Experimental Pragmatics (Palgrave Studies in Pragmatics, Languages and Cognition)

Rating:
Summary: Important confluence linguistics, psychology
Review: Excerpt: The book is divided into 3 parts devoted, respectively, to pioneering approaches (Chapters 2-6), to current issues in experimental pragmatics (Chapters 7-11), and to the special case of scalar implicatures (Chapters 12-15). Although this volume aims to develop and give a name to a budding field of inquiry, the chapters in Part I are devoted to researchers who have been working in this area all along.
Chapter 2, by Herb Clark and Adrian Bangerter, provides both a historical and a contemporary perspective on reference, which is the ubiquitous activity involved in picking out an object for an addressee. Consider the utterance Put the small coffee cup over there. One would have to pick out the cup (presumably from among other candidate objects) and know where over there is (presumably from a gesture). Their chapter describes how reference was initially viewed as autonomous and addressee-blind before it came to be viewed as an activity that requires the coordination of both speaker and addressee. Among the features of referring highlighted are: (a) the multiple methods of directing an addressee's attention to individual objects; and (b) speaker-addressee pacts to arrive at a reference (i.e., to agree to certain provisional names). The coordination involved in referring is extensive, Clark and Bangerter argue, leading them to conclude that it is far from being an autonomous act. In fact, it requires more than mere coordination, it is an act that requires the full participation of both initiator and addressee. The chapter highlights how armchair reflection, field observations and careful experimentation have combined to lead to a more profound understanding of this fundamental communicative act. The chapter also provides an opportunity to appreciate Clark's well-known contributions to discourse analysis (the Given-New contract, common ground) in the context of pragmatic theory-making.
For more than 20 years, Raymond W. Gibbs, Jr., has embodied the aim of this book, by specifically testing linguistic-pragmatic theories using experimental psychological methods. In Chapter 3, Gibbs describes how his experiments have constrained theories with respect to four areas that are at the heart of linguistic-pragmatics: making and understanding promises, under-standing definite descriptions, making and interpreting indirect speech acts, and the distinction between what is said and what is meant. In each case, he has - like most accomplished experimentalists - come up with one or more clever designs that, in the end, either elucidate a given theory (e.g., the short-circuited nature of indirect requests) or force one to rethink a theory's claims (e.g., Searle's speech act theory with respect to promises). The aim of Gibbs's chapter is to convince experimentalists of the value of linguistic-pragmatic theories and to convince linguists of the value of experimentation.
Metaphor is a classic pragmatic form whose understanding has been greatly advanced by psycholinguistic investigations. As Sam Glucksberg shows in Chapter 4, metaphor comprehension in psycholinguistics was initially viewed through a Gricean lens, in which the literal interpretation of a metaphor is given priority. According to Grice (or Searle), a metaphor renders an utterance 'defective' and prompts one to look for another meaning. In his chapter, Glucksberg argues that this standard pragmatic model persisted in the literature because its literal-first hypothesis resonates with an approachthat assumes that both semantics and syntax are primary while pragmatics is secondary, an assumption that is common in psycholinguistic circles. Through his and his colleagues' pioneering work on metaphor, Glucksberg demonstrates how metaphorical interpretations of sentences such as Some jobs are jails are carried out as automatically as other linguistic processes. He extends his analysis to other related phenomena (e.g., showing how novel features emerge in conceptual combinations like peeled apples) in order to show just how automatic pragmatic processes are in comprehension tasks. He concludes by suggesting that experimentation is needed to determine the correct division of labour between linguistic decoding and pragmatic inferencing, a central issue in current pragmatic theory. The pragmatic process, as shown by Sam Glucksberg, does not merit its 'stepchild' status; pragmatics is so automatic that it is arguably a module. In Chapter 5, Guy Politzer - who was often a lone voice underlining the importance of linguistic-pragmatics to the field of reasoning - provides a pragmatic analysis of both classic and modern reasoning tasks along with experimental results that stress the importance of the way individual premises, conclusions and task information in general are interpreted. For a notable example, consider Piaget's famous class-inclusion problem, in which children are shown a picture of five daisies and three tulips and then asked, 'Are there more daisies or more flowers?' After presenting a 'microanalysis' of the way the task's demands are interpreted, Politzer shows that young children (5-year-olds) fail to answer correctly (to say flowers) because they interpret 'flowers' to mean flowers-that-are-not-daisies. He also shows how a short series of disambiguating questions prompts even the youngest children to demonstrate their class-inclusion skills. Such microanalyses can be applied equally to many of Kahneman and Tversky's tasks (e.g., the Linda problem and the Engineer-Lawyer problem), Wason's tasks (the 2-4-6 problem and the Selection Task), as well as to individual terms like conditionals and quantifiers. The implications for this approach are clear: one cannot do reasoning work without linguistic-pragmatics.
