Thanks for the stimulating article not kept in a silo. This engagement is not a critique but musings upon your thesis.
Why Silos Matter
Everyone wants to tear down silos. Consultants treat them as obstacles to collaboration. University managers treat them as barriers to innovation. The phrase “silo mentality” has become shorthand for inefficiency and isolation. Yet this view misunderstands what a silo is and why civilisations built them in the first place. The original silo was not a wall. It was a preservation technology.
The word comes from the Greek siros, meaning an underground pit that protected grain. Ancient Egypt built its own storage systems. Tomb paintings in the Old Kingdom show sealed bins and raised granaries designed to keep air, moisture, and pests away from harvested crops. Without these structures, famine followed the flood of the Nile with brutal regularity. China developed parallel systems. Neolithic cultures used sealed pits. The Zhou and Han dynasties built elevated granaries (仓) and national storage networks to stabilise food supply. The Qing preserved grain for up to ten years using plastered walls and controlled ventilation. Civilisations survived because they protected the harvest.
A silo keeps contents stable when the outside environment is volatile.
The academic discipline works the same way. A field is a structure of preservation. It protects methods, evidence standards, and the accumulated knowledge of a community. You cannot do physics without mastering physics. You cannot edit a gene without deep biological training. Disciplines are silos because depth requires containment long enough for expertise to form.
But not all silos are benign. Apartheid was a silo of separation designed to privilege one group by excluding another. It controlled movement, education, and knowledge. Gender operated as a silo in the academy for centuries. Women’s work was often kept outside the main intellectual storehouse. Many fields built canonical silos that recognised male scholarship while discarding female contribution. These were structures of domination, not of preservation.
This distinction matters. A silo can protect a harvest, or it can deny one. The structure is the same. The purpose differs.
Modern geopolitics shows the same duality. The United States siloes advanced semiconductor technology from China. China siloes rare-earth processing from the United States. Patent law is a silo that protects invention. Export controls are silos that ration access. Some preserve value. Others preserve inequality.
The challenge for universities is to strengthen the silos that safeguard knowledge while dismantling the ones that shield exclusion. The critics are wrong to claim that disciplines cause stagnation. The real danger is the loss of methodological depth. AI increases this risk. Its breadth is impressive, but without human experts digging deep within their silos, there is nothing firm for the system to learn from.
Civilisations did not survive by leaving grain in the open. They survived by protecting what they could not afford to lose. The same is true of knowledge. Silos are not the enemy. The wrong silos are.
Thanks for this piece. I'm embarrassed to say how little I appreciated the invention of the silo. I've always taken them for granted.
Not to abuse the metaphor too much, but I think it's worth noting that silos are for preserving and protecting grain. The actual growing of it happens in the open, in the field.
To me there's a yes-and. Yes, let's enable silos to keep preserving and protecting specialized knowledge. AND let's recognize that knowledge is first cultivated in the world. To rephrase a point you made late in the piece, "[intellectual] breadth [becomes transformational with] the ongoing digging, complex theoretical modeling, and focus of researchers in disciplinary silos."
This is great. This helps me frame some things I've been seeing recently. In some universities with very good philosophy departments, there has been a lot of pressure from deans and admin to push the philosophy department to "integrate" more with the other humanities. The other humanities departments see the philosophy department as elitist and snooty, locked up in their silos. But perhaps one way to defend the philosophers is to say that, well, they're afraid of rot if they open up. The other humanities are not necessarily a picture of health, and in any case, the philosophers have a rigorous discipline and a long-developed tradition and they're not going to risk that for some trendy topics. (I'm massively generalizing and sketching here. The details in real universities will vary a lot case to case.)
Wonderful article on silos. “The work of knowledge production is the intellectual equivalent of feeding the world.” And shared perspectives about what the silos say with informational integrity is crucial to help “double check facts” because “breadth is shallow and useless without ongoing digging”.
My grandfather was a wheat farmer during the Dust Bowl and plowing behind four mules in a picture I have. To preserve that effort today in silos is because of the thinkers you bring to the fore in your writing. “Silos as the best sites of debate and self-correction, where old paradigms are challenged and understanding is refined” is well stated. The opportunity to connect in depth is preserved in silos, and breadth can bring either possibilities or pollution. The comment on the preservation of information needing to be siloed and the Manhattan Project prime examples.
