The Dialogue Exchange

Speech as Structure: What a Passage of Dialogue Must Do, and How to Test It

The Craft Behind the Story (Second Edition), Reference — Chapter 3. Part I: The Scene and Its Materials.

This chapter is a companion to Chapter 1 (The Purposes of a Scene), whose domains of purpose (including its proposed relational domain) and value-shift test it inherits and extends. It is also a companion to Chapter 4 (The Descriptive Detail), whose denotation/connotation pair is the descriptive analogue of the text/subtext distinction developed here.

All illustrative passages are quoted briefly for the purposes of criticism and study; full citations appear in §14. Where a work is read in translation, the translator is named, and the chapter observes the series’ standing distinction: structural claims survive translation, texture claims do not.

Contents

On this page: Chapter Abstract · Learning Outcomes · Key Terms · How to Use This Chapter

1. Introduction: The Transcript Problem
2. Definitions and Boundaries
3. Why Dialogue Is Not Speech — the theory section
4. The Domains of Purpose — five inherited, and a proposed sixth
5. Multiplicity: The Economy of the Exchange
6. Text and Subtext: The Two Channels
7. Diagnostic Method — the ten tests
8. Failure Patterns — with their exceptions
9. Worked Analyses — eight close readings and a revision case
10. Dialogue Across Forms · Concordance: concept to novel
11. Revision Checklist
12. Studio Work — exercises, seminar questions, assessment rubric
13. Glossary
14. Sources Cited
15. Further Reading
16. Coda: Against Mechanical Application

Chapter Abstract

Dialogue is the only unit of fiction with a rigorous empirical science attached to it. Linguists have been recording, transcribing and analysing actual human conversation for fifty years, and have established with considerable precision how turn-taking works, how meaning is generated by not saying things, how refusals are marked, and how status is negotiated in talk. Almost none of that has reached creative-writing pedagogy, which continues to teach dialogue by ear and by prohibition.

This chapter takes the empirical literature seriously, and finds that it confirms the craft tradition’s intuitions while explaining them far better than the craft tradition can. It proceeds in four movements:

1. Theory — why a passage of dialogue is not a transcript, grounded in H. P. Grice’s account of conversational implicature, in the conversation-analytic findings on turn-taking and preference organisation, in Erving Goffman’s work on footing and face, and in Bakhtin’s account of the novel as the form that stages competing social languages. A historical strand traces the contingency of quotation convention itself, and an empirical strand establishes the chapter’s principal negative claim: transcribed conversation and novelistic dialogue are different objects, and literary dialogue is a stylised convention that reads as real rather than a record of the real.
2. Taxonomy — the five domains of purpose (plot, character, relational, theme, texture), each anatomised in terms of its textual carriers and its characteristic under- and over-delivery. The relational is the fifth domain proposed in Chapter 1 (The Purposes of a Scene) and adopted here because dialogue makes it load-bearing. A sixth is then proposed: the Silence domain, covering what is not said, who does not speak, and who is not in the room.
3. Method — the principle of multiplicity applied to the exchange; a full account of the two channels of text and subtext, with a taxonomy of twelve named subtext mechanisms; and a ten-test diagnostic protocol for revision.
4. Application — twelve failure patterns, each with a published counterexample; eight close readings; a revision case worked on the household example that runs through this series; and an apparatus of exercises, glossary, sources and further reading.

The chapter’s governing claim is that a dialogue exchange is a transcript with an engine under it: two or more parties pursuing incompatible wants through speech, in which the balance of information, standing or feeling is measurably different at the end than at the start. Its most consequential single contribution is an identification: what craft writers call subtext is what Grice called conversational implicature, and the mechanism the two describe is the same mechanism. A character who answers a direct question by changing the subject is flouting the maxim of relation, and the reader computes the implicature. Naming that correspondence converts a matter of taste into a matter of procedure.

A note on the examples. The close readings are drawn chiefly from the European and anglophone novel, with one short story, Hemingway’s “Hills Like White Elephants”, used because it isolates the variable more cleanly than any novel can; the Core chapter adds post-2000 fiction and screen cases, listed in §14.6. A concordance mapping each concept to the novel that best displays it appears at §10.6.

A note on quotation. Dialogue is the element of fiction most damaged by being quoted at length out of context, because its effects are cumulative: a line lands because of the eleven lines before it. Passages here are therefore short, and are frequently described rather than quoted. Where the description does the analytical work as well as the quotation would, description is preferred, and students should read the surrounding pages before accepting any judgement made here.

Learning Outcomes

By the end of this chapter, a student should be able to:

• Distinguish a dialogue exchange from adjacent units — the speech, the conversation, reported speech, the interview, the monologue — and explain why the distinctions matter in revision.
• Explain, with reference to Grice, to conversation analysis and to Goffman, why meaning in dialogue is generated as much by what is withheld and how it is withheld as by what is stated.
• Identify the six domains of purpose in any published exchange, and mark which lines discharge which.
• Analyse the relation between the two channels of text and subtext, and name which of the twelve mechanisms in §6.4 a given line is using.
• Apply the ten-test diagnostic protocol to a draft exchange and produce a prioritised revision plan.
• Diagnose the recurring failure patterns by mechanism rather than by surface feature, and articulate the conditions under which each “failure” is the correct choice.
• Evaluate the historical and cultural specificity of the chapter’s own prohibitions, in particular the twentieth-century Anglo-American preference for compressed, indirect, subtext-heavy speech.
• Assess the ethics and the craft of phonetically rendered dialect, and the uneven history of the prohibition against it.

Key Terms

Dialogue exchange · the two channels · text · subtext · conversational implicature · the Cooperative Principle · flouting · adjacency pair · preference organisation · dispreferred response · turn-taking · repair · footing · participation framework · face-work · heteroglossia · polyphony · the relational domain · the silence domain · attribution · said-bookism · literary dialect · the gap-closing move

Full definitions appear in the Glossary (§13). Terms are bolded on first substantive use.

How to Use This Chapter

A caution before beginning. This chapter contains taxonomies, tests and checklists. None of them is a formula. They are instruments for noticing, and their proper output is a question, not a verdict. A protocol that returns the answer this exchange fails test seven has done nothing until the writer asks why the book might want it to. Every test in §7 is failed by at least one of the novels analysed in §9, and in several cases the failure is the reason the exchange is famous.