You Don't Negotiate with Terrorists: The Irrational Adversary Is Vermin

In one of the first more complex disputes I ran, we offered the other side a deal that was economically far better for them than the likely judicial outcome. The client was prepared to pay more than "what was at stake" simply to put the matter behind us. Logically, it was a compromise that gave the other side an above-average outcome and spared them the cost of further proceedings. They turned it down. And not only that — they responded with an amended claim adding new, incomparably more aggressive heads of relief. After five years of litigation, two appeals, and two cassation reviews, it was over: the other side ended up with less than what we had offered them at the outset. And yet to this day they think they won, because they "kept us in court for years."

This is not an anomaly. This is an entire category of disputes that the classical Harvard doctrine of negotiation cannot explain well. In the earlier text on psychology and tactics in legal disputes I worked through concepts like BATNA, principled negotiation, and tactical empathy. All of them stand on the assumption that the other side is a rational actor with some interests. But what happens when that assumption fails? When the person on the other side isn't playing chess, but is willing to slam the chessboard against the wall just so you can't keep playing?

This article is about that. About game theory with an irrational adversary. About the dark triad. About altruistic punishment. About the sunk-cost fallacy on steroids. And about which strategies make sense when the Harvard School isn't enough — and when the best move is not to keep playing at all.

¶ Part One: The Myth of the Rational Actor

Most standard negotiation theory — Roger Fisher, William Ury, Robert Mnookin and the rest — implicitly assumes that the other side behaves roughly according to expected utility theory. They maximise expected utility. They respond to incentives. If I sweeten the deal, they yield. If I impose costs, they show willingness to negotiate. This premise traces from von Neumann and Morgenstern (Theory of Games and Economic Behavior, Princeton University Press, 1944) through Nash and Schelling to today's behavioural game theory.

The problem: empirics keep breaking it.

The sharpest demonstration came from Werner Güth in the now-famous experiment An Experimental Analysis of Ultimatum Bargaining (Güth, Schmittberger & Schwarze, Journal of Economic Behavior & Organization, 3(4), 367–388, 1982). In the ultimatum game one player gets money and may offer the other any amount. The other can accept (both get their share) or refuse (both get nothing). On a strictly rational view, the second player should accept any positive offer — even one cent is more than zero. Reality: offers under 30% are overwhelmingly refused. People are willing to forgo their own gain simply to punish someone they perceive as unfair. Irrational from the standpoint of money, deeply rational from the standpoint of social norms.

The phenomenon is universal. Henrich, Boyd, Bowles and colleagues replicated it in fourteen small-scale societies across five continents (In Search of Homo Economicus: Behavioral Experiments in 15 Small-Scale Societies, American Economic Review, 91(2), 73–78, 2001). The same pattern everywhere: people sanction unfairness at their own cost. From an evolutionary standpoint this trait has good standing (it sustains cooperation in large groups), but in a legal dispute it leads to the other side refusing a deal that would materially benefit them, only because they don't want to grant you the "win."

Ernst Fehr and Simon Gächter built on this with the concept of altruistic punishment (Altruistic Punishment in Humans, Nature, 415(6868), 137–140, 2002). People are willing to bear costs in order to punish defectors — and evolutionary models show why this disposition is genetically advantageous to the group (Boyd, Gintis, Bowles & Richerson, The Evolution of Altruistic Punishment, PNAS, 100(6), 3531–3535, 2003). For your dispute it means one thing: if the other side subjectively perceives you as a defector — someone who breached, harmed, cheated, humiliated — they will be willing to punish you at their own cost. To which the personality profile of the individual adds its own variable: some people punish gladly and abundantly, others only reluctantly.

¶ Part Two: Schelling and the Paradox of Credibility

It looks absurd: is being irrational a negotiating advantage? In the 1950s Thomas Schelling brilliantly formulated this in The Strategy of Conflict (Harvard University Press, 1960), for which he was awarded the 2005 Nobel Prize in Economics.

Schelling described what is known as the commitment problem: if I want the other side to yield, I must convince them that I will not. But if I am rational, they know I will yield as soon as the cost of continuing exceeds the gain. My rational commitment is therefore not credible. What is credible? Something that destroys the later option to yield. Burned bridges. Madness. A publicly demonstrated impossibility of turning back. Schelling's favoured image: two drivers race toward each other in cars; the one who throws the steering wheel out the window first wins. Once thrown, he can no longer swerve. The other has to — because the alternative is mutual death.

