The tools we think with are not neutral. A word processor does not merely record thought — it shapes the thought that gets recorded. A bullet list does not merely organise ideas — it selects for the kind of ideas that fit inside bullets. Marshall McLuhan spent his career arguing that the medium is the message, and he was right in a way that the personal knowledge management community has almost entirely failed to reckon with. The structure of a thinking tool is also the structure of the thinking it produces.
Consider the format in which most people record their thinking about a complex problem.
They open a document. They begin at the top. They write in the direction the language runs — left to right, top to bottom — and they proceed through the material in sequence: first the background, then the context, then the argument, then the conclusion. If they are organised, they use headings. If they are very organised, they use numbered lists. The document grows in one direction: downward. When they are done, the document represents their thinking about the problem in the form of a linear sequence of propositions arranged in a hierarchy.
This format feels natural, because it is. Writing has always been sequential. Language unfolds in time. The sentence must come before the next sentence. The paragraph must end before the next paragraph begins. These constraints are deep in the nature of the medium, and working within them feels not like a choice but like the only available option.
But here is the question worth asking: is the thinking that the problem actually requires linear? Or has the tool selected for the subset of the thinking that is expressible in linear form, and quietly discarded the rest?
McLuhan's Claim and Why It Was Taken Seriously
Marshall McLuhan's most famous proposition — "the medium is the message" — is one of those ideas that is widely cited, widely misunderstood, and, in the domains where it matters most, widely ignored.
The misunderstanding usually runs in one of two directions. The first is to treat it as a claim about communication only: the medium through which a message is transmitted changes how the message is received. Television makes things visual and emotional; text makes them linear and rational; radio is intimate and imaginative. This reading is not wrong, but it is significantly weaker than what McLuhan actually argued.
The second misunderstanding is to treat it as a claim about bias: every medium has a particular slant, and understanding the medium helps you correct for the slant. This is closer but still incomplete.
McLuhan's actual argument, most fully developed in Understanding Media and The Gutenberg Galaxy, is more radical: the medium does not merely colour the message. The medium constitutes the field within which messages are possible at all. It shapes what can be thought, not just what can be said. The printing press did not merely make books faster to produce; it produced a different kind of consciousness — one oriented toward sequential argument, individual authorship, standardised text, and the kind of linear, analytical reasoning that the text-based book both required and rewarded. The literate mind that emerged from centuries of print culture is not a natural kind. It is the product of the medium.
This is a stronger claim, and a more unsettling one. It says not just that our tools shape how we express ideas, but that they shape which ideas we are capable of having. The structure of the medium becomes, over time, the structure of the thinking. We do not use our tools neutrally and then choose what to do with them. We are shaped by them in ways we cannot easily observe from inside the shaping.
The Specific Pathology of Linear Tools
The dominant thinking tools of the modern knowledge worker are overwhelmingly linear: the word processor, the email thread, the slide deck, the meeting agenda, the project management board with its columns representing stages in a linear process.
Each of these tools organises thinking in a sequence. There is a first item and a last item. There is a beginning and an end. Progress moves in one direction. The format rewards the ability to arrange ideas in a coherent order and penalises the presentation of ideas that do not fit cleanly into such an order — ideas that are circular, that are constitutively dependent on other ideas that are themselves under development, that exist as part of a network rather than a chain.
The pathologies this produces are familiar to anyone who has been in an environment where linear thinking tools predominate.
The slide deck forces ideas into bullet points. Bullet points are syntactically parallel: they expect items of the same type, at the same level of generality, with the same relationship to the claim they are supposed to be supporting. Real arguments do not have this structure. Real arguments involve premises that support conclusions that become premises for further conclusions, qualifications that complicate the main claim, objections that need to be addressed before the argument can proceed, and connections between apparently unrelated considerations that are essential to the reasoning but impossible to represent in a list.
The result is the presentation that looks organised and is actually hollow — a sequence of short, parallel assertions that has the form of an argument without containing one. The bullet points eliminated the connective tissue that makes an argument an argument. What remains is a skeleton with no bones.
The email thread forces ideas into a conversational sequence. Each reply responds to the previous message, which responded to the one before it. The structure of the exchange is determined by what was said last, not by the logical structure of the problem being discussed. Ideas that were introduced early get buried. Connections between ideas introduced in different parts of the thread are invisible. The conversation moves forward in time but not necessarily forward in understanding.
The word processor is more flexible than the bullet point or the email, but it still enforces linearity at the level of the document: there is a beginning and an end, and the reader moves through the text in the order the writer chose. Ideas that are genuinely simultaneous — that must be grasped together to be grasped at all — must be serialised. One comes before the other, even though their logical relationship is not sequential.
