Recently a student approached me after I gave a lecture on 'Interpersonal Communication' and asked a question about the course's textbook. I answered his question and we spoke for a while about this book. Yet, at a certain moment we realized he was talking about a Marketing textbook, whereas I was referring to the Communication textbook assigned for my course. It turned out that he was about to attend the next lecture, he had never seen his own lecturer and, given that I was standing in front of the lecture room messing with my papers, he assumed I was teaching his course. Most people have experienced situations quite like this. After talking for quite a while, it dawns on you that you are talking at cross purposes.

The fact that such miscommunications occur is understandable when you look closely at what people say in conversations. The things we say, namely, are quite vague and ambiguous. Take a simple statement like "I will be there soon". These words can either mean 'in a minute' or 'tomorrow'. Similarly, a statement like "One has to wonder about Obama", seems to imply a lot, but nonetheless it could mean just about anything, positive or negative. Moreover, many utterances, such as requests, are made indirectly. Rather than saying "close the door" one often says something like "it's getting pretty cold in here" or "I do not want the cats to escape". To make it worse, people may mean the exact opposite (e.g., in irony: "a fine friend you are") or something completely different than what they literally say (e.g., "you're the candy on my cake"). If words are so ambiguous, then how on earth do we understand what someone means?

Apparently, understanding the literal meaning of a string of words is only the beginning of interpreting what a speaker intends to say. One of the most important tasks of listeners is therefore to figure out the intentions behind a speaker's words. Speech act theory (Austin, 1962; Searle, 1969) was one of the first attempts to tackle how people do this. A listener needs to figure out what someone wants to achieve with saying "I do not want the cats to escape". This is quite crucial, because if this utterance is taken literally one may start lining up the cats, when it is recognized as an indirect request the listener may indeed close the door. To be able to understand, a listener first of all needs to recognize whether the literal meaning is intended, or rather some ulterior intention is communicated indirectly. How exactly people do this is a tremendously difficult question, and researchers from various disciplines are still debating on it. However, in general one can say that the intonation of an utterance, the context in which it is uttered, and background knowledge all contribute in various ways to interpersonal understanding (Holtgraves, 2002).

Apart from word meaning a large part of the intended meaning of an utterance is communicated nonverbally, for instance by tone of voice, gestures or facial expressions. People use verbal and nonverbal acts in concert to convey meaning (Bavelas & Chovil, 2000). When someone points at the door while saying "I do not want the cats to escape" it is quite clear that she wants the door closed, and when a meaningful smirk accompanies a sentence ("a fine friend you are") it is apparent that the message should be taken ironically.

In addition to nonverbal acts, the context in which an utterance is made helps to disambiguate words and is enormously important for understanding intentions. Paul Grice (1975) provided an influential theory that explains how people use the context of a conversation to convey meaning, and, in turn, how listeners use the conversational context to recognize what a speaker means to say. Suppose you ask someone "Where is Peter?", and you get the reply: "There is a red truck in front of Maria's house". How do you know what this means? Is it actually a reply to the question? The trick is that people are expected to abide to certain conversation rules. That is, in a conversation one is supposed to say things that are relevant, truthful, clear, and of the right size. Most likely, you will expect this reply to be relevant, true and an informative answer to your question. This assumption allows you to interpret its meaning (i.e., probably Peter owns a red truck and the fact that it's seen at Maria's house suggests he's there).

According to Grice (1975) listeners will search for an ulterior meaning when the literal meaning of an utterance appears to violate conversational rules. For instance, indirect and ironic utterances (e.g., "you're the candy on my cake", "a fine friend you are") are violations of the conversational rules, because they are not literally true. Yet, because listeners assume it must be relevant in some way, they interpret the meaning of these utterances such that it is understandable in light of the conversation context.

These examples demonstrate how complicated it is for listeners to understand what someone means to say, and, accordingly, how complex it is for a speaker to convey a simple message to someone. In order to understand, and to be understood, conversation partners must take the context and each other's viewpoint into account. Research has shown that perspective taking and determining common ground are crucial here.

