A Soldier Is Taken Prisoner

What continues to catch our attention about this matter of persuasion is that, by very different routes, other sciences have reached similar conclusions. Today, for some scientists, a system’s capacity to be persuaded is an indicator of where we should place that system on the broad spectrum of what we know as intelligence, which in turn is related in complex and no less confusing ways to its level of agency.

Once again, we’re not trying to find intelligences superior to ours, but ones close enough for us to build a bridge. What, exactly, is that bridge made of? If we believe it’s made of Jakobson’s structuralist theory, then our task is already complete and this book has no reason to exist. We’ve seen that dogs can be senders and receivers of a message, and that communication can be successful even when the code is hybrid or not fully shared. However, this idea leaves out something we might consider fundamental to carrying out the delicate art of conversation with greater complexity: inferences.

In the previous chapter, we saw that inferences are assumptions we draw from what the other person said or left unsaid, from their specific violations of the cooperative principle. They depend on a kind of metagame: the exchange of secret messages smuggled between the lines, while words go back and forth in broad daylight. They imply that both interlocutors share some understanding of the context they find themselves in and can make assumptions about one another.

Let’s leave the animals in peace for a moment and look at an example involving humans:

Second World War. An American aviator is shot down over Italian territory. He parachutes down and immediately takes off his uniform. Half-naked as he is, he begins fleeing north through the forests of Piedmont, hoping to find the Swiss border. But he doesn’t get far before he is captured by a fascist patrol. The Italians realize he is a soldier and take him to a cell, but throughout the entire journey, despite the abuse and threats, the American soldier makes sure not to say a single word. Imprisoned, he thinks about how he can get out of this situation. A brilliant idea comes to mind. When the officers come to interrogate him, he’ll speak German. That way, they’ll think he’s an Axis soldier and let him go. The only problem is that he doesn’t know a single word of German. Except, he remembers now, a few lines from a poem he was taught at school:

With nothing to lose, the soldier waits. When the enemy officers appear and begin interrogating him, all he does in response is recite lines from the poem in the best German accent he can manage. The Italians ask him which army he belongs to, and he answers with the country where the lemon trees bloom. They ask what mission he’s on, and he answers that a gentle wind blows from the blue sky. They tell him they’re going to execute him, and he confesses that he loves them and wants to travel with them.

The Italians don’t understand a word, so the ruse works. Convinced they’ve captured an Axis soldier, they dress him and let him go. The American soldier flees, crosses the Swiss border, and disappears from history.

He leaves behind a single question: did he actually manage to communicate with the Italians? The soldiers did what he wanted and understood his intentions, but they didn’t decode the meaning of his words. Is that communication? Or is it an event more like my cat coming over because it sensed that’s what I wanted, rather than because it understood the word let’s go?

I’m inclined to think the former: it was an act of communication because the Italians understood that there was a code, understood which code it was, and simply couldn’t decipher it, but consciously took part in the act of communication. Austin would say the poem was a felicitous utterance, which is how Austin defines an utterance that performs successfully. Regardless of whether its content was true or accurate, it turned out to be unexpectedly appropriate to the situation. Grice would be satisfied because the intention was understood and, therefore, the meaning was conveyed. But Searle, who came up with the parable of the captured soldier, asks: how did that act of communication work? On what basis did the Italians make exactly the inferences the American soldier wanted them to make?

The debate remains open, but we can take something valuable from it. If Searle is right, communication is impossible without access to a shared code. If Grice is right, we can continue the journey with a glimmer of hope because, even without ruling out the element of luck the American may have had, it was his ability to imagine the Italians’ mental states and influence them that unlocked communication.

When it comes to two human beings, we take for granted that there is a certain symmetry in the cognitive abilities at play. Saussure’s speakers A and B, Jakobson’s sender and receiver, the person who constructs implicatures and the person who infers them in Grice are all subjects of the same species: humans. But if we want to apply these ideas to a hypothetical interspecies communication, where there may be a profound asymmetry in the interlocutors’ cognitive nature, we’ll probably need another kind of validation. Another way of measuring the terms of the conversation. What term can we use instead of theory of mind when we don’t know whether the minds involved are comparable entities?

