um, Heated Rivalry. Is there something like a slower AI? Can we just slow down, just a bit? In our use, sure. But also to ask whether the panic isn’t telling us less about AI than about the university’s uncertainty about itself. And whether there might be sites (Visual Studies is one I know best) where iterative making, interpretive judgment, and genuinely dialogic back-and-forth with tools—old and new—are already happening without the need for grand claims. If that’s right, then the question isn’t how to save the university from AI, but how to notice, protect, and extend forms of work that have quietly been preparing students for this strange present all along.
I may not know, but I do have faith .