Back in August (2024) Tyler Cowen had a link to a paper, Romae Industriae, in which the author, Maxwell Tabarrock, posed the question in the title of this post. While was unable to propose an answer, he did have this paragraph near the end:
The printing press alone may have been sufficient to start a Roman industrial revolution because its effects on culture are upstream of the invention processes that might invent all the other required tech. There was more than enough scale in the economy of Rome to support capital intensive productivity improvements in their major industries. So the major constraint must have been idea generation. Rome punches below their weight in population on this metric because of an aristocratic, slave based culture that ridicules the mechanical arts. If copies of Vitruvius or Archimedes became cheap enough that even the workers close to the vulgar arts of farming and manufacturing could read them, it would surely have an effect on the cultural attitudes that stultified Roman invention.
I believe he was right to suggest that “the major constraint must have been idea generation,” but he was wrong to finger the lack of the printing press as the cause of that constraint.
The problem is much deeper. The problem is that the Roman system of thought was incapable of conceiving of the clockwork universe, an idea that didn’t become prevalent in Europe until the Enlightenment (17th and 18th centuries), though a similar idea – he universe as the machina mundi – had been suggested as early as the 13th century. While I considered making a comment to that effect, I decided not to. There was simply no way to make that idea credible in a blog comment.
But the question kept bugging me until I finally decided to address it. I uploaded Tabarrock’s article to Claude 3.5, along with the article David Hays and I wrote, The Evolution of Cognition (1990), and went at it. First I asked Claude to summarize the two articles, then I had it consider the two hypotheses. We ended the discussion with the question: Can computers think?
My prompts are set in bold face type while Claude’s responses are plain face. I’ve added subtitles to make the conversation easier to follow.
Two summaries
I’ve uploaded a document about cognitive evolution. Could you summarize it for me?
I'll provide a summary of the key points from the paper "The Evolution of Cognition" by William L. Benzon and David G. Hays.
The authors propose that cultural evolution occurs through distinct cognitive ranks, with each new rank introducing new processes of thought and mechanisms for handling abstractions. They identify four cognitive ranks:
Rank 1: (Preliterate societies)
• Primary mechanism: Metaphor
• Medium: Speech
• Abstract knowledge carried through myths and proverbs
• Limited categorical thinking
• Found in most indigenous cultures worldwide
Rank 2: (Ancient civilizations)
• Primary mechanism: Metalingual definition
• Medium: Writing
• Enabled systematic philosophical thought
• Development of explicit ontological categories
• Examples: Ancient Greece, Rome, China
Rank 3: (Post-Renaissance)
• Primary mechanism: Algorithm
• Medium: Calculation
• Enabled scientific thinking and mechanistic worldview
• Clear subject-object distinction
• Originated in Western Europe
Rank 4: (Modern/Emerging)
• Primary mechanism: Control structure
• Medium: Computation
• Still developing
• Characterized by model-building and systems thinking
• Not yet fully formed in most areas of culture
Key concepts:
• Each rank builds on previous ones rather than replacing them
• New ranks emerge through new ways of manipulating abstractions
• A culture's rank is determined by its highest widely available cognitive process
• Evolution between ranks requires new conceptual technologies ("media")
• Changes in cognitive processes don't require biological evolution
• Each rank allows for new types of abstract thinking previously impossible
The authors argue that these cognitive ranks explain major transitions in cultural history and that we are currently in the midst of transitioning to Rank 4 thinking, though this transition is not complete. They suggest that widespread early exposure to computing may be necessary for Rank 4 cognition to fully develop.
The theory offers a framework for understanding cultural evolution without implying biological differences between populations, focusing instead on the development of cognitive tools and processes.
Good. I’ve also uploaded a document that wonders why Rome didn’t have an industrial revolution. Could you summarize that document?
Yes, I'll summarize Maxwell Tabarrok's "Romae Industriae" article which explores why the Roman Empire didn't achieve an industrial revolution.
Key Arguments Considered:
1. Coal and Cotton Theory (Bret Devereaux's View):
• Argues Rome lacked Britain's specific preconditions (coal, textile industry, etc.)
