by Vern Poythress
[Published in the Scottish Journal of Theology 32/2 (1979) 113-31. Used with permission. Also as pdf.]
1.INTRODUCTION
SOMETIMES exegetes differ from one another not so much because they have seen different data but because they are in fact looking for different types of ‘meaning’. For this reason, it is useful to classify different types of analysis of a text, and different types of ‘meaning’ resulting from such analysis.
The need for such cl a s sification has increased with the progress of biblical scholarship. Biblical exegetes and theo- logians have long had to deal with questions like ‘What is the meaning of this word of Scripture?’ and ‘What is the meaning of this verse, paragraph, section, or book?’ The intrinsic difficulties of recovering meaning from dead languages and sometimes unfamiliar cultural settings are often challenge enough. But, as increasing refinement and exactitude are sought, another kind of difficulty can arise, namely a difficulty with kinds of ‘meaning’. Is it indeed true that there is always only one meaning which is the meaning of a text? Is this the case even in poetic passages that may suggest or allude to new perspectives and comparisons without explicitly teaching them?1 Moreover, supposing that someone has arrived at ‘the meaning’ of a text, how is he to communicate this to someone else? In a commentary? In a sermon? In a form like Kittel’s Theological Dictionary of the New Testament? Or perhaps even in the form of a painting or a new social and political organisation?
Philosophers have struggled with the meaning of ‘meaning’ for some time.2 Now various books and articles are appearing [p. 114] in theological circles, exploring the nature of ‘meaning’ and warning of confusions concerning ‘the meaning’ of the biblical text. The most well known of these is James Barr’s The Semantics of Biblical Language (Oxford: Oxford University, 1961). Since 1961, scholars have explored refinements in several different directions. First, from the philosophical side have come explorations of various phenomena in religious language, including language about ‘meaning’. For example, Gerald Downing distinguishes at least ten kinds of meaning, and suggests that though exegesis is more focally concerned with three or four of these, it cannot safely ignore any of the kinds.3
Second, from the linguistic side have come structuralist systems for dealing with or highlighting certain new levels of meaning. The journals Linguistica Biblica and Semeia are the obvious representatives.
Third, a more eclectic intermediate group draws on (a) the traditions of biblical critical methods, (b) modern linguistics, and (c) literary criticism. By means of these, it has introduced greater precision in distinguishing strata of tradition and methods for exploring these strata.4
Thus the types of analysis of texts, as well as the types of ‘meaning’, are proliferating. In view of this, I want to under- take a systematic classification of types of meaning. Of course, in one sense every attempt to give the meaning of a text is different; hence an exhaustive categorisation is impossible. Moreover, different types of classification are possible. One can classify a given exegesis in terms of the passage exegeted; or in terms of the style (popular or scholarly); or in terms of the methods used (philological, theological, literary, critical, etc.); or in terms of the tradition of exegesis in which the sample falls (Roman Catholic, Anglican, Lutheran, etc.). In this article I am interested only in one kind of classification, namely classifi- cation applicable to any kind of interpretation of language at any stage of history.
[p. 115]2. SYNCHRONIC ANALYSIS AND DIACHRONIC ANALYSIS
My first distinction is between synchronic and diachronic analysis. Some of the most stimulating theological discussion today involves the factor of time. We speak of the differences between the ancient Near East and our time. We reflect on the different meanings assigned to a biblical story in the different stages of tradition. Any such analysis which compares languages or meanings at two different temporal stages I shall call diachronic analysis. By contrast, analysis which takes a cross section of languages, meanings, people, and cultures at a given point in time I call synchronic analysis. Synchronic analysis takes into account people’s memories, including memories both of events and of literature, since those memories exist and influence understanding at the given point of time. However, in principle it ignores what former or later generations may have thought about the same events and literature.
The diachronic/synchronic distinction was first introduced explicitly by Ferdinand de Saussure,5 but it was to some extent implicit in the methods of biblical scholarship before he wrote. For example, the distinction is present any time we shift from (a) talking about how St. Paul understood and used an OT passage (synchronic analysis of Paul) to (b) talking about the meanings that this passage has in all stages of tradition (dia- chronic analysis).
The importance of synchronic analysis stems from a simple but far-reaching observation: namely, that the meaning of a discourse for St. Paul or for his hearers can depend only on what Paul and his hearers know and remember about their language and culture. It cannot depend on (say) the etymology of a Greek word, unless Paul or his hearers are aware of that etymology. Neither can it depend directly on (say) the events behind the Exodus story, but only on Paul and his hearers’ understanding of those events, an understanding influenced by the interpretation of Exodus by their contemporaries.
Elementary as such an observation may be, it is necessary to appeal to it periodically. Certain subtle pressures tempt biblical scholars to brush it aside. First, there is pressure from the [p. 116] modern cultural tendency, still vigorous so many years after Saussure, to regard genetic historical explanations as the definitive type of explanation. Hence a person may stop when his job is but half done. For example, suppose I trace a certain metaphor or theological idea in Paul back to his rabbinic background, or to mystery religions, or to apocalyptic, or to the Jewish wisdom tradition, or to teachings of Jesus. Have I thereby explained this element of Paul? In a sense, I have. I have explained where the idea came from. But I have not yet explained what part it plays in Paul’s own thought and life. I have dealt with the diachronic analysis of the idea, but not the synchronic analysis. Moreover, the synchronic analysis may, in a certain sense, be far more important for the understanding of Paul’s letters. When Paul used a given idea in a letter, he might not at the time have been thinking about where that idea came from. In fact, if asked he might not even have remem- bered where or how he had picked up the idea in the first place. In such a case, a genetic, diachronic analysis might still help us to obtain a first approximation to Paul’s synchronic use. But the synchronic use, not the diachronic one, would be the definitive context for understanding what Paul was doing as he actually wrote the letter.
