Speaker as accessor and discourse topic

head: God honors those who honor him ch. 1-6 God controls world rulers Ds life shows that God is sovereign in the affairs of men D friends honored God he honored them D accessing prophecy ch.7-12 step Figure 29: Updated spaces in the Book of Daniel Since Daniel is intrinsic to the role of both parts of the book in the global schema, which is headed by the macropredication “Daniel’s life shows that God is sovereign in the affairs of men.” Superimposed spaces, as well as consolidating a narrative unit to serve as a step in a further argument, make it possible to see Daniel as global topic for the book. If an accessed space has a referential theme, that is, a discourse topic, which is distinct from the accessor, then the paragraph will have two topics: an accessed topic and an accessor topic, in two distinct spaces. This happens in several places in the last part of the Book of Daniel. One is in Daniel 8:13–18, where the topic in the accessed space is apparently an angel: “Then I heard a holy one speaking.… So he came near to where I was standing, and when he came I was frightened and fell on my face; but he said to me, ‘Son of man, understand that the vision pertains to the time of the end.’ Now while he was talking with me, I sank into a deep sleep with my face to the ground; but he touched me and made me stand upright. And he said….” This discourse unit had two topics, Daniel and the angel, but they belong to different superimposed spaces see §3.4.5. Discourse units with superimposed spaces of this kind therefore furnish one possible answer to the question, raised in §3.4.5, of how discourse units can have two topics and discourse schemas can still have a single head. With superimposed spaces, each space has its own schema which could have a topic, and nothing suggests that one or the other schema has more than one head. In “Somewhere over the rainbow” and the last part of Daniel, the narrator is also text-internal accessor, which entails being “present” in the accessed space. The narrator, by remembering, talks about himself, and “uses” the remembered self in accessing other information in the same way that Conrad, in Lord Jim , uses Jim for that kind of access. The question arises whether the narrator himself could in some sense have an accessor space. The answer is “Yes,” and if that kind of analysis remains implicit for some texts, there are others for which it is necessary and needs to be recognized explicitly. It is to that possibility which we now consider.

3.6.5 Speaker as accessor and discourse topic

The existence of discourse units with two superimposed spaces validates the fact of discourse spaces, since they can have “a life of their own” that is somewhat distinct from the discourse units with which they are associated. But it also raises a question about coherence: If the coherence of a discourse unit has to do with the ability of the addressee to represent it in a single, unified space, how can discourse units with two superimposed spaces be said to be coherent? In addressing this question, it helps see similarities between text-internal access and speaker access. Text-internal access can be considered a “virtual” form of speaker access. The external narrator also has an accessor space—the encoding situation—which projects onto the text and can be realized within the text when the narrator wishes—in providing explanatory comments or evaluations, in relating something within the text to the current situation, etc. In the encoding situation, it may even make sense to think of the narrator as topic, just as a subjective character is topic in his internal accessor space, or an addressee in the encoding situation could also be topical. As evidence of that, we note that whenever elements of the encoding situation are realized in the text, they are coded as active, just as values for orientation dimensions for the accessed space are active §2.3.1. In the Labov and Waletsky oral narrative 4 Appendix E, for example, the narrator’s explanation in lines 05–08 indicates various elements of the encoding situation as active: you know 05, we call it 06, today 07, I 08. None of these elements has more than minimal coding. It would be possible to say that deictic centers in accessor spaces, whether in text-internal access or in the encoding situation, remain active in their respective accessed spaces. This includes the person of the accessor as well as centers for temporal and locational deixis. So there are strong parallels between superimposed spaces for a discourse unit with the “top” space being an internal accessor space and what could be thought of as superimposed spaces for a discourse unit with the “top” space being the external accessor space, the encoding situation. For written material and long-distance conversations, the decoding situation is separated from the encoding situation, and sometimes the speaker treats the decoding situation as a virtual encoding situation. That is, the speaker virtually inserts himself into the decoding situation, as if he were encoding the message as the addressees decode it. This brings about a virtual reunification of the encoding and decoding situations, perhaps compensating for the separation of speaker and addressees in a physical sense. At any rate, the speaker, or the participant set made up of speaker and addressees, can function as a discourse topic either in face-to-face communication or in a virtually reunified encoding-decoding situation. A common text type for virtually reunified encoding-decoding is, interestingly enough, academic writing. One example of this is found in the following expository paragraph from Unger 1996, which Unger himself analyzes in a later passage in a way that is different from the analysis presented here: Example text 44: Expository paragraph: “Hierarchical discourse structure” Unger 1996:420f. 01 The thrust of the last section was to argue that coherence relations are not part of the structure of discourse. 