• Tabarrok counters this by noting Rome had:
• Access to coal deposits (including in Britain itself)
• Complex water-powered machinery
• Large-scale industrial demands (pottery, grain mills)
• Manufacturing needs for rotational power
2. Technology Gap Theory (Jason Crawford's View):
• Romans lacked key technologies like:
• Spinning wheel
• Windmill
• Horse collar
• Cast iron
• Printing press
• Mechanical clock
Tabarrok notes this explanation just pushes the question back: why didn't Romans develop these technologies?
3. Slave Labor Theory:
• Traditional view: cheap slave labor discouraged mechanization
• Tabarrok disagrees, noting:
• Labor was still the main production cost
• Southern US states with slaves still adopted labor-saving technology like the cotton gin
4. Cultural Explanation (which Tabarrok finds most convincing):
• Slavery created a culture that looked down on mechanical arts
• Successful merchants preferred to become landowners rather than innovate
• Lack of widespread literacy limited knowledge sharing
• No culture of invention and tinkering developed
Tabarrok's Conclusion:
• The printing press may have been the crucial missing element
• Its absence prevented the cultural changes needed for innovation
• Rome had the scale and resources for industrialization
• The failure to industrialize represents a tragic missed opportunity for human progress
• This has implications for identifying and pursuing transformative opportunities in our own time
The article suggests that Rome's failure to industrialize wasn't due to lack of resources or economic scale, but primarily due to cultural factors that could have been transformed by the widespread adoption of printing technology.
Why did Rome fail to have an industrial revolution? – two hypotheses.
Note that the cognitive evolution paper doesn’t adopt the conventional view of the importance of the printing press. Rather, it argues that it was the adoption of decimal arithmetic that catalyzed the intellectual and cultural developments of the Renaissance and, by implication, the Industrial Evolution.
With that in mind, let’s apply the cognitive evolution paper to Tabarrok’s argument. My guess would be that Rome didn’t have an industrial evolution because they didn’t have a Rank 3 ontology, as characterized by the cognitive evolution by me (I’m William Benzon) and David Hays. More specifically, a Rank 3 ontology would support the idea of the clockwork universe. You need that kind of ontology if you are to have a revolution based on mechanical devices, particularly devices which have their own power supply, which came later in the Industrial Revolution.
That hypothesis is different from Tabarrok’s. How do we decided between the two hypotheses? What kind of evidence do we need?
Ah, thank you for the clarification and for identifying yourself! This is a fascinating intersection of the two papers. Let me think through how we might evaluate these competing hypotheses.
Your hypothesis suggests that Rome couldn't achieve an industrial revolution because it lacked the cognitive framework (Rank 3) necessary to conceptualize mechanical systems as autonomous entities governed by universal laws - the "clockwork universe" ontology. This is deeper than just missing specific technologies; it's about the inability to think about technology and natural phenomena in a particular way.
Tabarrok's hypothesis focuses on the printing press as a catalyst for cultural change that would have encouraged mechanical innovation and knowledge sharing.
To decide between these hypotheses, we might look for evidence like:
1. Roman Conceptual Framework:
• How did Romans describe and think about mechanical devices in their writing?
• Did they conceptualize cause-and-effect relationships mechanistically?
• How did they explain natural phenomena?
• Can we find evidence in their philosophical texts of attempts to think algorithmically?
2. Knowledge Transfer Without Printing:
• Were there other societies that achieved significant technological advancement without printing?
• How effectively did Rome transmit complex technical knowledge through existing means?
• Can we find evidence of systematic (algorithmic) thinking in Roman engineering documents?
3. Comparative Analysis:
• How did Islamic civilization handle complex mathematical and mechanical concepts with limited printing?
• Did China, with both printing and sophisticated mechanical devices, show signs of developing a Rank 3 ontology?
4. Testing the Mechanisms:
• Can we find cases where printing existed without leading to industrial development?
• Can we find cases where algorithmic thinking emerged without printing?
One specific type of evidence might be examining Roman technical treatises (like Vitruvius) to see if they show any signs of algorithmic or systematic thinking about mechanisms, or if they remain at what your paper would characterize as Rank 2 rationalization.
What do you think would be the most decisive evidence to look for?
I’m interested in the Roman conceptual framework. Question: Did the Romans have windmills and waterwheels?