A second, related pressure is the pressure from the emotional satisfaction of etymologising. This is most obviously true on the level of words. Take, for example, the Greek word paraklesis, ‘exhortation, comfort’. An etymologising explanation would say that paraklesis really means ‘calling along side’, from klesis‘calling’ and para ‘along side’. Synchronically, this is just plain false. In fact, the usual glosses ‘exhortation’, ‘comfort’ come far closer than does ‘calling along side’ to describing the actual senses of paraklesis in the NT. The etymologising gloss ‘calling along side’ could only be justified if the typical speaker of Koin( Greek were himself accustomed to analysing paraklesis into its constituent parts rather than treating it as a single unit.
Though the etymologising is at this point a false step, it is easy to be swept up with an impression of its profundity and insight- fulness. As James Barr has pointed out, etymologising eliminates the conventionality of word-meanings, and replaces them by supposed rationally ‘sensible’ meanings.6 For instance, why [p. 117] should the idea of ‘comfort’ be expressed by c-o-m-f-o-r-t in English and by p-a-r-a-k-1-e-s-i-s in Greek? It seems ‘arbitrary’. Hence the yearning for an ‘explanation’ in terms of etymology. Etymologising makes the meaning of paraklesis seem ‘sensible’, at least until one realises that the same questions can be raised concerning para and klesis. Then these have to be explained in terms of their etymologies, ad infinitum.
A similar attraction can exert itself in the treatment of concepts and phrases as well as words. Phrases like ‘Son of Man’ or ‘the last Adam’ can be explained in terms of historical origin (a sort of concept etymology). To trace the designation ‘Son of Man’ back to the Similitudes of Enoch, or to Daniel 7, or to Jewish Adam speculation, is in a certain sense satisfying, because it explains one thing in terms of reducing it to or deriving it from something else. The meaning of ‘Son of Man’ is made clear and sensible. On the other hand, if the same phrase ‘Son of Man’ is dealt with synchronically, either on the lips of Jesus or in one of the four Gospels, all the idiosyncrasies and nuances of usage come into play. For example, in Mark, the phrase ‘Son of Man’, is connected with Daniel 7.13 (Mark 14.62). But ‘Son of Man’ is also connected in an extremely complex fashion to many teachings of Jesus. Not all these connexions derive from Daniel 7.13. Hence the ‘sensibleness’ of explanation disappears. Indeed, an account of the various uses of ‘Son of Man’ in Mark may appear to be simply a description and not a ‘real explanation’ at all. It is just a fact that ‘Son of Man’ is connected in such and such ways with suffering, salvation, and authority. With such bare facts we tend not to be satisfied, even as we are not satisfied just to be told thatparaklesis means ‘exhortation, comfort’. Hence the search for genetic explanation tends to be renewed. In many cases, however, the search for genetic explanation has to end with the uniqueness of personalities or historical events. Jesus said some things when he could have said others, and said them in one way when he could have said them another.
Even when we can give genetic explanations, the explanations do not in fact explain why this idea was taken up rather than another, and why the idea was left the same rather than modified (or the reverse). For example, suppose that we attempt to explain Paul’s use of apocalyptic in language in [p. 118] 1 Thess. 4.13-18 in terms of the joint sources of (a) prepauline Christian tradition as in Mark 13 and (b) OT and pseudepigraphal apocalyptic. We have still to explain why Paul thought these ideas were significant, why he chose to bring the two sources together, why he did or did not modify them in certain ways in his own thinking, and why he did or did not use the given motifs at a given point in his letters.
Another reason for caution about genetic explanation is that it tends to introduce bias towards what is historically common, in contrast to what is historically unique. Synchronic description can notice and do justice to the unique and idiosyncratic in authors and their works, as well as to that which they have in common with their language and culture. There is no reason why diachronic analysis should not also do justice to both the unique and common. When, however, diachronic analysis is largely preoccupied with the problem of explanation, in a causal or semicausal sense, the unique will get swallowed up by the common. When it is a question of what events happened, familiar events will be preferred to odd events, because there are more rational connexions to explain the familiar. Similarly, when it is a question of what persons caused a given result, groups will be preferred to individuals because group behavior is much more predictable and causally explicable than individual behavior. Hence, a bias arises for explaining texts in terms of group traditions rather than individual innovation–even though in any community the principal controllers and creative handlers of tradition are likely to be comparatively few.
A third pressure leading to the dissolution of the synchronic/ diachronic distinction is the pressure from incomplete evidence. We do not have direct access to the situations in which biblical men spoke and wrote. In the absence of full information, it is necessary to reconstruct those synchronic situations using a good deal of diachronic argument. Suppose, for example, that Daniel 7 or the Similitudes of Enoch prove to be a diachronic source for the expression ‘Son of Man’ in the Gospels. This does have a good deal to say about a historian’s reconstruction of the linguistic and theological situation in which Jesus and his hearers lived. It says something about how ‘Son of Man’ would have been understood when it was first used by Jesus or his followers. In principle, all this can be most valuable. But there [p. 119] is a temptation to slide over from the business of historical reconstruction to the more questionable operation of providing the explanation of’Son of Man’, and swallowing up synchronic meaning in diachronic analysis.
In general, both synchronic and diachronic analysis prove useful for biblical studies no less than for other studies of human verbal behavior. Nevertheless, synchronic analysis is more fundamental in two respects, (1) Diachronic analysis always involves comparison of meanings and other linguistic phenomena from several points in time. As such, it already presupposes at least a preliminary synchronic analysis of the phenomena at each such point in time. (2) A speech is uttered or a sentence or a book is written over a span of time. But this span is almost always short in comparison to the time span for change in major features of the language structure and the cultural structure.7 Hence, language structure and cultural structure are properly regarded as a synchronic fixed point in the interpretation of a speech. The diachronic change in language and culture before and after the speech are only secondarily relevant to its interpretation.
3. SYNCHRONIC ANALYSIS: ORAL AND GRAPHIC
Now let me introduce some basic distinctions in synchronic analysis. One of the most obvious distinctions is the distinction based on the medium of communication: sound or writing. Mixed media are also possible. A speech can be supplemented by a written resume, an outline, or visual illustrations. A person may also introduce exotic media like Braille or sign language. However, for the ancient world I may pretty well restrict myself to sound and writing, with the occasional addition of things like symbolic actions (Jer. 27:2, 28:10) and gestures (Luke 1.22, John 13.24).
