A few years ago, an idea called Non-Violent Communication (NVC) became important to how I communicate, especially in situations of conflict or potential conflict. When I’ve mentioned the NVC book to friends, I’ve found that many of them had already read it and use the ideas—one of my high school friends said “That book changed my life.”

In this blog post, I summarize some of the basic ideas of NVC, and advocate for the ideas in scientific and technical communication. I describe it in a way that makes sense to me, and might be more approachable for some people; the way the NVC book is written really put me off when I first tried to read it. But, if you find this intriguing, I do encourage you read the book itself; this is not meant as a complete introduction to NVC.

Conflict is a natural part of scientific communication, not only because science involves resolving competing ideas, but also because it is a human process. Yet, I don’t think we’re often trained in how to deal with conflict. NVC offers tools for managing conflict, and, more importantly, communicating what is important to us in a way that prevents conflict in the first place.

I also draw connections between NVC and scientific concepts like “umwelt” in this post.

I think that NVC can be useful to everyone; feel free to ignore the science stuff if it’s not interesting to you.

This post was spurred by a very nasty paper reviews I received recently. In the next blog post, I’ll apply some of these ideas to reviewing, including discussing that nasty review, and describe simple, easy-to-follow guidelines that greatly improve paper reviewing. See that post for an extended example of these concepts.

Non-Violent Communication (NVC)

NVC provides communication techniques for managing and preventing interpersonal conflict. NVC can generally improve empathy and connection between people.

A key concept is to separate what you can observe from your judgments and interpretations, the way that a scientist ideally separates the data they measure from their intepretations of the data.

Think about a recent time when you got angry at someone important in your life, or they got angry with you. Maybe someone was late, or talked over you, or broke a promise—and it made you mad. In these cases, it’s tempting to describe your experience as objective facts about the world: “You’re always late” “That was really shitty of you.” “You didn’t even think about other people.” Or, perhaps you know that saying these things won’t help, so you don’t say anything, and continue to feel bad when the same thing keeps happening again.

The problem with those statements that they they describe your own interpretations of events and judgements of people. They’re subjective. The other person may not agree with them, and even feels insulted by your comments, and then things get worse.

NVC offers a recipe for communication in these difficult situations. A key idea is that you can only talk about things you yourself observe. That is, you can talk about what you saw, what you did, and what you felt.

In NVC, statements about things that you couldn’t observe are off-limits. You are not allowed to make judgements about other people, to say what they thought or believed, or to describe who they are. In NVC, these are all called “violent communication:” a violent communication is any statement that imposes an unwanted opinion onto someone else. Telling someone that they’re usually late, or that they’re rude or inconsiderate are violent communications.

Even compliments can be violent communications. For example, in some contexts, someone telling me that I’m very organized might sound to me like they’re saying that I’m uptight or uncreative. Moreover, in my experience, unsolicited advice is usually violent, since it often conveys a lot of wrong assumptions by the advice-giver about the unwitting advisee.

One goal is to find common ground, which means avoiding statements that the other person disagrees with. If you make judgements or interpretations about what happened—especially about them—they may not only disagree, they may get defensive or angry. On the other hand, if you state only things you observe, the other person almost certainly cannot disagree with them, and then hopefully you have at least improved your understanding of each other. Honestly describing your feelings in these situations is so hard, but so powerful.

NVC is about managing conflict, and just describing your feelings or hurt is not enough; you cannot expect the other person to know what you need. For this, once you’ve stated your observations, you can describe the more basic needs you have, and then make a request: something that you want the other person to do. The request needs to be concrete and actionable; not “treat me better” but something very specific that the person can do “when I cook dinner, do the dishes without being asked.”

In summary, the recipe for NVC in a conflict is: 1. say what you saw and observed (observable facts), 2. how it made you feel (internal facts), 3. describe your personal needs, and 4. your precise and specific request for what you want the other person to do. “You said X, it made me feel Y, I need to feel respected, and I request that you to not say that again.” They might say no to the request, but they should ideally have no reason to dispute the facts of what you experienced and what you need.

