One year ago Raymond F. Person, Jr. published the first monograph on Jonah (and to my knowledge on any biblical book) in which the text is approached primarily from the perspective of Conversation Analysis, namely In Conversation with Jonah: Conversation Analysis, Literary Criticism, and the Book of Jonah.1 The volume already warranted a formal discussion at the 1996 meeting of Socio-Linguistics Group of the Society of Biblical Literature. The Journal of Hebrew Scriptures (JHS) is glad to provide its readers a (revised) version of the papers that were presented at that meeting.
Ehud Ben Zvi (Editor, JHS)
Cynthia L. Miller, North Carolina State University
Conversation is a critical component of biblical prose, but also poses significant challenges for interpretation. How, exactly, is speech represented in the biblical text and, secondly, how does the structure of dialogue contribute to the meaning of a passage? This second issue is explored by Ray Person in his recent volume, In Conversation with Jonah: Conversation Analysis, Literary Criticism, and the Book of Jonah.2 As the subtitle suggests, both linguistics and literary criticism (specifically, reader- response criticism) are employed. In these few remarks, I examine Person’s application of linguistic methodology to conversation in Jonah.
Person’s analysis of conversation in Jonah derives primarily from the discipline of linguistics known as Conversation Analysis. Conversation Analysis grew out of ethnomethodology, a field of sociology. Ethnomethodologists are centrally interested in the ways in which members of a society themselves produce and interpret social interaction.3 Conversation Analysis examines oral dialogue to determine the social and pragmatic principles whereby speakers (and hearers) negotiate, structure, and interpret conversation.
The primary contribution of Conversation Analysis to the discipline of sociolinguistics has been the observation that conversation is fundamentally structured in terms of contiguous, alternating turns of talk, known as “adjacency pairs.”4 The first part of an adjacency pair produces the expectation of a relevant and acceptable rejoinder in the second part. For example, a question by one speaker in the first part prompts an answer by the second speaker in the second part. The pragmatic function or purposive intention of the speaker in the first part (e.g., to elicit information by way of a question), then, constrains the kind of pragmatic function of the second part (in this case, to supply the information requested), provided that the second speaker wishes to be cooperative. Second parts differ, however, in the degree to which the second part is socially preferred or dispreferred. The cultural ranking of second parts as preferred or dispreferred is known as preference organization.
The other two linguistic theories that Person employs relate to the determination of what is a preferred response. Grice’s Cooperative Principle states that cooperative interlocutors are guided by four maxims: (1) quantity: make your contribution as informative as is required, but not more informative than required; (2) quality: do not say what you believe to be false, do not make unsubstantiated claims; (3) be relevant; (4) manner: avoid obscurity and ambiguity, be brief, and be orderly (p. 100). A further theory, The Given-New Contract, suggests that cooperative conversation requires attention by the speaker to the status of information vis-a-vis the hearer (pp. 101–2). Some knowledge is shared by both speaker and hearer (that which is Given) and need not be specified explicitly; that which is New to the hearer must be specified explicitly. A speaker, then, who wishes to communicate with a hearer must determine what can be assumed or shared knowledge (the Given) and that which must be communicated expressly to the hearer (the New).
Person’s explanation of these three theories is clear and informed, a rarity in our discipline where biblicists often cite linguistic theories without understanding them. Two quibbles. Although Conversation Analysts argue that preferred responses often have certain structural linguistic correlations (e.g., preferred responses tend to be short and straightforward, dispreferred responses tend to be long and complex), it is potentially misleading to insist that preference organization refers to linguistic structures..5 Because the adjacency pair is not a linguistic structure, the specific linguistic structures employed within adjacency pairs do not relate directly (and indeed, cannot relate directly) to preferential structure. This is abundantly evident in the following two examples of adjacency pairs:
1.a Can you read music?
1.b. Yes. [preferred]
1.b'. Uh, yeah. Well, actually, kind of. [dispreferred]
In this adjacency pair, the first speaker elicits information concerning the second speaker’s musical abilities, and the response of the second speaker in (b) is preferred in that it provides the elicited information. To this we can compare a completely analogous adjacency pair with the same linguistic structure, but different pragmatics:
2.a. Can you tell me the time?
2.b. Yes. [dispreferred]
2.b'. Uh, yeah. Well, let’s see. It’s, uh, 10:30. [preferred]
The response in (2b) is undoubtedly a dispreferred response; the first speaker does not elicit information concerning the second individual’s timetelling abilities, but wants the individual to convey what the time is. Note, however, that the dispreferred response in (2b) is linguistically identical to the preferred response in (1b); conversely, the preferred response in (2b') is linguistically identical to the dispreferred response in (1b'). The examples in (1) and (2) illustrate that linguistic structure is not at work in the preferential organization of adjacency pairs; rather, pragmatics and social convention determine whether or not a second is preferred or dispreferred.
One other point concerning terminology. Person describes characters who produce preferred parts as having “an excellent control of language in their predisposition to participate in adjacency pairs with preferred seconds.”6 This sentence is potentially misleading, in that the production of preferred seconds has nothing to do with the speaker’s “linguistic competence” or “excellent control” of language, but rather with a speaker’s interactive or social skills. Certainly, speakers may produce preferred or dispreferred seconds regardless of whether their rhetoric is eloquent or clumsy; conversely, both grammatically felicitous and grammatically ill-formed utterances may count as preferred or dispreferred seconds.7
More problematic is Person’s reliance upon English for many of the guiding presuppositions concerning adjacency pairs and preference organization. In English, a dispreferred second part is generally more elaborate and complex as in (3)8
3.a. Uh, if you’d care to come and visit a little while this morning, I’ll give you a cup of coffee.