Hence I may distinguish two forms of analysis: oral analysis if the medium is sound, and graphic analysis if the medium is writing. To see that the distinction is theologically significant, one need only recall that the methodological line between form [p. 120] criticism and source criticism is often understood to be a line between oral sources and written sources. The transmission of oral material is different from the transmission of written material, because the sound is not available for repeated inspection. But more than this, each medium calls for its own style. A well-constructed written sermon will sound ponderous and artificial if read aloud to a live congregation.
However, the oral/graphic distinction is not really a distinction of types of analysis. It is a distinction between types of material to which analysis is applied. By contrast, synchronic analysis and diachronic analysis are different types of analysis, since both can be applied to the same unit of human verbal behavior. From now on, I shall be concerned mostly with distinctions of this type, applicable to both oral and graphic material.
The biblical material, in the form that we now have it, is all graphic. Nevertheless, in the subsequent discussion I shall introduce terminology taken from the oral sphere. This is partly because sound is the primary medium of language, from which writing is derived. But the terminology will also serve as a continual reminder that the same classification can be equally employed outside the field of biblical studies.
4. SYNGHRONIC ANALYSIS: SPEAKER, DISCOURSE,
AND AUDIENCE
In a typical case of human verbal behavior we find three elements, (a) a speaker, (b) the discourse which he produces, and (c) the situation in which it is produced. To these three elements there correspond three types of analysis: speaker analysis, discourse analysis, and situational analysis. In the case of graphic material, of course, the first might be called author analysis and the second literary analysis. But for convenience I use the same terms to cover both oral and graphic material.
Speaker analysis asks, ‘What did the speaker want to say? What did he intend to say? What did he think he was saying? What did he mean?’ Discourse analysis asks, ‘What did the speaker in fact say, and what does it in fact mean?’ Situational analysis asks, ‘How did the situation alter in response to the speaker’s discourse?’ Typically, the most important part of the [p. 121] situation is the speaker’s actual audience, those who hear him. As the principal subdivision of situational analysis we have therefore audience analysis, which asks, ‘What did the audience understand by the discourse? What did they think the speaker meant? How did they react to it?’
A little reflection shows that the above three types of analysis sometimes result in different answers.
First, audience analysis is not the same as discourse analysis. An audience may misunderstand a speaker’s discourse through inattentiveness or bias. Hence the meaning that the actual audience attributes to a discourse can be distinguished, in principle, from the meaning of the discourse itself. A case in point is the letter referred to in 1 Cor. 5.9. Let us assume that the letter itself was clear enough. Then the discourse meaning was similar to the meaning of 2 Thess. 3.6. But a dullness on the part of the recipients created a different audience meaning. The audience meaning was ‘Withdraw completely from the world’.
Sometimes a difference between discourse meaning and audience meaning may be a deliberate temporary effect in order to heighten dramatic impact. Thus the utterance of Jeremiah, ‘Every jar shall be filled with wine’ (Jer. 13.12) has a discourse meaning associated with judgment. But it is deliberately put in such a way that it will be at first misunderstood by the people as a symbol of blessing.
Next, speaker analysis is not the same as either discourse analysis or audience analysis. The speaker’s competence may not have enabled him to succeed as well as he hoped. Perhaps without noticing it he introduced ambiguities or even false meanings that he did not intend. Or, conversely, he may find afterwards that his speech was fraught with far more implication than he was consciously aware of at the time.
Of course, our knowledge of the biblical material is almost always inadequate to enable us to make such a delicate differentiation. But the speaker/discourse distinction does have some implications for redaction criticism. Suppose, for example, that we agree that Luke the Evangelist used the Gospel of Mark as one of his written sources. The analysis of his alterations of Marcan material will then tell us much about his writing style and perhaps about his theological interests and purposes. All [p. 122] this is a decisive contribution to speaker analysis. However, it tells us exactly nothing, at least directly, about discourse analysis or audience analysis. Consider first the audience. It cannot be expected that they had read Mark. Even if they had, they would not have diligently compared Luke with Mark, line by line, to reconstruct Luke’s theological intention. Hence, in assessing the impact of Luke on the audience, one must studiously ignore the literary relationship between Mark and Luke. Similarly for discourse analysis. The Gospel of Luke is a literary unity of its own, and as such it certainly does not ask readers to compare it with Mark. Again, the literary dependence on Mark must be ignored.
On the other hand, the speaker analysis of Luke does have secondary effects on discourse analysis and audience analysis. Its effects are fundamentally like those of short-range diachronic analysis. We have seen that diachronic analysis helps significantly in situations of incomplete information, by enabling us to perform a more thorough historical reconstruction of linguistic and cultural setting. Now, Luke shares language and culture with his audience, and so speaker analysis of Luke informs us indirectly about his actual and potential audience. Moreover, with greater knowledge of Luke’s intentions, we can make a more intelligent guess as to the nature of his intended audience and their problems. We may notice nuances in the discourse which escaped us before.
But certain types of redactional observation still fail to have any real effect on discourse analysis or audience analysis. First, some redactional observations can be made only by an appeal to sources. For example, Conzelmann claims that Luke sees John the Baptist not as a precursor of the Messiah but as the last of the prophets.8 Such a claim, even if true, can have no real effect on discourse analysis and audience analysis. Luke has not eliminated from his source in 1.15-17 the precursor theme. Hence this, rather than Conzelmann’s idea, will be picked up by all his readers.
Second, some arrangements, structures, or motifs of a speaker are too subtle or artificial to be noticed (even subliminally) by an ordinary hearer. These subtleties will not affect discourse [p. 123] and audience analysis. Some of the allegories and allusions that Austin Farrer finds in Mark are of this kind, as he himself admits.9
All this says nothing about whether interpretations like Conzelmann’s or Farrer’s are right. It says only that, if they are right, a difference exists in those particular cases between speaker meaning on the one hand and discourse and audience meaning on the other. Nevertheless, suspicions do arise. By and large, we may assume that biblical writers wrote to be understood, not to indulge a fancy. Hence the speaker’s meaning and the discourse meaning were, for the most part, quite close if not virtually identical. A modern interpretation which creates a large gap between the two is therefore prima facie unlikely. In the absence of anything like full information about the situation of a biblical speaker and his audience, the viability of an interpretation both as the speaker’s meaning and as the discourse meaning is a useful check to interpretive arbitrariness.