This is not easy—it’s not easy to dispassionately observe and discuss intense emotions. It takes a lot of practice and iteration to learn the skills of NVC: to filter out the judgements, to try to observe your feelings accurately, and to openly describe them. I still find it quite difficult.

I have not attempted to fully explain NVC, and I’m certainly not an expert. I recommend reading the first few chapters of the book to learn more. As with any skill, reading isn’t enough: you then need to practice difficult conversations in your own words, ideally together with someone else, and iterate. I promise it’s worth it.

Talking about feelings

At the core of NVC is something that sounds antithetical to science: talking about feelings.

Many of us were trained in childhood—perhaps unintentionally—to avoid talking about feelings and peoples’ needs. In technical, scientific, and pragmatic topics, feelings might seem irrelevant and distracting—one works hard to write objective, factual accounts and to make rational decisions. In interpersonal conflict, it’s tempting to focus on what promises were made, who is “in the right” or what people “deserve”.

If so, feelings and needs end up being neglected, when they’re sometimes what need the most attention. Often conflicts that seem to be about logistics or fundamental disagreements are really about miscommunication leading to hurt feelings. Whenever I see someone get angry in a work meeting (including myself), I believe that there’s a real issue of their feelings of being threatened or hurt, or not getting some basic needs met—not whatever practicalities the discussion is superficially about.

In, say, a faculty meeting, you might see a lot of charged discussion, which indicates a lot of difficult feelings. Some junior faculty feel insecure in their status and some senior faculty feel insecure in their continued relevance, and insecurity leads to people attacking colleagues and faculty candidates. Some faculty seem upset about everything, and this often comes from their own insecurity and fears.

At times, just saying how you feel can really help in tense situations. It feels so difficult to recognize your feelings at the core of a difficult situation, and then to acknowledge them out loud. Then when you do, you think “that wasn’t so hard.” (Actually, it is hard, and I am not good at this, especially in the heat of the moment.)

Several years ago, I was frustrated after receiving several rounds of very highly-critical feedback on a manuscript from a mentor. I couldn’t tell if they thought my work was even worthwhile at all. I finally wrote back that “This project is very important to me, … and I found your email to be really discouraging.” While they did not specifically comment on this, I found their subsequent comments to be much more considerate and balanced, even including the occasional compliment. Their comments and support were absolutely crucial to me eventually getting the work completed and published.

One of my favorite comedians, James Acaster, tells stories in his “Hecklers Welcome” tour about confronting conflict and fears. The show culminates in a story of how he disarmed scary bullies on a train by honestly describing how scared he was. It seemed like a textbook use of NVC (although rewatching it now, he doesn’t talk about needs or requests). (At the theater where I saw it in San Francisco, he complained that the hecklers were too nice, which got a laugh.) It’s a very funny show and I highly recommend the comedy special. (I also loved his prior special for different reasons, alternate link).

Even in everyday conversation, we can include discussion of feelings. These are little things: “I’m happy to see you” “I’m disappointed that they’re out of dessert.” But I think a willingness to mention feelings in everyday situations is good practice for more difficult situations; conversely, people who find this difficult might also have similar difficulties in difficult conversations.

Judgments as opinions

NVC says to avoid all sorts of external judgements. But, in everyday conversation, we make judgments all the time, like “The show is funny” or “That’s a good idea.” These judgments are generally statements of opinion, not facts, and different kinds of judgments can function in different ways. When I said that “Hecklers Welcome” is a funny show, what I really mean is that I laughed a lot and enjoyed it, and I think that other people would find it funny. That is, I used an objective-sounding judgment about the show as a shorthand for my own opinion of the show.

I think of most judgments as being shorthands for opinions. Explicitly expressing every opinion as an opinion would be painfully verbose and unfun.

Trouble arises when people treat such judgements as absolute. If you insist that a show is funny—contradicting anyone that says otherwise—then you seem to be saying there’s something wrong with people who didn’t laugh at the show. For this reason, I try avoid such judgments when there’s potential disagreement. For similar reasons, I advocate against evaluative definitions of art (“This is good enough to be art”), and that unnecessary judgements impede creativity and artistic practice.

One alternative is to replace generic judgments with specifics. Instead of “that’s a good movie,” or you comment on aspects of it, e.g., “I liked the acting but not the special effects.”