This insight from game theory explains a number of phenomena in litigation that look senseless from the economic optic. The other side that "behaves crazily" is, in a sense, throwing the steering wheel out the window. They are signalling: I will keep going whatever happens. And rational you must swerve. If you swerve, the "crazy" side gets more than they would by negotiating normally.

Side note: Richard Nixon and Henry Kissinger called this the madman theory during the Cold War (Sagan & Suri, The Madman Nuclear Alert: Secrecy, Signaling, and Safety in October 1969, International Security, 27(4), 150–183, 2003). Whether or how much it was a real strategy is contested, but the theoretical frame holds. The aim is not to be mad — the aim is to be credibly mad. And some people transmit that signal authentically, because they truly are.

This leads to a sobering conclusion: in a contest with someone who credibly demonstrates they have nothing to lose, classical game theory recommends yielding. The practical question is: when is it actually so, and when is the other side merely trying to make us believe it is?

¶ Part Three: Whom Are We Meeting — The Psychology of the Destructive Adversary

The irrational adversary is not a single uniform type. Clinical and personality psychology can distinguish individual profiles fairly well, and each calls for a different strategy.

The most-studied classification is the dark triad, introduced by Delroy Paulhus and Kevin Williams in The Dark Triad of Personality (Journal of Research in Personality, 36(6), 556–563, 2002). Three distinct but overlapping traits:

1. Narcissism — grandiosity, need for admiration, entitlement. A narcissist in a dispute hates losing face more than losing money. The dispute is a mirror. If they see a damaged self-image in it, they will keep the game going until the mirror shows something else — even at the cost of substantive self-harm.
2. Machiavellianism — strategic manipulation, instrumental view of others, willingness to deceive. The Machiavellian is functionally rational, but their utility function includes pleasure in the manipulation itself. Negotiating with one is possible, but requires constant verification and unceasing motive-testing.
3. Subclinical psychopathy — lack of empathy and guilt, impulsivity, risk-seeking. Robert Hare described it in clinical context in Without Conscience: The Disturbing World of the Psychopaths Among Us (Pocket Books, 1993); Paul Babiak and Robert Hare in Snakes in Suits: When Psychopaths Go to Work (HarperBusiness, 2006); and Martha Stout in The Sociopath Next Door (Broadway Books, 2005) — in the civil context of firms and families. The subclinical psychopath in a dispute does not experience most of the psychological costs we take for granted. Costs we put into the dispute do not stand in their way. Costs we impose on them do not deter them.

Kevin Williams and Delroy Paulhus later expanded the frame into the dark tetrad by adding everyday sadism (Buckels, Jones & Paulhus, Behavioral Confirmation of Everyday Sadism, Psychological Science, 24(11), 2201–2209, 2013). The sadist takes outright pleasure in the suffering of others — which has a devastating implication for a dispute: the greater the economic pain you imagine you are imposing on them, the happier they are. The standard incentive "come back to the table, this is going to keep hurting" works in reverse.

Beyond the dark tetrad there are other profiles we encounter in practice:

• Borderline traits — extreme emotional oscillation, idealisation flipping into devaluation, very difficult to build a stable working alliance. The DSM-5 (American Psychiatric Association, 2013) summarises the clinical picture; in litigation the chief feature is instability of stance — what holds today does not hold tomorrow, and it is hard to build strategy on it.
• Paranoid traits — interpret routine communication as attack, distrust even rational explanations, look for hidden motives. Aaron Beck in Prisoners of Hate: The Cognitive Basis of Anger, Hostility, and Violence (HarperCollins, 1999) described the cognitive structure of hostile thinking and its resistance to contradicting evidence.
• The "wronged hero" — a functionally neutral personality whom a particular event injured so deeply that identification with the victim role became part of the self-concept. Roy Baumeister in Evil: Inside Human Cruelty and Aggression (W. H. Freeman, 1997) described how the self-perception of having been wronged justifies disproportionate aggression toward the perceived perpetrator.

This typology is not for diagnosing the other side — that is not the lawyer's task and they have no tools for it. It is to keep us from applying a single strategy to different types of actors. What works on a narcissist (the option of saving face, an "elegant exit") does not work on a sadist. What works on the Machiavellian (a convincing demonstration that the trick will be exposed) does not work on borderline traits.

¶ Part Four: The Spiral of Escalation, and Sunk Cost on Steroids

Irrationality is not a constant. It is a trajectory. Many adversaries who look irrational today were not so at the start. Escalation has its own dynamic, described by Dean Pruitt and Jeffrey Rubin in Social Conflict: Escalation, Stalemate, and Settlement (with S. H. Kim from the second edition, McGraw-Hill, 1986/1994). A similar escalation logic was captured by Friedrich Glasl in his nine-stage model of conflict (Konfliktmanagement, 1980), which I cover in more detail in the piece on the psychology of the parties.

