About Constructive Collaboration

Last Updated: October 6, 2026
Contents

PFN aims for Participants to discuss, debate, and work together to develop policy solutions and content, such as Solution Reports, even when some of them don’t agree with the solutions being presented. To support these expectations, PFN believes that Participants should use certain techniques and mindsets that support constructive collaboration in every PFN project and activity, including comment areas, discussion forums, online and in person meetings, short articles, and Solution Report projects.

And when developing a Solution Report, Network Group Participants should follow certain baseline requirements that themselves support constructive collaboration and go further to assure the Groups are relying on a diversity of proponent-affirmed views. Also, PFN expects that each Solution Report will be structured to present the Report’s Solution and the key alternative solutions, and their various arguments and counterarguments, as explicitly and objectively as possible.

This document describes the constructive collaboration techniques and mindsets that PFN believes Network Groups and their Participants should use in all projects and activities. See About Solution Reports for more information about the baseline and other requirements that apply to Solution Reports.

Capitalized terms that this document does not define have the meanings given in the PFN Glossary. See Related Content below.

Key Features in Brief

PFN does not ask Participants to abandon or compromise their views about policy topics and solutions to reach consensus with other Participants. Instead, PFN expects that Participants with different views will hold their views while engaging in constructive discussions and content projects that will support Next-Level ideas, thinking, and solutions for current and future policy experts. Here are several of the key techniques and mindsets that support constructive collaboration:

  • Diversity of views. PFN believes that a team of engaged experts with different views should be able to produce better policy discussions, solutions, and content. To that end, PFN sets baseline research, contact, and other procedural requirements for every Solution Report project regardless of the size and composition of the Network Groups developing the Reports to ensure that all Reports incorporate a sufficient diversity of proponent-expressed and -confirmed views. These are particularly important for smaller, less diverse Network Groups, but even larger, more diverse Network Groups are expected to satisfy the baseline requirements. See About Solution Reports.
  • Steelmanning. Participants should use the steelman and similar techniques when discussing policy topics and solutions with others, particularly those with different views. The steelman technique calls for a Participant to restate a proponent’s view as strongly and affirmatively as possible, in a way the proponent would support, and then to suggest changes that improve that view in ways consistent with the proponent’s perspective. The test of a restatement and suggested improvements is whether proponents would recognize them as their own.
  • Empathy. Empathy makes discussions among Participants with different views more effective and productive. It is associated with listening to others and their views, giving others the benefit of the doubt, and exhibiting intellectual and emotional humility. Empathy helps make steelmanning possible.
  • Intellectual rigor. Participants should strive to meet high standards regarding the clarity, precision, specificity, focus, logic, coherence, consistency, honesty, and accuracy of their views. Intellectual rigor is not inconsistent with empathy.
  • Intellectual humility and curiosity. Participants should be open to the possibility their views may contain mistakes or could be improved. Participants should also seek information and views that may challenge their views.
  • Transparency. Participants, and the content PFN and Participants develop, should be open about the values, evidence, and other factors underpinning their views, about alternative views, and about the counterarguments, pros and cons, and tradeoffs of their own and others’ views. This makes a Participant’s views more understandable and checkable by others, and supports empathy and intellectual rigor.
  • Creativity. PFN strives for its discussions, projects, and activities to produce new, innovative views and content that will advance the understanding and thinking of other Participants and of policymakers. Next-level solutions and content depend on creative ideas, thinking, and views.
  • Honesty and integrity. Participants should represent their own views and others’ views accurately, acknowledge what they do not know, and treat other Participants as they would want to be treated.

PFN further believes that Network Groups and Participants should follow the policies in the PFN Code of Conduct, including the Chatham House Rule, which support Participants in thinking out loud and discussing their own and others’ views explicitly, directly, and professionally in private, restricted-access venues. These policies provide further support for constructive collaboration.

If a Participant refuses to engage constructively with others’ views, the PFN Code of Conduct applies and the Network Group Leader or the PFN Leader should be informed.

Diversity of Views

PFN values differences in what its Participants believe, what their views are, the arguments they find persuasive, and how they think about and discuss policy topics and solutions. PFN believes that these differences contribute to more constructive discussions and produce more innovative solutions.

Most policy topics do not divide into two or even three sides: for and against, this and that, or left, right, and middle. In the Retirement & Aging Domain, for example, there are multiple views, and shades of views, on the need and options around the future of Social Security, and the differences among them do not map cleanly onto partisan views or other factors. Treating a topic as having two sides can undercut a group’s effectiveness.