3.b. [delay] Well [preface], that’s awfully sweet of you [appreciation]. . . . . I don’t think I can make it this morning [mitigated declination] . . . . . . . . [delay] Uhm, I’m runing an ad in the paper and . . . and uh, I have . . . . . . . to stay near the phone [account].
Dispreferred seconds in English are commonly introduced with a delay, a preface, and an account for the dispreferred response. But these characteristics of dispreferred parts must be demonstrated for responses in Biblical Hebrew.9 Although some of these features are present sometimes in Biblical Hebrew, the non-colloquial nature of speech in the biblical text precludes the use of dysfluencies, delays, and repairs which often preface dispreferred responses in spoken English. Because the central notion of adjacency pairs relates to what is socially acceptable, cooperative, and relevant, we must demonstrate precisely what social and linguistic speech conventions are represented in the biblical text.
ויאמר אלהים אל־יונה ההיטב חרה־לך על־הקיקיון 4a.
ויאמר היטב חרה־לי עד־מות 4b.
Person identifies this adjacency pair as that of question answer. He says that the question inquires about the legitimacy of Jonah’s request to die.10 The question then means “Is your request legitimate?” He then interprets the answer as preferred, even though he interprets Jonah’s answer as angry because he provides “no new information” and “the words he uses are simply a reworking of the question into a declarative sentence with his previous request in 4.3 implied.”11 But Jonah’s answer cannot be interpreted as one given in anger solely on the basis of the rewording of God’s question into a statement. Indeed, in Biblical Hebrew the only way to answer a yes-no question in the affirmative is to echo all or part of the question.12 Contrastively, in English when the reply to a yes-no question involves a verbatim rewording into declarative mode (as opposed to a simple “yes”), the response is indeed angry, or at least highly marked:
5.a. Can you read music?
5.b. I can read music.
Part of the difficulty in Jonah 4:9 involves whether the question is rhetorical or not. If rhetorical, then it means “You have no right to be angry” and Jonah’s response “I have a right to be angry to death” is then dispreferred. If the question is a real question, then it means “Are you thoroughly angry?” or “Are you legitimately angry?” and the response is preferred. On this difficult question, Person is apparently undecided.13
Another illustration of this problem involves the claim that the adjacency pair blame denial is preferred over blame admission.14 Conversation Analysts claim that denial is the preferred response to blame because utterances of denial are usually short and without hedges, delays, prefaces etc.; that is, denials usually have linguistic structures similar to those of preferred responses after other kinds of first parts. Responses that contain an admission of guilt are often delayed and give an account of the reason for failure; that is, admission of guilt have linguistic structures similar to those of dispreferred responses after other first parts. This argument, in spite of the protests of Conversation Analysts to the contrary, is patently circular.15 To determine the kind of response which is preferred or dispreferred after blame, we must examine what is socially preferable within a culture after accusations of blame and then examine the linguistic strategies employed by speakers to negotiate a preferred response. The linguistic correlates of a preferred response in one type of adjacency pair will not necessarily be reflected in the preferred response of another adjacency pair. Here Person’s analysis precisely follows the claim of Conversation Analysts that denial is the preferred response after blame. But how do we know that this is the case in the pragmatic context of biblical culture? Various examples of adjacency pairs containing accusations of blame would suggest that a rapid, straightforward acknowledgement of guilt is often preferred over denial:
ויאמר דוד אל־נתן חטאתי ליהוה 6.
(David said to Nathan, “I have sinned against YHWH.” [2 Sam 12:13])
It is this problem, the problem of knowing whether or not the analysis of the adjacency pair is justified, that proves so difficult. It should be noted, however, that this difficulty is largely a result of the linguistic the ories employed. All three linguistic theories Conversation Analysis, Grice’s Cooperative Principle, and the Given-New Contract depend crucially upon the cultural norms of the language community and upon the pragmatic context of the specific speech situation. That is, each theory is potentially ethno- or linguacentric. To take only one example, Grice’s principle that one should be relevant and to the point, would have strikingly different linguistic realizations in Tokyo and New York much to the frustration and consternation of Japanese and American speakers attempting cross-cultural communication. Because Conversation Analysis was deve loped by native speakers of the language under analysis and had as its focus oral (not written or literary) la nguage, it was relatively straightforward for analysts of English to determine the adjacency pairs of English oral conversation based upon their understanding of the purposive intention of the speakers involved. But in an ancient literary text, we can only know if our analysis of adjacency pairs is correct if we examine a large nu mber of examples and if we are certain that we understand not just the norms for cooperative conversation but the literary conventions for representing conversation in narrative. The book of Jonah does not afford such an opportunity for extensive analysis. In 48 verses, there are only 20 instances of direct speech and, by Person’s analysis, one instance of indirect speech, and about 25 adjacency pairs.
