But that’s also how so many others in the past made grave moral errors.

Right now, it seems important to make an assertion that is so obvious that one shouldn’t have to say it: Even in a war, all militaries/armed groups should be very careful to avoid unnecessarily killing ordinary people who are just going about their lives.

Of course, there’s no one value that’s infinitely valuable, so in wild situations, there can hypothetically be times when killing is justified. But the vast majority of the time, when groups have claimed that not being careful to avoid killing innocent civilians is justified, they were wrong.

Sadly, being very careful not to kill innocent civilians is not the same as killing zero civilians. The horrible reality is that war in areas with lots of civilians nearby will almost always lead to dead civilians even when a lot of care is taken not to kill them. But taking great care should lead to many fewer civilians dead than would have occurred otherwise.

When groups PURPOSELY kill innocent people to achieve their ends, I believe that more than 99.9% of the time they were morally in the wrong taking that action.  And if you find yourself advocating for PURPOSELY killing innocent people, it is a very strong red flag that you’re advocating that other people do something very immoral. There are incredibly strong moral priors that say you should take great precautions to avoid murdering innocent people.

In my view, one of the only potentially-valid justifications for killing innocent civilians is to save a much larger number of other innocent civilians. But, also, that calculus should be used with great care – I believe that most of the time, when it ends up getting used in real life, it is actually misused, and ends up (accidentally) becoming a justification for immoral behavior.

This piece was first written on October 17, 2023, and first appeared on this site on October 26, 2023.

It has been estimated that over 70% of adults in the world have experienced at least one traumatic event in their lifetime. Consequently, the majority of friendships and relationships will involve at least one person who has experienced trauma.

This article investigates how one person’s trauma can “clash” with another person’s trauma to cause relationship breakdown and what you can do to resolve conflicts like this if you experience them yourself. You might find it helpful to read if you…

• have some trauma in your past that sometimes gets triggered by other people’s behavior;
• find that your friends or romantic partners sometimes suddenly hurt you for inexplicable reasons;
• struggle to understand other people’s hostile reactions to actions you take;
• want a new model that might help you understand conflict with people close to you.

Let’s look at some specific examples of what we mean by Clashing Trauma: 

1. Trauma from anger and abuse 

Bob is sensitive to teasing because of a physically and emotionally abusive ex that used to tease him incessantly. Anne lightly teases him in what she thinks is a playful way, and he blows up in anger. Anne is sensitive to anger due to trauma in her childhood. She stops replying to Bob’s messages, which makes Bob even angrier. 

2. Trauma from abandonment and unwanted sexual advances 

Bob is physically affectionate with his close friend Anne. It starts to freak Anne out – when male friends have acted this way in the past, they have usually tried to initiate a sexual relationship with her in a way that she found traumatic. To prevent this, Anne stops being warm to Bob. Bob, however, regards Anne as a dear and platonic friend. Because Bob had a traumatic experience of abandonment in the past, he finds Anne’s sudden coldness very upsetting. The fear of abandonment causes Bob to be more affectionate toward Anne as he tries harder to reconnect. This triggers Anne further, causing her to withdraw even more, leading to Bob feeling even more hurt and confused. 

3. Trauma from substance abuse 

Anne has a history of alcoholism that Bob doesn’t know about. Years ago, she resolved to stop drinking when she realized she was becoming like her parents, whose drinking would often lead to physical fights that she found distressing. Bob drinks around her and encourages her to drink, which makes her angry. Bob has protected himself from people’s anger in the past using humor, so he tries to downplay Anne’s reaction by joking around and making light of the situation. Anne finds herself even more triggered and upset by Bob’s behavior – she wants her anger to be taken seriously because alcoholism was a huge challenge in her past.

In all three examples above, one person’s reaction to their trauma being triggered in turn triggers the other person’s trauma response. Clashing Trauma is an example of a broader pattern of Clashing Reactions, where one person’s reaction to an action that upsets them triggers the other person to become upset or angry, which in turn intensifies the first person’s negative emotions, and so on. We’ve noticed that a surprisingly high proportion of fights between people who care a lot about each other fit this basic pattern. The cases of Clashing Reactions on which this essay is focused involve reactions that have their foundations in trauma, but be mindful that trauma doesn’t have to be at the root for similar patterns to occur.

What is trauma?