Your deep dive into the delivery of informational trust from preserved integrity is well delivered in this writing and is not going to be siloed but shared!
Nice piece. I think often of a friend of mine at Cambridge, who once told me that the way they do things at Cambridge changes very, very slowly. And he thought that was excellent. The admin/consultant fad of “let’s break things to show how innovative we are (and add that to our CVs)” is most likely to lead to rot and spoilage, as you aptly point out.
Interesting! Both can be true, also, other systems, such as the USA’s Old Republic, solved some of the short comings of both through deliberate redundancy, in so far as there was silos, each silo had several twins, each linked to each other — like they were with everyone else — because they were in the same system, but each was its own internal world, with very little common sourcing in personnel (at least in impactful ways), and each was fully resourced
Great piece. For better or worse, silos also influenced modern architecture. They were among the industrial structures that Le Corbusier praised as "the magnificent first fruits of the modern age."
I did not know that! Thank you! I was surprised silos were invented so "late," as it were. I was also delighted how universities played a role. Le Corbusier is a complicated character (as anyone who reads James Scott knows well) but I am glad to know he praised silos.
What a lovely metaphor - now I know, silos play a more important role in modern life that I hadn't thought of before. The history lesson was great. And what a punchline: 'Societies that cannot preserve their knowledge face starvation. Hunger follows decay.'
Maybe you call such attempts destruction vs preservation. The semantic difference depends on the person making the distinction, isn't it? Hunger follows decay seems more certain, though.
I am very sympathetic to what I take to be your basic inclinations in this piece, but I have been surprised to see so little pushback on certain matters in the comments, though it's not entirely absent.
To me, the problem is that disciplines go through their own extremely sterile purifications. As an economist, I think a lot of the excesses of the neoclassical paradigm illustrate this, and in English, the craze for theory pretty much transformed the discipline from what I can see, much of it not in a good way.
I think along with you that simply dragooning these disciplines into multidisciplinarity for the sake of it is hardly much of a response. Indeed, reading your piece, it seemed that the kinds of moves you were critiquing are the moves of the consultants and the generalists and the academic bureaucrats who, at least in my country (Australia), have captured virtually all the growth in employment and increased pay towards the top of management. Professors' salaries have flat-lined with Vice Chancellors now bringing home $1.5 million a year.
You write that "Silos weren’t perfect of course and there were bad silos that could use the light of new thinking but they allowed deep and sustained focus on problems that needed to be solved." But economics spent their energies solving a lot of problems that didn't need to be solved - like whether it was possible for an economy comprising perfectly competitive industries (there is no such thing) could achieve a stable equilibrium. The answer was that it could, so long as you made a bunch of other assumptions - like that it didn't have a financial sector! Then there's string theory, which I'm not qualified to judge, but which lots of people who are think is a continuing train wreck.
Great piece, much appreciated. I’m a physical scientist currently collaborating with an economist (but one who knows a lot of physical science herself and is used to working with physical scientists). She said to me that a discipline is when people agree on which assumptions or preconceptions they’re going to give each other a free pass on, which potential gaps in our thinking they’re not going to worry about. I thought this was brilliant in that it’s a nominally negative way of looking at it - what disciplines don’t do - but she very much meant it with love for disciplines, much as you’re expressing here and for the same reasons.
An old friend, Conor Friedersdorf recommended your newest piece and from there I found the link to this one. Both are fascinating perspectives. Thank you.
I’ve read over the last couple of years of a decline in the quality of published research as measured by the number of errors and inaccuracies that slip through the peer review process. I seem to remember that the blame was laid at pressure to publish more frequently rather than to do qualitatively better research. 1) do you think this is a fair critique of journals today? 2) does this mean that our “silos” have degraded or have begun to degrade structurally such that they no longer preserve knowledge and further cutting edge research?
Good questions and I see a combination of both. Silos weren’t perfect of course and there were bad silos that could use the light of new thinking but they allowed deep and sustained focus on problems that needed to be solved. There have always been Centers to bring scholars together across silos and journals that publish interdisciplinary scholarship, but yes, interdisciplinarity can be tops of waves and not substantial at all
I’m in agreement that silos are essential to depth of research and understanding.
