Abstract
Students are generally not that keen on cognitive psychology, but many believe that a special circle of hell is reserved for psycholinguistics. Students listen to lecturers on eye movements and lexical decision experiments and briefly wonder, “What is the point?,” before eagerly looking forward to their next class on serial killers.
I have some sympathy with these students because placing so much emphasis from the beginning on detailed findings from apparently artificial techniques puts the cart before the horse while simultaneously not being able to see the woods for the trees. Students need to be reminded why the study of language is important before they are force-fed details of how words might be processed. Under the weight of experimental details, they soon lose sight of what language does and therefore why its study is important. Teachers often fail to relate experiments to the wider context, which is that language is essential for communication, plays a fundamental role in human cognition, is pretty much unique to humans, and at least one language, and often more than one language, is acquired by human children without any apparent effort. So, it was a great relief to come across a book called “Understanding dialogue” by Martin Pickering and Simon Garrod (P&G hereafter). Dialogue, a conversation between two or more people, is self-evidently central to language as communication; indeed, it’s mostly the point of language. The authors should be congratulated on making this point clear.
What more do psychologists need to understand about dialogue, particularly after the work of Herb Clark and the ethnomethodologists in the seventies? I’d say we still don’t fully understand:
What people choose to talk about.
How language and context are represented and manipulated by the brain.
How the use of language depends on the social context.
How speakers make choices about selecting particular syntactic and lexical forms to express their thoughts.
To what extent are these choices tailored by the speaker to a particular listener and a particular situation.
Individual differences in eye gaze when speaking and listening.
How these ideas extend from dyads to larger numbers of people.
Would anyone looking for answers to these questions find them in “Understanding dialogue”? Yes, they will find answers to some of them and a significant contribution to answering all of them.
Although this book is new, the research and theory are not. The book is a summary and expansion of what the authors consider to be their most important work over the last decade or so. According to P&G, successful dialogue depends on the processes of alignment, synchrony, simulation, and prediction.
Alignment is defined in “two individuals are aligned at a given level of representation to the extent that they have the same representation as each other.” This definition really warrants a discussion longer than the book, and, after reading it, I’m still a little unclear as to what it means other than the extent to which people “see” the world in the same way. Alignment has short-term and long-term aspects, working from the very local aspects of a sentence or utterance, to the dialogue as a whole, although I don’t see why this division should be dichotomous. Alignment refers to both linguistic processes and the model of the situation, although, surely, the distinction is blurred. I feel that P&G like many other cognitive psychologists are too keen on presenting dichotomies where no simple dichotomy actually exists. Effective communication necessitates speakers talking about the same thing. Alignment is an emergent, what happens when a conversation runs efficiently, managed by micro-processes. P&G talk about alignment as a process, that is, a thing in itself rather than a result of other processes.
Two individuals are in synchrony when they construct and modify representations in time with each other. As P&G note, in practice, dialogue is sequential. Synchrony has many aspects including turn-taking and talking about what is in conversational focus. But, again, is it really a process in itself, or something that emerges with the micromanagement of conversation?
Simulation occurs when a person translates the perception of the performance of an action into an action-based representation: you represent the actions of others in the way that you would carry them out yourself, the most famous example being analysis-by-synthesis models of speech perception. Simulation is important for prediction: P&G argue that prediction, particularly prediction-by-simulation, plays a major role in dialogue and language processing in general. But it’s a big assumption to say that language and dialogue are just like any other physical action, and not one I agree with. Lifting a piano up the stairs and shaking hands obviously involve prediction and synchrony. The prediction is an essential component to completing the action successfully. It is a different matter with holding a conversation. What is the advantage to predicting what the other person is going to say? It takes time and effort, it is not essential, and you might be wrong.
The authors, however, present a superb body of empirical work. In addition, clearly the authors believe they have proposed a grand theory of understanding dialogue, but while it is an excellent description, I am less convinced it is a theory. Does the theory work in terms of making novel predictions? Does it work in terms of explanatory power, transcending levels? Like many accounts of language, P&G’s theory is, at its own level of description, a theory of language-use couched at the level of language use. They use a pseudo-representational language for talking about the way in which people think about objects. For example, they label objects and people discussed in the discourse in a way such that a particular object becomes “LAMP19-token-m”, and I found that these descriptions obscured rather than clarified. The pseudo-representational notation and the way they are used in diagrams give the impression of an account more scientific and more explicit than it actually is.
I do not think the book gives sufficient credit is given to the work of Paul Grice (e.g., Grice, 1975). Much of P&G expands on Grice’s work on his four maxims of conversational pragmatics. They reinvent the wheel on “saying just enough,” and there is very little discussion of relevance. Sperber and Wilson’s (1986) work isn’t mentioned at all, yet relevance is surely central to how we ensure that we take part in dialogue while ensuring that our contributions are timely, relevant, and at the most appropriate level of specification. Perhaps I’m just getting old, but I am increasingly finding that psychologists are dismissing philosophers—at their peril.
These reservations and observations notwithstanding, this book is very good, providing a snapshot of research into one of the most important areas of psycholinguistics. What though is the market? When I was 19 or so, I was amazed with Clark and Clark’s (1977) “Psychology and language: An introduction to psycholinguistics.” Of course those were simpler days, but it made language research look fun, and focused on things that were clearly important, such as conversation and language use. I don’t think my 19-year-old self would be similarly inspired by “Understanding dialogue.” Perhaps the 19-year-old me wasn’t their target, but the level of description is too advanced for a typical lay reader. I doubt if it is general enough for an undergraduate audience. On the contrary, it will be a key resource for an advanced undergraduate option or a postgraduate starting in the area of psychological models of dialogue.