Perspective taking and common ground

Perspective taking is one of the most important requirements for successful communication (Holtgraves, 2002; Krauss & Fussell, 1991). To get a message across a speaker must consider whether a conversation partner can understand what (s)he is saying. At the same time, a listener must consider the perspective of the speaker to understand the intention behind an utterance (what could he or she mean given his perspective?). Suppose your friend asks you "Isn't she from New York?". Before you can reply to this question you must understand who your friend is referring to. If you were in the middle of a conversation about the author of a book, you probably think he is referring to the author. However, maybe your friend strayed off and is referring to the singer of a song he hears on the radio. The question would be unclear to you if you were unaware of this sudden change in topic. To prevent miscommunication your friend would have to consider your perspective and indicate the change of topic by saying "This singer, isn't she from New York?", otherwise you may quite likely assume he is still talking about the author of the book (from Moskowitz, 2005).

Thus, to maintain mutual understanding conversation partners need to take each other's perspective. Yet, how does this work? An important area of research has revealed that conversation partners do this by constantly trying to determine what is common ground (Clark, 1996). Common ground refers to the knowledge that conversation partners share; the information conversation partners (have come to) mutually know, believe, and recognize. It has been shown that people rely on heuristics, or rules of thumb, that allow them to asses what is part of common ground (Clark & Brennan, 1991).

One of these rules of thumb that people rely on is the heuristic of physical copresence. If two people are physically together and attending to roughly the same aspect of their environment, they assume that what both perceive is common ground. When someone is passing by in front of you, you can assume that the person's behaviour and appearance is mutually known, and you consequently don't need to describe this to your conversation partner.
A second rule of thumb is the so-called heuristic of linguistic copresence. This means that people generally assume that information previously introduced during a conversation is part of the common ground. This simply means that once you have mentioned something, you don't need to explain it again. 

A third and very interesting finding is that people draw immediate assumptions about common ground from their conversation partner's community membership. The categorization of a person in a certain category or community (e.g., nationality, age, gender, occupation) activates a stereotype that allows one to make reasonable assumptions about what the other person knows or beliefs. When strangers meet they will immediately try to categorize each other and make generic assumptions about the other person's knowledge. Based on each other's appearances (a police officer, man in expensive suit, a little kid) or the situation in which they happen to be in (behind a post office desk, in a soccer stadium) people make immediate inferences. Based on these assumptions people adapt their utterances. Without contemplation you know you do not need to explain a postman about stamps, two English speakers can assume they are both familiar with the meaning of English vocabulary, and one can reasonably assume that a police officer is knowledgeable about traffic rules. The student in my previous example assumed that I was his lecturer simply because I was standing in front of the lecture room, and consequently inferred that I would be able to answer questions about a textbook.

Thus, based on immediate inferences about what they share with a conversation partner, people adapt their utterances. That this may occur quite unconsciously and on the basis on very subtle cues is shown in a classic field experiment by Douglas Kingsbury (1968). Kingsbury asked randomly selected pedestrians on a Boston street for directions to a department store several blocks away. However, he asked his question in either the local Boston dialect, or employing a dialect spoken in rural Missouri - one seldom heard in downtown Boston - or he prefaced his request with the statement "I'm from out of town". The results showed that when the request was made in the exotic dialect, the directions given were significantly longer and more detailed, compared to when he requested directions using the local dialect. In fact, the directions given to the exotic dialect request where quite similar to when it was explicitly stated that he was from out of town. Apparently, on the basis of his foreign dialect alone, people assumed that his level of local expertise was low and, without being asked, provided additional information. This, and other research, demonstrates that people categorize others on the basis of very subtle cues. This enables them to infer what the person is likely to know and to formulate a message that is understandable in light of such knowledge.

Of the individuals we know well, we also carry around quite some information about their expertise and the knowledge we share with them. For instance, if you have lived together for years with your husband, you assume that your shared vacations, preferred meals, disagreements, and private jokes are mutually known and can be referred to.

It should be clear that the amount of common ground between conversation partners has a huge impact on their utterances and the course of their conversation. People adapt their utterances based on the perceived common ground. But common ground is not static, it changes on a moment-to-moment basis. The conversation continues to unfold, new topics are introduced and even the environment and context may change. Whether something is already part of common ground or not influences how people refer to it.

A straightforward example is that with accumulating common ground, people change from indefinite descriptions, like 'a lawyer' or 'some houses', to definite descriptions, like 'the lawyer' or 'those houses' (Linde & Labov, 1975). The first mention of something, when it is not yet part of the common ground, is usually done with indefinite articles (a, an). Later, when a topic has been grounded (i.e., added to the common ground; Clark & Brennan, 1991) definite descriptions are used. This can also be seen in how stories are build up: "This is a story about a girl. The girl lived in a big castle".

