Communication games

The cooperative nature of communication

This post is the first of a series on communication games. If you want to understand the rich and subtle games that underlie your daily communications, tag along.

As humans stand out in many ways from other animals, many have pondered what makes us special. Aristotle is famous for pointing out the highly social nature of human life. And at the basis of this social life is a key ability: the capacity to communicate.

The anthropologist Robin Dunbar proposed an adaptive explanation for the link between language and social life: Language “allows a significant increase in the size of the interaction group.”¹ This idea makes sense because language allows us to exchange useful information. In particular, language allows us to manage larger social networks by exchanging information about others, their actions, and intentions.

You could however ask a simple question: why does communication work? If some people have useful information, why do they share it, instead of keeping it to themselves? When people disagree, why do they keep talking to each other? How can people trust what they are told when deception and inaccuracies are often present in communication? We can answer all these questions by unpacking the principles of the communication games we play in our everyday lives.

Communication games are mixed-motive games

Game theory is the study of social interactions—games—where the outcomes for each participant depend on the actions of all involved. As I have argued in an introductory post on the topic, games are everywhere in our lives, and game theory sheds light on how and why we behave the way we do.

In his classic The Strategy of Conflicts (1960), the game theorist and Nobel Prize winner Thomas Schelling proposes classifying games based on the degree to which participants' interests align. At one end of the spectrum are pure conflict games often called zero-sum games. In such games, what one player wins, the other loses. There is no common ground that players would want to reach. There can be only one winner. Such games include tic-tac-toe and duels to the death. You can win, lose, or tie; there is no win-win possibility.

At the other end of the spectrum are pure common interest games, where people’s interests are perfectly aligned but they need to coordinate their actions to achieve their shared goals. These games are also called pure coordination games. An example is a situation in which two people want to meet in a city without any preference for where. They just need to agree on a location together. If they do, it is a win-win solution.

A key insight of Schelling is that most real situations fall between these two extremes. They are mixed-motive games: situations where people’s interests are somewhat—but not perfectly—aligned. In such situations there is always some opportunity to cooperate, but also some inherent conflict of interest.

Why does cooperation work in communication games?

The notion of mixed-motive games is very useful for understanding the nature of communication games. Information is valuable for making decisions, so people can cooperate by exchanging information that benefits everyone involved. The exchange of useful information such as possible opportunities and risks pervades our daily interactions. Particularly important for a social species like ours is information about other people: their record of actions and intentions that inform us about their trustworthiness and willingness to cooperate with us and side with us, or to side with other people in disagreements.²

However, people do not have perfectly aligned interests when they exchange information. Let’s consider some examples. John and Jane are a couple who want to spend the evening together, but John prefers to watch an action movie while Jane prefers a romantic comedy. Jane and Jade work together to complete a project, but Jane prefers for Jade to do most of the work. John and Jack want to go on a hiking trip together, but John prefers a risky path while Jack prefers a safe one.

The challenge of cooperation when interests are not fully aligned

Whenever there is an element of disagreement, as in these examples, you may wonder why people would not just say whatever they can to get what they want instead of telling the truth. However, if people were always saying what’s most convenient for them, you’d expect that listeners would learn not to be gullible. They would end up ignoring what people say, whenever the point being discussed involves some disagreement. This means that communication could only work to solve pure coordination situations in which people’s interests are perfectly aligned.

But that is not the case. People do communicate even when they do not perfectly agree on the possible outcomes of the discussion. How is this possible? A famous answer provided by game theorists Vince Crawford and Joel Sobel (1982) is that communication can take place and be mutually beneficial between people who have different incentives, as long as these incentives are not completely opposed.³

Repeated interactions

Another answer comes from one of the most interesting results from game theory, which I have discussed in a previous post: repeated interactions create incentives for cooperation. In short, when people are repeatedly engaged in social situations with opportunities to communicate, they may benefit from choosing to be truthful, even when it is not in their short-term interest.

When Jane and John communicate truthfully with each other in repeated exchanges, they can both benefit from getting accurate and, therefore, valuable pieces of information. This cooperation can last as long as they both trust each other. However, if one of them lies to the other and is discovered, the mutual trust may be broken, and the beneficial exchanges might end. Thus, whenever Jane or John is tempted to be disingenuous, the potential for future interactions casts a shadow on their present decision.