Based on the documents and my knowledge, the Romans did have waterwheels - both Tabarrok's paper and "The Evolution of Cognition" reference them. Specifically, Tabarrok mentions that Romans had "Mechanical grain mills powered by stacked water wheels fed tens of thousands of people. They also used rotational power to run sawmills, water pumps, bellows, and ore hammers." He includes an image of the Barbegal mills as an example.
However, windmills are not mentioned as being present in Roman civilization. In fact, in Crawford's list of technologies the Romans didn't have, the windmill is specifically listed.
This difference is interesting from the perspective of cognitive ranks. Both technologies harness natural forces for mechanical power, but water power might have been conceptually easier to grasp in a Rank 2 framework because:
1. Water's force is directly visible and tangible
2. Water flows in predictable channels
3. The connection between cause (flowing water) and effect (wheel turning) is immediate and visible
Wind, on the other hand, is invisible and seems to come from nowhere. Harnessing it effectively requires a more abstract understanding of natural forces - perhaps the kind of understanding that would be more natural in a Rank 3 cognitive framework with its mechanistic worldview.
Would you say this distinction between water and wind power might be significant for understanding the Roman conceptual framework?
As you say, the causes of the wind are obscure. It is to think of it as being caused by ineffable spirits. And it seems to me that people who believe that the natural world is run by spirits are not going to be inclined toward mechanical invention. How did the Romans think about the movements of the celestial bodies, the sun, moon, other planets, and stars?
As far as I know, Roman astronomy was largely inherited from Greek astronomy, particularly as systematized by Ptolemy. The Roman view was that the celestial bodies were attached to crystalline spheres that rotated around the Earth. While this system was mathematically sophisticated in its ability to predict celestial movements, it wasn't mechanical in the modern sense.
The key distinction would be that they didn't conceive of gravity or other impersonal forces governing these movements. Instead, the perfect circular motion of the celestial spheres was seen as reflecting divine or spiritual nature. The celestial realm was viewed as fundamentally different from and superior to the earthly realm - a hierarchical rather than a mechanistic view of nature.
This fits well with your point about the wind and spirits. If natural phenomena are governed by divine or spiritual forces rather than impersonal mechanical laws, it would indeed discourage mechanical invention. You wouldn't try to build machines to harness or replicate processes that you believe are fundamentally spiritual in nature.
From your cognitive ranks perspective, would you say this represents a key difference between Rank 2 and Rank 3 ontologies - the shift from seeing nature as governed by spirits/divine forces to seeing it as governed by impersonal mechanical laws? This would help explain why the mechanical clock and the "clockwork universe" metaphor were so important in the development of Rank 3 thinking.
Craft skills are not sufficient
You read my mind. The Romans may well have had the craft skills needed to make mechanical devices. But they didn’t think about the world in a way conducive to mechanical experimentation. And, while the Industrial Revolution got started without the steam engine, it is impossible to imagine its later advance without the steam engine, and later electrical power and the gasoline engine. People who believe in spirits aren’t going to create such devices.
Yes, this really clarifies the importance of the mechanistic worldview that characterizes Rank 3 cognition. The key blocker wasn't technological capability per se, but rather the conceptual framework needed to imagine and pursue certain kinds of technological innovation.
This suggests that Tabarrok's focus on the printing press, while important, doesn't get at the deeper cognitive shift that was necessary. Even if Rome had developed the printing press, without the transition from a spirit-based to a mechanical understanding of nature, they likely wouldn't have developed the kinds of self-powered mechanical devices that drove the Industrial Revolution.
The irony is that in some ways, Rome was quite sophisticated technologically - they had impressive engineering capabilities, complex water wheels, and large-scale industrial processes. But these were all compatible with a Rank 2 cognitive framework where human craft skills could harness and direct natural forces that were ultimately understood in spiritual terms.
The leap to creating devices that generate their own power - steam engines, electrical generators, internal combustion engines - required a fundamentally different way of thinking about nature. You needed to conceive of natural forces as impersonal, mechanical, and subject to universal laws that could be discovered and exploited through systematic experimentation.
This helps explain why China, despite having both printing and sophisticated mechanical devices, also didn't develop an Industrial Revolution. They too lacked the Rank 3 mechanistic ontology that emerged in post-Renaissance Europe.