Similarly, when speakers are trying to explain what they are talking about, they first tend to offer a conceptualization that is tentative or provisional, using hedges such as sort of, kinda, looks like (e.g., "the guy that acted kinda like Mr. Bean, you know?"). Once conversation partners agree to a conceptualization, they drop the hedges and use more definite descriptions (e.g., the Bean guy; Brennan & Clark, 1996; Clark & Bangerter, 2004).
Related to this is another important finding, namely that with accumulating common ground conversations gain in efficiency. Conversation partners need less words, less turns, and less time to get mutual understanding on a topic (Wilkes-Gibbs & Clark, 1992). You will not need to tell your husband the whole story about your mutual holiday to France in great detail. When referring to this mutual experience you can simply say "just like France", and he will understand. In other words, when common ground between conversation partners increases, they can understand each other more easily (Schober & Brennan 2003).

Miscommunication and grounding

Just as the presence of common ground is enormously important in reaching mutual understanding, many forms of miscommunication are the result of a failure to consider common ground. One thing is that relying on general rules of thumb or heuristics in assessing common ground can lead to errors. First, the reliance on the heuristics of physical and linguistic copresence doesn't guarantee a correct judgment. A conversation partner may have missed the person passing by, and have forgotten about or missed something you previously mentioned. In a similar vein, community membership doesn't guarantee knowledge on a subject. General expectancies based on categorization to a category or community (i.e., stereotypes) do not always fit particular instances. We have beliefs about what elderly people are like in general, but this doesn't mean that each individual older person is exactly like that. As a result, people may address an elderly woman using a loud voice, even when she is not deaf at all. Similarly, if someone speaks with a particular foreign accent, they may have lived in your city for a decade, and may not need extensive directions to the supermarket.

Other types of miscommunication arise from a type of egocentrism in which the other person's perspective is simply not taken into account. In everyday conversation people often speak about ideas that pop into their mind (e.g., about a singer you hear on the radio), yet they fail to realize that they have not informed their conversation partner about the shift in topic. Also, taking another person's perspective requires the right knowledge, ability, and motivation that people often do not have (Horton & Keysar, 1996;Schober & Brennan, 2003). The student in the textbook example simply couldn't know I was teaching another course, because he hadn't seen his own lecturer.

Luckily, depending on the communication channel, conversation partners can immediately repair miscommunications once they become clear. In face-to-face conversations we provide each other feedback about whether something is understood or not, and if needed we can immediately correct or provide additional information. Feedback, using vocal (e.g., uhuh, yeah, ok) and nonverbal signals (e.g., nodding, a smile or frown) are crucial for having a smooth conversation (Clark & Brennan, 1991; Clark & Krych, 2004; Clark & Wilkes-Gibbs, 1986). This process in which listeners inform speakers about their level of understanding occurs mostly automatic and outside awareness. Yet, these subtle nonverbal reactions are very important, and even in situations in which one person is talking and the other is merely listening, speakers constantly monitor their conversation partner and change their story depending on the listeners' replies and feedback (Bavelas, Coates & Johnson, 2000; Beukeboom, in press;Kraut, Lewis & Swezey, 1982).

An interesting demonstration of the effect listeners have on speakers is provided in an experiment in which participants were instructed to tell a close call story to another participant (Bavelas, Coates & Johnson, 2000). Half of the participants pairs were allowed to behave naturally, yet in other experimental conditions the listeners were distracted (unknown to the speaker) by asking them to count the number of t's in the speaker's story. This distraction task made it virtually impossible for listeners to show their usual listener responses, which allowed the researchers to test whether this changed the speaker's behavior. Interestingly, switching off the responses of listeners by distracting them had a huge impact on how speakers told their story. In particular, when listeners were distracted, speakers' stories were told less well and particularly poorly at what should have been the dramatic conclusion. The endings were abrupt and choppy, they circled around, or were told more than once. Also, speakers tended to needlessly reiterate the problem, add irrelevant information, and tended to justify or explain the plot. This research clearly demonstrates that conversations are a joint activity, in which conversation partners collaborate moment by moment to create and maintain mutual understanding.