I wasn’t so much thinking about interdisciplinary scholarship, although perhaps an issue I’m unaware of. I think the article below is one of the ones I read. Perhaps I don’t know enough of the process, but it seems like the “silos” are not ensuring that research is new or accurate and instead are just granting prestige. And then disseminating information that is incomplete, unhelpful, or even flat out wrong or fake. Which if true, would defeat the purpose of the “silo”.
Of course the answer would be to build new silos with integrity and safeguards rather than just smash them and advocate everyone “do their own research”.
I'm curious, do you believe that the academic siloes in the social "sciences" including economics, political science, and psychology have proven themselves to be as valuable as those in the hard physical sciences?
I appreciate this take on academic silos. In addition to the safeguard in relations to AI, it’s also deeply relevant to the pushes for interdisciplinarity stemming from budget cuts rather than actual intellectual interest/investment “across silos.”
In information security I need silos all the time - but the appropriate term is different - compartment. Lets say I have sensitive health personal data - I can NOT comingle that data with other data. Ditto a lot of personal data (Look up GDPR if you want books full of regulations and penalties). Ditto financial data or corporate financial data that reveals how a company is doing. .....
Yes, compartmentalization of data hinders operations - but it also allows survival.
The idiots who what flat access don't understand the consequences of compromise.
Love this perspective. I will never look at a silo (real or organizational) the same way again. Protecting wisdom in the age of AI is more important than ever.
Thanks for the stimulating article not kept in a silo. This engagement is not a critique but musings upon your thesis.
Why Silos Matter
Everyone wants to tear down silos. Consultants treat them as obstacles to collaboration. University managers treat them as barriers to innovation. The phrase “silo mentality” has become shorthand for inefficiency and isolation. Yet this view misunderstands what a silo is and why civilisations built them in the first place. The original silo was not a wall. It was a preservation technology.
The word comes from the Greek siros, meaning an underground pit that protected grain. Ancient Egypt built its own storage systems. Tomb paintings in the Old Kingdom show sealed bins and raised granaries designed to keep air, moisture, and pests away from harvested crops. Without these structures, famine followed the flood of the Nile with brutal regularity. China developed parallel systems. Neolithic cultures used sealed pits. The Zhou and Han dynasties built elevated granaries (仓) and national storage networks to stabilise food supply. The Qing preserved grain for up to ten years using plastered walls and controlled ventilation. Civilisations survived because they protected the harvest.
A silo keeps contents stable when the outside environment is volatile.
The academic discipline works the same way. A field is a structure of preservation. It protects methods, evidence standards, and the accumulated knowledge of a community. You cannot do physics without mastering physics. You cannot edit a gene without deep biological training. Disciplines are silos because depth requires containment long enough for expertise to form.
But not all silos are benign. Apartheid was a silo of separation designed to privilege one group by excluding another. It controlled movement, education, and knowledge. Gender operated as a silo in the academy for centuries. Women’s work was often kept outside the main intellectual storehouse. Many fields built canonical silos that recognised male scholarship while discarding female contribution. These were structures of domination, not of preservation.
This distinction matters. A silo can protect a harvest, or it can deny one. The structure is the same. The purpose differs.
Modern geopolitics shows the same duality. The United States siloes advanced semiconductor technology from China. China siloes rare-earth processing from the United States. Patent law is a silo that protects invention. Export controls are silos that ration access. Some preserve value. Others preserve inequality.
The challenge for universities is to strengthen the silos that safeguard knowledge while dismantling the ones that shield exclusion. The critics are wrong to claim that disciplines cause stagnation. The real danger is the loss of methodological depth. AI increases this risk. Its breadth is impressive, but without human experts digging deep within their silos, there is nothing firm for the system to learn from.
Civilisations did not survive by leaving grain in the open. They survived by protecting what they could not afford to lose. The same is true of knowledge. Silos are not the enemy. The wrong silos are.
Thank you for this! I appreciate the distinctions here and the work you put into making it!