Chapter 6 by Tony Sanford and Linda Moxey reviews their previous work on the psychological processing of quantifier understanding and demonstrates how experimental approaches can inform linguistic-pragmatics. They begin by pointing out that not all quantifiers are alike. A large set of 'non-standard' quantifiers, such as few, many and most, convey much more than a rough notion of quantity or proportions; they have communicative functions as well. For example, polarity plays a determinative role in quantifier interpretation. A negative quantifier like few and a positive quantifier like a few have quite different effects on the interpretation of sentences. Compare few versus a few of the MPs attended the meeting. Few is more likely than A few to place the focus on the complementary set, those MPs who did not show up. Their findings show that the interpretation of quantifiers goes well beyond the semantics of these terms.
Current issues in experimental pragmatics
The chapters in this section extend both the range of topics one can investigate in Experimental Pragmatics and the techniques one can use. The chapters here cover inter alia disambiguation, metaphor and joke comprehension, promise understanding, the import of saying even-if, and the telling of time. All these topics are addressed using various experimental paradigms from neuropsychology, developmental psychology, reasoning, psycholinguistics and anthropology.
In Chapter 7, Jean-Baptiste Van der Henst and Dan Sperber review experiments that test central tenets of Relevance Theory and in particular the cognitive principle of relevance ('human cognition is geared to the maximization of relevance'), and the communicative principle ('every utterance conveys a presumption of its own relevance'). Some of these experiments draw on two standard paradigms in the psychology of reasoning: relational reasoning and the Wason Selection Task. Others investigate the behaviour of people asked the time by a stranger in public places. All involve manipulating separately the two factors of relevance, effect and effort. These experiments illustrate how a pragmatic theory that is precise enough to have testable consequences can put previous experimental research in a novel perspective and can suggest new experimental paradigms.
Orna Peleg, Rachel Giora and Ofer Fein give an account of the role of the context in accessing the appropriate meaning of ambiguous terms in sentence comprehension in Chapter 8. They argue against: (a) a modular view which assumes that lexical access to all meanings of a word are automatic and encapsulated only to be refined by an independent non-modular system; and against (b) a direct access view which relies largely on just the context to arrive at a word's intended meaning. Rather, they propose the graded salience hypothesis, which assumes that: (a) more salient meanings are accessed faster from the start; and that (b) context also affects comprehension on-line. Their chapter presents four experiments whose results lend strong support to their claims.
In Chapter 9, Seana Coulson provides a review of the way Evoked Response Potentials' (ERP) methods can be applied to language comprehension, with a focus on what this technique has to offer pragmatics. The chapter is instructive in that it describes ERP's various dependent variables (P300, N400, P600 etc.) and the aspects of comprehension with which these measures are associated. Coulson cites studies of pragmatic import - for example, on joke comprehension and metaphor integration - including many that come from Coulson herself. She works from a model that predicts that processing difficulty is related to the extent to which comprehension requires the participant to align and integrate conceptual structure across domains. She goes on to suggest ways in which ERP experiments could be exploited to investigate other linguistic-pragmatic issues, such as prosody and thedistinction between explicatures and implicatures. Overall, her chapter shows very clearly how imaging can be exploited and indicates what one should expect from this technique in the future.
In Chapter 10, Josie Bernicot and Virginie Laval focus on children between the ages of 3 and 10 and their developing understanding of promises, based on the theoretical framework of Speech Act Theory (Austin 1962; Searle, 1969, 1979; Searle and Vanderveken, 1985). The authors summarize a programme of research that has been investigating promise comprehension among children from the point of view that language is a communication system and that language competence is the acquisition and use of that system. What counts as a promise? Here the authors present two experiments investigating the extent to which interlocutors' intentions (listener's wishes about the accomplishment of an action) and textual characteristics of utterances (verb tense) play a role in understanding that a promise was made.
