Too Much to Process

In our increasingly complex world, we need new tools and frameworks to understand and manage knowledge.

The theme of the 2019 re:publica, a prominent European conference on digital culture, perfectly aligned with this call. It recognizes that we live in complex times and that we need tools to illuminate issues from different angles. In their own words:

Things are complex. Things need to be thought through, discussed, and viewed from different angles. That’s why we’re dedicating the next re:publica to long-form content, the fine print, the footnotes, the power of research, the power of controversy, and the urgency of addressing the topics that divide us (or unite us!).

As a communication designer, I approach the challenge of creating better tools to work with knowledge through the lens of metaphorical thinking. This essay explores how we can better share abstract mental concepts, use design and technology to understand our reality, discover and explore information more easily, and identify patterns in the noise of information to enrich our worldview.

A Framework for Understanding

A concept that has proven particularly helpful in thinking about knowledge management is that of mental filters. These are linguistic or visual concepts that we can “install” on our cognitive “bio-hardware” to see things more clearly.

To illustrate this, consider the medieval practice of ship navigation. Sailors would use sunstones held up to the clouds to locate the sun or moon, providing a fixed point for navigation. Mental filters function similarly, allowing us to discern patterns and meaning in the complex information landscape around us.

One of the first mental filters I encountered came from discussions within the Chaos Computer Club, a prominent European hacker organization. They introduced a crucial distinction between “truth” and “what really happened.” This separation acknowledges that when we talk about events, there isn’t just one truth, but multiple perspectives. Making this distinction is critical in our approach to information and knowledge.

When we observe the world, we receive a constant stream of data through our senses. Our brains break down these observations into individual schemas. For instance, when looking at an apple, we might note its shape, outline, color, and any distinctive features. We then embed this information into our existing mental framework through associations, connecting the new concept with things we already know.

The Double-Edged Sword of Abstraction

The human mind has a remarkable ability to take these abstract connections and reduce them to symbols. This process of bundling knowledge areas together and compressing them into symbols is both powerful and problematic.

On one hand, compression allows us to handle vast amounts of information efficiently. We can abstract complex ideas into simple symbols that are easy to communicate and manipulate mentally. This is how we move from raw data to information to knowledge.

However, this compression comes at a cost. As we abstract and compress, we inevitably lose some of the original information. This is similar to how image compression works in digital technology. Areas of an image are grouped into pixel clusters that approximate the original, resulting in an image that is recognizable but not identical to the source.

The benefit of this compression is that we can convert our observations into signs that can be easily communicated. We can name things, creating linguistic concepts or codes for what we observe. However, this process is not perfect. When we communicate these compressed symbols to others, there’s always a chance of misinterpretation or loss of meaning.

This imperfect transmission is due to the process of contextualization. Each person interprets information based on their existing knowledge, life experiences, and personal biases. As a result, the same symbol or concept can be understood differently by different individuals.

The Subjective Nature of Perception

The process of compression and contextualization doesn’t just happen on a small scale; it occurs fractally at all levels of our understanding. We package multiple connections into abstract symbols to communicate more efficiently, but this leads to a fundamental challenge: the theory of reality tunnels.

This concept, popularized by Robert Anton Wilson, suggests that reality is ultimately a mental construct, and we can’t all share the same reality. Our perceptions are, in essence, games of chance, and what we consider reality is merely our interpretation of sensory input filtered through our personal experiences and beliefs.

While this subjectivity is manageable in small communication structures, it becomes problematic when applied to larger networks. In today’s interconnected world, opinion formation largely takes place through networked exchanges. If we accept that individual perceptions of reality can only ever be communicated gradually and imperfectly, we begin to see the challenges this poses for shared understanding and consensus.

The Emergence of Shared “Truths”

To navigate this complexity, we often rely on what science fiction author Bruce Sterling calls “Consensus Narratives.” These are opinions that emerge within large groups and are eventually seen as truth, regardless of their relation to actual events or divergent opinions.

Consensus narratives gain traction because they are easy to tell and often confirm what we already believe. In our social interactions, we tend to gravitate towards stories that align with our existing worldviews rather than those that challenge them.

However, the landscape of consensus narratives has become more complex in the modern era. We now have actors with economic and political interests actively shaping these narratives. Social media platforms, for instance, use algorithms designed to keep users engaged, often by reinforcing existing beliefs and preferences.

The Challenge of Asymmetric Communication

The digital age has introduced new complexities to information management and communication. We now have actors with significant resources and advanced tools that can manipulate opinion formation on a massive scale.

For example, political entities might use bot networks to flood platforms like YouTube with slightly altered versions of videos, bypassing content filters and artificially amplifying certain messages. This creates an asymmetry in communication power, where some actors can concentrate opinions much more effectively than individuals.