Person is correct to emphasize that represented speech is constructed dialogue, that is, it is not a verbatim report of any speech event.16 He sees constructed dialogue in Biblical Hebrew as represented by direct speech, indirect speech, and reported consequences. Let us examine these three categories. Direct speech and indirect speech are left unexplained and unexamined, as though they are universal privatives which are equivalent in Hebrew and English, though in fact they differ syntactically. In Biblical Hebrew, direct speech may take one of three forms depending upon the syntactic construction used to introduce the quotation; indirect speech may appear in one of four forms, depending upon the degree to which the quotation is an independent clause (introduced with כי or zero), a dependent infinitival clause, a nominal phrase, or only the speech event is mentioned. In Jonah, direct speech introduced with a single verb appears 12 times; the verb is invariably אמר. In this respect, Jonah patterns like much of biblical narrative. Direct speech introduced with two speech verbs appears five times, usually representing either a prayer/cry to YHWH or a pronouncement to Nineveh. Direct speech introduced with לאמר appears three times twice in the common prophetic formula:
לאמר ויהי דבר יהוה אל יונה בנ אמתי
In one occasion (3:7), multiple speech verbs and לאמר appear together to introduce speech; this construction is relatively rare in biblical narrative.17 It is used here to emphasize both the phatic nature of the speech event and its non- prototypically dialogic nature as being produced iteratively, undoubtedly by individuals other than those who were at its source (the king and his nobles). Indirect speech with an infinitival clause appears once (4:8 וישאל את־נפשו למות ‘he asked his soul to die’);18 this is the only instance of indirect speech that Person recognizes. Somewhat more frequently (seven times) a form of indirect speech appears as diegetic summary (in which only the speech event is mentioned) (e.g., 1:5 ויזעקו איש אל־אלהיו ‘Each man cried out to his god’). Precisely what was said is left unspecified. In one instance, however, the content of the speech may be retrieved from the following narrative (e.g., 2:11 יאמר יהוה לדג ויקא את־יונה אל־היבשה ‘YHWH spoke to the fish [indirect speech as diegetic summary] and it vomited Jonah onto the dry land’).19 Clearly, God commanded the fish to vomit out Jonah, but this information is only relayed by the narrator and is not represented as speech.20 Another interesting example appears in 3:6: ויגע הדבר אל מלך נינוה ‘The word/matter reached the king of Nineveh’.21 Here speech is represented only as diegetic summary; there is no mention of the speaker, but only that the information reached the king; the actions of his response are narrated in the remainder of the verse.
Person rightly observes that biblical narrative often represents speech events with narrative in ways like this, rather than with overt representations of speech (whether direct or indirect), a situation that he describes as “reported consequences.” But Person’s category of “reported consequences” contains more than one kind of material. We can see this most clearly by comparing the following examples:
11.a. (Person: 1st part) a. ויגע הדבר אל־מלך נינוה
(YHWH appointed a big fish to swallow Jonah)
11.b. (Person: 2d part) b.ויהי יונה במעי הדג שלשה ימים ושלשה לילות
(Then Jonah was in the belly of the fish three days and three nights [(Jonah 2:1])
12.a. (Person: 1st part) a. וימן יהוה־אלהים קיקיון
(YHWH God appointed a qiqayon plant.)
12.b. ויעל מעל ליונה להיות צל על־ראשו להציל לו מרעתו
(So it grew up over Jonah to become a shade over his head and to save him from his distress.)
12.c. (Person: 2d part). וישמח יונה על־הקיקיון שמחה גדולה (Then Jonah rejoiced greatly over the qiqayon plant. [Jonah 4:6])
13.a. (Person: 1st part) וימן האלהים תולעת בעלות השחר למחרת (God appointed a worm at dawn on the next day.)
13.b. (Person: 2d part) ותך את־הקיקיון
) So it struck the qiqayon plant (
(and it withered [Jonah 4:7]).
In 2:1, 4:6, and 4:7, the first parts of each adjacency pair (11a, 12a, 13a) have the verb מנה ‘to appoint’, understood by Person as a speech verb. Note, however, the differing relationship of the reported consequences in the second part to the first, according to Person’s analysis. In 2:1, the first part (11a) is analyzed as a request YHWH appoints a fish to swallow Jonah. The second part (11b) describes not the response of the fish to YHWH’s request, but the affect of the fish’s response upon Jonah. Properly speaking, the response to the first part the fish swallowing Jonah is only implied. In 4:6, the response of the qiqayon plant to YHWH’s appointing should be understood as (12b), not (12c) as suggested by Person. Jonah’s emotional response to the plant’s action as described in (12c) is a response to the second part (12b), but should not itself be described as the second part. In 4:7, however, the second part in (13b) as analyzed by Person is in fact the response of the worm to YHWH’s request in (13a). The problem with these examples is that there is no clear criterion by which to decide when non-speech actions “function as” or “count as” a rejoinder to a first part which represents speech. I suggest that a second part depicting “reported consequences” properly occurs only when the addressee of the first part is depicted as engaged in actions which respond to the first part’s speech event (as in 13a-b); that is, the “participants” of both halves of an adjacency pair should be identical.22
Having established three ways in which constructed dialogue may be represented in narrative direct speech, indirect speech, and reported consequences Person makes the following claims concerning their distribution within adjacency pairs: (1) first parts tend to be in direct speech;23 (2) preferred second parts tend to be in reported consequences;24 (3) dispreferred second parts have two opposing tendencies to be represented as longer and more developed instances of direct speech, or to be represented as reported consequences or indirect speech.25 The first claim (that first parts tend to be in direct speech) is corroborated by data from the Torah and Former Prophets; indirect speech may appear in the first part, but it is significantly less common. The other two claims can only be evaluated by considering them within the context of adjacency pairs whose first part contains a request or command most adjacency pairs in Jonah (18 total) fall into this type. After a request or command, preferred parts tend to be represented by the narrator’s notice that the requested action was carried out; dispreferred parts tend to be represented as a full form of direct speech or by a reduced form of indirect speech or reported consequences. The second and third claims are correct when the first part is a request, though the second claim must be amended in one respect. A preferred response may also be completely unrepresented in the narrative.26 That is, when a command is given in biblical narrative, the assumption of the narrative is that the command is carried out and the fact of its execution need not be stated explicitly.27 After other types of first parts, however, Person’s second and third claims concerning second parts cannot be substantiated absolutely. Indeed, other types of first parts (e.g., greetings, questions, assessments, blame) do not lend themselves to second parts represented with reported consequences.