In sum, it is clear that speaker analysis, discourse analysis, and audience analysis really are distinct types of analysis, not only in the questions that they ask, but at least sometimes in the answers they obtain. But, so far, I have spoken of each of these three types in a vague way. Is it possible to introduce greater precision concerning their characteristics?
Let me begin with speaker analysis. I have said that speaker analysis asks about the meaning that the speaker himself attributes to his discourse, and the intentions that he has in giving it. But now we are confronted with a certain inevitable vagueness about the words ‘meaning’ and ‘intention’. Does ‘meaning’ have to do with ‘what is going on in the speaker’s mind at the time’? But ‘what is going on in his mind’ may include feelings of hunger or sleepiness, reminiscences about events of the day, and other material only vaguely related to the subject of his discourse. Let us therefore try again. Is the ‘meaning’ what the speaker thought about the discourse? This is closer to what is wanted. But how do we find out about what he thought, except from the discourse itself? Well, speaker analysis is interested in any verbal or other information that can be obtained from the speaker at the time–or could have been obtained had the proper questions been asked. Even if the [p. 124] speaker cannot be interviewed, pieces of information can sometimes be picked up about his personality and his sources.10
To be precise, I should distinguish between the speaker’s understanding at the time of his utterance, and understanding some time later (or before). For example, some time after an author writes a discourse, he may read his own work, and so become part of the actual audience. Similar possibilities are available with oral discourse. The speaker may later remember, forget, or distort what he said earlier. As William Wimsatt has pointed out, the speaker does not necessarily stand in a privileged position when he interprets his own utterances.11 The speaker is likely to have more and better information about the context of a previous utterance. But from time to time, through lapse of memory or the pressure of self-justification, he may misinterpret his own words.
Next, consider audience analysis. For convenience, I confine audience analysis to the analysis of the meanings discerned by the actual audience, not the potential audience or the audience intended by the speaker. The difficulty with discussing a potential audience is in deciding just how far the scope of ‘potentiality’ or possibility is to extend.
Now, the actual audience consists of those people who in fact hear (or read) the discourse first-hand. Second-hand reports of the discourse, or written copies, are new discourses, which require separate synchronic analysis. The comparison of these discourses with the original will be one kind of diachronic analysis. The matter is complex in the case of written material intended to be read aloud. For example, technically speaking the actual audience of the autograph of 1 Corinthians consists of those who read the document. The illiterate among the Corinthians were obliged to rely on second-hand discourses, namely oral readings of the autograph. This is nearly the same. But of course it is not quite the same. One has only to imagine one of the trouble-makers at Corinth, reading the letter aloud with subtle innuendoes, to realise that it might not be the same.
[p. 125] So far I have treated the audience as a unit. In many cases this is accurate enough as a first approximation. The audience can be assumed to share a more or less common linguistic background and common understanding of the situation and the speaker. However, in some cases significant divisions exist within the audience. Paul must reckon with a Corinthian church split into parties (1 Cor. 1.12, 11.18). Prophets must expect a diverse reception of their message (Jer. 26.16-24, Mal. 3.16; cf. 2 Chr. 30.10-11). How does this affect the nature of audience analysis?
If I wish, I can define audience meaning as the sum total of meaning or meanings actually attributed to the discourse by the audience. Or I can confine it to the ground common to all the interpretations of individual members of the audience. Or I can do a separate audience analysis for each individual member, or for various small subgroups of the audience. All of these count as instances of audience analysis.
Finally, what about discourse analysis? It concerns the meaning of the discourse itself, not what the speaker or the audience make of it. As I have observed, the audience may misunderstand, and the speaker may not actually fulfil his intention. But if neither speaker nor audience are infallible, to whom shall we go to find out ‘the’ meaning of the discourse? Who is to decide, ultimately, what a discourse really means? Do we end in a pure relativism, where each insists on that meaning which is ‘right in his own eyes’?12
This is a real problem. It has led some forms of literary criticism virtually to identify the ‘real’ meaning either with speaker intention (‘the intentional fallacy’) or with audience reaction (‘the affective fallacy’).13 The intentional and affective fallacies at least have the advantage of tying the meaning down to some specified persons. According to the intentional fallacy, the speaker becomes the oracle by whose pronouncements meaning is judged. According to the affective fallacy, the audience becomes the oracle.
But another way can be found by asking how we normally decide when the speaker or the audience has given a mistaken [p. 126] or biased interpretation. Suppose, for example, that we were to discuss with Paul himself and his Thessalonian audience the meaning of ‘saints’ in 2 Thess. 1.10.14 Suppose further that one or all of them were to argue that it meant ‘angels’ (not ‘God’s people’). If we thought that they were mistaken, we would proceed to argue it by reference to the range of meaning of hagios (‘holy’) in Koin( Greek, by possible reference to the semitechnical meaning of the plural hagioi already familiar to Paul and the Thessalonians, and by reference to the immediate context of 2 Thess. 1. 10. The argument is fought on the grounds of the publicly ascertainable facts about (a) the Greek language, (b) the discourse context of the word hagiois, and (c)the situational context in which the second Thessalonian letter first appeared in the world.
