Language as Knowledge

When reading intense academic papers, I am often confused and disoriented by the language. There is a great challenge in deciphering the knowledge that is ingrained in the text, especially if you are largely unfamiliar with the literature surrounding the subject matter.

But many times, while I can follow the discussion surrounding the theories, the mathematical formulas for me are very hard to understand and decipher. I often feel like a student that I am tutoring in math, one that is largely disoriented by the technical language and techniques.

Teaching Language

Currently, I am tutoring two students. One is a high schooler who does not understand math, the other is a “junior higher” who does not speak English. The high schooler is actually quite good at “in-the-head” math — he can multiply and add numbers in his head very well. However, he has a difficult time understanding the language and process of math. Using the example of addition, he tries to add several 4 digit numbers in his head instead of laying it out on paper in the standard vertical structure and adding them.

This is just a structural representation to understand how to accomplish a process. It is the same with complex formulas and breakdowns. They are just symbols to represent precise mathematical operations. It is a language that requires translation.

As I learn a language (such as mathematical expressions) or words/types of language (like the wording and terminology in research papers) I begin to gain a quick heuristic of what this symbol means. There is a representation of this linguistic entity in my head that I no longer have to work to formulate.

I find that I tutor both of these students in a similar manner. This manner is one that employs the use of similes, metaphors, and translations to effectively communicate the content of the language. This allows them to build a structural understanding of the grammar in their head.

The Mindset of a Beginner

Since a beginner does not have the structural representation built in their head, they will have to ask many questions to build it properly if there is no easy metaphor for a foundation. The perspective of a beginner is essential in the learning process for the expert, as they will consider many things that the expert has not addressed in many years.

Some of the hardest questions that I have attempted to answer have come from the youngest pupils. When one asks you what multiplication is and what its purpose is, it forces you to think deeply and exercise muscles that have not been used in a while.

But in order to run at the speed of expertise, you cannot pause to walk through all of these fundamental foundations of knowledge. You have to use the building blocks that you have amassed to build new theories — and new ideas. But when reconsidered from the perspective of the beginner, some of these building blocks may need rebuilding.

The Economy of Knowledge

I think that language is the foundational building blocks of knowledge, as it is the medium in which we tend to build the representation in our head. It is the way we communicate how to build these ideas, learning them and transmitting messages to others.

If there is an economy of knowledge — language is its currency. It is the output that is produced and the data that is consumed. Sometimes we do not share our knowledge, we keep it to ourselves. This can be beneficial in some circumstances, but detrimental in others. Sharing knowledge and communicating establishes trust. Without this trust, you can harm your ability to operate in certain environments.

Sharing language allows us to establish heuristics, frameworks, and building blocks that can be shared amongst one another. From this language, experts can be built and beginners can be taught. It creates a pathway of knowledge that can be followed by another person.

Language can also seem to be overvalued when it lacks internal content. We all have met people who sound really smart and know how to employ the tactics of language — but the content that they are speaking is actually not substantial. Or similarly, they can repeat/regurgitate knowledge of other people through their speech — but the knowledge is not their own. They truly don’t understand it, they are just speaking it.

The Future or Knowledge and Automation

We are now entering an era when knowledge will be consumed and produced by artificial agents. And as we enter this era, we must consider the harms that may befall us when we start to optimize for output, rather than the content of knowledge itself.

For instance, in the domain of marketing we have started to see automated content generators. These AIs can produce blog posts and social media content that can seem knowledgeable — under the guise of another person or organization. This is optimizing for the appearance of knowledge, for the purpose of establishing trust in the broader community. But this sort of trust is superficial — it does not say anything about the person/organization if they do not truly contain that knowledge themselves. In order to exhibit ethical responsibility, we must consider that we are using content automation to speak about ourselves and our own ideas effectively to establish the right sort of trust.

This can also be seen in the rise of academic automation, where we start to optimize for our output as researchers. Relying upon automation to help us synthesize ideas is one method, but riding on AI to write and prove results on our behalf will end up self-destructing our own ingenuity. The work will not be our own, and our own expertise will decay.

Being an Active Contributor to the Economy of Knowledge

When it starts to become so easy to rely upon automated sources of knowledge, we should attempt to dig in our heels and remember to exert ourselves like a beginner. The beginner is the one who seeks to learn, understand, and ask the right questions. Some of the best innovators are the ones that are difficult beginners — they fight against the status quo and ask many questions. They are often labeled as the ones with a learning disability because they think differently or ask too many questions.

And these difficult beginners are incredible experts because they have built the building blocks of knowledge with their own hands. They have questioned every corner, know every piece, and often have invented the theories themselves. And when we attempt to throw together building blocks with AI, without truly knowing what they are, we start to separate ourselves from true expertise.

We can’t learn everything. And it will be incredibly beneficial for us to rely upon AI to speed up our research development in certain ways. But we must not get so lost in the pursuit of progress and output that we lose sight of our own content, one day discovering ourselves to be trapped up on top of our apparatus of science — with no one who knows the way down.