Herbert H. Clark

When people communicate, they take for granted certain shared knowledge, beliefs, and assumptions - their common ground. And in conversation they update their current common ground with each spoken or gestural contribution. But what is common ground, and how is it established and used? Catherine Marshall and I proposed answers to these questions in an extensive chapter called "Definite reference and mutual knowledge." My colleagues and I followed these up in later papers and chapters, including the work on "grounding."
Publications on Common ground

To communicate is to "make common," but when two people talk, how do they make sure that what they have said has truly been made common? The answer several colleagues and I have proposed is this: People try to establish, as they go along, the mutual belief that they have understood each other well enough for current purposes. This process is called grounding. The main proposals are found in papers by Deanna Wilkes-Gibbs, Ed Schaefer, Susan Brennan, and myself, and it has been followed up in more recent work.
Publications on Grounding

Referring to things has traditionally been something speakers do on their own, with noun phrases such as "Napoleon," or "the man you just met," or "me." That, indeed, had been my initial assumption. But it changed when Deanna Wilkes-Gibbs and I analyzed the references to Tangram figures in a task in which one person tried to get the other to arrange the figures in a particular order. There it became clear that referring was a process jointly carried out by speakers and addressees. That paper was called "Referring as a collaborative process," and it was followed up in many other papers both from my lab and from elsewhere.
Publications on Referring as a collaborative process

People divide their utterances, roughly, into given information, which is inferable from common ground, plus new information, which is not inferable from common ground. Listeners, in turn, assume that speakers do this, and that shapes the inferences they draw from what is said. Susan A. Haviland and I published our first work on this issue in several papers, in particular "What's new" and "Comprehension and the given-new contract." Another important paper in this series is "Bridging."
Publications on Given and new information, and bridging inferences

When I tell you "My car isn't red", I take for granted that at this moment you might assume my car is red, and I'm saying that such an assumption is incorrect. I take nothing like that for granted when I tell you "My car is white". Bill Chase and I did a long series of experiments on the process of understanding affirmations and denials, which we published in four papers. I summarized this and other unpublished research in a 1976 monograph called Semantics and Comprehension in. This work I saw as closely related to the next category.
Publications on Negation and denial

The words long and short aren't simply opposites. Long doubles as the name of the dimension, as in the expressions "length" and "two feet long," whereas short does not. Long is the positive end of the dimension, and short the negative end. Beginning with "Linguistic processes in deductive reasoning," I showed how implicitly negative adjectives like short take longer to understand, and that has profound effects on how people reason.
Publications on Adjectives and reasoning

We have no trouble creating and understanding novel words, such as the verb porch in "Did the newspaper boy porch the newspaper?" But how do we do that? Eve Clark and I took up this question in "When nouns surface as verbs," and she, I, and our colleagues followed up these ideas in later research. Dan Morrow and I looked at interpretations of the verb "approach," and I took up color terms in "Words, the world, and their possibilities."
Publications on Novel words and novel meanings

I have long been interested in how space, time, and the body are represented in language. This led to research on prepositions such as above and below, spatial adjectives such as long and short, and people's understanding of space and time in relation to the body in a chapter called "Space, time, semantics, and the child." It also led to work with Hiram Brownell and Bill Banks on the link between perception of verticality and terms like up and down, higher and lower. Later, I took up demonstrative reference and pointing, and placement as a communicative act.
Publications on Language, perception, and the body

Using language is a joint activity, and it is itself used to coordinate other, more basic activities. My interest in joint actions grew out of our work on common ground, grounding, and collaboration, but it took on a life of its own. The book Using Language, Cambridge University Press, 1996, is entirely about using language as a joint activity. But other work has bolstered or expanded on those ideas in several directions.
Publications on Joint actions

As John Austin put it, people "do things with words." They make statements, ask questions, apologize, greet others, and make commitments. But how do people make clear which speech acts they are performing? My colleagues and I have examined this issue from several angles. We carried out experiments on how people make so-called indirect requests, and how do they do so politely. We studied how people perform so-called ostensible invitations, such as "Let's do lunch sometime." We analyzed how people ask questions and influence answers, and how people ask questions and answer them without using words. Finally, Richard Gerrig and I studied irony, which led to an extensive paper on quotations as demonstrations.
Publications on Speech acts: Requests, questions, irony, quotation

The way we understand speakers depends on our status as listeners. If we are addressees ("you"), we can count on the speaker having designed utterances for us to understand. Not so if we are bystanders. Speakers have no obligations to bystanders and may even want to conceal what they say from them. So the process of understanding changes depending on whether we are: (a) a participants vs. non-participants; (b) among participants, addressees vs. a side-participants; and (c) among non-participants, overhearers vs. eavesdroppers. Thom Carlson and I began this line of research with an extensive paper called "Hearers and speech acts," and we and others have followed it up in later papers.
Publications on Addressees, overhearers, audience design

Spontaneous speech tends to riddled with disfluencies. Although these have been studied for a long time, my first research, with Vicki Smith, was "On the course of answering questions." We examined the causes of disfluencies when people were trying to answer difficult questions such as "In what sport is the Stanley Cup awarded?" Jean Fox Tree and I then studied when and why speakers pronounced the as "thee," and when and why speakers produce "uh" and "um" as signals of delay. Tom Wasow and I also looked at repeated words, as in "I uh I wouldn't be surprised at that." I incorporated these and other ideas in several theoretical accounts of disfluencies.
Publications on Disfluencies

Scientists often carry out experiments in which they sample not only participants (such as college students or providers on Mechanical Turk), but also items (words, sentences, pictures, etc.). The problem is that when these scientists computed inferential statistics, they assumed that they had sampled the participants but not the items. This paper demonstrates that doing that leads to statistical claims that are not just false but often wildly exaggerated. The paper offers several solutions for this problem in analyses of variance.
Publications on Statistics

Standard theories of quotations assume that quotations are linguistic objects - mentions or reproductions of linguistic expressions from another source. The prototype for these theories is reported speech or the mention of words. An alternative theory, proposed in Clark & Gerrig (1990), is that quotations are demonstrations - that is, performed depictions of linguistic or non-linguistic events. The papers by Clark & Gerrig and Wade & Clark provide systematic evidence for the demonstration theory and against the reproduction theories. Depicting itself is a general method of communicating, as described in two papers by Clark. And as Clark & Fischer argue, social robots are depictions of social agents.
Publications on Quotations, demonstrations, depictions

People are said to "communicate" with virtual agents on the telephone and with robots, but how do they do that? This paper outlines an answer that is related to the work on imagination and depictions.
Publications on Virtual partners