In general terms, then, we may say that the discourse meaning is that meaning that can be arrived at by competent judges with sufficiently extensive knowledge of the linguistic context, the discourse context, and the situational context shared by the speaker and his intended audience.15 In the case of veiled or unsuccessful discourse, ‘discourse meaning’ includes also the ambiguity of meaning or lack of meaning that may be present. The linguistic context involves not only the language as a whole (e.g., Koin( Greek), but the speech peculiarities [p. 127] shared by the speaker and his intended audience (e.g., a semitechnical sense of hagioi). This discourse context is the larger context of human verbal behavioural units in which the given verbal unit occurs. For example, the word hagiois ‘saints’ occurs in the phrase tois hagiois autou ‘his saints’, which in turn occurs in the clause vs. 10a, in the sentence vss. 3-12, and in the Second Thessalonian Epistle as a whole. Finally, the situational context is that complex structure of personal preferences, evaluations, attitudes, and behavior in which the discourse occurs. It includes social, political, economic, and religious presuppositions, standards, and constraints. Paul and the Thessalonians, for example, shared a common knowledge of the OT background of the language of 2 Thes. 1.10 (Ps. 88.8 LXX, 67.36 LXX).
There are at least two attractive alternatives to the above definition of discourse meaning. First, I could have defined discourse meaning in terms of judgments made on the basis of context known to the speaker, instead of context shared by the speaker and the intended audience. But I choose to classify this as one kind of speaker analysis.
As a second alternative, discourse meaning could be defined in terms of the context shared by the speaker and the actual audience (rather than the intended audience). The difficulty with this is that the actual audience may not share much with the speaker. For example, we can imagine one of Paul’s letters becoming sidetracked en route, and later being read by someone with poor knowledge of the situation or poor knowledge of Greek. In that case the shared meaning would be very shallow.
An added benefit of my first proposed definition is that it allows me to include in the picture all the Thessalonian Christians, or all the ‘real’ recipients of a Pauline letter, rather than merely those who happen to have perused the autograph by eye.
Before leaving this subject, I should note that there is some difficulty in deciding who are ‘competent judges’ and what counts as ‘sufficiently’ extensive knowledge of the shared context. Competence is in part a matter of verbal ability and sensitivity to nuances and connections. In most ordinary cases, the decision about who is (relatively) competent is in fact not too difficult. Suppose, for example, that a discourse is judged [p. 128] by those who are (a) native speakers of the language in question, (b) friends or acquaintances of the speaker or audience, and (c)people relatively well acquainted with the situational context in which the discourse occurs. Then, unless the discourse is of a highly allusive or abstract nature, there is usually no difficulty in obtaining reasonable agreement about the basic facts concerning the meaning of small units of the discourse. Nor will there be much difficulty in deciding who is most insightful in picking up nuances and noticing less obvious connections. But greater difficulty arises in discussing the more remote implications of the discourse, or in resolving marginal ambiguities.
However, it must also be acknowledged that, at times, the ‘judges’ can end in radical disagreement about a discourse. Then they challenge each other’s competence. Particularly with regard to religious discourse, such response is a possibility, as Mark 2.7, Luke 13.17, John 9.34, 7.48-52 remind us. Mutual challenges of competence can in fact become a basis for division of a religious community.
Now consider a second condition: ‘sufficiently extensive knowledge of the shared context’. Ideally this is knowledge possessed by someone (a) who is a native speaker of the language and (b) who is not too remote from the cultural situation of the speaker and the intended audience. In practice, it is sufficient to define this knowledge as knowledge extensive enough so that further increases in knowledge do not lead to significant alterations or improvements in analysing the discourse.
In interpreting ancient texts, most of the difficulty lies right here. We are unable to obtain sufficiently complete knowledge of either language or cultural context to provide a really thorough analysis of discourse meaning. In the absence of such thorough analysis, a general, less precise analysis of meaning is still possible. For languages like Hebrew and Greek, our knowledge is sufficiently extensive so that most of the time the general area in which meaning must lie can be determined with confidence. However, when we press for details, disagreements multiply because of the fragmentary character of the evidence. In such cases, the question of ‘competence’ on the part of the judges looms much larger. The possibility of mutual challenges of competence becomes greater. Historical as well as linguistic judgments are frequently required of the discourse analyst, and [p. 129] then criteria for competence in historical reconstruction must also be included.
5. SYNCHRONIC ANALYSIS: SIZE OF UNITS
Up to this point, I have used the vague word ‘discourse’ without specifying more precisely what size unit is to be analysed. The discussion has in fact centered on the level of monologue. A monologue usually consists of a continuous, temporally connected chunk of human verbal behaviour, whose beginning and end can be roughly identified by participants native to the culture.16 It is the largest connected unit produced by a single speaker, and accordingly it is bounded on both sides by silence or the sounds of other speakers and nonspeech sounds. The corresponding graphic unit could be called a monograph (e.g., a Pauline epistle or the Gospel of Mark).
A full analysis of a monologue can only take place in connexion with analysis of smaller units: section, paragraph, sentence, clause, phrase, word, and morpheme. To do justice to these complexities, one can distinguish monologue analysis, sentence analysis, and so on. Each paragraph, sentence, etc., can be considered from the point of view of speaker analysis, discourse analysis, and audience analysis. Thus one asks, ‘What did the speakerintend to say in this paragraph?’, ‘What did this paragraph say?’, and ‘What did the audience understand from this paragraph?’
Moreover, the monologue is not the largest unit of human verbal behavior. One also finds dialogues, conversations, conversational groups, and perhaps still higher levels.17 I have chosen to stop at the monologue level because in the biblical materials we do not have available significant pieces of higher level units. We do not have, say, the Corinthians’ message to Paul or a letter from Amaziah to Amos (Amos 7.10). What we do have are Paul’sallusions to the Corinthian messages and Amos’s record of Amaziah’s speech. Dialogues and conversations are indeed inserted in biblical material, as in Amos 7 and in the [p. 130] Gospels. But these reported conversations are, by means of quotation, embedded in a larger monologue. The largest available unit is still the monograph (e.g., the Book of Amos or the Gospel in question).