When focusing on the different mechanisms people rely on in conversations, one may realize that it is in fact quite a remarkable achievement that we usually are able to understand each other quite easily. Generally, we don't (need to) think about how we actually do this. What we may experience, however, is whether a conversation runs smoothly and feels pleasant or not. Interestingly, the ability of conversation partners to reach and maintain mutual understanding is strongly related to how the interaction is experienced. People always strive to get understanding with as little effort as possible (i.e., principle of least collaborative effort; Clark & Brennan, 1991) and we generally prefer to get acceptance and agreement on our utterances (i.e., preference agreement; Pomerantz, 1984). Consequently, pleasant conversations are conversations in which both parties quickly understand what the other intends to say and possible misunderstandings are smoothly corrected. It is simply nice to learn that you share common ground with someone. In contrast, it is annoying when it is difficult to reach common ground; when you keep disagreeing, or when it takes too many words or turns to get the other to understand something. Talking at cross purposes for a minute about a different textbook may be annoying simply because it is a waste of energy. The best conversations are most likely those in which you understand each other without words. But don't take this literally.

References

Austin, J. L. (1962). How to do things with words. Cambridge: Cambridge University Press.

Bavelas, J. B., & Chovil, N. (2000). Visible acts of meaning. An integrated message model of language use in face-to-face dialogue. Journal of Language and Social Psychology, 19, 163-194.

Bavelas, J. B., Coates, L., & Johnson, T. (2000). Listeners as co-narrators. Journal of Personality and Social Psychology, 79,941-952.

Beukeboom, C. J. (in press). When words feel right: How affective expressions of listeners change a speaker's language use.European Journal of Social Psychology.

Brennan, S. E. & Clark, H. H. (1996). Conceptual pacts and lexical choice in conversation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 22, 1482-1493.

Clark, H. H. (1996). Using language. Cambridge: Cambridge University Press.

Clark, H. H., & Bangerter, A. (2004). Changing conceptions of reference. In I. Noveck & D. Sperber (Eds.), Experimental pragmatics . Basingstoke, England: Palgrave Macmillan, pp. 25-49.

Clark, H. H., & Brennan, S. E. (1991). Grounding in communication. In L. B., Resnick, J. M., Levine, & S. D. Teasley (Eds.),Perspectives on Socially Shared Cognition (pp. 127-149). Washington, DC: American Psychological Association.

Clark, H. H., & Krych, M. A. (2004). Speaking while monitoring addressees for understanding. Journal of Memory and Language, 50, 62-81.

Clark, H. H., & Wilkes-Gibbs, D. (1986). Referring as a collaborative process. Cognition, 22, 1-39.

Grice, H. P. (1975). Logic and conversation. In P. Cole, & J. Morgan (Eds.), Syntax and semantics (pp. 41-58). New York: Academic Press.

Holtgraves, T. M. (2002). Language as social action: Social psychology and language use. Mahwah, NJ: Erlbaum.

Horton, W. S., and Keysar, B. (1996). When do speakers take into account common ground? Cognition, 59, 91-117.

Kingsbury, D. (1968). Manipulating the amount of information obtained from a person giving directions. Unpublished Honors Thesis, Department of Social Relations, Harvard University.

Krauss, R. M. & Fussell, S.R. (1991). Perspective-taking in communication: Representations of others' knowledge in reference.Social Cognition, 9, 2-24.

Kraut, R. E., Lewis, S. H. & Swezey, L. W. (1982). Listener responsiveness and the coordination of conversation. Journal of Personality and Social Psychology, 43, 718-731.

Linde, C., & Labov, W. (1975). Spatial networks as a site for the study of language and thought. Language, 51, 924-939.

Pomerantz, A. M. (1984). Agreeing and disagreeing with assessment: Some features of preferred/dispreferred turn shapes. In J. M. Atkinson & J. Heritage (Eds.), Structure of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press.

Searle, J. (1969). Speech Acts. Cambridge: University Press.

Schober, M. F., & Brennan, S. E. (2003). Processes of interactive spoken discourse: The role of the partner. In A. C. Graesser, M. A. Gernsbacher, & S. R. Goldman (Eds.), Handbook of discourse processes (pp. 123-164). Hillsdale, NJ: Lawrence Erlbaum.

Wilkes-Gibbs, D., & Clark, H. H. (1992). Coordinating beliefs in conversation. Journal of Memory and Cognition, 31, 183-194.