Trauma refers to enduring changes in your emotional reactions or self-protective behaviors that are the result of painful or frightening incidents in your past. Examples of trauma that people are usually familiar with include experiences with war, abuse, or sexual assault. For example:

• A soldier who was shot at in combat may find that loud noises that sound like gunshots cause her to panic, even in completely safe environments.
• A man who was sexually assaulted by someone with an English accent may feel anxious around people with those same accents.

But it is important to remember that trauma occurs on a spectrum; many people have emotional reactions to reminders of upsetting experiences in their past, even when those experiences were not life-threatening or severely upsetting. For instance:

• Someone who received bad news over the phone may experience a sense of dread and anxiety when they receive a call from an unknown number.
• Someone who was teased by a school teacher for stuttering during a class presentation may find that they get nervous when asked to speak in front of people.

Trauma is the result of a mechanism your brain uses to protect you. This mechanism is often helpful – by sensitizing you to patterns that are similar to ones where you were hurt or frightened in the past, your brain tries to protect you from getting hurt again. Unfortunately, this protective mechanism sometimes goes too far, leading to reactions that can seriously impact people’s welfare (at which point we call it “trauma”).

The soldier, who is now in essentially no danger of getting shot, may have recurring panic attacks that reduce her quality of life and avoid places she used to love just because there are loud noises there. Trauma can manifest in relationships, too; many people are hurt by those close to them (unintentionally or not), and this can make us extra vigilant about avoiding similar painful experiences in the future.

What is going on when people’s trauma clashes?

Here is what we think is going on in situations where friendships or relationships break down due to Clashing Trauma:

1. Anne takes an action that seems normal and reasonable to her. Anne doesn’t know it, but her close friend (or partner) Bob is sensitive to this kind of action due to trauma in his past. Consequently, Bob experiences intense negative emotions as a result of her action.

2. Bob instinctively tries to protect himself from this negative emotion using whatever approaches seemed to help him during past traumatic events (e.g., by avoiding Anne, yelling at her, getting very distraught, or trying to control her behavior). But he doesn’t explain clearly to Anne what is happening. Anne is caught off guard by Bob’s reaction—she doesn’t understand why he’s behaving this way. Even worse, due to trauma in her own past, she gets triggered by Bob’s reaction and starts experiencing intense negative emotions of her own.

3. Anne now acts instinctively to try to protect herself from the way that Bob is making her feel (e.g., by lashing back out at him, avoiding him, shutting down, or trying to control him). This reaction results in Bob feeling even more upset, causing him to intensify his self-protective behavior.

You can now see how Clashing Trauma works:

• From Bob’s perspective, Anne suddenly did something that triggered his trauma, and when he tried to protect himself, she doubled down and hurt him more.
• From Anne’s perspective, Bob suddenly started acting in a strange and hurtful manner, and when she acted to protect herself, he hurt her even further.
• From the outside view, it’s clear that neither person intended to hurt the other, yet both feel hurt, and the relationship suffers.

Incidentally, if you think that you might be Anne or Bob, you’re not alone. When we shared a draft of this article, one friend of the author reached out asking if Anne represented them, and another reached out asking if they were Bob. But the post wasn’t actually about either of them! We are trying to describe a general pattern that we’ve observed many times. Or, put another way: if you think you are Anne or Bob, then this post may well apply to you, but it’s also about lots of other people as well.

How can you prevent Clashing Trauma?

Strategy 1: Prevention

The best way to prevent this pattern of conflict is to discuss in advance with trusted friends and romantic partners what triggers you each have, how to avoid these triggers, and what the best actions to take are if you trigger each other by accident.

For instance, Bob could have told Anne in advance that he can feel very attacked when someone is critical of him and that he tends to lash out in anger as a defense mechanism. Then Bob and Anne can work together to plan how Anne can give Bob feedback in ways that won’t trigger him.

Strategy 2: Interception

Barring Prevention, upon being triggered the first time, Bob would ideally wait until he’s calm and then tell Anne (without blaming her) that her action triggered negative emotions for him, explaining what the trigger was and how to avoid it. (See our post on non-violent communication for advice on how to communicate this kind of thing without making the other person feel bad.)

Then, Anne, caring about Bob, would ideally apologize for inadvertently hurting him and commit to trying to avoid those triggering actions in the future. But behaviors can take time to alter—Anne should set realistic expectations about how quickly she can make that change.