Excellent coherent explanation
Thanks for this piece. I'm embarrassed to say how little I appreciated the invention of the silo. I've always taken them for granted.
Not to abuse the metaphor too much, but I think it's worth noting that silos are for preserving and protecting grain. The actual growing of it happens in the open, in the field.
To me there's a yes-and. Yes, let's enable silos to keep preserving and protecting specialized knowledge. AND let's recognize that knowledge is first cultivated in the world. To rephrase a point you made late in the piece, "[intellectual] breadth [becomes transformational with] the ongoing digging, complex theoretical modeling, and focus of researchers in disciplinary silos."
Yes and yes!
This is great. This helps me frame some things I've been seeing recently. In some universities with very good philosophy departments, there has been a lot of pressure from deans and admin to push the philosophy department to "integrate" more with the other humanities. The other humanities departments see the philosophy department as elitist and snooty, locked up in their silos. But perhaps one way to defend the philosophers is to say that, well, they're afraid of rot if they open up. The other humanities are not necessarily a picture of health, and in any case, the philosophers have a rigorous discipline and a long-developed tradition and they're not going to risk that for some trendy topics. (I'm massively generalizing and sketching here. The details in real universities will vary a lot case to case.)
Wonderful article on silos. “The work of knowledge production is the intellectual equivalent of feeding the world.” And shared perspectives about what the silos say with informational integrity is crucial to help “double check facts” because “breadth is shallow and useless without ongoing digging”.
My grandfather was a wheat farmer during the Dust Bowl and plowing behind four mules in a picture I have. To preserve that effort today in silos is because of the thinkers you bring to the fore in your writing. “Silos as the best sites of debate and self-correction, where old paradigms are challenged and understanding is refined” is well stated. The opportunity to connect in depth is preserved in silos, and breadth can bring either possibilities or pollution. The comment on the preservation of information needing to be siloed and the Manhattan Project prime examples.
Your deep dive into the delivery of informational trust from preserved integrity is well delivered in this writing and is not going to be siloed but shared!
Thank you so much for this! I had fun researching silos and I am glad it resonated.
Nice piece. I think often of a friend of mine at Cambridge, who once told me that the way they do things at Cambridge changes very, very slowly. And he thought that was excellent. The admin/consultant fad of “let’s break things to show how innovative we are (and add that to our CVs)” is most likely to lead to rot and spoilage, as you aptly point out.
Interesting! Both can be true, also, other systems, such as the USA’s Old Republic, solved some of the short comings of both through deliberate redundancy, in so far as there was silos, each silo had several twins, each linked to each other — like they were with everyone else — because they were in the same system, but each was its own internal world, with very little common sourcing in personnel (at least in impactful ways), and each was fully resourced
Great piece. For better or worse, silos also influenced modern architecture. They were among the industrial structures that Le Corbusier praised as "the magnificent first fruits of the modern age."
I did not know that! Thank you! I was surprised silos were invented so "late," as it were. I was also delighted how universities played a role. Le Corbusier is a complicated character (as anyone who reads James Scott knows well) but I am glad to know he praised silos.
What a lovely metaphor - now I know, silos play a more important role in modern life that I hadn't thought of before. The history lesson was great. And what a punchline: 'Societies that cannot preserve their knowledge face starvation. Hunger follows decay.'
I'm unsure, though, how to distinguish what you've in mind vs echo chambers & misinformation. I try to deal with this related dilemna in https://rajeshachanta.substack.com/p/the-humpty-dumpty-lineage.
Maybe you call such attempts destruction vs preservation. The semantic difference depends on the person making the distinction, isn't it? Hunger follows decay seems more certain, though.
I love this — there are ways that language never stays in its silos does it. Meaning escapes and bleeds or collaborates with words in other silos.
Thanks for this, Hollis.
I am very sympathetic to what I take to be your basic inclinations in this piece, but I have been surprised to see so little pushback on certain matters in the comments, though it's not entirely absent.
To me, the problem is that disciplines go through their own extremely sterile purifications. As an economist, I think a lot of the excesses of the neoclassical paradigm illustrate this, and in English, the craze for theory pretty much transformed the discipline from what I can see, much of it not in a good way.