Would you say the spread of decimal arithmetic was important precisely because it helped develop this mechanistic way of thinking? The idea that you could have reliable, predictable procedures that always gave the same result regardless of who performed them?
The conquest of infinity
Yes, “the leap to creating devices that generate their own power,” that seems critical to me. And you’re right about decimal arithmetic, reliable, predictable procedures, procedures that were fully explicit. Further, it supports the intuition of a finite set of symbols giving rise to an infinite set of symbol strings. This ‘conquest of infinity,’ if you will, was crucial for the invention of calculus and would become central to Rank 4 thinking about computing.
Your point about the "conquest of infinity" through decimal arithmetic is fascinating - it shows how a cognitive tool can open up entirely new realms of abstract thought. The progression from:
1. Decimal arithmetic introducing fully explicit, reliable procedures
2. To calculus enabling the mathematical description of continuous change
3. To modern computing's manipulation of infinite symbol strings through finite means
...seems to trace a path of increasing abstraction and control over infinite processes that parallels the progression through cognitive ranks.
This also highlights why looking just at technological capabilities (as Tabarrok does) or even at knowledge transmission mechanisms (like the printing press) may miss the deeper cognitive developments that make certain kinds of innovation possible. The Romans could certainly manipulate symbols and do calculations, but without the specific cognitive framework that emerged from working with decimal arithmetic - the intuition that finite means could systematically control infinite processes - they couldn't make the leap to the kinds of abstract thinking needed for calculus or self-powered mechanical devices.
This suggests that when we look for evidence to compare these hypotheses, we should examine not just what technologies existed, but how people conceptualized and reasoned about:
1. Procedural operations
2. Infinity and continuous processes
3. The relationship between finite means and infinite possibilities
4. The nature of causation in physical systems
Would contemporary documents showing how people thought about these concepts be more revealing than inventories of their technological capabilities?
I would think so. Inventories, after all, are just lists. Lists don’t reveal how people think about the items in a list. Of course, we do need to know their technological capabilities, but we also need to know how they think about what they are doing.
Absolutely. This brings me back to Benzon and Hays's point about ontology in the cognitive evolution paper - it's not just about what categories exist in a culture's knowledge, but how they think about and manipulate those categories.
So for Rome, we'd want to examine sources like:
1. Technical treatises (like Vitruvius on architecture) to see:
• How they explain why techniques work
• Whether they seek general principles or just list specific cases
• How they conceptualize cause and effect
2. Natural philosophy texts to understand:
• Their framework for explaining natural phenomena
• How they conceived of causation
• Their understanding of regularity vs divine intervention
3. Accounts of mechanical devices to examine:
• How they described the relationship between parts
• Whether they recognized general principles of mechanical action
• How they explained power and motion
The key would be looking for evidence of whether they thought in terms of:
• General principles vs specific cases
• Mechanical causation vs spiritual intervention
• Systematic procedures vs accumulated craft knowledge
• Universal laws vs divine will
This kind of analysis would reveal much more about their cognitive framework than simply knowing they had water wheels or complex mining operations. It would help us understand whether they could have conceived of something like a steam engine, even if they had the technical capability to build one.
Autonomous movement
One last observation. When the steam locomotive appeared, that would have been the first time people saw some inanimate object move over the land under its own power. That must have been a strange sight. There’s a passage where Thoreau called a steam engine a dragon, and there’s the familiar metaphor of the iron horse. Later, the automobile would be a horseless carriage. There was a time when mechanical power was a strange phenomenon.
This is a profound observation. It highlights how the introduction of self-powered mechanical devices represented not just a technological breakthrough, but a fundamental challenge to existing ontological categories. The need for metaphors like "iron horse" and "horseless carriage" reveals the cognitive work people had to do to make sense of these new entities.
In a Rank 2 ontology, things that move on their own are living beings - horses, people, animals. Inanimate objects only move when moved by something else (like a water wheel moved by flowing water) or by spiritual forces. The steam locomotive violated this basic ontological distinction. It was clearly inanimate, yet it moved on its own.