PFN does not attempt to mirror the distribution of views in the country, in the profession, or in the literature. PFN attempts to ensure that each significant and credible view relating to a topic has at least one capable proponent.

PFN does not ask Participants to be neutral or to compromise their views. Participants should continue to hold their individual views while understanding, restating, improving, and engaging with opposing views in constructive ways. PFN believes that a group of committed experts who have different views, understand each other, and work together effectively is a powerful contributor to more valuable discussions and solutions.

If a Network Group lacks a Participant who holds an important view relating to its Domain or a particular project, PFN would rather reach out to someone who holds that view than have someone who doesn’t argue it secondhand. In these cases, PFN will consider bringing in an outside expert who holds the view, including an expert who is not a PFN Participant, for the project or activity. A Network Group may also undertake significant research of non-PFN proponents’ statements and even ask them to review draft restatements of their views.

On projects to develop Solution Reports, PFN goes further by directing that all Network Groups must meet certain baseline requirements to assure that diverse proponent-affirmed views inform their work. For example, each position in a Solution Report must be built from what its own proponents have expressed, the arguments for and against it, and its track record, whether or not those proponents are PFN Participants. The Report’s statement of each position is then checked by someone who holds that position and did not write the statement. PFN further directs that each Solution Report be structured to explicitly address, for each solution, the underpinning views, arguments and counterarguments, alternative solutions, pros and cons, tradeoffs, and other relevant items. Network Groups that have more Participants with greater viewpoint diversity can do more in their Solution Reports, such as relying more on Participants that hold particular views and having direct discussions that allow to deeper discussions about what proponents might concede or trade off, on what issues a group of proponents might split, and which arguments move and don’t move viewpoints. See About Solution Reports.

Steelmanning: What It Is

Steelmanning is a collaboration technique that advances understanding, discussions, and policy solutions by calling for each Participant to restate and improve strong versions of a proponent’s views, including views the Participant might disagree with, before the Participant conveys objections and presents preferred views. After these steps, the Participant should be ready for the proponent to steelman the Participant’s objections and preferred views in return.

There are two parts to effective steelmanning:

1. Restate. A Participant should be able to present a proponent’s view and arguments in a way the proponent recognizes and supports. A Participant shouldn’t provide a weak or a merely fair summary, but should present the proponent’s view as strongly as the Participant can, even if the restatement is stronger than what the proponent set forth. And the proponent should acknowledge that the restatement is accurate.

2. Improve. A Participant should suggest changes to strengthen the proponent’s view and arguments. The Participant should consider how the proponent’s view is vulnerable and how that vulnerability might be addressed in ways consistent with the proponent’s view. The Participant should also identify evidence, arguments, and considerations the proponent may not have put forward but that would support or strengthen the proponent’s view. The Participant should consider modifications that would elevate the view by strengthening its goals, making their achievement more likely, or meeting key objections. As with the restatement, the proponent should agree that the suggested improvements are likely to strengthen the view, and in return, the proponent should steelman the suggested improvements.

PFN believes the improvement part carries special significance. Being able to suggest improvements to a different view depends on a strong understanding of the view, what it is trying to achieve, and the arguments that support it, and that understanding is what produces new ideas and thinking. A Participant who disagrees with a view will also see it with different eyes and may be better able to suggest changes that elevate it beyond what its own proponents have identified.

PFN believes that Participants should apply the two parts separately and should apply both before sharing their questions, objections, counterarguments, and alternatives. Keeping separate things separate supports constructive discussions by increasing understanding, reducing confusion and misunderstanding, and clarifying responsibilities.

This is particularly important when the discussions are occurring via writing and other approaches that include time gaps between comments. In these cases, a proponent may not be able to react to a Participant’s restatement or suggested improvement until later, so it’s important that in the meantime the Participant should strive to imagine how the proponent might react to the restatement and suggested improvement and thus to fashion these comments accordingly. When a proponent later can react to a Participant’s restatement or suggested improvement comments, the proponent should address whether they are accurate.

PFN expects the steelman technique to be used in multiple directions during policy discussions and the development of policy solutions. Participants should state a strong version of a view they do not agree with and suggest how it could be improved; Participants who support that view should state strong versions of the objections to it and suggest how those objections could be improved; and Participants whose views are being steelmanned should engage constructively with the restatements and suggested improvements.

