by Vern Poythress
[Published in the Scottish Journal of Theology 32/4 (1979) 319-31. Used with permission. Also as pdf.]
l. INTRODUCTION
WHEN an exegete approaches a biblical text, he wants to find out ‘what it means’. But which meaning is he after? Is he interested only in what the speaker (or author) intended? Or is he interested also in the speech product (which may not fully succeed in embodying the speaker’s intention)? Is he interested in what the audience thought of the discourse? And which of these meanings (if any) is the appropriate starting point for a modern sermon?
In this article I will explore some of these different types of meaning, and the different types of analysis aiming at finding these meanings. The most important distinction for my purposes is that between speaker, discourse, and audience. A fuller discussion of this distinction can be found in my earlier article, ‘Analysing a Biblical Text: Some Important Linguistic Distinctions’.1 Roughly speaking, speaker analysis asks what the speaker means by a given discourse. Discourse analysis asks what the discourse means to a competent hearer-evaluator who is familiar with the situation in which the discourse is uttered. Audience analysis asks what the audience takes the discourse to mean. These three types can now be further subdivided in various ways. First, I want to discuss what happens when the analysis approaches a discourse with special interest in a particular topic.
2. PROJECTED ANALYSES
An analysis of a discourse D that is wholly interested in one particular topic I call a projected analysis of D or a projection of D onto that topic. As an example, consider the sayings of Jesus. Can we make these, and these alone, into a ‘canon’? At first blush, this might seem to some to be an attractive possibility. But we do not have direct access to such sayings. What we have are copies of Gospels D which include reports of such sayings. The preresurrection utterances of Jesus of Nazareth are sources D’ behind the Gospels. Obtaining the sayings involves his- torical reconstruction from the Gospels. In the process, one would analyse the Gospels with an interest in certain discourses behind them. Such analysis is projected analysis or a projection of the Gospels.
To a certain extent the possible types of projected analysis can be catalogued. Of any discourse D0 we may ask what it says about any given subject of interest. We may ask what it says (or implies) about economics, politics, religion, aesthetics, psychology, etc. This could be called a projection of D0 onto economics, politics, and so on. But since we are concentrating on human verbal behavior, it is of special interest to look at the projections of D0onto discourses. We may ask, (1) What does D0 say about its source discourses D’? (2) What does D0 say about its transmission discourses (copies and reports of D0? (3) What does D0 say about discourses with which it has no causal connexion? D0 may, of course, speak predictively in areas (2) and (3); but on the average more will be said in area (1), concerning D0‘s sources. Luke 1.1-4 and references in Chronicles (2 Chr. 16.11, 20.34, etc.) are the most obvious examples.
Here I should distinguish two kinds of approach to D0‘s sources. On the one hand, there is analysis of D0‘s own view of the sources, as it is expressed both directly and indirectly. On the other hand, there is analysis in which an analyst uses D0 to construct his own view of the sources. In the latter case ‘third degree’ techniques may be used to compel the discourse to yield information beyond what a sensitive hearer receives. As an example, take Luke 5.12-16. The Gospel of Luke’s view of the sources of Luke 5.12-16 is something like ‘this story came from a reliable source’. Such is what the reader of Luke ‘hears’ Luke saying. But the analyst’s view, based on the ‘third degree’ method of minute comparison with Mark, may be ‘This passage is literarily dependent on Mark 1.40-5’. Thus the analyst has learned more than the discourse D0 told him.
Inquiry concerning D0‘s view I shall call internal projection of D0. Construction of an analyst’s view I shall call external projection of D0 (because the viewpoint chosen is ‘external’ to the discourse itself).2 External projection might, if one wished, not be classified as ‘projection’ at all, but simply as one facet of the general process of historical reconstruction. But I have supplied a special term for convenience.