The biologist Michael Levin proposes the term persuadability:

Besides the question of scale—which is no small matter—this was also the problem faced by the scientists of Solaris. How persuadable is the ocean? How persuadable is an insect? How persuadable is my cat? The answer to the last question was always a firm “not very,” except for that one day when I said “let’s go” and he came.

In a passage from El nudo de la conciencia, Enzo Tagliazucchi explains that there are different ways of approaching a system. For instance, although we might casually say that the water heater is intelligent because it heats the water when it detects that it’s cold and turns off the flame when it detects that it’s hot, the truth is that when it breaks, we can’t convince it to change its ways. We have to take it apart and micromanage it: repair the parts, put the system back in order.

Communicating with other species, no offense intended, forces us to take similar considerations into account. I need to know where my interlocutor falls on the persuadability spectrum to know whether I’ll manage to convince them to do something—especially to convince them to respond to me with a complex signal like a word—or whether I need a whole different set of tools. For example, Levin works to persuade organic tissue to grow. Usually, that would be done by modifying the cells’ DNA (taking the water heater apart and repairing the parts); Levin, however, is trying to interfere with the electrical signals they send to one another. He intervenes in the communication to alter the content of the messages, as in any decent spy movie. According to him, physics uses tools with low agency, which is why it always sees mechanisms. “If you want to see minds, you need to treat them as minds,” he says. It’s another way of saying that if I want to talk to my cat, I need to take for granted that he has a mind. That he can use it. Not treat him like a water heater, but, in the terms of Daniel Dennett (which, in turn, I’m borrowing again from Tagliazucchi’s The Knot of Consciousness), adopt the “intentional stance.” Assume he is persuadable. That we can play a higher-order game together without needing to understand what the abyss beneath us is made of.

But watch out. Because there are two problems. On the one hand, what I’ll call the problem of rebellion: as beings acquire psychological dimensions and constructions of the “self,” they may have fewer reasons to obey an order. This is undoubtedly true of my cat. According to pragmatics, uttering something has an effect on the real world. But there’s no guarantee that the effect will always be the one we’re after. That’s the risk the American soldier was taking.

The second problem is what we might call the Bible salesman problem. In a famous short story by Flannery O’Connor titled “Good Country People,” the following story is told:

Hulga, an intellectual, cynical woman with a wooden leg, lives trapped in the conviction that her nihilism makes her superior to the simple people in her rural community, including her mother. When Manley Pointer, a young Bible salesman who appears to have unshakable faith and complete naivety, arrives at the farm, Hulga decides to use him to prove her superiority. She plans to seduce him and break his heart, stripping him of his illusions and proving to him, to her own religious mother, and, why not, to God himself that the world has no meaning. She arranges to meet him in the barn. Everything is set for turning the saint into a sinner; however, in a brutal twist, Hulga discovers—too late—that the roles had been reversed from the start: Manley isn’t the simple young man she thought he was, but a cynical predator whose goal is not sexual but purely sadistic: to steal her wooden leg, leaving her helpless and revealing that her intellectual arrogance has blinded her to a far more sophisticated and dangerous form of evil.

The point is this: the most persuadable system can often be the most persuasive. Or, put another way, we—the sapiens, the rational ones, the pinnacle of animal thought—can also be persuaded. Or, to put it one final way, for the sake of absolute clarity: no one comes away unscathed from a conversation worth having.

The point is that, by now, to fulfill our dreams we no longer need only a species that shares our spatial scale, has a certain kind of consciousness and self-awareness, has a grammar that lets it create new sentences from existing ones, has some ability to provide false, abundant, scarce, or irrelevant information, prompting its interlocutor to make inferences, is persuadable, and has some kind of motivation to communicate with us. We also need to take a risk. Oh, and a translator.

The animal kingdom is vast, but our options have just shrunk dramatically.