6. DIACHRONIC ANALYSIS: STAGES OF TRANSMISSION
Now I turn to diachronic study of human verbal behavior. Synchronic analysis, I said, permits attention to memories of the past and precognitions of the future. But in diachronic analysis attention also extends to human verbal behavior of the past and of the future. In diachronic analysis, a given discourse can be compared with other discourses from its past or its future, irrespective of whether there is a literary connexion between the discourses. However, when there is no literary connexion, the other discourses can at best be used to build up the general linguistic, sociological, and ideological insights which will be brought to bear on any discourse. Such general insights can become part of the framework in which synchronic analysis takes place. Hence, properly speaking, diachronic analysis compares discourses between which there is (or might be) some causal connexion. It considers both the sources of a discourse (prehistory) and the later transmission of the discourse (posthistory). In the case of biblical texts, tradition-historical criticism ((berlieferungsgeschichte) is the usual name for analysis of sources of the canonical text, and text criticism the name for analysis of subsequent transmission. In keeping with the general system of terminology that I am building up, I call these source analysis and transmission analysis respectively. In contrast to the usual definition of ‘source criticism’, I include the study of both written and oral sources under ‘source analysis’. Similarly, ‘transmission analysis’ is broader than the usual definition of ‘text criticism’. In the case of biblical materials, it includes study of oral transmission, commentaries, sermons, and the like. In the case of folk tales or myths, the transmission might be predominantly or exclusively oral.
The transmission process itself can be analysed in terms of speaker-discourse-audience interaction. Suppose we have an original discourse D0, uttered by a speaker S0 to an actual audience A0. Transmission begins when a member of the actual audience A0 becomes a speaker S1 uttering another discourse D1 [p. 131] to an audience A1,. In cases of genuine ‘transmission’, D1 must be somehow modelled on or influenced by discourse D0. In the case of scribal transmission, the modern ideal is for D1 to be an exact (‘diplomatic’) copy of D0. But other kinds of transmission are possible: paraphrase, precis, expansion, oral reading of a written text, written recording of an oral discourse, etc.
The transmission can be continued when a speaker S2, belonging to audience A1, utters a discourse D2 to audience A2. At stage n + 1, some member of the actual audience An becomes a speaker Sn + 1 of a discourse Dn + 1 to an actual audience An + 1. Sn + 1 need not be part of the intended audience of speaker Sn, but only of the actual audience An. He may overhear a discourse not directed to him.
The transmission of discourse may be traced through as many stages as necessary to reach the present-day sermon or commentary with its living speaker and audience. For that matter, one may speculatively extend one’s analysis into the future, predicting the future transmission and influence of the discourse. Note that each discourse Dn can itself become the object of fresh synchronic analysis. The analysis remains diachronic analysis of D0only if Dn is compared with D0, with or without reference to the intermediate forms from D1 to Dn – 1.
Of course, the situation is further complicated if there are multiple lines of transmission or if the lines of transmission and the agents of transmission are incompletely known. Such is almost always the case in both oral and written transmission over a significant period of time. One must then content oneself, for the most part, with general observations about the tendencies and concerns of scribes, story tellers, and other hearers of tradition.
In general, transmission analysis studies the historical unfolding of all subsequent discourses Dn derived in some fashion from D0. n-th stage transmission analysis designates the analysis of the transmission from speaker Sn via discourse Dn to audience An. In cases of conflation, Dn can be designated as n-th stage only if it shows no contact with material already designated as stage n-2, and some contact with material designated n-1. Whether or not the stages can be exactly enumerated, one may distinguish speaker transmission analysis, discourse transmission analysis, and audience transmission [p. 132] analysis. The first focuses on the interests and intentions of the speakers (story tellers or scribes). The second focuses on the alterations of the discourses themselves in the course of time. The third focuses on the class of people for whom the stories are told or texts copied. Text criticism must take some account of all three types of analysis (though it is chiefly interested in discourse transmission analysis).
7. DIACHRONIC ANALYSIS: STAGES OF SOURCES
The history of a discourse can be traced backwards as well as forwards in time. The result is source analysis. I have come near to speaking about source analysis already. For suppose one starts with a discourse Dn at the n-th stage of transmission. If this discourse rather than the original discourse D0 is treated as the starting point, the discourses Dn – 1, Dn – 2, …, D0 form a sequence of progressively more remote sources.
If we relabel and start with D0, we can in general suppose that the speaker S0 was a part of the audience A-1 of a discourse source D-1 uttered by speaker S-1. S-1 in turn was part of audience A-2 for discourse D-2 uttered by speaker S-2. In general an indefinitely long sequence of discourse sources D-1, D-2, . . ., D-n is obtained. Of course, as in the case of transmission, so here, there will often be multiple sources. The scribe may conflate two or more manuscripts, or the Gospel writer may use many written and oral materials available to him (Luke 1.1, cf. 1 Chr. 9.1, 2 Chr. 13.22, 16.11, 24.27).
By analogy with transmission analysis, we can define n-th stage source analysis as the analysis of the sources D-n, speakers S-n, and audiences A-n n stages removed from the original discourse D0. Even when the sources are not separable into clearly defined stages, we can distinguish speaker source analysis, discourse source analysis, and audience source analysis. The Table summarises the types of analysis so far obtained.
How many stages of the sources need to be considered? In theory, the sources of discourse D0 might be traced back an indefinite number of stages, and the transmission might be traced forward an indefinite number of stages. Hence it would seem that the analysis of discourse cannot stop short of analysis of the totality of the historical time scale. This is but the linguistic form of the philosophical problem of the meaning of [p. 133]
TABLE OF TYPES OF DIACHRONIC ANALYSIS
| speaker analysis | discourse analysis | audience analysis | ||
| … | … | … | … | |
| source analysis |
n-th stage (n-th stage source analysis) |
S-n (speaker source analysis) |
D-n (discourse source analysis) |
A-n (audience source analysis) |
| … | … | … | … | |
| 1st stage | S-1 | D-1 | A-1 | |
| synchronic analysis |
S0 | D0 | A0 | |
| 1st stage | S1 | D1 | A1 | |
| … | … | … | … | |
| transmission analysis |
n-th stage (n-th stage transmission analysis |
Sn (speaker transmission analysis) |
Dn (discourse transmission analysis) |
An (audience transmission analysis) |
| … | … | … | … |
history. According to some, the meaning of history cannot be fathomed without having seen the whole of history. Likewise it might be claimed that a discourse be fully appreciated without having seen all its ramifications in time.