The scale and complexity of these interactions make it incredibly difficult for individuals to navigate and understand the information landscape. The entanglements between various actors, platforms, and messages create a web that’s nearly impossible to fully comprehend.

The Gravity of Ideas

Joscha Bach, an AI researcher, introduces the concept of “Belief Attractors” to explain how certain ideas can become deeply entrenched. Using an analogy from physics, Bach suggests that some beliefs gain a kind of gravitational pull, making it difficult for individuals to escape their influence once they’ve been drawn in.

This concept applies not just to religious beliefs, but to any strongly held idea or worldview. The social and psychological costs of changing one’s beliefs create “valleys” that require significant energy to escape. This phenomenon makes it challenging for individuals to integrate contrary opinions or change their worldviews, even in the face of new information.

Rethinking Our Approach to Information

Given these challenges, we need new tools and metaphors to help us navigate the complex information landscape. However, many of our current interface metaphors come from a pre-network age, rooted in institutional and hierarchical thinking.

For instance, the desktop metaphor, which replaced physical filing cabinets with digital folders, still encourages us to think in terms of classification and hierarchy. While this approach worked well for physical documents, it falls short in representing the interconnected nature of digital information.

The Vision of Networked Information

The idea of networked information isn’t new. In the 1960s, Ted Nelson introduced the concept of hypertext, envisioning a way to make cross-references between different pieces of information. Nelson recognized that simulating human thinking through paper-like simulations was inadequate and sought ways to connect ideas across different hierarchical levels.

This network-centric view of information aligns with patterns we see throughout nature and the universe. From the distribution of dark matter to the structure of cities, network structures appear at all scales. Hierarchies, in this view, are simply isolated parts of larger networks.

A Better Metaphor

Almost 40 years ago, in the 1984 novel Neuromancer, William Gibson, who coined the term “cyberspace”, described it as looking like “clusters and constellations of data” resembling cities at night. This metaphor captures the networked nature of information better than hierarchical models.

When we engage with information, we naturally select what interests us, moving from broad concepts to specific details. However, our current systems often prevent us from making connections across different hierarchical levels. We need tools that allow us to connect individual words or concepts with other books, ideas, or information sources regardless of their position in a hierarchy.

Challenges of Visualization

As we move towards more network-centric models of information, we face new challenges in navigation and visualization. When dealing with complex networks containing thousands of data points, traditional visualization methods often fall short.

While we can reduce visual complexity through abstraction, such as representing nodes as simple points, this approach has limitations. Even with color coding and other visual aids, it becomes nearly impossible for humans to navigate and understand such complex systems visually.

The Promise of Machine Learning

Machine learning algorithms offer a potential solution to this challenge. These algorithms can project concepts, texts, and words into vector spaces, allowing for more nuanced clustering and visualization of information.

By repurposing these algorithms away from simply feeding us content and towards exploration, we can create more intuitive ways of navigating complex information spaces. We can project these vector spaces onto two- or three-dimensional spaces, creating cluster structures that humans, as natural pattern recognizers, can more easily interpret.

Towards Exploratory Knowledge Management

To truly advance our ability to manage and understand knowledge in the digital age, we need to embrace more exploratory and spatially-oriented approaches. By viewing our knowledge as interconnected islands or worlds rather than flat, hierarchical structures, we can better understand the paths we take through information and create more personalized knowledge landscapes.

This approach aligns with the idea of becoming “mature trackers of ourselves” and puts individuals on interesting paths that enable them to escape the gravity traps of entrenched beliefs. By bringing an exploratory character back into our tools, we can expand our understanding within our micro-societies and beyond.

Looking at this talk from today’s vantage point, several themes remain highly relevant and have even become more pressing:

1. Information Overload: The challenge of navigating vast amounts of information has only intensified with the continued growth of digital content and social media.

2. Algorithmic Influence: The role of algorithms in shaping our information diet and beliefs has become more pronounced, with increasing concerns about filter bubbles and echo chambers.

3. Misinformation and Disinformation: The spread of false or misleading information has become a major societal concern, highlighting the need for better tools to verify and contextualize information.

4. Spatial Computing: With the advent of augmented and virtual reality technologies, the idea of spatial representations of information is becoming more tangible and relevant.

5. AI and Knowledge Management: Advancements in artificial intelligence, particularly in natural language processing and knowledge graphs, offer new possibilities for managing and exploring complex information landscapes.

6. Digital Literacy: The need for improved digital literacy skills to navigate the complex information environment has become increasingly recognized in educational and policy circles.

7. Decentralized Knowledge: The rise of blockchain technology and decentralized systems has introduced new ways of thinking about information ownership and verification.

These ongoing developments underscore the continuing relevance of rethinking our approaches to knowledge management and the metaphors we use to understand and interact with information. As our digital landscapes evolve, so too must our tools and conceptual frameworks for navigating them.