The steelman technique will be useful many times, in many contexts and directions, over the course of policy discussions and the development of a Solution Report. PFN expects the development of a Solution Report to involve hundreds of steelman and similar exchanges. These would not be set out in the published Report, but their results should be visible in it.

PFN believes that Participants who are steelmanning others’ views should help the proponents of those views be open to the restatements and suggested improvements, and that the words Participants use can make a positive difference. Examples for introducing a restatement include “Let me try to state your view as strongly as I can…” and “In the following, I’m trying to describe the proponent’s views and arguments as accurately and constructively as I can….” Examples for introducing improvements include “Let me suggest some ideas that may improve your view and arguments…” and “In the following, I’m suggesting some changes to the proponent’s views and arguments that I think will make them better or more effective….” In all instances, the Participant should invite the proponent and others to react, including in later writings.

Steelmanning: Benefits

Steelmanning can take significant time and effort. But it promotes fairness and respect in discussions, and more importantly it improves the quality of the discussions and the resulting solutions. Participants who know they are expected to restate and improve others’ views pay more open, focused, and thoughtful attention; apply more rigorous and creative thinking; and perform better as constructive collaborators. That benefits the other Participants and the content, and it benefits the Participants themselves, because understanding the thinking behind other views strengthens understanding of one’s own.

Here is a list of some of the benefits that PFN believes should come from using the steelman technique in collaborations among Participants with different views:

  • Moves the goal of the discussion from winning an argument to improving an idea.
  • Reduces defensiveness. People who feel accurately understood engage more productively.
  • Promotes empathy, humility, discussion, and flexibility among people with different viewpoints.
  • Exposes the real disagreement earlier. Once both sides state a view the same way, people are better able to see whether they differ over facts, analysis, or values. Differences are typically easier to address when they’re identified earlier.
  • Guards against confirmation bias. Restating a viewpoint fairly forces you to engage with evidence and reasoning you’d otherwise skip past.
  • Tests others’ views. A view that survives contact with the strongest case against it rests on firmer ground.
  • Strengthens critiques. Criticism aimed at a strong version of a viewpoint holds up much better than criticism aimed at a weak version.
  • Surfaces risks and errors earlier, when people are likely to be more open to changing their viewpoints.
  • Builds credibility with everyone reading or listening, including people who hold neither position.
  • Invites reciprocity. A person who receives a fair restatement and supportive improvements is more likely to offer the same to others.
  • Builds habits that productive collaboration and disagreement depend on, including listening, empathy, giving people the benefit of the doubt, staying open to being wrong and to changing one’s views, transparency, and trust.

Empathy

Empathy, particularly cognitive empathy, is an essential mindset because it supports, in the short term, the effectiveness of discussions and solutions and, in the long term, the building of enduring, trusting, and productive Network Groups made up of experts with diverse views.

For a Participant who disagrees to restate views and suggest improvements in ways the proponents would support, the Participant has to understand the proponents’ views from the inside: what are the proponents worried about, what are their values and priorities, what problems are they trying to address, what are they trying to achieve, what are the alternative views, why have they chosen one view over the others, what are they protecting, what is their reasoning, and what do they think other views fail to consider. PFN believes Participants often have difficulty suggesting meaningful improvements to views if they don’t understand the thinking of the people who hold them.

With empathy, Participants are also more likely to treat each other respectfully and professionally, to give each other the benefit of the doubt, and to treat each other as people trying to solve a hard problem together rather than as opponents to be defeated. In some situations, a Network Group Participant who holds a minority view may become less engaged and more likely to leave PFN. Other Participants should watch for this and act constructively, since losing a minority view undercuts the diversity of views the Network Group depends on.

Empathy is not agreement. Participants should be able to hold their views with full conviction and still steelman others’ views effectively. Also, empathy does not require or imply that Participants hold weaker convictions or soften their conclusions. A well-understood and effectively argued counterargument can still be firmly rejected. And empathy does not depend on a proponent’s good faith. It is about a Participant understanding a proponent’s view well enough to explain and argue for it, even if the Participant has questions about how sincerely the proponent or others hold the view.

Intellectual Rigor

Intellectual rigor refers to meeting high standards, similar to those recognized in the best academic, legal, and think tank work, for the clarity, precision, specificity, focus, logic, coherence, consistency, integrity, honesty, and accuracy of a view and of how the view is expressed and communicated. Intellectual rigor is consistent with empathy.