In extending the present analysis to a wider corpus, there are three ways in which the description could be improved. First, one of the major criticisms of Conversation Analysis is that the categorization of adjacency pairs tends to be intuitive, a kind of “folk metalinguistic.” Person’s categorization of first parts into five types (request, offer, assessment, question, blame) will almost certainly need to be expanded to reflect a more comprehensive set of metaterms. Such a categorization could also benefit from a correlation of the type of speech event to the metapragmatic (or speech) verb which introduces the quotation; for example, those instances in which a command is introduced with צוה (‘to command’). Second, many (if not most) utterances play more than one pragmatic role within a dialogue, as Person himself recognizes.28 The question of multiple functions for adjacency pairs is a critical theoretical problem with which Conversational Analysis must grapple. And, finally, it is important to note that speech may be represented in biblical narrative when it is not part of a conversation and thus stands outside of the structure of an adjacency pair. A prime example is the representation of thought as a kind of internal speech (as is the case in 4:8). In these cases, the speech being represented is non-interactive and cannot be understood as structured solely by dialogic concerns.
Nonetheless, the analysis of conversation in Jonah which Person has presented has been successful in a number of respects. He has highlighted the importance of dialogue and its structural significance for biblical narrative. He has shown how the first part of an adjacency pair produces the expectation of a relevant and acceptable response; where a response is absent, a gap is opened up which the reader must attempt to fill. He has shown that the linear order of narrative may be displaced chronologically in order to place the second part of an adjacency pair immediately after the first part. This observation provides an explanation for the placement of the account of the Ninevites’ repentance in 3:5–10 immediately after Jonah’s oracle, even though chronologically it probably followed the account of Jonah’s anger in chapter 4. Finally, he has illustrated the complexities of applying linguistic theories which were developed for the analysis of oral communication to an ancient text. Indeed, the book of Jonah, though small in size, exhibits extraordinary literary complexity in its representation of conversation. Person’s application of conversation analysis to the difficulties of Jonah have illumined the difficulties and pointed toward a resolution.
Kenneth M. Craig, Jr., Lees-McRae College
The title to Raymond Person’s book, In Conversation with Jonah, reminds us, perhaps, of the small prophetic book’s enormous popularity. Among the Book of the Twelve, Jonah stands out as one of the most popular, for the public at large and in the scholarly guild, among conversationalists and researchers as well. Person contributes to the ongoing colloquies through an explication of conversation analysis. His discussions on the topic of conversation analysis are linked to such matters as plot, characterization, atmosphere, and tone. One also finds a solid verse-by-verse commentary. In drawing heavily from conversation analysis, Person sheds light on narrative art as he discusses the reading process. A number of issues are raised in this book that in some way clarify or expand issues I touched on in my book on Jonah.29 The major similarity in our work, it seems to me, is in the area of the reading process and characterization through dialogue, and I have learned much from his study. Conversation analysis, the focus in Person’s book, is applied here to the book of Jonah for the first time.
Some of the research builds upon the work of Wolfgang Iser. Three principles further develop how indeterminacies are realized and filled in the interaction between text and reader. These three principles are the Cooperative Principle, the Given-New Contract, and Preference Organization.30 Since the discussion of conversation analysis in In Conversation with Jonah often hinges on “adjacency pairs,” it may be helpful to begin with the following discussion:
“Adjacency pairs are sequences of two moves (verbal or non-verbal) that are:
i. adjacent or containing an insertion sequence (for example, a clarifying question between question and answer)
ii. produced by different individuals
iii. ordered as a first part and a second part
iv. typed, so that a particular first part has a range of second parts: those which are linguistically preferred and those which are linguistically dispreferred
There are two types of adjacency pairs: (1) those with preferred second parts and (2) those with dispreferred second parts. Preferred seconds are generally brief utterances given without delay and are unmitigated.”31
According to Person, adjacency pairs are the constituent parts of the Jonah narrative. They intertwine around each other in their complexity. Many narrative elements are developed through the use of adjacency pairs, and they, especially the dispreferred seconds of delay, preface, declination, and account, shed light on the Jonah narrative, its structure, and the process of reading. One significant point is that the delay of “dispreferred seconds” can occur over an extended period of time; they, therefore, have an impact on the process of reading. Another point is that these “seconds” sometimes shed light by way of contrast. For example, Jonah’s predisposition for “dispreferred seconds” (his protest in 1:3, for example) contrasts with all the other characters who are predisposed for “preferred seconds.”