In Chapter 11, Simon J. Handley and Aidan Feeney develop a psycho-logical account of the way in which people reason with even-if, working in a mental models' framework (Johnson-Laird, 1983). According to the mental model approach, many errors of reasoning arise because people represent only one or a few of all the models of a given set of premises and leave the other models implicit. They then draw their conclusions on the sole basis of the explicitly represented models. Handley and Feeney compare two possible ways in which this partial representation of problems might arise. In one, all models are represented before being pared down by extra-logical, namely pragmatic, factors; in the other, which the authors advocate, initial representations are limited to one model while pragmatic considerations add new models. They present two experiments based on inference making from even-if premises that lend support to their account. They discuss the implications of their work for experimental pragmatics in general.
The case of scalar implicatures
The chapters in the third section of the book focus on one pragmatic phenomenon, scalar implicature, which is at the heart of ongoing debates in pragmatic theory. As described earlier, there are two main accounts of these inferences. One assumes that such implicatures are automatically associated with the use of a weak term (as exemplified by Levinson, 2000) and the other assumes that the implicature is drawn out effortfully (as exemplified by Relevance Theory). In these chapters, four authors (or groups of authors) present experimental findings that lend support either to Relevance Theory or to some form of the default view.
In Chapter 12, Anne Bezuidenhout and Robin Morris first describe how they operationalized the two theoretical accounts into testable pragmatic-processing models. This is less obvious than it might seem because it is hard to do justice to the rich and detailed accounts that have been offered by these rival theories on the topic of scalar implicatures. They then report on two eye-movement experiments that test predictions generated from the models as participants read a series of sentence-pairs such as Some books had colour pictures. In fact all of them did, which is why the teachers liked them. One can determine whether Some in the first sentence readily prompts Not all by investigating potential slowdowns and look-backs when processing the second sentence. They argue that the weight of the evidence favours the Underspecification Account (which is the one inspired by Relevance Theory); however, they argue that their Default Model (the one inspired by a neo-Gricean account) could be modified to accommodate their results.
In Chapter 13, Gennaro Chierchia, Maria Theresa Guasti, Andrea Gualmini, Luisa Meroni, Stephen Crain and Francesca Foppolo present a novel account of implicatures based on the Semantic Core Model, which challenges a way of interpreting Grice's proposal that has become dominant in the field. According to the dominant view, one first retrieves the semantics of a whole root sentence and then processes the implicatures associated with it (in a strictly modular way). The Semantic Core Model proposes, instead, that semantic and pragmatic processing take place in tandem. Implicatures are factored in recursively, in parallel with truth conditions. They go on to present experimental evidence from adults and children that support this new model. One of the novel findings from this work demonstrates how particular grammatical contexts predict the non-existence of scalar implicatures.
In Chapter 14, Ira A. Noveck reviews the two rival accounts and the processing predictions they engender, before summarizing his laboratory's findings from experiments investigating those logical terms (i.e., might, some, or and and) that could be interpreted either minimally (i.e., with just their linguistically encoded meanings) or as pragmatically enriched. His developmental studies show how children are less likely than adults to pursue pragmatic inferences, leading to a robust experimental effect in which children actually appear more logical than adults. Follow-ups show how task-demands, and not just age, can affect the production of pragmatic inference making, pointing to the important role of context in these paradigms. The adult studies, which include an ERP investigation, primarily explore the time-course of scalar inferences. Whereas participants' pragmatic treatments of underinformative statements (e.g., the time taken to respond False to Some cows are mammals) are very time consuming, True responses are not. Furthermore, time pressure encourages True responses. Noveck presents his findings as support for Relevance Theory.
In Chapter 15, Anne Reboul presents a novel task, which she calls Koenig's puzzle, as promising ground for testing between the two rival theories. Imagine that after being handed a glass of wine, a speaker says Better red wine than no white wine. The puzzle consists in determining the speaker's wine preference and inferring what she was actually given. While referring to the two sides of the debate as localists and globalists (for the Default and Relevance accounts, respectively), Reboul describes Koenig's puzzle in detail and proposes a solution to it. Reboul then explains why such sentences may be used to test between the two accounts. Finally, her paper reports two experiments whose results show how implicatures are actually involved in the puzzle. Her results are presented as support for global over local theories for this specific pragmatic phenomenon.
How to approach the book
The chapters are representative of what we are calling Experimental Pragmatics. Each summarizes previous experimental work or presents original experiments that address topics central to pragmatic theory - metaphor, quantifier interpretation, scalar inference, disambiguation, reference and promise understanding, to name a few. Many of the chapters share common themes, especially the last four, but each can be read and appreciated separately. Our intention has been to illustrate how Experimental Pragmatics may contribute to linguistics and psychology, and to the cognitive sciences in general.