The risks and costs of indirectness

↵ Back to class homepage

One of the main hallmarks of pragmatic meaning is that it is often indirect: people can mean something without saying it literally, explicitly, or directly. Most of this class has been focused on understanding and analyzing the many ways this can happen, through things like illocutionary force, implicatures (whether conventional, conversational, or something else), presuppositions, and metaphors.

Our ability and propensity to communicate indirectly, and to understand indirect communication (and to talk in ways that are based on expecting our listeners to understand our own indirect communication) might even be part of what makes human language unique. Nonhuman animals do have communication systems (although no nonhuman animal communication system has yet been found that displays the special features of human language), but it is not clear that nonhuman animals do pragmatic, Gricean sorts of reasoning that humans regularly do (see, e.g., Scott-Phillips, 2017 and Woensdregt & Smith, 2017). Pragmatics also sets human language apart from computer programming languages and formal logic (so when people make arguments about how language should be logical—e.g., if someone tries to use "logic" to criticize the way someone else uses or does not use double negatives, the Oxford comma, stuff like that—those arguments have no basis in linguistics or reality, and are usually just a way for people to try to use prescriptive grammar to try to act like they're better than someone else.)

At this point we can ask why people talk this way. Why did humans develop, and continue to use, this uniquely human pragmatic capacity, instead of just saying what they mean?

This is not a trivial question, because there are some serious problems that can arise with indirect pragmatic communication, as we will see shortly. In spite of these problems, people still communicate indirectly very often, so it seems like there probably is some reason it's worthwhile to use indirect communication in spite of those problems. In this module we will see some of the problems with indirect communication, and in the next module we will see some of the reasons why indirect communication is often useful and worthwhile anyway (so please don't misunderstand this module as something telling you not to use indirect communication; this module is not prescriptive, it's not telling you what to do, it's just trying to understand what the problems of indirect communication are so that in the next module we can even better appreciate why indirect communication is so valuable in spite of those problems).

Moving on to the problems with indirect communication: Thomas (chapter 5; citing Dascal, 1983) describes indirect communication as "costly and risky". Let's examine each of these characteristics in turn, starting with risky.

Indirect communication is risky

The most obvious drawback of indirect communication is that it carries the risk of being misunderstood. If you say something indirectly instead of saying it literally, people might not pick up your indirect meaning, and they might even pick up a different meaning than what you intended. The webisode below includes an example (an English translation is provided just below here):

Here is the key piece of dialogue (plus internal monologue) from the beginning of the clip:

It's clear that this guy wanted to know if the woman has a boyfriend. But instead of just asking her whether she has a boyfriend, he instead made an utterance which presupposes that she has a boyfriend ("You don't need to hang out with your boyfriend?"), in the hopes that she would answer his question by responding to the presupposition (e.g., by saying "Actually I don't have a boyfriend.") This is a super common use of presuppositions (another example, from a scene in BoJack Horseman: two characters, Emily and Todd, are developing an app, and are also clearly romantically interested in each other; Emily asks Todd, "So, do you think your girlfriend would get jealous if we started this project together, or...").

But instead, she responded to his literal meaning without addressing the presupposition. The question he was actually interested in remains unanswered, and thus all his confused thoughts. This illustrates a risk of indirect utterances: the person you're talking to might take the utterance in a different way than you intended. (In this case we can't even say the woman "misunderstood" what he wanted to ask; it's probably even more likely that she understood exactly, and avoided the question on purpose, perhaps because she doesn't want to give information about her personal life to some creepy guy.)

Is indirect communication costly?

Thomas (chapter 5) asserts that indirect utterances are "costly" because they take longer for the speaker to say, and also take more time and effort for a listener to understand. Both of these claims, however, are debateable.

It is not clear to me that indirect utterances always take longer to produce than corresponding direct utterances. (Note, though, that I am already oversimplifying the question; "indirect utterances take longer than direct utterances to produce" is a generic statement, and as we have seen, the meaning of generic statements is unclear; "indirect utterances take longer" might not necessarily mean that all indirect utterances take longer!) For example, asking a question by using a presupposition (as in the example we examined above) does not always take more words than asking the same question directly; for instance, "Where's your boyfriend?" would take fewer words (3) than "Do you have a boyfriend?" (5). Furthermore, as we saw in the "What's a stronger alternative?" module, utterances can be shorter than what they implicate (e.g., "I have lived in France and the UAE" is shorter than "I have only lived in France and the UAE").

It's also not clear to me that mere number of words (or any more fine-grained unit, like number of morphemes, number of syllables, number of letters, number of sounds, etc.) is an accurate measure for how difficult something is to say or understand. Study of psycholinguistics will quickly reveal that there are many ways that two sentences of identical length may differ in difficulty. Furthermore, sometimes a shorter utterance can be much harder to understand than a longer utterance that explains things more clearly, and likewise it may be harder to find a clear way to produce a short utterance than a longer one (ever tried to write a 200-word abstract or give a 1-minute speech?).

I think Thomas's main claim, however, is about the psycholinguistic processing costs of indirect utterances, rather than their actual linguistic length or complexity. She is saying that they take more time and effort for listeners to understand, and they are more difficult for listeners to understand; even for speakers, even if indirect utterances aren't actually longer, they might take speakers more time and mental effort to plan and execute the utterance.

Nevertheless, even those latter sorts of claims have also come under question. Thomas points out that many psycholinguistic experiments have confirmed this claim, but she was writing in 1995, before the heyday of experimental pragmatics began. As Noveck (chapter 14, inter alia) describes, psycholinguistic findings on this question are quite mixed. While there are some phenomena where people seem to often require extra effort to understand indirect or implicit meaning (for example, most scalar implicature researchers believe scalar implicatures require processing effort—and Noveck, chapter 6, reviews a lot of convincing evidence that children take a long time to learn this aspect of meaning), there are many others where people generally seem to get the indirect/implicit meaning more easily and more quickly than the direct/literal meaning (metaphor is one of the main cases of this). Ultimately, whether or not pragmatic meanings engender processing difficulty is still a pretty open question. Even for scalar implicature, where there is more widespread agreement on this, I am still not convinced.

While it is not clear yet whether indirect communication is costly (it's not clear what exactly would be evidence for or against this claim, how this sort of "costliness" could be measured, or whether the experimental research so far supports the claim), I think it should still be uncontroversial that indirect communication is, at least, risky, as discussed above.

As we have discussed above, though, speakers still do use indirect communication all the time, even though it is risky. Why do they do that? In the next module we will examine some of the reasons indirect communication may be useful or beneficial.

Video summary

In-class activities

An utterance that communicates something indirectly, like "You don't need to hang out with your boyfriend?", may sometimes be longer than another utterance that communicates the same thing directly, like "Do you have a boyfriend?" This is basically Thomas's point that we discussed above.

We could, however, argue that this utterance isn't longer relative to what it communicates. The indirect utterance communicates more than one thing (in addition to asking whether the woman has a boyfriend, it also communicates its literal meaning). So, even though it's slightly longer, it also conveys more information; maybe its information-to-length ratio is actually bigger than that of the direct utterance!

To be able to test whether or not this is true, and to be able to test Thomas's other claims, we would need some way to meaningfully measure indirectness. Particularly, we need to measure how much information is in an utterance (how many ideas it communicates), and how much linguistic material is in the utterance (how long it is).

Have students brainstorm a way to measure these things. Thomas (chapter 5.4) discusses several possibilities.


⟵ Politeness
Benefits of indirect utterances ⟶

by Stephen Politzer-Ahles. Last modified on 2022-03-27. CC-BY-4.0.