Interception is hard, but appreciating the negative consequences of unresolved Clashing Trauma might help motivate you to implement this strategy.

Strategy 3: Repair

If it’s too late for Interception, try Repair: once Anne is hurt by Bob’s response to her behavior, Anne could wait until she feels calm and then initiate a conversation with Bob.

During the conversation, she could explain how Bob’s behavior seemed (to her) to suddenly and mysteriously change and how this had hurt her, while expressing interest in hearing Bob’s experience of the situation. She would try to genuinely understand Bob’s experience (non-violent communication could also be a helpful way to employ this strategy).

Bob could then explain what he experienced in the situation and what made him react that way towards Anne. They could each commit to new behaviors to reduce the chance they trigger each other in the future.

Understanding trauma

Each of the Prevention, Interception, and Repair strategies involve Anne or Bob explaining their triggers (and/or the trauma that underlies them) to each other. These strategies will therefore be harder to implement if you don’t understand your own trauma or what can trigger it. One way to build an understanding of your trauma and triggers is to discuss it with a trained therapist or trusted friend (or partner). Conveniently, this is a great segue into setting up Prevention strategies with a trusted friend or partner so that you can avoid triggering each other!

Did you find this model of relationship breakdown helpful?

The next time you feel suddenly and unexpectedly hurt by a close friend or romantic partner, it may be worth asking yourself – could this be an instance of Clashing Trauma? Or, more generally, could it be a case of Clashing Reactions? In either case, consider the strategies of (1) Prevention, (2) Interception, and (3) Repair. A lot of great friendships and romantic relationships end for preventable reasons. We hope you can use these strategies to stay close to the people that you love.

This essay was first written on April 17, 2022, was turned into an article on Clearer Thinking (coauthored with Holly Muir) on May 5, 2022, and first appeared on this site on May 6, 2022.

1. Soldier Altruists think it’s obvious how to improve the world and that we just need to execute those obvious steps. They see the barriers to a better world as:

(i) not enough people taking action (e.g., due to ignorance, selfishness, or propaganda), and

(ii) bad groups blocking things (e.g., corrupt politicians or greedy corporations).

2. Scout Altruists think it’s hard to figure out how to improve the world – and most attempts either don’t work, only slightly help, or make things worse. They see the barriers to a better world as:

(i) not enough understanding of causal mechanisms (e.g., due to a lack of high-quality evidence, not enough attention to the evidence we do have, not enough careful reasoning, etc.), and

(ii) too much investment in bad solutions (e.g., due to people jumping to conclusions, doing what feels good emotionally rather than what is effective, ideological blindspots, inertia, etc.)

Soldier Altruists say we need to DO and FIGHT more. Scout Altruists say we need to THINK and TEST more. Soldier Altruists are more likely to think that if we could just get people to be less selfish and more motivated to act, we would make a lot of progress towards a better world. Scout Altruists are more likely to think that if we could just get people to pay more attention to evidence and to have more good-faith debates with strong norms around the quality of argumentation, we would make a lot more progress.

Soldier Altruists may think Scout Altruists are far too reluctant to act and are wasting their time on research and debate. Scout Altruists may think Soldier Altruists are far too confident in their conclusions and are wasting their effort pushing for changes that aren’t going to help much (and which, in some cases, might even make things worse). Of course, in reality, there is a continuum between these two positions. So, on a scale from 0 (Soldier Altruist) to 10 (Scout Altruist) where do you fall? I’m probably a 7 or 8.

As some commenters have pointed out, there is a relationship between this distinction and “Conflict Theory” vs. “Mistake Theory.” I think it is related – but also distinct in important ways. Conflict theory says that there is a giant zero-sum struggle (groups fighting over fixed resources). Whereas in this case, we’re operating from a framework of altruism: “the world can be made a lot better – what’s the big barrier to that happening? Is it that we know what to do and we’re not doing it enough/with enough energy, or is it that we don’t really know what to do?”

Also, to clarify another important point brought up in the comments: I’m not asking, “do you think it’s obvious how we should improve the world if we had a magic wand that could change whatever we wanted?” – instead, the question is: “is it obvious what to do to improve the real world, given that we don’t have a magic wand?” Do we just need to put more money/time/effort/people into executing the current “obvious” strategies because they will work well if we just scale them up? Or is it pretty unclear what strategies we should even be putting more resources into (meaning that a lot of thinking, research, debate and/or evidence evaluation will typically be necessary to even figure out what is worth scaling up)?