Key escalation mechanisms:

• Reciprocal escalation — sides answer the last move one notch tougher. "I was just reacting to what they did." In practice this means each exchange ratchets the conflict upward until someone deliberately intervenes.
• Commitment escalation — the sunk-cost fallacy (Arkes & Blumer, The Psychology of Sunk Cost, Organizational Behavior and Human Decision Processes, 35(1), 124–140, 1985). The more the adversary has invested, the more psychologically they need to keep going — otherwise the investment was "wasted."
• Identity fusion — over time the dispute fuses with identity. William Swann and colleagues described identity fusion in Identity Fusion: The Interplay of Personal and Social Identities in Extreme Group Behavior (Perspectives on Psychological Science, 7(5), 441–456, 2012). When the adversary's identity weaves together with the position in the dispute, surrendering the position means surrendering the self.
• Moralisation — Linda Skitka showed in a series of works (e.g. The Psychology of Moral Conviction, Annual Review of Psychology, 72, 347–366, 2021) that morally framed conviction is more resistant to compromise than purely instrumental conviction. Once the other side morally frames their position ("it's about justice!"), compromise feels like betraying a value, not making a sensible trade.

Combined, these mechanisms produce dependence on the conflict — a state in which the other side has more reasons to continue than to stop, regardless of economic calculus. And in this state, classical Harvard doctrine fails as a whole.

¶ Part Five: Asymmetric Defence — What Can Be Legitimately Done

When standard game theory falls short, what doesn't? From various corners of the literature — military strategy, FBI crisis negotiation, the sociology of conflict — a set of recommendations converges that share certain features. Call it asymmetric defence. Its essence: don't play the game the adversary is offering you, and instead play a different one they aren't ready for.

Strategy 1: Structural isolation. Instead of direct interaction with the adversary, you move through formal channels that "dehydrate" their destructive dynamic. In a legal dispute that means: maximum use of procedural formality, elimination of ad-hoc communication, everything via written documentation and, where it helps, via mediator, expert, or even the court as a filter. William Ury in Getting Past No: Negotiating in Difficult Situations (Bantam Books, 1991) called this "go to the balcony" — a deliberate stepping out of the immediate dynamic into a reflective position.

Strategy 2: Asymmetry of costs. If the adversary plays on costs (your money, time, emotional energy), arrange the scenario so that each next move costs them more than you. Interim relief, security order, an insolvency petition, enforcement security — these are all tools that change the cost structure. The aim is not "to win quickly" (often impossible), but to reshape the economy of the fight. There is a deterrent effect too: an adversary who is playing on your exhaustion loses momentum the moment they discover that costs land mostly on them, not on you (Pruitt & Rubin, 1994).

Strategy 3: Tit-for-tat with forgiveness. Robert Axelrod in the classic The Evolution of Cooperation (Basic Books, 1984) showed that in iterated prisoner's dilemma the simple tit-for-tat strategy wins: start by cooperating, then repeat the opponent's last move. The later refinement Pavlov / Win-Stay, Lose-Shift (Nowak & Sigmund, A Strategy of Win-Stay, Lose-Shift That Outperforms Tit-for-Tat in the Prisoner's Dilemma Game, Nature, 364(6432), 56–58, 1993) adds a mechanism for forgiving mistakes. For a dispute: mirror the adversary's hardness, but keep the door to de-escalation open so it doesn't have to be ritually re-earned. Important: tit-for-tat does not work in a one-shot game. In a dispute that is plainly the last interaction (e.g., before a definitive end of the relationship), the adversary has no reason to cooperate even at the end, so tit-for-tat's space narrows.

Strategy 4: Avoid the blow they prepared for themselves. This one comes from judo logic and the Chinese military tradition (Sun Tzu, The Art of War, 5th century BCE, available in countless translations). If the adversary invests massive resources in a move designed to break you, don't oppose it head-on. Instead, step out of the contact zone and let their punch land in empty space. In legal terms: don't react to a provocative filing where it isn't procedurally necessary, and instead think about which line of the dispute is strategically important to us, and defend that. Often an adversary expecting a frontal clash runs out of breath when they don't get one.

