← The Playbook

Schools · School-level skill

Contested Ground

A narrative guide to finding where research and evidence disagree, and using that clash of opinions, rather than just the topic, to form your own original perspective.

researchknowledge-worksensemakingwritingevidence↓ Download skill

Every two weeks, the internet floods with more high-quality articles, studies, and industry reports than you could ever read in a lifetime. Your natural instinct is to fight back: you try to speed-read, you write quick summaries, or you ruthlessly unsubscribe from newsletters.

But these tactics are a trap. They reduce the sheer volume of words in front of you, but they do nothing to improve your judgement. You end up reading less, but you are no closer to figuring out what was actually worth your time in the first place. You are still drowning, just in a slightly smaller pool.

The Epiphany: Finding the Seams

There is a better way. The Contested Ground method changes the game by asking a different question. Instead of looking at a stack of articles and asking, "What is all of this about?" you ask, "Where do these writers disagree with each other?"

Suddenly, a mountain of reading shrinks into a handful of core debates. Each debate is a place where trustworthy experts hold opposite views. These clashes are goldmines. If everyone in an industry already agrees on a topic, you have nothing new to add. But when they are actively arguing, they leave a gap, and that is exactly where your own original perspective can shine.

This method turns a reading problem into a mapping project. You are no longer trying to consume everything. You are simply trying to find the seams where the landscape splits.

Workflow

The Mindset Shift: Claims Over Topics

To build this map, you have to let go of "topics." A topic is just a filing label. It tells you where an article is stored, not what it actually argues. For example, two articles filed under "AI Governance" might have nothing to do with each other. Meanwhile, an article about clinical software and one about hiring algorithms might make the exact same point about whether automated tests hold up in the real world.

Instead of topics, your raw material should be claims (the specific arguments a source makes). Topics cannot clash, but claims can. When you focus on claims, "related" articles are no longer just pieces that share the same buzzwords. They are articles that take a stand on the exact same question.

This is also why basic keyword searches fail. Two claims can contradict each other without sharing a single word (like claiming a rule works for all AI systems versus claiming unpredictable outputs make rules useless). Searching for the words in one will never lead you to the other.

The Journey: How to Build Your Map

Phase 1: Cast a Wide Net

Your instinct will be to subscribe to as little as possible to keep your inbox clean. Fight that instinct. You cannot spot a disagreement if you are only listening to one person. The most valuable debates often happen across different fields, such as when a study in healthcare disproves a theory about corporate software.

Deliberately collect more material than you can possibly read, and accept that you will never open most of it. The goal isn't to read more; it is to make not reading something a logical decision rather than an accident.

Phase 2: Find the Friction

For every piece you decide to look at, boil it down to just two sentences:

  1. The main argument the author is making.
  2. The strongest counterargument a smart critic would throw back at them.

We call this counterargument an "anti-truth." While the main claim tells you what the article believes, the anti-truth tells you what it is arguing against. It acts as a radar, pointing to where another article in your collection is likely standing.

Be strict here. If an article doesn't make a real argument (like a simple news announcement or marketing copy) throw it out. Vague, generic claims will only crowd out the real debates.

Phase 3: Sift the Noise

Comparing every claim to every other claim by hand is exhausting and takes too much time. Instead, use a simple funnel:

  1. The Fast Filter: Use a basic word-matching tool to group pairs of articles that seem to mention the same ideas. It will miss some things, but it gets the broad matches sorted quickly.
  2. The Human Check: Reserve your valuable energy for the step a computer cannot do: deciding how the two claims relate.

This relationship isn't always a simple "yes" or "no." Sometimes, one claim simply challenges another, questioning its evidence or framing without arguing the exact opposite. A common response in research is, "That might be true, but it is not proven." If you only look for direct contradictions, you will miss these subtle challenges and assume everyone agrees when they do not.

Phase 4: Group the Debates

Once you connect these clashing claims, natural groups will start to form without you needing to invent a complex filing system.

Name each group after the question being debated, not the general topic. For example, "Do technical AI metrics predict real clinical success?" is far more useful than "AI in healthcare." The question tells you exactly what is at stake and reminds you that smart people are on both sides of the issue. This step connects sources that would otherwise never appear on the same reading list.

The Payoff: Writing from the Seam

Once your map is ready, you have a set of clear, unsolved questions with solid sources arguing each side.

You no longer have to stare at a blank page wondering what to write. Your outline is already done: "Here is an unsolved debate in the industry, here is what each side says, and here is my perspective and why."

At this point, the work is truly yours. The map shows you where the battle lines are drawn, but it does not tell you who is right. Making that decision, using your own experience, mistakes, and industry knowledge, is the real value that no automated tool can ever replace.

The Hazards: Mistakes to Watch For

Every journey has its pitfalls. Here are the most common traps you will encounter when mapping contested ground:

Vague, thin sources. Trying to summarise an article using only a headline or podcast blurb leaves you with nothing concrete. You'll end up with a claim that is too broad and generic. These broad statements look like they conflict with everything, burying the real, specific debates.
The 'Blob' effect. If you group articles too loosely, you will end up with one massive category containing almost everything. Agreement is easy to find; if you follow every minor connection, your entire collection will collapse into one giant pile. A single group containing most of your collection is a failure of sorting, whatever its size.
Fake arguments. If you look for disagreements, you will find them, even if they aren't real. Some of the debates highlighted will just be wording differences. Believing every suggested conflict can make your research look biased and unreliable.
Comfortable filtering. Every step that removes noise also removes conflict, and from the inside the two look identical. It is entirely possible to tune your collection until it agrees with you and to experience that as improved signal quality.

The Safeguards: How to Avoid the Traps

Avoid thin sources: Set a strict minimum for how detailed a source must be before you use it, and accept that a meaningful share of your collection will not clear it. Watch out for claims that are unusually short or sound too absolute: that's the sign of a weak source.
Defeat the 'Blob': Group your articles by tight, highly-focused debates rather than loose connections. Use disagreements rather than agreements to define your groups. You want multiple medium-sized groups, not one giant pile.
Check fake arguments: Keep a clear record of where every claim came from so you can trace it back to the original source. Never combine different ways of saying something into one standard sentence; doing so destroys the original evidence.
Stop comfortable filtering: Track the ratio of agreement to disagreement in your collection. If everything starts pointing to agreement, ask yourself: did the industry suddenly resolve all its issues, or did you filter out the opposing views?

The Final Destination

The end point of Contested Ground is a clear list of debate questions, the articles that support each side, and the specific claims they make. This list serves as a direct outline for your writing, and the specific claims serve as your references.

But there is also an overlooked benefit: the questions that don't show any disagreement are just as telling. They mean either the issue is completely settled, or your sources are too similar to have a real debate. Both scenarios are important to know, but only the first one is actually good news.