Extract from Axelrod's The Evolution of Cooperation (1984)

Reputation as a communicator

We can extend these insights from interactions between two people to interactions between many individuals. I explained in a recent post how cooperation between strangers can be sustained if their past actions generate a reputation that determines whether others want to interact with them or not. This insight is succinctly summarised by psychologist Hugo Mercier:

Humans have developed a great way of making incentives fall in line: reputation. - Mercier (2020)

Incentives to be honest

Once we understand that communication has an element of cooperation, it is easy to appreciate that reputation also plays a role in that domain and provides incentives for honesty. When you lose the reputation of being a trustworthy communicator, it has real consequences: people stop caring about what you have to say. This is the lesson of the tale of the boy who cried wolf too many times—an Aesop fable commonly used to teach children the consequences of lying.

Illustration of the fable of the boy who cried Wolf, by Francis Barlow (1687). As the wolf takes away one of the boy's sheep, a farmer who is not intervening is seen in the background. The boy has previously cried too many times “wolf” while none were there. Source

Adam Smith, always so insightful about human psychology, had also noticed in The Theory of Moral Sentiments that the fear of losing one’s reputation keeps people trustworthy.

The prudent man is always sincere, and feels horror at the very thought of exposing himself to the disgrace which attends upon the detection of falsehood. - Smith (1759)

There are reasons why one may feel horror at the thought of having been exposed to have purposefully conveyed falsehood. Losses in reputation have a real impact on our social prospects, and this is also the case for losses in our reputation as an honest communicator.

The loss of reputation that can result from dishonesty is a major cost in a highly social species like humans, who continually monitor and gossip about each other’s behaviour and form judgements accordingly. - Scott-Phillips (2014)

Losing the trust of others shrinks our ability to cooperate with them in the future.

Incentives to be diligent

The importance of reputation extends beyond just honesty; our reputation as a communicator also encompasses the diligence with which we communicate. Are we striving to convey the most accurate and precise information? A trustworthy communicator provides well-calibrated signals (unbiased and not over/under estimating uncertainty). Communicators who are not fully diligent may make unnecessarily imprecise or unduly confident statements given the evidence available to them.

Being a diligent communicator is a crucial trait of a good cooperation partner. Receivers should be able to keep track of who is diligent and who isn’t, and adjust their future behavior on this basis, so that they are less likely to listen to and cooperate with people who haven’t been diligent. - Mercier (2020)

The origin of hedging

Consequently, we have an interest in looking diligent in our communication. It is a major reason for hedging, the use of expressions to qualify the degree of our confidence in the information we give to others.

Consider John making the following statement: “Jane is pregnant”. This statement is either true or false. In case of error, John could be seen as carelessly making an inaccurate statement. To prevent this, he can qualify this statement with expressions indicating different degrees of confidence, like: “I’m sure”, “I believe”, “I think”, “I guess”, “I suspect”, “I have the feeling that”, or “I have heard that”, and so on. Expressions indicating a degree of uncertainty and pointing to the source of the information can help preserve John’s credibility later on. Even if proven wrong, he may be seen as having made a reasonable assessment of the evidence at the time.⁴

Deception with plausible deniability: paltering

Reputational concerns do not fully eliminate differences in interests. Indeed, it is well-known that people lie. Studies have found that people typically lie at least once a day.⁵ However, flat-out lies—factually wrong statements—are hard to deny when found out. As a consequence, people more often have recourse to deceptive statements that provide the cover of plausible deniability.⁶

Between truth and falsehood, there is the area of half-truths and imprecise statements whereby people try to mislead others or hide the truth without openly making untrue statements. Here again, Adam Smith had identified this issue in communication. His quote about the “prudent man” I presented above was followed by these words:

But though always sincere, he is not always frank and open; and though he never tells any thing but the truth, he does not always think himself bound, when not properly called upon, to tell the whole truth. - Smith (1759)

One interesting type of half-truth is when accurate statements are used to deceive. This practice has been termed paltering by Schauer and Zeckhauser (2009). Consider the following example: Jane told John she would bake a cake for his birthday. Short on time, she goes to a bakery and buys a birthday cake. When John sees the cake, he compliments her on how good it looks. At this point, let’s imagine she just says “thank you”. With this statement, she implicitly confirms John’s assumption that she did make the cake herself.