The concept of projection can be applied also to speaker analysis, discourse analysis, and audience analysis. Suppose that D is the original discourse and D’ a source discourse. There are nine possibilities for internal projection. We can ask (1) what the speaker meaning of D says about the speaker meaning (and hence the speaker) of D’, (2) what the speaker meaning of D says about the discourse meaning (and hence about the discourse D’ itself), (3) what the speaker meaning of D says about the audience meaning (and hence the audience) of D’, (4) what the discourse meaning of D says about the speaker meaning of D’, (5) what the discourse meaning of D says about the discourse meaning of D’, and so on.
The same ninefold division can be carried out even when D and D’ are the same discourse rather than one being a source of the other. First, speaker analysis can be projected onto the speaker, the discourse, or the audience. Speaker analysis can ask: (1) What did the speaker intend his discourse to accom- plish for himself at the time? (2) What was the speaker’s perception of the discourse itself, and its organisation, at the time he gave it? Was he, for example, aware that it did or did not fully succeed in expressing his intention? (3) What potential audience did the speaker have in mind and how did his ideas about the audience affect what he said? How did he suppose that the audience would respond to what he said? (1), (2), and (3) represent speaker analysis projected onto speaker, discourse, and audience, respectively
Next, discourse analysis may be projected onto speaker, discourse, or audience, (1) In the projection onto speaker, one asks, ‘What did the discourse actually communicate concerning the speaker and his intentions, wishes, beliefs, reactions?’ (2) In the projection onto discourse, one asks about the discourse’s commentary on itself. How does discourse structure reinforce and modify the parts of the discourse? (3) In the projections onto audience, one asks, ‘What did the discourse communicate about its intended audience?’
Audience analysis can also be projected in three ways. The analyst asks, (1) ‘What did the actual audience think the speaker intended by what he said?’, (2) ‘How did the audience interpret the discourse itself?’, and (3) ‘What response did they interpret the discourse as expecting from them?’ Table 1 summarises the resulting divisions of synchronic analysis.
TABLE 1
Types of Projected Synchronic Analysis
The Viewpoint:
| The Topic: | Speaker Analysis | Discourse Analysis | Audience (and Situation) Analysis |
| speaker | speaker analysis projected on the speaker |
discourse analysis projected on the speaker |
audience analysis projected on the speaker |
| discourse | speaker analysis projected on the discourse |
discourse analysis projected on the discourse |
audience analysis projected on the discourse |
| audience | speaker analysis projected on the audience |
discourse analysis projected on the audience |
audience analysis projected on the audience |
The concept of projected analysis enables us to clarify certain problems in biblical exegesis. Our knowledge of the Ancient Near East is limited and piece-meal. In particular, we know little about most biblical writers and audiences, beyond what the text explicitly tells us (internal projection) and what we can infer from the text (external projection). For example, speaker analysis of the Gospel of Mark is hampered by the absence of an opportunity to interview the author personally or at least to work from a full-scale biography. Hence speaker analysis tends to boil down to discourse analysis projected onto the speaker, approaching the speaker’s intentions by the only available means: the extant text. Nevertheless, such analysis may serve as a preliminary to deeper and more thorough discourse analysis projected onto the discourse. This can fill some of the gap left by our inability to perform full-scale speaker analysis. Similar remarks hold for audiences such as the Colossian or Corinthian church, about whom little is known apart from Paul’s letters. But the danger of circularity is much greater when we must rely on discourse analysis projected onto the speaker or discourse analysis projected onto the audience rather than on-the-spot case study of speaker or audience.
In at least one instance the distinction between speaker analysis and discourse analysis becomes more important. That is the case where, because the author has left more than one extant work, we have an opportunity to penetrate into his thinking to a greater extent than a single discourse permits.