What Non-Linear Thinking Actually Looks Like
The intuition that human thought is not inherently linear is not merely philosophical. There is a body of cognitive science and educational research that supports it directly.
Concept mapping — the practice of representing ideas as nodes connected by labelled relationships — was developed by Joseph Novak at Cornell in the 1970s as a research tool for studying children's developing understanding of scientific concepts. Novak's finding, replicated many times since, was that the structure of a student's concept map — the density of the connections, the accuracy of the relationship labels, the presence of cross-links between ideas from different branches of the map — was a far better predictor of genuine understanding than performance on propositional tests that could be passed by memorisation.
The concept map externalises the structure of understanding in a form that the linear note cannot. Understanding, as it develops in the mind, is not a sequence of true propositions. It is a network of interrelated concepts in which meaning emerges from the relationships between nodes rather than from the nodes themselves. The concept "gravity" is not a proposition; it is a node in a network that connects mass, force, acceleration, distance, orbital mechanics, general relativity, and dozens of other concepts, with relationships between them that constitute what it means to understand gravity rather than merely know the word.
A linear note about gravity can capture the propositions — the mathematical relationships, the definition of the field. It cannot capture the network. It can gesture at connections, but the gesture is not the connection. The connection is what exists in the mind of the person who understands.
This is why experts in a domain think differently from novices, even when they are given the same information. Experts have a denser, more accurately structured network of concepts. When they encounter a new problem, they perceive it as an instance of a familiar type because their concept network recognises the deep structure beneath the surface features. Novices, whose networks are sparser and less accurately connected, perceive the surface features and miss the deep structure. The difference is not in the propositions they know. It is in the network that gives those propositions meaning and allows them to be used.
Linear thinking tools build propositional knowledge. They do not build, and cannot represent, the conceptual network that distinguishes expertise from familiarity.
The Gutenberg Mind in the Age of Complex Problems
McLuhan's analysis of the printing press produced a specific account of what kind of mind the print era developed: linear, analytic, individual, sequential. The ability to follow an extended argument through hundreds of pages of text; to suspend judgement while building up a complex chain of reasoning; to work through a problem methodically, one step at a time, before arriving at a conclusion.
These are genuine intellectual virtues. The culture of print produced extraordinary things: the scientific revolution, the Enlightenment, the novel, the legal system, the research university. The linear, analytic mind that print cultivated was genuinely powerful for the kinds of problems it was applied to.
But there is a growing recognition — in complexity science, in systems thinking, in network theory, in the biological and social sciences — that many of the most important problems of the present are not the kind that linear analytic thinking solves well.
Climate change is a systems problem. Economic inequality is a systems problem. The dynamics of social media and political polarisation are systems problems. They involve feedback loops, emergent properties, non-linear causation, and network effects that linear chains of reasoning cannot adequately represent. The mind trained exclusively on linear tools will approach these problems with tools that are inappropriate for them — will try to identify cause-effect chains in phenomena that are constitutively about circular causation, will try to find single solutions to problems that have no single solutions, will try to reason sequentially about things that are structurally simultaneous.
This is not a failure of intelligence. It is a failure of tools. The linear thinker is not stupid; they are using an instrument optimised for a different kind of problem. And the persistent use of linear thinking tools in domains where non-linear thinking is required is producing exactly the kind of systematic analytical error that McLuhan's framework would predict.
The Network as a Thinking Structure
There is a reason why the Directed Acyclic Graph — the mathematical structure that underlies much of modern systems thinking, database design, and knowledge representation — is structurally more adequate to the representation of complex understanding than the linear document.
A graph can represent relationships that a sequence cannot. It can represent bidirectional dependence: the sense in which understanding A requires understanding B and understanding B requires understanding A — a circularity that linear tools can only gesture at, but that the mind navigates constantly in the development of understanding. It can represent multiple paths to the same node: the fact that the same concept can be approached from many different directions and that the approach matters for what is understood. It can represent the density of connection that distinguishes genuine expertise from surface familiarity: an expert's knowledge graph is dense with short paths between nodes; a novice's has the same nodes connected by longer, sparser paths.
Most importantly, a graph can grow in directions that were not anticipated at the outset. A document grows in one direction, and that direction is determined at the beginning: there is a first sentence, and the second sentence is constrained by the first. A network grows from any existing node, in any direction that a new connection warrants, without reorganising what already exists. It can accommodate the fact that intellectual development is not a linear progression from ignorance to knowledge but a multidirectional exploration that doubles back, discovers connections across widely separated terrain, and occasionally finds that two nodes it assumed were far apart are, in fact, immediately adjacent.