I think along with you that simply dragooning these disciplines into multidisciplinarity for the sake of it is hardly much of a response. Indeed, reading your piece, it seemed that the kinds of moves you were critiquing are the moves of the consultants and the generalists and the academic bureaucrats who, at least in my country (Australia), have captured virtually all the growth in employment and increased pay towards the top of management. Professors' salaries have flat-lined with Vice Chancellors now bringing home $1.5 million a year.
You write that "Silos weren’t perfect of course and there were bad silos that could use the light of new thinking but they allowed deep and sustained focus on problems that needed to be solved." But economics spent their energies solving a lot of problems that didn't need to be solved - like whether it was possible for an economy comprising perfectly competitive industries (there is no such thing) could achieve a stable equilibrium. The answer was that it could, so long as you made a bunch of other assumptions - like that it didn't have a financial sector! Then there's string theory, which I'm not qualified to judge, but which lots of people who are think is a continuing train wreck.
Great piece, much appreciated. I’m a physical scientist currently collaborating with an economist (but one who knows a lot of physical science herself and is used to working with physical scientists). She said to me that a discipline is when people agree on which assumptions or preconceptions they’re going to give each other a free pass on, which potential gaps in our thinking they’re not going to worry about. I thought this was brilliant in that it’s a nominally negative way of looking at it - what disciplines don’t do - but she very much meant it with love for disciplines, much as you’re expressing here and for the same reasons.
Thank you and thank your economist friend for the discipline description!
An old friend, Conor Friedersdorf recommended your newest piece and from there I found the link to this one. Both are fascinating perspectives. Thank you.
I’ve read over the last couple of years of a decline in the quality of published research as measured by the number of errors and inaccuracies that slip through the peer review process. I seem to remember that the blame was laid at pressure to publish more frequently rather than to do qualitatively better research. 1) do you think this is a fair critique of journals today? 2) does this mean that our “silos” have degraded or have begun to degrade structurally such that they no longer preserve knowledge and further cutting edge research?
Good questions and I see a combination of both. Silos weren’t perfect of course and there were bad silos that could use the light of new thinking but they allowed deep and sustained focus on problems that needed to be solved. There have always been Centers to bring scholars together across silos and journals that publish interdisciplinary scholarship, but yes, interdisciplinarity can be tops of waves and not substantial at all
Thanks for the quick response.
I’m in agreement that silos are essential to depth of research and understanding.
I wasn’t so much thinking about interdisciplinary scholarship, although perhaps an issue I’m unaware of. I think the article below is one of the ones I read. Perhaps I don’t know enough of the process, but it seems like the “silos” are not ensuring that research is new or accurate and instead are just granting prestige. And then disseminating information that is incomplete, unhelpful, or even flat out wrong or fake. Which if true, would defeat the purpose of the “silo”.
Of course the answer would be to build new silos with integrity and safeguards rather than just smash them and advocate everyone “do their own research”.
https://www.theguardian.com/science/2025/jul/13/quality-of-scientific-papers-questioned-as-academics-overwhelmed-by-the-millions-published
Excellent article
Glad I stumbled across this article!
I'm curious, do you believe that the academic siloes in the social "sciences" including economics, political science, and psychology have proven themselves to be as valuable as those in the hard physical sciences?
It's a good questions! This is a good conversation to have....
I'm heavily biased towards thinking that the answer is no, but hearing opposing arguments and having a healthy convo around this topic would be great.
I appreciate this take on academic silos. In addition to the safeguard in relations to AI, it’s also deeply relevant to the pushes for interdisciplinarity stemming from budget cuts rather than actual intellectual interest/investment “across silos.”
In information security I need silos all the time - but the appropriate term is different - compartment. Lets say I have sensitive health personal data - I can NOT comingle that data with other data. Ditto a lot of personal data (Look up GDPR if you want books full of regulations and penalties). Ditto financial data or corporate financial data that reveals how a company is doing. .....
Yes, compartmentalization of data hinders operations - but it also allows survival.
The idiots who what flat access don't understand the consequences of compromise.
Love this perspective. I will never look at a silo (real or organizational) the same way again. Protecting wisdom in the age of AI is more important than ever.