Thus in the case of Paul’s letters, a real distinction can arise between speaker analysis and discourse analysis projected onto the speaker. In the former case one uses the background information from all the genuine pauline epistles to inform one’s understanding of Paul’s intention at a given point. In the latter case, one asks how far the text itself reveals Paul’s in- tention, apart from a reader’s knowledge of the other epistles. This latter form of analysis is a useful exercise. After all, Paul’s original readers did not have the rest of the pauline corpus available to them. In the case of the Corinthian and Thessa- lonian churches, of course, they had heard Paul’s preaching, which must have covered some of the same ground that we see covered in the other pauline epistles. Hence the other pauline epistles help us to reconstruct the otherwise irrecoverable common situational context of the Corinthian and Thessalonian letters. But what if we examine the epistles to the Romans and to the Colossians (assuming that the latter is genuinely pauline)? The audiences here are two churches that Paul had never visited. They could not possibly have approached the epistles with an understanding of the background of pauline theology such as the modern New Testament scholar has. Hence, in determining discourse meaning and audience meaning, one must beware of importing too much from other pauline epistles. Interestingly, Colossians and Romans do not represent so very different a style from the Corinthian and Thessalonian epistles. What are we to conclude? A number of possibilities present themselves, (a) In Romans and Colossians Paul in fact seriously overestimated his readers and a number of important points escaped them. (b) Paul was not so concerned that his audience grasp every nuance to his message, as long as the main points were understood, (c) Paul expected that those friends in the Roman and Colossian church who did know him would explicate the difficult points. (d) Paul did not use much specialised vocabulary or many specialised concepts, and under- standing his letters did not in fact require specialised knowledge of his ideas. All of these options may have a degree of truth to them. (d) is the most interesting, from my point of view, because it shows that there is a danger of oversubtlety at those points where an analyst knows more about the speaker than does his audience.
3. DESCRIPTIVE, PRESCRIPTIVE, AND NORMATIVE ANALYSIS
In defining discourse analysis in my previous article, I introduced the idea of competence and standards for com- petence in interpretation. This idea can, in fact, also be applied to speaker analysis and audience analysis. For example, I can distinguish, at least in principle, between (a) what the audience in fact understood from a discourse and (b) what they were warranted in understanding from it. (b) can be more closely defined as what competenthearers in the audience could be expected to understand. The first of these (a) I shall call descriptive audience meaning. The second (b) I shall call prescriptive audience meaning. More precisely, I can say that the prescriptiveaudience meaning of a discourse is the meaning that could be arrived at by competent judges who analyse the discourse on the basis of sufficiently extensive knowledge of the linguistic, discourse, and situational context available to the audience. Parallel definition can be constructed for prescriptive speaker meaning and prescriptive discourse meaning. Thus we obtain the following definitions.
Prescriptive speaker (audience; discourse) meaning of a discourse is the meaning that could be arrived at by competent judges who analyse the discourse on the basis of sufficiently extensive knowledge of the linguistic, discourse, and situational context from the speaker’s viewpoint (from the audience’s viewpoint; shared by speaker and audience).By contrast, descriptive speaker (audience, discourse) meaning of a discourse is the meaning in fact attributed to the discourse by the speaker (by the audience; by people sharing the language of speaker and the intended audience who are informed about the situation and asked to judge the discourse meaning). In general, speaker analysis is understood to include discussion of both prescriptive and descriptive speaker meaning (what did the speaker under- stand and what was he warranted in understanding). Similarly, discourse analysis and audience analysis are understood to include discussion of both descriptive and prescriptive aspects.
Most of the time descriptive and prescriptive meaning will be very similar. In addition, in the case of biblical studies, we have so much less information than the original participants, that we are hardly in a position to challenge their interpretation. Hence, except in a few cases of obvious misunderstanding, such as the letter of 1 Cor. 5.9, we will seldom if ever be able to challenge the descriptive meaning attributed to a discourse by a native speaker.
However, there is a kind of challenge that takes place much more frequently, namely a challenge of the historical, scientific, or ethical validity of parts of a discourse (Sachkritik). I shall call normative analysis any attempt to determine, what the speaker ought to have said (normative speaker analysis), what in the discourse ought to be held as true (normative discourse analysis), and what the audience ought to have believed and done in response (normative audience analysis). All such judgements will be made in terms of historical, scientific, ethical, and other norms of the analyst. The main types of descriptive, prescriptive, and normative analysis are indicated in Table 2.