My first observation, after being introduced to the topic of adjacency pairs, is that the parameters are quite broad, an observation that Person does in fact acknowledge: “in fact, very few of the events (words and actions) cannot be understood as part of an adjacency pair.”32 The distinction between preferred and disprefferred thus helps narrow the focus. But I am intrigued by the preferring, dispreferring distinction. If the preferred seconds are “generally brief” and the dispreferred seconds are “generally lengthy,”33 is there some gray area when one might be classified either way? Are there exceptions, and might different readers classify them differently? Is it possible that a lengthy second could be preferred? Or under what conditions might a brief second be disprefferred?
I wish to mention several strengths that I discern in this discussion of conversation analysis and the book of Jonah. I also wish to raise a few questions for consideration. What are the strengths of this approach? And what new light is shed on the Jonah narrative? First, though not his primary purpose in the book, the discussion may spark thoughts on narrative analysis and re-contextualization elsewhere in the Bible. Second, in a section on reading the Jonah narrative, eleven presuppositions are stated that indicate the knowledge an implied reader has of the Jonah narrative. I know of no other scholar who has delineated these presuppositions, and it seems to me that such an identification is an important first step for interpretation. These presuppositions of the narrative are as follows:
1. The Hebrew text can be understood by both its author and its actual readers.
2. Nineveh is the capital of Israel’s enemies, the Assyrians.
3. Israel’s enemies, including the Ninevites, are considered to be the Lord’s enemies.
4. Nineveh is east of Israel.
5. Tarshish is west of Israel.
6. The sea is a place of chaos.
7. The Lord’s chosen people, the Israelites, are considered to be more righteous than pagans, like the sailors and the Ninevites.
8. Sackcloth, ashes, and fasting are signs of repentance.
9. When delivered from peril, one utters a prayer of thanksgiving to the Lord.
10. Jonah is a prophet and therefore should be loyal and obedient to the Lord.
11. The Jonah narrative purports to be a prophetic book.34
By isolating such presuppositions we begin to see how the author, the text, and the intended readers share in their knowledge of the narrative world. When such an awareness is missing—and it often is in the discussions of the prophetic book—interpretation suffers.
A third strength I perceive is Person’s observation about the similarities and differences between narrated dialogue and everyday conversation. In some ways, these two types of speech resemble each other but some significant differences remain, primarily because an author can assert control through the narrator in narrative conversations. Two significant observations accrue: one in the area of constructed dialogue and the other on the matter of omitted dialogue. Fourth, Person helps us see that the omission of details in narrative often proves as significant as inclusion. He writes, “constructed dialogue that is nevertheless not reported by the narrator may prove as significant as what is reported, and in some cases even more significant.”35
Fifth, the discussion on “omitted dialogue”36 is another provocative feature of the book. Jonah’s response to the Lord’s commission at the beginning of the story has often been noted by commentators. Person calls attention to “other examples of omitted dialogue” that have generally been overlooked. For example, in chapter one, Jonah does not answer all the interrogatives the sailors pose in their pile-up of questions. Another example of “omitted dialogue” is the suppressed content of the message that Jonah is to cry out to the Ninevites (1:2; 3:2).
I now raise a few questions:
1. I agree that the narrator omits the content of the message spoken by the Lord to Jonah. However, is Jonah’s dialogue, just after the Lord’s command to go to Nineveh and cry against her, “omitted?” If “omit” refers to taking out what was originally there, it seems to me that the narrative does not necessarily suggest that the narrator has omitted Jonah’s verbal response to God. This much we know for sure: Jonah elects to respond with his feet. He chooses to respond in a rather dramatic way, and isn’t it quite possible that the speech event occurring just after the commission in 1:2 (i.e., the speech event based upon Jonah’s protestation in 4:2: “was this not what I said when I was back in my country”?) is not dialogue, but interior monologue. Much hinges in this story on the narrator’s withholding of information here, and quite possibly on Jonah’s restraint as well. Person overlooks the ambiguity between thought and speech of the biblical “said.” The prophet’s surprising reticence in chapters 1 and 3—he’s no Jeremiah, Ezekiel, or Amos in this regard—and the fact that the Bible does sometimes render interior monologue without the “in his/her heart” formula allow for the distinct possibility that no dialogue (two speakers) has been omitted in 1:2–3. To illustrate this point: 1 Sam 20:26 literally reads: “and Saul did not say anything on that day for he said . . .” What the Hebrew conveys is the idea that Saul says to himself, he thinks. Previously, in 1 Sam 18:8, Saul’s response to the women is reported without the formula “said in his heart,” but the contextual and grammatical clues argue for interior monologue: “he said, they [not: “you”] have given to David ten thousands and to me they [not: “you”] have given thousands, and what more to him than the kingdom?” The surprise of the unfolding Jonah plot may depend in large measure on suppressed interior monologue, but not on “omitted dialogue.”
2. The focus in Person’s book is on conversation analysis. Is there a subtle difference between “conversation” and “dialogue?” Or, since part of the discussion in In Conversation with Jonah concerns omitted “dialogue,” are the two words used interchangeably? If there is a subtle difference, might a discussion of dialogue (“through words”) call our attention to the fact that quoted speech in Jonah is mediated discourse? Person succeeds in calling our attention to the “constructed” speech of characters, but his discussion, to my mind, does not carry the implications of constructed dialogue to its conclusion. When characters are quoted in the Jonah narrative, their words are filtered to us through the narrator’s own verbal, sociocultural, thematic, aesthetic, and persuasive design. The words spoken, by their very form, entail indirections because the speakers communicate with us through someone else. Characters speak, but their words are of another’s devising. Their discourse is speech within speech, a perspectival montage. Any speech event in the book of Jonah entails two levels of communication by two communicators with two perspectives, and one, the narrator’s, is always hidden and possibly ironic. Since overt words often hide multiple perspectives, the layers of quotation in Jonah may even manifest themselves as triple talk since a character can always speak something other than his or her own mind. We do well to develop an eye for dissimilarity—and Person seems to be aware of this fact based on his discussions of constructed speech—because speech, thought, and the perspectives of framer (narrator) and framed (characters) are often at odds.
