COM/SOC 125 - Steven Clayman



Steven E. Clayman & John Heritage

Question design as a comparative and historical window

into president–press relations

In Markku Haakana, Minna Laakso & Jan Lindström (eds.) (2009) Talk in Interaction: Comparative Dimenstions, pp. 299-315. Helsinki: Finnish Literature Society (SKS).

Introduction

This chapter synthesizes recent research on how speakers formulate questions in a particular institutional environment, namely journalism as practiced in U.S. presidential news conferences (Clayman & Heritage 2002b; Clayman et al. 2006; Clayman et al. 2007). The research focuses on variation in question design, exploring how the act of questioning presidents has changed over time, and how it varies under different social circumstances. It is thus an exercise in applied conversation analysis in a comparative mode, where question design serves as a window into the institution of journalism and its evolving relationship to the state.

The theme of this chapter and of this volume is 'comparative analysis.' Perhaps the dominant impulse underlying this theme is that of cross-linguistic and cross-cultural comparison. What does our conversation analytical research tell us about commonalities and divergences in the ways in which speakers of different languages address common interactional contingencies? Startling similarities are beginning to emerge across very diverse sets of languages (Stivers et al. 2009), while intriguing differences inhabit languages as closely related as British and American English (Jefferson 2002).

A second level of comparison is represented by Curl and Drew (2008). This level is language and culture-internal. Curl and Drew investigate the possibility that different ways of designing requests are specific to informal conversation on the one hand, and more task-based 'institutional' contexts on the other. Ultimately, the authors argue that this possibility is not the case, and that different request forms encode differing balances between the entitlement of the requester and the contingencies that may attend the granting of the request, and that this distinction holds across contexts.

Curl’s and Drew’s conclusion for requests notwithstanding, there clearly are significant and major differences in the ways that talk is organized in different settings within a single culture. There are notable differences in the management to turn-taking in the law courts (Atkinson & Drew 1979), news interviews (Greatbatch 1988; Clayman & Heritage 2002), and classrooms (McHoul 1978; Mehan 1979), and all of these differ from ordinary conversation (Sacks, Schegloff & Jefferson 1974; Drew & Heritage 1992). These differences may scale down many levels of granularity. Twenty years ago, Heritage (1985) argued that a specific practice – formulating or summarizing prior talk – could be heard as withholding and disaffiliative in ordinary conversation, but appropriate and professional in the news interview context. Subsequently Drew (2003) broadened the range of comparisons to show how contextually specific and multifunctional these summative formulations can be.

Between these two forms of comparative analysis – cross-linguistic and language-internal – lies a profound methodological divide between "etic" and "emic" perspectives. Conversation analysis is, of course, an emic enterprise which grounds its claims about interactional practices in "the demonstrable orientation and understanding of the parties to the interaction as displayed in their consequent conduct" (Schegloff frth: 8). Cross-linguistic comparisons can, of course, arise from this form of emic analysis in separate treatments of each language. However, by their very nature, cross-linguistic comparisons cannot be directly grounded in the "demonstrable orientations" of the parties. Language- and culture-internal comparative analyses, by contrast, can be and are generally thoroughly 'emic' enterprises.

We introduce these distinctions in part to problematize them. Our study of question design within American presidential news conferences is located within a single institutional and cultural context. It is nonetheless a comparative study encompassing both historical variation on the one hand, and diverse socio-political circumstances on the other. Although data are drawn from what is ostensibly a single linguistic and cultural domain, the scope of our comparison – a nearly fifty year span of news conferences (1953–2000) – may raise doubts that we are examining a 'single' homogeneous linguistic and cultural entity. Our study indeed documents long-term changes in the style and substance of presidential questioning, as well as dramatic turning points in the language and culture of the news conference (Clayman et al. frth). It has often been observed that language undergoes a process of slow incremental change and, correspondingly that, although speakers act in their daily lives on the assumption that they are speaking the same language, imperceptibly the language changes to the point that its users are not able to understand its earlier or earliest incarnations. Given this process of change which can be very much more rapid in institutional contexts (Clayman & Heritage 2002a), together with the changes in the presidential news conference that we will document, there are grounds to question whether our study is indeed monocultural in its focus, and thereby amenable to an essentially 'emic' treatment or whether, alternatively, we are engaged in what amounts to a cross-cultural and unavoidably 'etic' comparison.

A further complication arises from the fact that our historical comparison necessitates the use of quantitative methods. The categorization of questioning practices used in this study was developed in fully 'emic' fashion from case by case analysis of questioning in a variety of journalistic contexts (Clayman & Heritage 2002a). Yet the aggregation of cases in quantification inevitably removes us from the specificities of the participants' orientations in any particular case. Moreover, aspects of context to which our work makes reference – historical change, the state of the economy etc. – are broad, and their relevance and impact are diffuse. Accordingly, it is much less easy to demonstrate their salience within singular cases of question design than it is, for example, to demonstrate the salience of the identities of journalist and president. We will return to these themes in the concluding section of this paper.

The Phenomenon

The main axis of variation examined here is between modes of questioning that are (1) polite, cautious, or deferential, as opposed to (2) vigorous, aggressive, or adversarial. To illustrate this distinction, consider how the issue of the federal budget was put before two U.S. presidents spanning almost three decades – Dwight Eisenhower and Ronald Reagan.

(1) [Eisenhower 27 Oct 1954: 9]

1 JRN: Mr. President, you spoke in a speech the other night of

2 the continued reduction of government spending and tax cuts

3 to the limit that the national security will permit.

4 Can you say anything more definite at this time about

5 the prospects of future tax cuts?