As an example of the distinctions, take Paul’s appeal in 1 Cor. 15.5-7 to witnesses to the resurrection. Descriptive speaker analysis discusses what Paul thought he was saying. Prescriptive speaker analysis discusses what he should have thought he was saying. These two will differ only in the case that the speaker blunders without realising it. In the case of 1 Cor. l5.57 both of these uncover the meaning: Christ appeared to Cephas (and others). However normative speaker analysis asks whether Paul ought to have made this appeal to witnesses. Bultmann here says no.
TABLE 2
First Order Distinctions of Types of
Synchronic Analysis of the Monologue
| speaker analysis | discourse analysis | audience analysis | |
| uses information about the speaker |
uses information common to speaker and the intended audience |
uses information about the actual audience |
|
| descriptive | What did the speaker understand? |
What did evaluators knowing the language understand? |
What did the audience understand? |
| prescriptive | What did the speaker have a right to understand? |
What did evaluators have a right to understand? |
What did the audience have a right to understand? |
| normative | What ought the speaker to have said? |
How valid is the discourse? |
How ought the audience to have responded? |
Next descriptive discourse analysis and prescriptive discourse analysis ask what 1 Cor. 15.5-7 means to judges contemporary with Paul. Again one gets the meaning: Christ appeared to Cephas. But normative discourse analysis asks, ‘Is the tradition of the Cephas-appearance historically trustworthy?’ Non- Christians would say no. Finally, normative audience analysis asks whether the readers should have believed that Christ appeared to Cephas. It is possible to say yes to this and still disbelieve in the Cephas-appearance oneself-if one thinks that only modern research in human psychology or in biology have made the Cephas-appearance unworthy of belief. Thus the three types of normative analysis are distinct from one another, and they are clearly distinct from descriptive and prescriptive analysis.
Much of the heat in modern church controversy centers on normative analysis. It is never enough to simply describe what the biblical documents say. There is always the further ques- tion, ‘What shall we believe?’ And what ought a modern sermon to say on the basis of the biblical discourse? Funda- mentalism and conservative orthodoxy have had a ready answer. Because the Bible’s discourses are God’s word, their prescriptive meaning is normative for us. Hence the transition from prescriptive to normative occurs fairly easily.
However, forms of liberal and neo-orthodox Christianity are obliged by their own premises to make a more complex transition from the descriptive and the prescriptive to the normative. A good deal of the time, the problem is ignored simply by so constructing the final result of the prescriptive analysis that the analyst can in good conscience put it forward also as a normative analysis of what ought to be said. Or at least it is what ought to form the basis for what is to be said in the modern sermon. When, however, one asks why a transition can be made from this prescriptive analysis to the normative, diffi- culties arise.
One of the contributions of Bultmann’s demythologising programme and the new hermeneutic has been to bring the problem more into the open. Bultmann, for one, has put in the center of the stage the question of norms for the use of the N.T. in modern preaching. It is obvious that as a descriptive analyst he can explain well enough what the N.T. is saying. He gives an undemythologised account of it at the beginning of his programmatic statement.3 The demythologizing program is then an attempt to describe (descriptive and prescriptive analysis) what the N.T. says about human existence and self- understanding; but to say it in the way in which it ought to be proclaimed to modern men. The debate then goes back to whether Bultmann has eliminated something central and essential from the gospel: redemptive history, or a historical resurrection, or Christian faith as including doctrinal content. But, of course, the selection of some element or theme as ‘central’ involves a normative judgement by the analyst. How can any amount of description enable him to say, ‘Because this theme is a “center” in such-and-such a way, we ought to preserve it as central and derive our norms from it’? Where does the ‘ought’ come from? Most scholars hope that the ‘ought’ will come from the Bible itself. But, even if some portion of the Bible were to say, ‘Such-and-such theme is the central element needing to be preserved’, such a statement would be part of the prescriptive meaning of the Bible. Normative analysis would still be free to challenge it or incorporate it as ‘basically true, but needing qualification or reinterpretation’ Hence, as long as the norms of modern analysts have free rein no resolution of major theological and hermeneutical differences is in sight.
