Can we still create knowledge?
Have you ever felt overwhelmed by the amount of information that exists out there? Sometimes, it seems like it would take several lifetimes to get a good grasp of any topic, and contribute something relevant.
Despite being intangible, knowledge is a form of capital that can accumulate over time. It can be transferred, it can depreciate and it can be productive by allowing for the accumulation of more knowledge.
Although many organisms produce, maintain and transfer knowledge, humans are unique in the extent to which knowledge plays a crucial role in their survival. Our large brains — largest of any mammals relative to body size— allow us to explore our environment and recombine elements to produce new thoughts. Our robust memories allow us to maintain large databases of information over long periods of time. And our ability to communicate through speech allows us to efficiently transfer our knowledge to other individuals. As a result, new generations are endowed with the knowledge developed by their ancestors, leading to a process of accumulation.
Human memory is quite impressive, yet the brain is subject to physiological limits for its processing and storage capacity. In addition, like the rest of our organs, a brain is senescent, meaning that its cells degrade over time. Eventually, when an individual dies, all the information stored in their brain also vanishes. This can be problematic for the accumulation of knowledge in a society. Imagine a tribe where only one individual knows which mushrooms can be eaten and which are poisonous. If an accident kills that individual before he has time to transfer his expertise to anyone, the entirety of the tribe’s mushroom knowledge disappears.
Such limits to our cognitive abilities have over time been compensated by a number of artifacts. The first human beings who engraved symbols in dried earth, effectively extended their memories to the physical environment. People were no longer limited to their brain capacities, they could simply write down information to store it and transfer it to other individuals.
Such inventions made it easier to create, transfer and maintain knowledge over long periods of time. As the population size grew, more individuals could contribute to the total knowledge capital, leading to unprecedented knowledge accumulation. Newton famously said in 1675 “If I have seen further it is by standing on the shoulders of Giants.” As knowledge accumulates, scientists can benefit from their predecessor’s achievements to make new discoveries or improve a technique even further.**
Take the story of the bicycle. The first version of the modern bicycle can be traced back to 1817 Germany, where the Baron Karl von Drais invented a steerable two wheeled contraption. The next attempt to make a bicycle was a bit more successful, with the addition of pedals which allowed for individuals to propel themselves. The individuals following in the Baron’s step benefited from previous discoveries and built on top of them to improve the bicycle.
Yet, as the total amount of knowledge grows, society faces a new problem. The more accumulated knowledge there is, the longer it takes for an individual to reach its boundaries. To understand this more intuitively, think of the total accumulated knowledge as a big tower. At birth, each individual starts climbing all the way up the tower. Once they arrive on top, they can build a new floor. At some point however, individuals face a practical problem: it takes them so long to climb up the tower that they cannot reach the top in their lifetime. Today, an individual who wants to improve our knowledge of the universe must first spend decades understanding the intricacies of relativity and string theory.
Thankfully, humans invented tools to climb the tower of knowledge faster — what I call efficiency innovations. Over decades, we have improved our teaching techniques such that individuals can learn faster, therefore compressing the time that it takes to reach the top of the tower. Modern schooling is just an example of a way to transfer knowledge very efficiently to young students. Technology can also help us climb faster. Think of how much easier it is to learn a new language thanks to language apps and instant translators. In addition, many domains are siloed enough such that one needs only very specific knowledge to understand them. For example, a data scientist does not need to know how a computer physically encodes information to be able to develop new algorithms. The division of knowledge into different fields allows to break down the big tower into a number of small towers, that one individual can climb easily in a few years.
Such efficiencies have allowed our total knowledge capital to continue growing and become orders of magnitudes larger than what a single brain can process. What does this mean for the future of innovation?
First, as time goes by, innovation will slow down. As it takes longer to reach the top of the knowledge tower, it takes more and more time for scientists to make discoveries. Such an observation has already been made by professor Nicholas Bloom, although he invokes different reasons for the slow down (namely the apple tree model). Productivity growth has slowed down progressively in the United States from its peak at 4% per year in the 1950’s to only 1% per year today.
However, if new efficiency innovations are made, there is a chance that the speed of scientific discoveries may pick up again. As a PhD student, I spend a lot of time climbing up the tower of knowledge through painfully long literature reviews. Sometimes it can take me several months before getting a grasp of what has already been said on a specific topic. Progress in AI could one day automatize the process of literature reviews so that it would take a scientist only a couple hours to map out what has already been done in a particular field. Such advancements appears closer every day, with the significant improvement of natural language processing programs such as GPT-3.
Second, the division of knowledge will become more intense. In the 18th century, it was common to be an expert in several domains. Newton is often described as a mathematician, physicist, astronomer, and theologian. Adam Smith is famous for his work in economy theory, but he also extensively wrote about social psychology and morality. Today, the story is quite different. An academic claiming that she is a philosopher, a mathematician and a psychologist would raise a few eyebrows.
To become an expert, individuals will have to choose increasingly narrow subjects. This specialization may make it harder for experts to learn from other fields, as they evolve in isolated bubbles with their own tools and vocabulary. Yet, many breakthroughs in the past centuries came from the transposition of a discovery from one field to another. For example, the 2019 Economics Nobel Prize was given to Esther Duflo and Abhijit Banerjee, famous for taking the clinical trial approach to the field of economics. In a world where knowledge is increasingly siloed, people who maintain the ability to build bridges across fields will be increasingly in demand.
Finally, innovation will be powered by teams, not individuals. As knowledge continues to accumulate, one way to keep up is to bring several individuals together to work on a project. It took only two men, Wilbur and Orville Wright, to invent the first powered aircraft. We can however expect that the invention of the solar-powered flying car will be the work of a large team of individuals, all contributing very specific skills to the project. This trend is already visible in research, where the average number of authors per published paper keeps rising.
As more people collaborate on a single project, new challenges arise. It is often easier to coordinate the work of 2 people than that of 50 people. The next big discoveries may be delayed as a result of poor coordination between individuals, just like a brain whose neurons were not properly connected to one another.
** Note: Not all knowledge is cumulative. Sometimes new theories and techniques replace older ones, without building on top of them.
I am incredibly indebted to Nicolas Baumard and Jean-Baptiste Andre for all our conversations about knowledge as a form of capital. I strongly recommend their paper “Cultural evolution by capital accumulation”.