Julia’s book (which I highly recommend): https://www.amazon.com/Scout-Mindset-Perils-Defensive-Thinking/dp/0735217556/ref=nodl_

This piece was first written on April 23rd, 2021, and first appeared on this site on January 7th, 2022.

With that definition in mind, there are many reasons a work team may have low productivity. To make a team more productive, I suggest first trying to pinpoint the predominant causes of inefficiency, since different failure points typically have different solutions. The key is to identify and then focus on just the 1-3 of these causes that seem to be the biggest recurring blockers of team productivity.

Once these biggest blockers are identified, an analysis needs to be made of why they are occurring (e.g., using a method like the “5 whys”, where you try to uncover the root cause of the issue). Then, strategies need to be developed to try to resolve the blockers, and one or two strategies that have sufficient team buy-in need to be selected for implementation. Finally, re-evaluation needs to occur to make sure the problems are actually improved by those strategies.

Below is a list of these potential failure points or blockers.

There are four main facets of team productivity (Planning, Effectiveness, Resources, and Communication), which break down further into specific reasons for low productivity. You can use this list by considering each potential reason and scoring the degree to which you think it’s a problematic factor on your team (or having all team members anonymously score each potential reason for how big a problem it is, and then aggregating the responses to find the few most important seeming issues).

—

Reasons For Low Team Productivity:

—

(A) Planning

(1) Wrong Goal – the work is focused on achieving a goal that is not actually particularly valuable.

      e.g., the team is building a new feature that customers don’t want

(2) Poor prioritization – rather than focusing on the very most important things, the team focuses on things of secondary importance.

      e.g., the team lead keeps pushing team members to make tiny improvements to existing features, even though there are really major features where implementation hasn’t even begun.

(3) Slow Path – the planned route for achieving the team’s goal is not as efficient as other routes to achieving the same goal.

      e.g., the team could use a well-tested plugin to help them solve their problem, but the team lead insists they build their own alternative to this plugin.

(4) Incompletion – there is no pressure to finish projects, or team members work on far too many projects at once, so work doesn’t get completed, or takes much longer than it should.

      e.g., when a crisis comes up, the projects already in progress are abandoned, and rarely are they picked up again.

(5) Feature Creep – the project takes a really long time to be released because its scope or feature set keeps unnecessarily increasing, repeatedly delaying the launch.

      e.g., the client keeps making more and more requests about what to include in the next version.

(6) Missing Buy-in – team members don’t actually care about achieving the team’s goals (i.e., they don’t view the goal as valuable, and don’t feel like a genuine team, such that they care about the team’s success).

      e.g., Bill doesn’t think the product they are building is worthwhile, Sally doesn’t feel invested in the success of her team members or the team overall.

(7) Incentive Misalignment – some team members have incentives that cut against the project going well, or are focused on optimizing for personal goals at the expense of project goals.

      e.g., team members know that if they succeed at the goal, then someone else will take all the credit, and if they fail at it, there will be no negative consequences, and Bob actually wants the project to fail because then he’ll get to move on to a different project he likes more.

(8) Missing Skills – certain skills are needed to accomplish the goals, but no one on the team has these skills, and there are no external consultants that can be called on.

      e.g., no one on the team has a good eye for design, so the interface you’re building has the right features but looks bad.

(9) Reactivity – the team is constantly reacting to crises or imminent deadlines, which means there isn’t time to focus on achieving the long-term goals.

      e.g., every month or two, there is a major system failure, followed by weeks of scrambling simply to get things back to normal.

(10) Groupthink – team members do not feel comfortable sharing their unique ideas, or challenging the team lead, or getting their ideas heard above the most vocal person, or contradicting the group consensus, so the best ideas don’t filter to the top.

      e.g., the first proposed solution to a problem was accepted because no one wanted to challenge the person who proposed it, and since no one spoke up, everyone assumed the other group members must be in consensus.

(11) Guessing – the team doesn’t have enough data or information to solve the problems they are working on, but rather than gathering this data or information, they guess at solutions that aren’t very likely to work.

      e.g., the team has the goal of reducing customer churn, but they haven’t conducted customer interviews or carefully analyzed the churn data, so they are merely guessing at why churn is occurring and developing inadequate solutions based on those guesses.