4. TYPE-TOKEN DISTINCTIONS
Three final refinements will complete my classification of types of linguistic analysis. The first is the introduction of the type-token distinction. Each occurrence of the word nomos ‘law’, is a token of nomos. The totality of all occurrences is the type nomos. Similarly, any morpheme, word, phrase, clause, etc., or even the discourse as a whole, can be treated as either a token (a unique utterance of a single speaker at a localised time’ and place) or as a type (a class consisting of all occurrences or perhaps potential occurrences of the unit in question in a variety of contexts). So far I have basically treated the dis- course and its subordinate parts as a token. However, a careful interpretation of the token can never dispense with some consideration of the token’s paradigmatic relationships to other units. What units of the same size does it contrast with? In what range of contexts does one expect it to occur? What are the expressions most closely related to it in one way or another? In particular one may find oneself inquiring about the meaning of a type: what range of meaning does this morpheme, word, phrase, etc., have in the language as a whole? Though such questions are primarily in the domain of exegesis, they are related in a complex way to the continuing discussion of historical revelation vs. timeless truths.4 The tokens are anchored in a definite literary and historical context, and so cannot be used without further ado as the starting point for syllogistic reasoning.
5. PHONOLOGICAL, GRAMMATICAL, AND SEMOLOGICAL ANALYSIS
Linguistic analysis of a discourse can concentrate on any of the distinct ‘strata’, aspects or structural hierarchies of language distinguished by descriptive linguistics. For example, language has phonological, grammatical and semological (semantic) structure. Speaker analysis or audience analysis or any of the types of analysis in Table 1 can focus on any one of these strata. The exact nature of the strata depends, of course, on the linguistic theory that one uses. Linguists do not yet agree even about the number of strata, much less the exact boundaries between them and the nature of the structure within one stratum. One can say at most that there is a general consensus that some meaningful distinction can be made between the sound structure (phonetics and phonology), the grammatical structure, and a semological structure (meaning, reference, paraphrase).
As an example, let us take discourse analysis projected on the speaker. Such analysis of phonology asks what the phonological structure of the discourse tells us about the speaker. Does his voice quality at a given point indicate anger, excitement, contentment? Why does the speaker emphasise certain words? Does he delight in certain rhetorical effects due to rime or alliteration? Next, discourse analysis of grammar, projected on the speaker, asks what the grammatical structure of the dis- course tells us about the speaker. Are there grammatical marks of the speaker’s class origin? What purpose does he have in cases of anacoluthon? Why does he choose predominantly active clauses or passive clauses? Finally, discourse analysis of semology projected on the speaker asks what the semological structure of the discourse tells us about the speaker. What information does he want to communicate? What do his state- ments show about his understanding of the world? His ethics? His attitude to social and political institutions?
It is at this point that my classification touches on the work of the structuralists. Structuralists have distinguished various ‘strata’ of discourse meaning by using linguistic and quasi- linguistic criteria. Roman Jakobson argues that communica- tion is organised in six ways: emotive, conative, referential, poetic, phatic, and metalingual.5 Roland Barthes distinguishes five ‘codes’, the hermeneutic, semic, symbolic, proairetic, and cultural.6 These do not correspond in any simple way to Jakobson’s categories. Still different categories can be found in the approaches of Claude Bremond, A. J. Greimas and Gerard Genette.7
The differences between the structuralist approaches dem- onstrate the hazardous character of present attempts to dis- tinguish such strata.