This structural property of the network — its capacity to grow in multiple directions simultaneously, to represent connections across domains, and to accumulate without requiring reorganisation — is not a technical nicety. It is an epistemological virtue. It reflects the actual structure of how understanding develops, rather than the convenient fiction that it develops in one direction.
The Tools We Choose Are the Minds We Build
The choice of thinking tool is not, as it is often presented, merely a matter of personal preference or workflow optimisation. It is a choice about what kind of thinking you are cultivating.
The person who thinks primarily in documents builds a document-shaped mind: one that is good at producing sequential arguments and poor at perceiving non-linear structure. The person who thinks primarily in bullet lists builds a bullet-list-shaped mind: one that excels at parallel enumeration and struggles with the kind of connected, qualified, multi-directional reasoning that complex problems require. These are not natural kinds; they are products of sustained engagement with particular structures.
This is not a reason for despair. If the mind is shaped by its tools, it can be reshaped by different tools. But it requires taking seriously, at the level of design, what kind of thinking a tool actually promotes — not just what kind of content it can store, or how elegantly it organises that content, but what cognitive activities it makes easy and what cognitive activities it makes hard.
The history of thinking tools is largely a history of tools that were designed around what was technically feasible or commercially attractive rather than what was cognitively optimal. The linear document was the natural product of the codex, which was the natural product of the scroll, which was the natural product of writing as a technology. The bullet list was the natural product of the typeset page. These tools were not designed to optimise thinking. They emerged from technological constraints and became so naturalised that they seem like the only way thinking can be externalised.
They are not. The range of possible structures for externalised thought is much wider than what has become conventional, and the choice between structures is consequential in ways that the conventional tools render invisible precisely because they are so conventional.
What This Demands of Tool Designers
The implication of McLuhan's framework, applied to thinking tools, is that the designer of a thinking tool is, in a significant sense, designing the minds of its users.
This is a large responsibility. It requires asking not just "what can this tool do?" but "what kind of thinker does this tool make the person who uses it?" It requires being honest about what the tool makes easy and what it makes hard, and about whether what it makes easy is actually what the user needs. It requires resisting the pressure to optimise for satisfaction — for the feeling of organised, productive thinking — in favour of optimising for the actual development of intellectual capacity.
It requires, above all, a willingness to build tools that do not fit the existing expectations of the market, because the existing expectations were formed by tools that were not built to optimise thinking. The person who has used linear tools all their life will find a non-linear tool unfamiliar and possibly uncomfortable. The discomfort is not evidence that the linear tool was better. It is evidence that the habit has been formed, and that forming a better habit will require a period of unlearning.
McLuhan argued that every new medium creates its own environment, and that the content of the new medium is always the previous medium — which is why the first films were shot plays, and the first television shows were radio programmes with pictures. The first thinking tools of the digital era were digital word processors and digital filing systems: the previous media, implemented in new technology. What has yet to be widely built is a tool genuinely native to the network — one that thinks in connections rather than in sequences, that grows in multiple directions rather than one, and that externalises the actual structure of understanding rather than the convenient fiction of its linear representation.
The linear thinking tools we have built have produced the linear thinkers we are. The question is whether the tools we build next will produce something different — something adequate to the complexity of what we are trying to understand, and honest about the structure of how understanding actually works.
The Honest Implication
McLuhan's most disturbing argument — the one that gets lost in popular summaries — is that the users of a medium are typically the last to perceive its effects. The fish does not notice the water. The person who thinks in linear tools does not perceive the linearity of their thinking, because the tool has shaped the perception as well as the thought.
This means that the problem is not correctable simply by pointing at it. Knowing that linear tools produce linear thinking does not immediately free a person from linear thinking. The habit runs deep, the tools are ubiquitous, and the alternatives are underdeveloped and unfamiliar.
What it does is create a different kind of responsibility. For the person who builds tools: the responsibility to design for cognitive development rather than cognitive convenience. For the person who uses tools: the responsibility to examine honestly whether the way they are thinking about a problem is a function of the tool they are thinking with, and whether a different structure might reveal what the current structure is hiding.
These are both uncomfortable responsibilities. They require questioning practices that feel natural and comfortable. They require investing effort in unfamiliar structures whose benefits are not immediately apparent. They require a long-term orientation that the culture of productivity optimisation does not easily accommodate.
But the alternative is to continue building linear minds for non-linear problems and to be surprised, repeatedly, when the tools that felt productive produce thinking that is not adequate to the complexity of what they were supposed to help us understand.
The medium is the message. And the message, right now, is that we have been choosing our media carelessly.
Aethel is built on a Directed Acyclic Graph because a graph is the honest structure for the representation of developing understanding. Every connection is real. Every relationship is explicit. The structure grows in the direction that understanding grows — outward, not downward.