In practice, the analysis of a discourse may terminate at any of a number of convenient points. It may, first of all, terminate with synchronic discourse analysis. Or it may include synchronic speaker, discourse, and audience analysis, leaving out diachronic analysis. Or it may terminate at the point where available information about sources and transmission fails and speculation reigns. Or it may terminate where the sources and transmission are no longer primarily verbal.
But any of these dividing lines produces a certain amount of arbitrariness. From one point of view, it is enough to say that the Apostle Paul responded to reports from Timothy about the Thessalonian church (1 Thes. 3.6). These reports were based on conversations and observations of Timothy in Thessalonica. We can stop at this point where the verbal sources end. But [p. 134] from another point of view it is interesting to try to understand the social and religious circumstances at Thessalonica which contributed to the rise of the situation about which Paul heard. Such analysis goes beyond verbal sources. Similarly with tramission analysis. We can stop with a consideration of what Mennonites have said about the Sermon on the Mount (verbal analysis), or we can go on to look at what they have done about it and how it has altered the structure of their communities (socioeconomic analysis).
8. CANON
The preceding diachronic discussion provides a useful framework for considering some questions about biblical canon. Every social group has means of appealing to oral or written tradition to substantiate and reinforce authority. It also has means of identifying those who are competent to interpret the tradition and exercise authority. In some cases the competent interpreters are formally recognised officials, and they exercise an authority more or less independent of the tradition, so that they can dominate it and manipulate it. Such is the case with the tribal shaman, the modern scientist, the Hindu guru, and the prophet in the Mormon Church. In such cases no stratum of traditional verbal material with specially marked authority may crystallise. In other cases, however, there are no formally recognised officials, or else these officials are expected by the social group to adhere to the tradition rather than dominate it. Then, an identifiable ‘canon’ is needed to separate authentic tradition from inauthentic. The heavier the group obligation to preserve and submit to the tradition, and the more important are the disputes that may arise, the more carefully must the boundaries of the canon be marked. Thus arise constitutions, Robert’s Rules of Order, contracts, legal codes, rule books for games, and religious canons like the Bible.
Now, a social group using a canon must be able to specify reasonably accurately which discourses are its canon, and in what sense each is a canon. Moreover it must be able to specify reasonably accurately which of the meanings uncovered by analyses as in the Table are to function authoritatively. This may not be a serious problem, if all or at least a good number of the meanings from the Table are close to one another. But [p. 135] it will become a problem to the extent that the results of different types of analysis diverge.
A good deal of the history of modern biblical scholarship can be understood in terms of attempts to answer the question of canon. Protestant orthodoxy of the sixteenth and seventeenth centuries made official pronouncements about which discourses were its canon.18 But only in the nineteenth century did questions about the types of meaning in the Table begin to become critical. First, advances in text criticism made it possible to identify autographs with increasing exactness. Of course, in the case of tradition in general, it is not always the case that an autograph exists. The tradition may derive from a number of historically interwoven strands which never happened to crystallise in a single monologue from which the bulk of subsequent tradition uniquely derived. But it would be generally admitted in the case of most of the biblical texts that an autograph lies behind the extant variants.
Text criticism, then, answered the canon question more precisely by providing established means of adjudicating among extant variants. But source criticism and later form criticism reopened the canon question by indirectly raising questions about the unique authority of the autograph. Let D0 denote a biblical autograph. Then sources D-n were supposed to have greater historical authority than D0. At the same time, they might or might not have greater theological and ethical authority. If D-n represented words of Jesus in contrast to the Gospels (D0), they had higher authority in the classical liberal view. But if D-n represented an OT source, it might be classified as a more primitive stage and so be accorded lower authority. The attributing of decisive historical authority to some stages and decisive theological authority to others resulted in an inevitable Geschichte/Historie tension or dichotomy.
What are the available present-day options? First, most would admit that all types of analysis in the Table might, as a scholarly exercise, be applied to biblical texts as well as to nonbiblical material.19 Writings of church fathers, Masoretic [p. 136] tradition, Qumran material, rabbinical midrash, and a host of other material can be treated as formally on a level with autographs and sources. The real question, then, concerns the authority of the results of such analysis.
One option is the abandonment of any ‘real’ canon, allowing one to pick and choose among the results so obtained. Rough boundaries can still exist to indicate the general direction in which material of value is likely to be found. For example, the classical biblical canon plus a few apocryphal books provide the natural direction in which people with some ‘Christian’ orientation will tend to go for their religious values. Such a position is spelled out explicitly in James Barr’s The Bible in the Modern World (London: SCM, 1973), 162-4; it is implicit in the procedure of a large group. This position is compatible with a high general respect for ancient tradition. For example, the person committed to Platonic philosophy or to Vedantic Hinduism will not lightly reject any saying of the masters, even though he may admit their fallibility in principle. Likewise with a person committed to the Christian faith.
A second option is the acceptance of the traditional church canon as one’s own–either in Protestant or Roman Catholic form. By itself, this does not answer all the questions. It still remains to decide in what sense the autographs are to be authoritative. Is one to canonise speaker meaning or discourse meaning or both? Does audience meaning have any special status?
A third, intermediate option is the selection of a new canon, either (a) by the selection of a new set of canonical discourses, or (b) by the selection of some overarching principle or theme against which the relative authority of each exegetical result can be measured. Instances of (a) would be a new list of canonical books, or the attribution of ultimate authority to the sayings of Jesus. Examples of (b) would be measurement of authority in terms of the principle or theme of justification by faith, or redemptive history, or proclamation of authentic human existence, or proclamation of Christ, or relevance to human salvation. But preference for a theme has its own problems. It tends to create a nebulosity of its own even greater than the original nebulosity about canonical discourses. For one thing, a variety of themes compete with one another for the office of [p. 137] authority, and each theme can be understood from a variety of perspectives. The themes themselves do not offer us a canonical shape of sufficient fixity to provide the foundation for group unity in time of serious moral tension. In other words, a biblical ‘theme’ apparently cannot serve as a functionally effective canon in a sociological sense. Some of the issues arising here will be explored in a future article which will ask what we are after in analysing a biblical text.