Interpreting sensations: umwelt and NVC

Now I turn to how NVC relates to scientific theories.

It seems like we experience the world as it is: you see a picture and know what’s in it. A stranger is aloof and unfriendly, and refuses to make eye contact with you. A dog sniffs some boring weeds but misses a spectacular view from a trailhead. Much of the time, these kinds of descriptions work.

But perceptions can mislead us. Visual illusions are not what they appear. Someone who seems aloof and unfriendly might just be shy. A dog experiences a wealth of sensations from smells that we humans cannot experience, whereas dogs are red-green colorblind and overall have worse visual acuity than humans.

This is because we don’t see the world as it is—we interpret our sensations. In biology, this idea is expressed by the concept of umwelt: no organism has access to the true facts of the world, it can only access its senses. Many of us organisms interpret our sensations. Different organisms have different experiences—me and my dog have different senses, and so we have really different experiences when we’re out on a walk.

Scientific research has analogous limitations. A scientist running an experiment cannot know the true causes of what happened in the experiment; they can only take measurements and form theories based on them. A scientist aims to report, as carefully as possible, what they did, and what they measured, and then to cautiously describe their interpretation of the results. Conversely, some cognitive-science theories liken organisms to scientists, measuring the data of their senses, performing experiments, and forming theories in their head, such as “Bayesian brain” theories and Gopnik et al.’s The Scientist in the Crib.

Here’s the complete analogy between umwelt, scientific communication, and non-violent communication. In each, there is an actor (an organism, a scientist, a person), who can observe specific things (senses, experimental data, conversations and feelings). Organisms and scientists form interpretations of their data (the state of the world, scientific theories). In communication, scientists report their data and their interpretations, at some dispassionate remove (separately their feelings from the data), in order to come to a shared understanding of science; people using NVC report their direct experiences dispassionately (including reporting on their feelings) in the hopes of coming to shared understadning.

This is just an analogy, not a literal equivalence; communication would be impossible if we always had to report every sensation on our retinas instead of “the clock said 3pm” or just “it was 3pm.” We can take shortcuts when there is shared understanding, and no real uncertainty about the meaning of our sensations.

Scientific communication typically involves a posture of objectivity, that one can make statements without a point-of-view or positionality. Even if we acknowledge that objectivity is technically impossible, the key idea here is that we always have some shared understanding. We can treat shared understanding as facts (no need to treat “the Earth is round” or “we went to a movie together yesterday” as opinion), and be more careful about uncertain or contested ideas.

Using NVC in scientific and technical communications

Having tried to incorporate NVC into my personal life—it’s difficult, but worth it—I find myself using those habits in some of my writing as well. I don’t advocate employing the full NVC recipe in all scientific and technical communication, but I advocate using some of the ideas. At the very least, one should be very careful with judgments, acknowledge feelings where relevant, and, in debates, avoid violent communication.

I’m currently co-authoring a technical paper on a controversial topic in “AI” art, and a draft used judgemental, emotionally-charged language. I found the language upsetting (I disagreed with some implications of it), and it did not seem possible to concisely summarize the issues—any judgments we made would likely be upsetting to some group. But we also didn’t need to take a position or to make these judgements. I reframed the introduction in factual terms that we could support with citations, so that we could motivate our technical problem without making controversial assertions. This might sound just like careful technical writing, but to me it felt like NVC, since we avoided judgmental language and limited ourselves to factual observations.

Now, when I do discuss controversies directly, I do try to acknowledge the conflicting fears and angers of the different stakeholders, and the sources of those feelings. For example, in a discussion of “AI art,” I discussed the anger that traditional artists feel as a product of their fears for their own livelihoods and identities, many of which are quite understandable. (Although I’ve also discussed it in terms of the value systems of different kinds of artists, which are themselves tied to feelings.) Some artists might find this discussion violent if they do not relate to it—different people have different opinions and feelings—but it seems better than not attempting to acknowledge the feelings.

In my next blog post, I discuss violent communications in paper reviewing (including some really bad reviews I've gotten), and describe a simple guideline for better, non-violent reviews

Many thanks to Rich Radke for comments on this blog post.