(2) [Reagan 16 June 1981: 14]

1 JRN: Mr. President, for months you said you wouldn't modify

2 your tax cut plan, and then you did. And when the

3 business community vociferously complained, you changed

4 your plan again.

5 I just wondered whether Congress and other special

6 interest groups might get the message that if they

7 yelled and screamed loud enough, you might modify

8 your tax cut plan again?

Although both questions concern budgetary matters and tax cuts, the question to Eisenhower is in various ways more deferential. Its agenda is essentially benign – indeed, it is framed as having been occasioned by Eisenhower's own previous remarks, and contains nothing that is argumentative or oppositional. It is also non-assertive – it displays minimal expectations about what type of answer would be correct or preferable, and is thus formally neutral on the subject of inquiry. Finally, it is cautiously indirect – it exerts relatively little pressure on the president to provide an answer, and even allows for the possibility ("Can you say anything…" in line 4) that he may be unable to answer.

Reagan's question, by contrast, is more aggressive. This question is similarly occasioned by the president's previous remarks (lines 1–4), but here the journalist details contradictions between the president's words and his actual deeds, contradictions that portray the president as weak and beholden to special interests. This prefatory material thus sets an agenda for the question that is fundamentally adversarial. Moreover, the adversarial preface becomes a presuppositional foundation for the question that follows (lines 4–7), which assumes that the preface is true and draws out the implications for the president's general susceptibility to pressure. And far from being neutral, the preface assertively favors a yes answer, thereby pushing the president to align with the adversarial viewpoint that the question embodies.

So these questions are indeed massively different. Moreover, they index very different stances toward the chief executive, ranging from deferential to adversarial. The study of question design here thus has ramifications beyond interaction per se. It engages what is perhaps the central issue in research on the news media in democratic societies, namely the relationship between journalism and the state.

Various models have been proposed to capture the journalism-state relationship. The model of journalism as an independent watchdog competes with other models emphasizing either subservient (e.g., Herman & Chomsky 1988) or oppositional (e.g., Patterson 1993) relations. Since journalistic conduct is circumstantially variable, such static models should give way to a more dynamic conception of the specific conditions under which journalistic vigor rises and falls. When, exactly, does the journalistic "watchdog" bark? Answering this question requires some way of tracking journalistic aggressiveness in a systematic way, and this is where the study of question design comes into play.

The Question Analysis System

Our question analysis system, grounded in prior conversation analytic research, decomposes the phenomenon of aggressive questioning into five dimensions:

1. Initiative – the extent to which questions are enterprising rather than passive in their aims

2. Directness – the extent to which questions are blunt rather than cautious in raising issues

3. Assertiveness – the extent to which questions invite a particular answer and are in that sense opinionated rather than neutral

4. Adversarialness – the extent to which questions pursue an agenda in opposition to the president or his administration

5. Accountability – the extent to which questions explicitly ask the president to justify his policies or actions

Each dimension is operationalized in terms of features of question design that serve as indicators. Below is a brief sketch of the coding system (for a fuller discussion, see Clayman and Heritage 2002b; Clayman, et al. 2006).

Initiative. Journalists exercise initiative when they (1) preface their question with statements that construct a context for the question to follow, (2) ask more than one question within a single turn at talk, or (3) ask a follow-up question. Each of these practices embodies a more enterprising posture on the part of the journalist.

Directness. Directness is measured by the absence of various practices that embody an indirect or cautious stance toward the question. Journalists are markedly indirect when they precede their questions with self-referencing frames (eg., "I wonder whether..., "I want/would like to ask...," "Can I/Could I/May I ask...") invoking their own intentions or desires before launching into the question proper. Indirectness is also manifest through the use of other-referencing frames that invoke the president's ability (eg., "Can you/Could you tell us...") or willingness ("Will you/Would you tell us...") to answer the question, and hence allow for the possibility that he may be unable or unwilling to answer. Both self- and other-referencing frames are optional choices in question design that reduce the level of coercion encoded in the question. Moreover, in some instances such frames provide an escape route – a way of not answering the question that is licensed by the design of the question itself ("Can you tell us…" -> "No, I can’t tell you"). Conversely, the absence of such frames represents a more forceful way of putting issues before the president.

Assertiveness. Assertiveness is measured only for yes/no questions, which are easier to assess. Yes/no questions can be designed to invite or favor either a yes- or no-type response in two distinct ways: (1) through a prefatory statement (ie., "Unemployment rose sharply last month. Are we in an economic downturn?"); or (2) through the linguistic form of the question itself, which can be negatively formulated and thus tilted in favor of yes (ie., "Aren't we in an economic downturn?").

Adversarialness. An oppositional stance can be encoded (1) in the preface to the question only, or (2) in the design of the question as a whole. Prefaces are coded as adversarial if they disagree with the president or are explicitly critical of the administration. It is also noted whether the subsequent question focuses on the preface (ie., "What is your response to that") and thereby treats it as debatable, as opposed to presupposing the truth of the preface. In the former case, only the preface is adversarial; in the latter case, an adversarial posture runs through the question in its entirety.

Accountability. Accountability is operationalized as questions that explicitly ask the president to explain and justify his policies. Because such questions decline to accept policy at face value, they are to some extent aggressive, although the degree of aggressiveness depends on the linguistic form of the question. Why did you-type questions invite a justification without prejudice, whereas How could you-type questions are accusatory, implying an attitude of doubt or skepticism regarding the president's capacity to adequately defend his actions.

Table 1 summarizes the question analysis system. For the dimensions involving multiple indicators, composite measures or scales were constructed with higher values corresponding to more aggressive practices or multiple practices used in combination (see Clayman et al. 2006). Are these scales, which we treated as ordinal variables, valid? A test of the assumption of proportional odds confirms (p ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download