6. ETIC, EMIC, AND UNIVERSAL-COMPARATIVE ANALYSIS
Linguists are well acquainted with the difference between phonetic analysis and phonemic analysis. Phonetic analysis describes speech sounds with reference to abstract reference points chosen for the convenience of the analyst. By contrast, phonemic analysis, or phonological analysis, undertakes a description of the sound system of a language, ignoring any sound differences that are not relevant to the system. Kenneth L. Pike has argued that the phonetic/phonemic difference can be generalised to an etic/emic difference characterising two approaches to human behaviour.8 Etic analysis is analysis in terms of outside reference points, reference points of the analyst. Emic analysis is analysis in terms of reference points ‘inside’ the behavior analysed, reference points of which the ‘natives’ are at least tacitly aware, and in terms of which they themselves operate. Bultmann’s undemythologised and demythologised accounts of the kerygma are examples of emic and etic analyses, respectively. An emic analysis may be valid or invalid; and so may an etic analysis. But emic analysis is not per se more or less ‘valid’ than etic analysis; it is simply different from etic analysis.
Etic analysis may itself be subdivided into two types. Suppose an analyst approaches a discourse with which he is radically unfamiliar. At first he must use his own categories and reference points, because he does not yet know those endemic to the discourse. Thus any preliminary analysis may be termed initial etic analysis. The analyst may then progress to an under- standing of the ‘system’ of the discourse, and present a proposed emic analysis. Finally, this emic analysis may be set within the larger framework of a system of universal reference points intended to be adequate for describing all languages and discourses. This set of reference points is once again ‘outside’ the discourse in question. Hence the result may be called a terminal etic analysis or (as I prefer) universal-comparative analysis.
For example, Bultmann’s demythologised kerygma should probably be understood as a version of universal-comparative analysis. To arrive at his demythologised gospel, Bultmann has long ago gone through the preliminary stages of an initial etic analysis (not free, presumably, from existentialist presupposi- tions) and some emic analysis.9 The final form of the demy- thologised gospel is set in a supposed universal framework derived from existentialist anthropology. The discussion of the impossibility of ‘presuppositionless’ hermeneutics would, I think, become clearer if distinctions were more often made between (a) the necessity of a starting point outside the system (initial etic analysis), (b) the possibility for success or failure in achieving an explanation in ‘native’ terms (emic analysis), (c) the possibility or desirability of integrating one’s interpretation into a more universal framework (universal-comparative analysis).
7. CONCLUSION
Above I have delineated a host of interlocking and intersect- ing distinctions of types of analysis. In practice, the types of analysis will frequently be used in fusion, in mutual reinforce- ment, or perhaps in tension. A decisionbetween instances of analysis can only be made on the basis of some standards already at hand for the analyst. Hence, the question, ‘Which analysis is valid?’, could more profitably be rephrased, ‘Which analysis (or analyses) is useful for what purposes?’ and ‘Which practices of analysis are valid according to what standards?’
VERN S. POYTHRESS
Westminster Theological Seminary
Chestnut Hill
Philadelphia, Penn. 19118
1 S.J.T., vol. 32 (1979), pp. 113-37.
2 My “internal projection’ corresponds to Hirsch’s discussion of the ‘implications of meaning’. (E. D. Hirsch, Validity in Interpretation (New Haven-London: Yale,1967)> pp.140ff.) According to Hirsch, the ‘implications’ are part of the verbal meaning. By contrast, my ‘external projection’, corresponding to Hirsch’s ‘significance’, goes beyond the ‘meaning’.
3 Bultmann, ‘New Testament and Mythology’, Kerygma and Myth, ed. Hans W. Bartsch (New York: Harper and Row, 1961), pp. 1-2.
4 See Paul Helm, ‘Revealed Propositions and Timeless Truths’, Religious Studies,viii (1972), pp. 127-36, especially the notes on sentences (types) which change truth value when uttered at different times (tokens) (pp. 128, 134).
5 Robert Scholes, Structuralism in Literature; An Introduction (New Haven-London: Yale, 1974), pp. 24-6.
6 Barthes, S/Z (New York: Hill and Wang, 1974), pp. 18-20; Scholes, pp. 153-55.
7 ibid., pp. 91-111, 157-67.
8 Kenneth L. Pike, Language in Relation to a Unified Theory of the Structure of Human Behavior, 2nd ed. (The Hague: Mouton, 1967), Ch. 2.
9 Bultmann, ‘New Testament and Mythology’, pp. 1-2.