The second half of chapter 6 faces the biggest puzzle of linguistic communication: why doesn’t lying spread?
The hearer’s side: epistemic vigilance (6-4)
Ostensive communication brings in useful and false information alike, and hearers must sift it. If they can detect a malicious or incompetent speaker, they can ignore what that speaker says (pp. 275–276).
The important point is that understanding is not accepting. If told “There’s cake for dessert,” I can understand that the speaker wants me to believe it without believing it. There are two reasons: I may not trust the speaker, or I may think the speaker is mistaken. The code model leaves no room for this. On the ostensive view, the hearer can respond to the communicative intention while not responding to the informative intention. This is epistemic vigilance (p. 276).
Vigilance is not always on full. I do not suspect my parents of trying to deceive me. The opposite of vigilance is not trust but blind credulity (p. 278).
The speaker’s side: argumentation
How can speakers get hearers to believe them? Force cannot work. What works is offering reasons the hearer can accept, that is, argument (p. 278).
This is the starting point of the argumentative theory of reasoning (Mercier and Sperber). On it, the proper function of reasoning is not to deepen knowledge but to build and evaluate arguments for persuasion (p. 278). This changes how confirmation bias looks.
- Traditional view: confirmation bias is a flaw in human reasoning.
- Argumentative view: for a speaker, seeking evidence that fits one’s claim serves the purpose; for a hearer, vigilance applies. Argument and vigilance are two sides of one coin (pp. 279–281).
The example is Wason’s 2-4-6 task. Participants must find the rule behind three numbers, and they mostly test sequences that fit their own hypothesis; few try to falsify it. Traditionally this was read as a reasoning flaw. The argumentative theory rereads it as an ability designed to supply evidence for one’s own claim to others (pp. 279–281).
The author notes this has the same structure as the motion-induced blindness argument in article 282: what looked like a flaw looks, from the standpoint of function, like an adaptation.
The boy who cried wolf (6-5)
The author cites an episode of The Simpsons and links it to Aesop’s “The Boy Who Cried Wolf”: a boy who lies is not believed when real danger comes (pp. 281–282).
In evolutionary biology this is a problem of evolutionary stability. If lying paid, individuals sending false signals would increase. Hearers would then do better by ignoring signals. Indifference to signals would evolve, and the system would collapse. Yet human language does not collapse. Why (p. 282)?
The author notes that the question “why do people hand so much information to others?” has been treated as a paradox, but he says the question is framed slightly off. Humans do not speak only to those who are reliably trustworthy; they speak because they want to affect the hearer’s mind (pp. 283–284).
The handicap principle
Biology’s famous answer is the handicap principle (Zahavi, Spence): a signal that costs a lot is evidence of honesty just by being costly. A peacock’s long tail is a burden, but it shows the bearer has survived the burden. An individual without spare capacity cannot imitate it (p. 284).
The author says, however, that the principle is often misunderstood. For the handicap to work, the cost of the signal must correlate with the meaning of the signal: better individuals bear the cost more lightly. Without that gap, the handicap does not guarantee honesty (pp. 284–285).
Does it apply to language? He doubts it.
- The cost of uttering words does not depend on whether the content is true. Lying and truth-telling take the same effort (p. 292).
- Proposals that, for example, “polite expressions are handicaps” also fail to show that the cost-gap condition is met (pp. 290–291).
Indices and deterrence
There are other ways to stabilize signals (pp. 286–289).
- Index: a physical causal link between the signal and what it indicates makes it unfakeable. For example, the formant dispersion of a red deer’s roar tracks body size, and a small animal cannot fake a large one’s roar.
- Deterrence: lying is punished. The size of a bird’s throat patch is not a real measure of strength, but individuals of similar strength come into conflict, so deterrence operates.
- Shared interest: if speaker and hearer interests coincide, there is no reason to lie.
The author says which of these applies to human communication is an empirical matter.
In humans: social reputation (6-6)
His answer is reputation (pp. 290–294).
Someone known as a liar will be ignored and shunned in the future. For a species as social as ours, which watches others and gossips, losing one’s reputation is very costly. The author says Cassio’s lament over lost honour in Shakespeare’s Othello conveys the lesson (p. 292).
So honest speech is held up not by the cost of the signal itself (a handicap) but by deterrence imposed by public reputation. This predicts that when people think their reputation will not suffer, they lie more readily. The author says he tested this prediction in a controlled experiment with the help of an evolutionary biologist, and the results supported the view. The data are unpublished (p. 293).
Is language a major transition in evolution? (6-7)
At the end of the chapter the author takes up Maynard Smith and Szathmáry’s “major transitions in evolution”: eight events (replication of genes, multicellular life, sex, and so on) in which the way information is passed changed. The origin of language is counted as the last one (pp. 295–296).
He is cautious. One of the criteria is that lower-level units that once replicated independently join into a higher-level unit, irreversibly. He is not sure the origin of language fits. But he says what is truly new is not language itself but the start of ostensive-inferential communication. It is uniquely human, and thanks to it any behaviour can become a means of communication. Language makes that novelty more special still (pp. 297–298).
Chapter summary
If there is a cognitive ability specific to language, its evolutionary function should be to improve the quality of linguistic communication. More generally, the direct function of ostensive communication is reading minds (the hearer’s side) and manipulating them (the speaker’s side), tied to the evolution of social intelligence. The most important mechanism keeping it honest is social reputation (p. 298).
Epilogue: answering the big questions (pp. 299–302)
The author collects from five sources a list of problems any theory of language origins should answer, sets it out in table 7-1, and adds his answers. In brief:
| Problem | The book’s answer (summary) |
|---|---|
| Uniqueness: why only humans? | Only humans began living in large, complex social groups. Advanced social cognition (especially recursive mindreading) was selected, making ostensive-inferential communication possible |
| Honesty: why stable? | Social reputation deters lying |
| Function and environment | Mindreading and psychological manipulation; evolved as a means of surviving complex societies |
| Expressive power | Ostensive-inferential communication allows free creation and flexible use of signals |
| Patterns | Individual languages move, through cultural attraction, toward patterns that suit human dispositions |
| History | Ostensive communication → conventional codes → these fixed into today’s languages |
| Signs tied to the world | Early ostensive-inferential communication relied on metapsychological ability; behaviours that gave meaning cues became shared norms and turned into words |
| Selfishness | Speakers speak to affect hearers’ minds |
| Coherence | Relies only on well-understood evolutionary mechanisms |
The author’s closing summary: only humans came to live in large, complex social groups, and social cognition was strongly selected. Ostensive-inferential communication then became possible, and people began to share various conventional codes. These became fixed as today’s languages. That is why we alone have language (p. 302).
From the translator’s afterword
The afterword (from p. 303) says the Japanese book is a full translation of the original. It also touches on the big question of whether language is for thought or for interpersonal communication, supplementing the book’s view: the understanding that language’s primary function is to help the hearer read the speaker’s intention, not to display thought, extends to the semantics of natural language, the translator says (pp. 304–306). I have not seen the bibliographic page.
Note
This feature is a summary of the whole book and relays the author’s claims as he makes them. Many points, such as the assessment of the handicap principle, the (unpublished) reputation experiment, and the placement of language among the major transitions, are contested among researchers. Check them against the original papers and other literature.