—

(B) Effectiveness

(12) Insufficient Time – people not working enough hours.

      e.g., a culture of showing up at 10 am and leaving at 4 pm.

(13) Waste – spending time engaged in unnecessary processes or pointless meetings.

      e.g., a culture of constant meetings where you have little time to get your actual work done, or a requirement to tediously document all your work.

(14) Distraction – people may not be able to get “in the zone” on their work because of frequent distractions, or an environment where it is hard to focus.

      e.g., a noisy office environment where colleagues continuously interrupt you to ask you questions.

(15) Burnout – people may feel stressed, depressed, disinterested, bored, or exhausted, and find it is psychologically difficult to get their work done.

      e.g., a culture where bosses regularly yell at, blame, and fire employees for things that are not their fault.

(16) Bad Foundations – the work may be building on other work that was not well made, slowing down additional progress.

      e.g., programmers inherit a massive, buggy, and poorly written spaghetti codebase (i.e., high levels of technical debt).

(17) Disempowerment – team members are not allowed to do certain things (or make certain decisions) that would move the project forward, or they are required to follow a bureaucratic or standardized process that is not an efficient or effective process.

      e.g., a user experience researcher is not allowed to pay customers for doing interviews, but customers won’t do the interviews for free.

(18) Intrinsic Difficulty – the work may genuinely be intrinsically difficult, with progress speed inherently limited.

      e.g., work involving proving new theorems that others have failed to prove.

—

(C) Resources

(19) Slow Platforms – the systems or platforms on which the work is performed make progress more difficult, slower, or more complex than it needs to be.

      e.g., teams are forced to work on Windows 98 and experience regular computer crashes, or teams are forced to work using old and out-of-date software.

(20) Lacking Tools – the team members don’t have the best tools to do their work.

      e.g., a construction team is forced to use a saw that is inappropriate for the job because they don’t have access to the right kind of saw.

(21) No Training – there is a lot of relevant information about how to do the job well that team members are not told, and have to laboriously figure out on their own, or struggle to get by without.

      e.g., programmers are expected to figure out how the undocumented API works on their own via trial and error.

(22) Insufficient Assistance – when team members are stuck, they have no one knowledgeable they can go to (or that they feel comfortable going to) for help.

      e.g., a machine learning researcher can’t get her neural network model training properly and has no one she can ask for advice.

(23) Insufficient Skill – team members are not sufficiently skilled at the tasks they are trying to do.

      e.g., the team members responsible for writing documentation are not skilled at clear communication.

—

(D) Communication

(24) Blocking – team members are waiting on each other to do things before they can get their own work done.

      e.g., Sam needs approval from Sally to move forward, but Sally requires approval from Samantha in order to give Sam permission.

(25) Conflict – team members dislike each other, don’t trust each other, or have clashing personalities.

      e.g., Bill refuses to work with Sam, Jill thinks that Jenny is trying to undermine her, Tim and Tammy get into heated arguments.

(26) Discordant Goals – different team members are working towards different long-term goals, because they are not on the same page about what the purpose, mission, or primary goal of the team is.

      e.g., the product manager is trying to add features, whereas the engineering team is trying to do a feature freeze so they can focus on stability and efficiency.

(27) Ambiguity – the goals are not clear enough, leading to confusion, false starts, and stops, or circling around problems without directly solving them.

      e.g., there is agreement that the user interface needs to be “easier to use,” but little clarity on what that means.

(28) Duplication – team members end up repeating work that has already been done (or that is in the process of being done) by others, either because they don’t know that other work exists or because they don’t trust the quality of it.

      e.g., the new lead engineer doesn’t trust the code developed by the prior lead engineer and so decides to rewrite it from scratch.

(29) Unassigned work – it is not clear who is supposed to tackle certain work (e.g., it is not in any particular person’s job description or current assignments), so some important work just doesn’t get done by anybody.

      e.g., nobody has been assigned the responsibility of bug testing, so bug testing simply doesn’t get done, and therefore, the product is highly buggy.

(30) Siloed Information – team members need information from each other to do their work well, but this information is not transmitted reliably.

      e.g., the team member doing customer interviews isn’t reliably communicating what they learn to the team member planning UX changes.

A big shoutout goes to Eddie Liu, who helped develop this framework.

This piece was first written on July 3, 2019, and first appeared on my website on January 22, 2026.