In any case, the important point is that all three options can be clarified by means of the linguistic and interpretive distinctions introduced in sections 1-7. The options can be more accurately distinguished from one another; and the question of where authoritative meaning is found can then be explored in greater detail within each option. I suspect that the distinctions within the options and between them will help to explain some of the differences among exegetes. Exegetes are likely to emphasise and favor those types of meaning that they regard as authoritative or canonical. Exegetes without any ‘real’ canon may simply follow a scholarly consensus about what is most important. Or they may strike out on their own, perhaps in search of esoterica, perhaps in search of whatever they regard as morally and religiously most commendable.
Distinguishing different types of meaning can therefore be useful. But by itself, it will not tell us which meaning or meanings are to be treated as ‘canonical’. That decision is bound up with one’s most basic ethical, religious, and philosophical commitments.
VERN S. POYTHRESS
Westminster Theological Seminary
Chestnut Hill
Philadelphia, Penn. 19118
1 cf. the Roman Catholic discussion of sensus plenior, R. Brown, ‘The Sensus Plenior in the Last Ten Years’, CBQ, XXV (1963) pp. 262-85.
2 cf. the foundational remarks of John L. Austin, Philosophical Papers (Oxford: Clarendon, 1961), pp. 23-43; Ludwig Wittgenstein, Philosophical Investigations, 2nd ed. (Oxford: Blackwell, 1968); and work in philosophical hermeneutics like Hans-Georg Gadamer, Philosophical Hermeneutics (Berkeley: University of California, 1976), Truth and Method (New York: Seabury, 1975); E. D. Hirsch, Validity in Interpretation (New Haven-London: Yale, 1967).
3 Downing, ‘Meanings’, What about the New Testament?, ed. Morna Hooker and C. Hickling (London: SCM, 1975), pp. 127-42.
4 cf., e.g., John F. A. Sawyer, Semantics in Biblical Research (London: SCM, 1972),pp. 4-27; George Fohrer et al., Exegese des Alten Testaments. Einf(hrung in die Methodik(Heidelberg: Quelle & Meyer, 1973); Wolfgang Richtcr, Exegese als Literaturwissenschaft. Entwurf einer alttestamentlichen Literaturtheorie und Methodologie(Gottingen: Vandenhoeck & Ruprecht, 1971).
5 Saussure, Course in General Linguistics (New York: McGraw-Hill, 1966) (French original, 1915).
6 Barr, ‘Etymology and the Old Testament’, OTS, XIX (1974), pp. 21-22, 28.
7 Sometimes, of course, a situational factor does lead to a change even in the middle of a speech. A speaker may adjust to puzzled looks from his audience. As some have argued, Paul may have adjusted to new information from Corinth when he began 2 Corinthians 10. Such factors are properly included under synchronic analysis of the situation.
8 Hans Conzelmann, The Theology of St. Luke (London: Faber and Faber, 1960),pp. 24-25.
9 Austin Farrer, A Study in St. Mark (London: Dacre, 1951), p. 347.
10 cf. Hirsch, pp. 1-23, 218, for an attempt to specify more precisely the speaker’s meaning. ‘Verbal meaning is, by definition that aspect of a speaker’s “intention” (in a phenomenological sense) which, under linguistic conventions, may be shared by others’(p. 218). Unconscious implications are included (pp. 220-4).
11 William Wimsatt, The Verbal Icon (London: Methuen, 1954), pp. 3-18.
12 Hirsch’s preference for author meaning is based on a lively feeling for this problem (Validity in Interpretation, pp. 3, 25).
13 Wimsatt, pp. 3-18, 19-39.
14 I assume here that 2 Thessalonians is Pauline.
15 Some reply is due to Hirsch’s arguments that discourse meaning must be speaker meaning (Validity in Interpretation). Hirsch is largely occupied with defending the idea of an original historically-fixed meaning over against later interpretations. My use of the diachronic/synchronic distinction shows that I am fundamentally sympathetic with him. Moreover, I would stress that discourse meaning takes into account everything that the intended audience would know about the speaker. Hence their estimate of the speaker’s intention is very much in the foreground. If the speaker is an effective speaker, the audience will understand him well enough and we can ignore the difference between speaker meaning and discourse meaning. But Hirsch is aware that a speaker may ‘bungle’ (pp. 233-4).He admits that in this case a difference exists between the speaker’s meaning and ‘What the speech community takes it to mean’ (p. 233). This distinction is all that I’m concerned for. But Hirsch argues further that, even in this case, since the parole is the parole of the author, the meaning must be the meaning of the author. This is a petitio principii which assumes rather than demonstrates that meaning must be located primarily with the author. If we accept Hirsch’s own theory that meaning is a ‘type’ distinguishable from any particular act of ‘intending’, there is no reason why we cannot describe ‘What the speech community takes it to mean’ as also a ‘meaning’. The problem is more general than what Hirsch considers. Even if the author does not ‘bungle’ seriously, yet he may commit infelicities. Even when a speaker is basically successful, he may not succeed in communicating all the nuances that he wanted, even though those nuances could have been communicated had he chosen his language more carefully.
16 On the identification of borders of ’emic’ units, see Kenneth L. Pike, Language in Relation to a Unified Theory of the Structure of Human Behavior, 2nd ed. (The Hague-Paris: Mouton, 1967), pp. 76-8.
17 This has been theoretically grounded by K. Pike (ibid., pp. 25-149) and confirmed by work on sociolinguistics of conversational interchange (Thomas P. Klammer, ‘The Structure of Dialogue Paragraphs in Written English Dramatic and Narrative Discourse’, Ph.D. dissertation. University of Michigan, 1971).
18 The French Confession of Faith (1559), III; the Belgic Confession, IV; The Thirty-Nine Articles of the Church of England, VI; The Westminster Confession of Faith, I, 2.
19 Sawyer, Semantics in Biblical Research; Brevard S. Childs, Exodus (London: SCM, 1974).