Strategy 5: Change the field. If the adversary plays destructively in one field (media pressure, personal attacks, parallel proceedings), it is sometimes most effective to neutralise that field altogether. Change jurisdiction. Change forum. Bring the matter back to a place where structural discipline applies. Galula in his classic Counterinsurgency Warfare: Theory and Practice (Praeger, 1964) showed in another context that what beats an asymmetric adversary is not primarily hardness but control of the field. Whoever defines where the dispute takes place dramatically shapes who wins on it.

Strategy 6: Fatigue as ally, not enemy. This one is counter-intuitive. In every long dispute the adversary tires over time — even the dark-triad type. Irrationality has its energetic capacity. The question is who tires first. If structural isolation, asymmetry of costs, and field change keep your costs low and theirs high, time works for you. Sun Tzu: "Attack the enemy's strategy" — in modern terms: degrade their ability to fund the fight, to focus, to maintain the narrative.

¶ Part Six: When to Walk Off the Battlefield

Running through all of the above is one knot the lawyer must be able to face: when to walk off the battlefield.

Robert Mnookin returned to this question specifically in Bargaining with the Devil: When to Negotiate, When to Fight (Simon & Schuster, 2010), which I cite in the manifesto from a different angle. Read it differently here. Mnookin acknowledges there are situations in which not negotiating is the right answer: a moral frame rules it out, negotiating would legitimise the unacceptable, or there is no real counterparty to negotiate with. But alongside that he describes much more frequent cases in which people refuse to negotiate for the wrong reasons — out of insult, ego, "principle." Mnookin's radical honesty: most people who refuse to negotiate should negotiate. And most people who negotiate with the devil should walk away.

The practical frame I use myself has three questions:

1. Is it converging? Is the gap between me and the adversary narrowing over time, or widening? If it is widening despite reasonable offers, you are in a zone where standard negotiation is no longer working. That is not yet a reason to stop — it is a reason to change tools.
2. Am I the only one here? If the client has other voices around them telling them to keep going for the wrong reasons (a lawyer in the background whispering "fight," a manager who wants to win in front of the board), a non-fight strategy has limited odds until that "second bench" is cleared.
3. What is the price of continuing vs. the price of ending? Not just money. The client's time, mental health, family, capacity to focus on their own business. The price of continuing is holistic. The price of ending is usually one-off.

A classical English saying nails the point better than any theory: "Don't wrestle with a pig. You both get dirty, and the pig likes it." You can't win a wrestling match in the mud with a pig. The pig enjoys it. And you'll leave the match either with revulsion or as the pig — but the pig leaves much later, because it's having a good time and you are not.

¶ Part Seven: The Ethics of Isolation and Hard Strategies

One last thing that cannot be skipped. The strategies described — structural isolation, asymmetry of costs, avoidance of head-on clash — work precisely because they are tough. Their toughness is justified where a standard approach has failed and where the adversary, by their own conduct, has crossed into the irrational zone. But the same tools, applied to a rational actor, are unjustified and professionally untenable.

This matters for the client who asks: "So can we play the way they play?" The answer: you can change the field, but you cannot adopt their values. The moment you start playing for harm rather than for protection, you lose your ethical position (and in law, the procedural one too). Asymmetric defence is not revenge. It is a cold recalibration of strategy in a world where Harvard isn't enough.

Bertrand Russell put something usable verbatim: "The trouble with the world is that the stupid are cocksure and the intelligent are full of doubt." Replace doubt with prudence and intelligence with reasonableness, and what you get is a mandate: be less slow, not less reasonable.

And once again Voss from the manifesto: tactical empathy without integrity becomes cynical manipulation. Asymmetric defence without integrity makes you what you are trying to defeat. The discipline is to know where your limits are — and to keep them even when the adversary has crossed theirs.

¶ Conclusion: When Harvard Isn't Enough, Reach for the Older Books

After fifteen years of practice I have learned there are two kinds of dispute. The first is competitive — rational actors who care more about the outcome than about each other look for a balance between maximising gain and preserving the relationship. There Fisher, Ury, Mnookin, and Voss work flawlessly. That's 80% of cases, and that literature deserves its gold-standard status.

And then there's the second. The one in which on the other side sits someone whose calculus contains a variable Harvard didn't name. Someone willing to destroy themselves to take you with them. Someone who has built their identity on fighting you. Someone who does not honour reciprocity. That case — sometimes 5%, sometimes 20% of practice depending on the field — calls for an entirely different library. Schelling, Axelrod, Glasl, Pruitt, Mnookin (Bargaining with the Devil), Sun Tzu. And a great deal of humility about the fact that sometimes the best strategy is to walk away.