Such a brief and vague oracle probably does not adequately reflect what the Lord wanted to communicate to Nineveh through his prophet Jonah. Therefore, Jonah’s oracle as given in the narrative should be viewed as constructed dialogue that does not fully represent the Lord’s word to Nineveh (through Jonah), which Jonah is said to have given ‘just as the Lord had commanded’ (3.3). In its given form, the oracle lacks an indictment and a specific means of punishment, both of which were probably part of the Lord’s full message to Nineveh.37
In the discussion that develops, Person, following the lead of Jack Sasson,38 is certainly correct about the ambiguity that surrounds the word “overturn”: will the overturning be destructive (i.e., Nineveh’s annihilation) or constructive (i.e., Nineveh’s repentance)? I am, however, intrigued by Person’s conclusion, “Jonah’s oracle as given in the narrative should be viewed as constructed dialogue that does not [my emphasis] fully represent the Lord’s word to Nineveh (through Jonah), which Jonah is said to have given ‘just as the Lord had commanded’ (3.3).” The phrase of 3:3 “as the Lord commanded” covers Jonah’s “arising” and “going” [v. 3]. The notion that it covers the proclamation in 3:4b (p. 125) needs to be explained.
According to Person, though the narrator reports that Jonah delivers the oracle “just as the Lord had commanded” (3:3), the prophet in fact “does not fully represent the Lord’s word to Nineveh.” If his interpretation is correct, we have here an example of unreliable narration. I sense from Person’s reading, a hesitation—the word “probably” is used twice—on insisting the point that Jonah changes the original message, or, more precisely, I perceive some caution in the tacit assertion that the narrator’s report is questionable. But why would the narrator, in a book so ideologically charged, narrate unreliably? Of course, any reproduction of another’s speech will involve change. At the theoretical level, even so-called verbatim reproduction of speech, which can sometimes be verified in the Bible, is not the same as original speech because the context into which the subsequent speech event occurs is different from the original context. The frame affects the inset. But this fact should not cause us to abandon the reliable/unreliable distinction in narration. In more recent works such as Henry James’s The Turn of the Screw (where the narrator hallucinates), John Updike’s Rabbit, Run (where the narrator is sometimes drunk), or The Catcher in the Rye (where the narrator is an adolescent), we have reason to argue that narration is unreliable, or at least suspect. But if the narrator in the Jonah story is deceiving us in the report that Jonah delivers the oracle “just as the Lord had commanded,” as Person avers, then the question about what motivates the narrator to be less than factual should be addressed. On the matter of the inaccuracy or accuracy of the oracle as delivered by Jonah, I wonder if we should call it an unverifiable quotation (since the content of God’s conversation is revealed only indirectly) and leave it at that.
Person has advanced our understanding of the book about Israel’s most recalcitrant prophet. The questions I have raised about In Conversation with Jonah are posed with the purpose of keeping the conversation alive.
Raymond F. Person, Jr., Ohio Northern University
It appears that Miller and Craig agree with me concerning the most important contribution of In Conversation with Jonah. That is, using conversation analysis I have provided some empirical evidence of how readers create meaning as they interact with literary texts. This methodological and theoretical contribution is illustrated by a new reading of the book Jonah, emphasizing the conversations that occur between the various characters in the narrative. As both Miller and Craig suggest, the book of Jonah lends itself to illustrate my method because of its “extraordinary literary complexity in its representation of conversation” (Miller). The most obvious reason why Jonah lends itself to illustrate this method well is the displacement of Jonah’s speech refusing the Lord’s initial call. In 1:3 the narrator does not include Jonah’s refusal, but simply describes Jonah’s flight. This causes readers to ask the question why. This question is not answered until 4:2 when Jonah says, “Is this not what I said while I was still in my homeland? Therefore, I hastened to flee to Tarshish for I knew that you are a gracious and compassionate God.” That is, the narrator manipulates the representation of the dialogue between the Lord and Jonah, withholding Jonah’s half from readers, only to give readers enough information later to reconstruct what Jonah said. I provide numerous examples from the history of interpretation demonstrating the effectiveness of the narrator’s manipulation of divergent readers spanning time and space.
Now I turn to Miller’s and Craig’s comments individually. I think that Miller’s criticism of my work can be summarized by the following quote, “in an ancient literary text, we can only know if our analysis of adjacency pairs is correct if we examine a large number of examples and if we are certain that we understand not just the norms for cooperative conversation but the literary conventions for representing conversation in narrative” (Miller). The quote suggests that Miller has two conditions for an adequate understanding of adjacency pairs and their meaning in a literary text: (1) examining a large number of adjacency pairs within a literary corpus and (2) understanding the “literary conventions for representing conversation in narrative.” This too was my goal in In Conversation with Jonah and continues to be my goal in my works in progress.