Here are some key characteristics of intellectual rigor:

  • Distinguish among what is known, what is claimed, what is believed, what is interpreted, what is an assumption, what is contested, what is predicted, and what is a value-based judgment. Be precise about what is what.
  • Factual claims typically have sources with citations, and a reader should be able to check the source to see whether the claim is accurate.
  • Where facts, evidence, analyses, interpretations, or values are uncertain, unsettled, or cut against a Participant’s views, the Participant should acknowledge the shortcomings and, where possible, cite the sources. A Participant should also acknowledge when a view depends on context and assumptions. Participants should be open about counterviews, pros and cons, and tradeoffs, and about when different views reflect differences in values.
  • Avoid words and phrases that overclaim or underclaim. A change that would make a difference at the margin should be described as making a difference at the margin. A correlation should not be portrayed as causation. And while precise numbers are often unnecessary and can be misleading, general words like “all,” “everyone,” “most,” “many,” and “some” can mislead too. That most or many members of a group hold a view does not mean the group or everyone in it holds it. A group’s mean or median figure for some attribute does not mean that everyone, most, or even many members share it.
  • Evaluate and discuss a policy view and the arguments for or against it, but not the people who favor or oppose it or their motives.
  • Understand what would change your mind. A view held with no recognition of what evidence or other factors would count against it is harder for others to engage with.

Intellectual Humility and Curiosity

Participants should be open to the possibility that their views, and the thinking around them, contain mistakes or could be improved. They should not assume their views are the only or the best views. Participants should not only be open to, but should seek out, information and views that may challenge their own. It’s good to be curious about open or uncertain facts or analyses. And Participants should react positively to being shown that their views may include mistakes or could be better.

Transparency

Transparency is related to intellectual rigor. Participants should be willing to convey the underpinnings, weaknesses, tradeoffs, and alternatives of the views they are discussing. Transparency helps make the assertion of a view more open, credible, and verifiable.

PFN content should reflect the same mindset. For example, a Solution Report is expected to name its authors, reviewers, and contributors, subject to each person’s attribution election; disclose the use of generative AI; disclose the relevant interests and affiliations of the Participants who worked on it; and cite the sources behind its factual claims and its analysis. See About Solution Reports and About Using Generative AI.

While published PFN content is transparently attributed, the discussions that produce it are protected. That protection matters, because Participants must be able to think out loud, change their minds, and try out views they are not sure about. The rules are in the Terms of Use (under Restricted Areas and Confidentiality) and the Code of Conduct. In short, Participants may use what they learn in a restricted discussion, but they may not say who was part of the discussion or who said what and they may not distribute drafts of PFN content and Solution Reports beyond the Network Group developing them.

Creativity

PFN does not aim simply to catalogue what people already think. It strives for next-level policy solutions that will be seen as new, innovative, and effective.

PFN collects, summarizes, and reports existing views by building a resource collection for Participants, not because this is PFN’s goal, but because a well-curated resource collection supports Participants in producing new ideas, analyses, and solutions.

Suggesting improvements as part of steelmanning a proponent’s view also favors creativity, because it can lead to ideas the proponents had not considered. Similarly, the interaction that flows from having people with diverse views work together on a particular view or solution can generate new ideas, and new combinations of ideas, that elevate the original.

Participants should keep an open mind to the possibility that new ideas will emerge from collaborative discussions among Participants with different views, and that some of them will improve views and solutions. Participants should not assume there are only two or three possible views on a policy topic. They should be particularly attentive to views that at first seem unrelated or out of left field. This matters most while views are still Exploratory Content, in a Gated Venue with a Gated Audience. Poorly explained and half-formed ideas bounced around within a protected group can lead to real improvements and new views.

PFN Content. About Network Groups · About Solution Reports · About Using Generative AI · Chatham House Rule · PFN Code of Conduct · PFN Glossary.

Acknowledgements

Contributors. Harry Conaway (PFN Leader) is the author and writer of this content and is solely responsible for its substance.

Disclosures. Claude (Opus 5.5, Anthropic) was used to assist in editing this content. All substantive and writing decisions were made by the author and writer noted in the Contributors section above.

PFN Content Type. This content is classified as PFN about and governing content, and thus is not subject to PFN’s Content and Statement Types policy for substantive policy content.

Version History

Current Version. To be added