A client able to ask "what kind of dispute am I actually in?" is a step ahead of all those who apply one doctrine to everything. And a lawyer who can put that question to the client does more for them than the one drafting the sharpest claim.

Related reading: on the psychology of the participants (judge, client, lawyer, counterparty) see The Psychology of the Parties. On the fact that the legal route is not always the right route — the companion piece on triage. On the fact that an irrational adversary is typically playing a different game from the one you think — Monopoly, where it isn't about arguments but about exhausting resources — see the text on the four operational frames of a dispute. The scientific background of the standard negotiation doctrine is in Statutes Are Just the Foundation: Psychology and Tactics in a Legal Dispute. And for a practical take, I recommend Not Every Battle Is Worth Fighting.

¶ Further Reading

• Arkes, H. R., & Blumer, C. (1985). The Psychology of Sunk Cost. Organizational Behavior and Human Decision Processes, 35(1), 124–140.
• Axelrod, R. (1984). The Evolution of Cooperation. Basic Books.
• Babiak, P., & Hare, R. D. (2006). Snakes in Suits: When Psychopaths Go to Work. HarperBusiness.
• Baumeister, R. F. (1997). Evil: Inside Human Cruelty and Aggression. W. H. Freeman.
• Beck, A. T. (1999). Prisoners of Hate: The Cognitive Basis of Anger, Hostility, and Violence. HarperCollins.
• Boyd, R., Gintis, H., Bowles, S., & Richerson, P. J. (2003). The Evolution of Altruistic Punishment. PNAS, 100(6), 3531–3535.
• Buckels, E. E., Jones, D. N., & Paulhus, D. L. (2013). Behavioral Confirmation of Everyday Sadism. Psychological Science, 24(11), 2201–2209.
• Fehr, E., & Gächter, S. (2002). Altruistic Punishment in Humans. Nature, 415(6868), 137–140.
• Galula, D. (1964). Counterinsurgency Warfare: Theory and Practice. Praeger.
• Glasl, F. (1980). Konfliktmanagement: Ein Handbuch für Führungskräfte, Beraterinnen und Berater. Verlag Freies Geistesleben.
• Güth, W., Schmittberger, R., & Schwarze, B. (1982). An Experimental Analysis of Ultimatum Bargaining. Journal of Economic Behavior & Organization, 3(4), 367–388.
• Hare, R. D. (1993). Without Conscience: The Disturbing World of the Psychopaths Among Us. Pocket Books.
• Henrich, J., Boyd, R., Bowles, S., Camerer, C., Fehr, E., Gintis, H., & McElreath, R. (2001). In Search of Homo Economicus: Behavioral Experiments in 15 Small-Scale Societies. American Economic Review, 91(2), 73–78.
• Mnookin, R. H. (2010). Bargaining with the Devil: When to Negotiate, When to Fight. Simon & Schuster.
• Nowak, M. A., & Sigmund, K. (1993). A Strategy of Win-Stay, Lose-Shift That Outperforms Tit-for-Tat in the Prisoner's Dilemma Game. Nature, 364(6432), 56–58.
• Paulhus, D. L., & Williams, K. M. (2002). The Dark Triad of Personality: Narcissism, Machiavellianism, and Psychopathy. Journal of Research in Personality, 36(6), 556–563.
• Pruitt, D. G., Rubin, J. Z., & Kim, S. H. (1986/1994). Social Conflict: Escalation, Stalemate, and Settlement. McGraw-Hill.
• Sagan, S. D., & Suri, J. (2003). The Madman Nuclear Alert: Secrecy, Signaling, and Safety in October 1969. International Security, 27(4), 150–183.
• Schelling, T. C. (1960). The Strategy of Conflict. Harvard University Press.
• Skitka, L. J. (2021). The Psychology of Moral Conviction. Annual Review of Psychology, 72, 347–366.
• Stout, M. (2005). The Sociopath Next Door. Broadway Books.
• Sun Tzu. The Art of War (5th century BCE; multiple English translations).
• Swann, W. B., Jetten, J., Gómez, Á., Whitehouse, H., & Bastian, B. (2012). Identity Fusion: The Interplay of Personal and Social Identities in Extreme Group Behavior. Perspectives on Psychological Science, 7(5), 441–456.
• Ury, W. (1991). Getting Past No: Negotiating in Difficult Situations. Bantam Books.
• von Neumann, J., & Morgenstern, O. (1944). Theory of Games and Economic Behavior. Princeton University Press.