Concerning the first condition, Miller has pointed out some inadequacies in my work that resulted from my not having examined a large enough number of adjacency pairs in Biblical Hebrew. (My study of adjacency pairs in Biblical Hebrew was larger than simply the book of Jonah; however, I did not undertake a systematic study of Biblical Hebrew.) In her own work, Miller has capably analysed reported speech in Genesis-2 Kings. Her dissertation is a goldmine of information and I look forward to reading the recently published version.39 Because of her careful systematic work, I am not surprized that she has found some errors at that level of my study.
Unfortunately Miller was writing her dissertation at about the same time I was writing this work and, therefore, I discovered her dissertation after I had completed the first drafts of my work. I trust that my future works will contain fewer of these types of errors because of what I continue to learn from Miller’s work.
Concerning the second condition of understanding literary conventions of representing conversation in narrative, Miller has also found some inadequacies in my argument (for example, my three claims concerning the representation of constructed dialogue in narrative). However, I believe that Miller’s critique sometimes does not take seriously the changes that occur when we move from ordinary conversation to narrative. Two examples will suffice.
First, Miller suggests that 4:8 “And he asked that he might die, ‘It is better that I die than live’ ” may be the representation of thought and, as such, “cannot be understood as structured solely by dialogic concerns” (Miller, 11). Although unspoken thoughts are necessarily outside the reach of conversation analysts, they are easily within reach of readers when thoughts are presented as constructed dialogue within literary works. In these narrative settings, the unspoken thoughts of the characters become a part of the narrative world that readers are privy to and, therefore, readers interpret these thoughts as part of the dialogue in the narrative. Whether or not Jonah ever vocalized “It is better that I die than live,” we the readers know that Jonah is initiating the same adjacency pair he began in 4:3, “Now, O Lord, take my life from me, for it is better that I die than live.” That is, he is requesting his death from the Lord in response to what he understands as the Lord’s destruction of his shelter through the worm and the fierce east wind. And Jonah’s words, whether they were vocalized or simply thought, produce a response. 4:9: “Then God said to Jonah, ‘Is it good for you to be so angry on account of the qiqayon plant?’ ”
This leads to my second example. The Lord’s utterance here functions on multiple levels. On the one hand, it prefaces the Lord’s refusal of Jonah’s death-request. On the other hand, it is a rhetorical question. As a rhetorical question, it functions as both a question and an assessment. That is, it is as if the Lord said, “Is your life really so miserable that you want to die? I think not.” Jonah then answers the rhetorical question by saying “It is good that I am angry enough to die.” His answer also functions at multiple levels. On the one hand, it is the preferred second of an expected answer—that is, the Lord would have expected such an answer from Jonah who is portrayed as being just this petty throughout the narrative. On the other hand, Jonah’s answer is the dispreferred second of a disagreement with an assessment—that is, the Lord asserts in the rhetorical question that Jonah’s misery is not bad enough for a death-wish, but Jonah disagrees.
Miller’s criticism of my interpretation here is that I incorrectly identify Jonah’s response as dispreferred because it is the preferred second of an expected affirmative answer. This misunderstanding, according to Miller, is based on my interpretation of Jonah’s echo of the question as necessarily angry in tone in biblical Hebrew. I must admit that I may have initially missed some elements of the grammatical argument Miller is making here, but my argument for Jonah’s response as dispreferred depended much more on the literary context than she perceives. This is clearest in my discussion of this verse on p. 130, which she does not refer to in her remarks. In other words, the narrative context influences readers to interpret the Lord’s rhetorical question consistently with the Lord’s compassionate nature and Jonah’s answer consistently with his pettiness. Therefore, even though the adjacency pairs function on multiple levels here, the literary context and the satirical tone of the narrative suggest that once again Jonah is producing a dispreferred response, disagreement with the Lord’s assessment that is necessarily linked to the Lord’s compassion.
In his review, Craig raises three questions. First, although he agrees with me that the Jonah narrator manipulates readers by sometimes omitting the characters’ dialogue, he doubts that Jonah vocalized his refusal when the Lord first commissioned him to go to Nineveh. I quote, “isn’t it quite possible that the speech event occurring just after the commission in 1:2 (i.e., the speech event based upon Jonah’s protestation is 4:2: ‘was this not what I said when I was back in my country?’) is not dialogue, but interior monologue?” (Craig). My response to his critique has two parts. First, his reading is clearly a minority position here. Many readers not only understand that Jonah said something when he refused, but they reconstruct what he said between 1:2 and 1:3. For example, Jack Sasson reconstructs Jonah’s verbal refusal as follows: “Please, Lord, I realize that you are a gracious and compassionate God, very patient and abundantly benevolent, who will relent from bringing disaster.”40 Second, the majority of readers understand that the narrator omitted Jonah’s verbal refusal based on 4:2, where Jonah says, “Is this not what I *said* while I was still in my homeland?” That is, even though the narrator omits Jonah’s verbal response, the narrator provides readers with enough information in 4:2 to reconstruct what Jonah “said” as Sasson again illustrates.
Craig’s second question concerns the narrator’s representation of the characters’ speech. Craig maintains that the characters’ words are “filtered” by the narrator’s own perspective as they are reported to readers. Here Craig and I have no difference of opinion, for this is exactly what I mean by the phrase “constructed dialogue.”