Jane’s answer is misleading John but it is harder to pinpoint as intentionally deceiving than a flat-out lie. Formally she has not lied. If John finds out that Jane did not make the cake, it is not enough to be sure that she intended to deceive him. Such an intention would depend on Jane's beliefs about John's understanding, which he cannot observe. Jane could for instance claim that she thought she had mentioned to him that she would get the cake at the bakery. To prove that Jane wanted to deceive him, John would need to know what Jane believed about his interpretation of her answer at the time. Jane’s explanation may be more or less credible, but most often she may benefit from having plausible (enough) deniability.

Our daily communications, both formal and casual, are communication games. These are mixed-motive games where participants have opportunities to gain from cooperating but where tensions between different interests can also be present. This perspective explains why we care about our reputation as communicators, and why we manage this reputation carefully through our discussions. By understanding the underlying elements of cooperation and competition in our daily communications, we can navigate the subtle communication games that shape our social lives with an enriched perspective.

This is the first of a series of posts on communication games. In the next posts, I’ll discuss a range of insights this perspective gives on our daily communications. Related reading: Dan Williams just published a post on how communication games shape our intellectual and political discussions.

References

Axelrod, R., 1984. The evolution of cooperation. Basic Books.

Crawford, V.P. and Sobel, J., 1982. Strategic information transmission. Econometrica: Journal of the Econometric Society, pp.1431-1451.

DePaulo, B.M., Kashy, D.A., Kirkendol, S.E., Wyer, M.M. and Epstein, J.A., 1996. Lying in everyday life. Journal of Personality and Social Psychology, 70(5), pp.979-995.

Dunbar, R. I. Gossip in evolutionary perspective. Rev. General Psychol. 8, 100–110 (2004).

Mercier, H., 2020. Not born yesterday: The science of who we trust and what we believe. Princeton University Press.

Schauer, F. and Zeckhauser, R., 2009. 2 Paltering. in Harrington, B. (ed.), Deception: From ancient empires to Internet dating. Stanford University Press.

Scott-Phillips, T., 2014. Speaking our minds: Why human communication is different, and how language evolved to make it special. Bloomsbury Publishing.

Serota, K.B., Levine, T.R. and Boster, F.J., 2010. The prevalence of lying in America: Three studies of self-reported lies. Human Communication Research, 36(1), pp.2-25.

Smith, A., 1759. The Theory of Moral Sentiments.

Sobel, J., 2020. Lying and deception in games. Journal of Political Economy, 128(3), pp.907-947.

1. Dunbar (2004)

2. This type of social information, gossip, has been estimated by Dunbar (2004) to account for around two-thirds of our social exchange.

3. For the interested reader, here is a description of the intuition of their results with a practical example. John, an entrepreneur, has private information about the true amount of money he needs: $500,000 to fund his start-up, which typically needs anywhere between $0 and $1 million. He approaches Jack, a venture capitalist, to secure funding. John's ideal amount of investment he would like to receive is slightly higher than what he needs: $600,000. However, Jack wants to invest the exact amount required for the start-up to succeed (he doesn’t know that this amount is $500,000). Jack is aware that John wants to ask more than what he needs.

How can they work things out given that they disagree to some extent on the sum? If John gives a number, Jack would want to give less since it is likely an inflated number. This could incentivise John to ask even more. So how can Jack believe anything John says? Crawford and Sobel show that there is an equilibrium where John doesn’t send a precise number but a range that includes the true value to Jack, for instance, $400,000-$700,000. So John is able to send some information to Jack. The smaller the difference in interests between John and Jack the smaller the range will be (the more information John will be able to convey to Jack). For instance, if John’s preferred amount is only $550,000 he may ask for an amount in the range of $500,000-$600,000.

4. This example is adapted from one discussed in Speaking our Mind by

Thom Scott-Phillips

(2014).

5. DePaulo et al. (1996) have found that people lie in one in five conversations and Serota et al. (2010) found that 40% of people lie at least once a day.

6. Economists will be interested to look at the formal treatment by Sobel (2020) of what lies and deception are.