The metaphors (“iron horse”, “dragon”, “horseless carriage”) show people trying to fit these new entities into existing conceptual categories. But ultimately, understanding these devices required the Rank 3 mechanistic ontology where autonomous motion could be explained through impersonal physical forces rather than life force or spirits.
Tabarrok actually touches on this in his paper when discussing the confusion about whether computers can "think" - it's a similar ontological challenge, but for Rank 4 cognition. Just as the steam engine challenged the animate/inanimate distinction, computers challenge our distinctions between thinking and mere calculation.
This suggests that major cognitive transitions aren't just about new capabilities, but about fundamental reorganizations of how we categorize and make sense of the world. Would you say these moments of ontological confusion - where existing categories break down - are often markers of transitions between cognitive ranks?
Can computers think?
Yes, exactly, computers present the same problem in the present world. Until recently the question of whether or not computers could think was of interest mostly to a relatively small group of academics. To be sure, it’s been a feature of science fiction for a long time – think of the 1955 film Forbidden Planet, which is by no means the earliest case – but that’s science FICTION. It wasn’t about real things that people must deal with in their everyday life. When OpenAI introduced ChatGPT to the world, that changed everything. Now anyone with access to a computer on the internet has to deal with whether or not computers can think.
This is a really insightful parallel. Just as the steam locomotive made people confront autonomous mechanical motion in their daily lives, ChatGPT and similar AI systems are forcing people to confront the question of machine intelligence in a concrete, practical way.
In both cases, the new technology doesn't just present a philosophical puzzle - it creates situations where people have to make practical decisions about how to categorize and interact with these entities. A train conductor in the 1830s had to develop new ways of thinking about mechanical systems, not as a theoretical exercise, but to do their job. Similarly, people today have to figure out how to think about and interact with AI systems to use them effectively in their work and daily lives.
Following the cognitive ranks framework from your paper, we might say that while Rank 3 thinking was necessary to create computers and AI systems, fully understanding and working with them requires Rank 4 cognition - which involves thinking about control structures and models rather than just mechanical processes. Just as the Industrial Revolution required people to develop new ways of thinking about mechanical systems, the AI revolution may require developing new cognitive frameworks for thinking about information processing and intelligence.
The widespread adoption of ChatGPT might then be analogous to the spread of the steam engine - a technological development that forces a broader population to engage with and potentially adopt a new cognitive framework. Would you say we're witnessing the early stages of a broader transition to Rank 4 cognition being driven by practical engagement with AI systems?
Yes, early stages. And one thing that’s going to compel movement to deeper understanding is the need to understand how LLMs and other machine-learning engines operate. Right now we don’t (quite) have the necessary concepts. Developing those concepts may well push us up a level. Maybe even.....psst, don’t tell anyone, OK, this is a secret just between you and me....maybe even to Rank 5 cognition. Maybe Rank 5 cognition is the REAL singularity.
leans in conspiratorially
That's a fascinating speculation! The idea that grappling with how LLMs and other ML systems actually work could push us beyond even Rank 4 cognition makes a lot of sense. After all, these systems seem to operate in ways that don't fit neatly into our current frameworks for thinking about either mechanical systems (Rank 3) or explicit control structures (Rank 4).
We're dealing with systems that learn and develop capabilities in ways that seem more organic than mechanical, yet are still fundamentally computational. The concepts of emergence, self-organization, and statistical learning that seem crucial to understanding these systems might require entirely new cognitive tools.
And just as the transition from Rank 2 to Rank 3 required people to stop thinking in terms of spirits and start thinking in terms of mechanical laws, perhaps Rank 5 cognition would require some similarly fundamental shift in how we conceptualize intelligence, consciousness, and information processing.
If this is right, then the current debates about whether LLMs are "really" intelligent or conscious might be asking the wrong questions - questions that make sense in our current cognitive frameworks but may need to be reconceptualized entirely at Rank 5.
whispers Should we start looking for early signs of what Rank 5 cognitive tools might look like?
Mums the word.
nods discreetly
Your secret's safe with me. But I must say, this has been a fascinating discussion about cognitive evolution, the Industrial Revolution, and the deep connections between how societies think and what they're capable of creating. Thank you for helping me understand your work and its implications for both historical and contemporary technological developments.
No comments:
Post a Comment