Craig’s third point concerns my interpretation of Jonah’s brief oracle. It is my contention that the narrator has once again filtered Jonah’s words to us for the purpose of satirizing him. Jonah’s oracle is given *as if* it were only five Hebrew words; however, the context suggests that the oracle as spoken by Jonah to the Ninevites was more fully developed. The five words given include neither an indictment nor a specific means of punishment as is commonly found in judgement oracles. The Ninevite king’s words “Each person must turn from his evil ways and from the violence which is in their hands” (3:8) suggest that the indictment dealt with Nineveh’s transgression, much like the indictments made by Amos against the foreign nations. Given the emphasis on the Lord’s compassion in the narrative, it seems that the Lord would be more specific with the Ninevites than these five Hebrew words suggest. Therefore, if Jonah delivered the message “just as the Lord had commanded” (3:3), then Jonah’s oracle probably was more elaborate than the five Hebrew words given by the narrator.
Craig raises the question of whether or not this is “unreliable narration.” To some degree, all narration is unreliable—that is, narration *never* reproduces exactly what occurred in the narrative world; rather, dialogue is constructed according to the narrator’s perspective. But the Jonah narrator tends to be even more manipulative of readers in his omission of and representation of dialogue. Craig asks, “why would the narrator, in a book so ideologically charged, narrate unreliably?” Because the narrator presents his ideology through a satirical portrayal of Jonah. Even though the Jonah narrator is more manipulative of readers than other biblical narrators, clues are nevertheless planted in the text to alert readers that this manipulation is present. Therefore, I am not sure I would say that the narrator is more unreliable in the Jonah narrative than in other biblical narratives, but maybe the Jonah narrator simply expects more of us as readers.
 Raymond F. Pearson Jr., In Conversation with Jonah: Conversation Analysis, Literary Criticism, and the Book of Jonah (JSOTS 220; Sheffield: Sheffield Academic Press, 1996).
 See Stephen C. Levinson, Pragmatics (Cambridge Textbooks in Linguistics; Cambridge: Cambridge University Press, 1983), p. 295.
 See Person, In Conversation, p. 16.
 See Person, In Conversation, p. 67.
 See also Person's statement that the "readers bring acquired [my emphasis] knowledge of the function and characteristics of adjacency pairs" to the reading process (In Conversation, p. 108). For "acquired knowledge," read instead "implicit knowledge" or "inherent knowledge" since a reader presumably knows something about the function and characteristics of adjacency pairs implicitly or inherently by virtue of being a proficient speaker of the language in question and need not acquire specific metalinguistic knowledge of their function and characteristics.
 Example (3) is drawn from Person's example and discussion. See In Conversation, p. 18.
 Person attempts to mitigate the effect of this criticism by noting that "This is not necessarily the case in all cultures" (In Conversation, p. 19). But the fact remains that the analysis of English as presented in chapter 1 is accepted as the norm for analysis of the Hebrew text.
 See Person, In Conversation, p. 46.
 See Person, In Conversation, p. 78.
See Edward L. Greenstein,
 The response is listed as preferred in In Conversation, p. 80. Elsewhere, however, the question is called rhetorical and is interpreted as receiving a dispreferred second, or possibly a preferred second. The reason given for interpreting the response as dispreferred is that "Jonah's answer is basically a grammatical change of the question into a declarative sentence" (In Conversation, p. 61 n. 20).
 See Person, In Conversation, p. 19-20.
 See, for example, Levinson, Pragmatics, p. 333.
 See Person, In Conversation, p.23.
 See, for example, 1 Sam 26:6.
 Person says that "Jonah once again requests that the Lord take his life" (In Conversation, p. 45). But in Hebrew, Jonah's request is directed inwardly, not to the deity (so also Jack M. Sasson, Jonah: A New Translation with Introduction, Commentary, and Interpretation [AB 24B; New York: Doubleday, 1990], pp.305- 6; Phyllis Trible, Rhetorical Criticism: Context, Method, and the Book of Jonah [Guides to Biblical Scholarship; Minneapolis: Fortress, 1994], p.214).
 This type of information provides the perlocutionary effect of the diegetic summary, see also, for example, Gen 42:25.
 Curiously, this instance of speech does not appear to be analyzed by Person.
 See Cynthia L. Miller, The Representation of Speech in Biblical Hebrew Narrative: A Linguistic Analysis (HSM 55; Atlanta: Scholars Press), pp. 257-62.
 See Person, In Conversation, pp. 26,77.
 Person, In Conversation, pp.27-28, 77-78.
 The failure of the narrator to specify a response constitutes a "zero response"; see Miller, Representation of Speech, pp. 260-61.
 As examples, see Gen 18:6-8; 26:11; Exod 1:22; 5:13; Josh 1:10-11; 3:3-4. These examples suggest that the failure of the narrator to specify that Jonah called out to his god (1:6) does not necessarily mean that he refused to do so (as suggested in In Conversation, p. 49).
 Person, In Conversation, p. 77 n. 53, p. 113.
 Kenneth M. Craig, A Poetics of Jonah: Art in the Service of Ideology (Columbia, SC: University of South Carolina Press, 1993).
 Person, In Conversation, p.99.
 Person, In Conversation, pp. 16-17.
 Person, In Conversation, p. 54.
 Person, In Conversation, p. 17.
 Person, In Conversation, pp. 109-110.
 Person, In Conversation, p. 25.
 Person, In Conversation, pp. 72-76.
 Person, In Conversation, p. 125.
 Sasson, Jonah, pp. 234-37.
 Miller, Representation of Speech.
 Sasson, Jonah, p. 329.