02 Rather, the perception of coherence relations results from the way relevance is optimized over the discourse. 03 Indeed, it sheds doubt on the assumption that discourse is hierarchically structured beyond the sentence level. 04 However, the claim that discourse segments are units because they are linked by coherence relations of the kind considered in the last section is only one argument in favor of hierarchical discourse structure. 05 Almost more prominent in discussions of discourse structure are views of coherence as linking the parts of a unit for example, clausessentences which talks about a common topic, understood as a set of topical participants, a topical concept, or a topic framework Brown Yule 1983:75, 94–95. 06 It is claimed that topic coherence is reflected in structure, especially in pronominalization structure. 07 It is also claimed that paragraph breaks or, in more general terms, breaks between discourse segments are reflected structurally in the discourse by various means. 08 It is to those issues that we now turn. This paragraph transitions between two sections of Unger’s article, from the preceding discussion of coherence relations to the question of whether discourse is hierarchically structured. It begins with the first issue and uses that to introduce the second one. It could therefore be analyzed with the second issue as a propositional theme “discourse is hierarchically structured beyond the sentence level”. My use of Unger’s wording here is a mere convenience; in general, a theme need not be formally stated in a discourse unit. The paragraph makes various comments about this thematic proposition but stops short of actually arguing for or against it. The schema for this paragraph is apparently something like the following: a doubt has been cast on the thematic proposition by the earlier discussion of coherence relations; b other arguments have been adduced in favor of it, namely b1 arguments involving topic coherence and b2 arguments involving structural signals of paragraph breaks; c these arguments in favor of the proposition are what we will now consider. The identity of the paragraph’s theme—the proposition that “discourse is hierarchically structured beyond the sentence level”—is not immediately apparent at the start, but only begins to be manifest in line 03. In lines 01–03 Unger accesses it in a stepwise fashion see §2.3.1: mention of “the last section” 01 gives access to its “thrust” 01–02, which in turn accesses the propositional theme for the current paragraph 03. When this proposition is mentioned in line 03, it is still not obvious that it is actually the paragraph theme, nor is it obvious where the author is headed in the paragraph his schema. The direction and theme only become clear as the paragraph proceeds. Its transitional function is realized primarily by using the theme of the previous section to access the theme of the new one. What is more relevant for the present discussion is the fact that this paragraph reflects a virtually reunified encoding-decoding situation: Unger joins the addressees in the decoding situation. This is made explicit in we now turn line 08, although the deictic center for the last section line 01 is ambiguous and could instead reflect Unger’s own deixis at time of writing. 112 The topic status of the participant set ‘we’ in this paragraph—indeed, throughout much of Unger 1996 as well as in the present treatment—can be seen not only in the occurrence of we in line 08 as already activated, but also in the possibility of referring to ‘we’ in other places. In line 01, for example, The thrust of the last section was to argue that … could have been stated In the last section we saw that … . There are doubtless good reasons for the existing wording, but the point is that ‘we’ would have been equally accessible there. 113 Example text 44 therefore can be represented with three discourse spaces as shown in Figure 30: 112 In either case, Unger is citing hierarchical discourse structure; see §2.6.6. 113 Later on in the same article Unger 1996:427, Unger cites the earlier paragraph Example text 44 above in order to show that the notion of discourse topicality is not viable: “the paragraph as it stands has two concepts which can be said to be topics: COHERENCE RELATIONS WITHIN DISCOURSE UNITS and COHERENCE WITH RESPECT TO A TOPIC WITHIN DISCOURSE UNITS . It is therefore not one topic-relevant unit.” His argument against topicality here is therefore seen to rest on the assumption that if discourse topicality is a viable notion, then every discourse unit will have exactly one topic, an assumption which the present treatment does not share. encoding situation accessed space ‘we’: speaker + addressees virtual encoding- decoding accessor space together virtually processed material in last section are about to process new material together temporal orientation propositional theme in last section doubt was cast upon it 2 positive claims have been made about it Figure 30: Superimposed spaces for expository paragraph “hierarchical discourse structure” The encoding situation is represented in two forms: a “real” form addressees know that the encoder wrote the text before they process it and the virtual encoding-decoding accessor space. The mental representation corresponds to the addressees’ understanding, and could be different if addressees understand the encoding situation differently. The encoding situation in these diagrams is represented as a type of base space that includes all of the discourse spaces, but actually it would probably best be represented as a top-level accessor space. That being the case, all texts which are perceived as having a speaker or any kind of text producer would have at least two levels of spaces: the accessor space corresponding to the encoding situation and an accessed space. There could also be other intermediate accessor spaces, such as a text-internal accessor space that corresponds to a subjective character or a virtual encoding-decoding accessor space. This means that even “ordinary” texts commonly have superimposed spaces, which raises the question of how any text could be coherent under our definition of coherence §2.2.1. I think that one answer is that coherence is principally about “internal contextualization” what the speaker is talking about rather than with “external contextualization” under what circumstances the text comes about. That is, coherence applies primarily to the accessed space, and then also to any kind of access that is indicated via the text, such as text-internal access. In all examples we have examined, the accessed space is embedded in the lowest-level accessor space, and any multiple accessor spaces are related by embedding as well. Neither of these conditions need apply, however, in order for a text to be coherent. An accessed space could conceivably have two or more accessors, whose accessor spaces would not be related by embedding. In such a situation, however, coherence would simply require that both accessor spaces be in the same text world on some level, and that temporary incoherences be resolved at some point with just-in-time coherence. These requirements are the same as those for any other text. On this view, superimposed spaces need not, in principle, destroy coherence. This does not mean that superimposed spaces are simple. Their conceptual complexity is, I believe, one reason why narratives often have them. Text-internal point of view differs from “ordinary” texts in that a discourse unit has two spaces that are constructed within the narrative , whether the “top” space is presented as fictive via imagining, as in Lord Jim or factual via remembering, as in “Somewhere over the rainbow”, or some mix of the two. The existence of two text-internal spaces in itself makes for more complex processing. Further, there are two sources of access for information in the accessed space: speaker access and text-internal access. Keeping those sources of access straight involves higher processing cost; not only do addressees need to know “who did what to whom,” but whose epistemics is responsible for what information. The fact of two accessors can, of course, be exploited for literary effect: by presenting a subjective character in first person, the narrator encourages addressees to imagine him as the narrator. “Easy comprehension practices” ask that the distinction between accessors be made clear; literary techniques may have reasons to leave it vague. 114 For such reasons, text-internal access is fundamentally more complex than pure speaker access. The complexity of a text can increase when the accessor space and the accessed space are both prominent, since a particular referent can be interpreted as being in either. Psalm 7 is a typical example of this, but it is not unusual or extreme: this kind of complexity is common in Old Testament material. Example text 45: Psalm 7 New American Standard Bible. 1995 1 ‘O LORD my God, in You I have taken refuge; Save me from all those who pursue me, and deliver me, 2 Or he will tear my soul like a lion, Dragging me away, while there is none to deliver. 3 ¶O LORD my God, if I have done this, If there is injustice in my hands, 4 If I have rewarded evil to my friend, Or have plundered him who without cause was my adversary, 5 Let the enemy pursue my soul and overtake it ; And let him trample my life down to the ground, And lay my glory in the dust. Selah. 6 ¶Arise, O LORD, in Your anger; Lift up Yourself against the rage of my adversaries, And arouse Yourself for me; You have appointed judgment. 7 And let the assembly of the peoples encompass You; And over them return on high. 8 ¶The LORD judges the peoples; Vindicate me, O LORD, according to my righteousness and my integrity that is in me. 9 O let the evil of the wicked come to an end, but establish the righteous; For the righteous God tries the hearts and minds. 10 My shield is with God, Who saves the upright in heart. 11 God is a righteous judge, And a God who has indignation every day. 12 ¶If a man [this NP is not in the Hebrew] does not repent, He will sharpen His sword; He has bent His bow and made it ready. 13 He has also prepared for Himself deadly weapons; He makes His arrows fiery shafts. 14 Behold, he travails with wickedness, And he conceives mischief, and brings forth falsehood. 15 He has dug a pit and hollowed it out, And has fallen into the hole which he made. 114 In Piñon 1984, as discussed by Mey 2001:789–795, in the last sentence of the book the first person subjective character is fictionally represented as being the real narrator …tomorrow I will start to write the story of Madruga. 16 His mischief will return upon his own head, And his violence will descend upon his own pate. 17 ¶I will give thanks to the LORD according to His righteousness, And will sing praise to the name of the LORD Most High.’ In this text both levels are highly prominent: the “top” or access level, that is, the encoding situation, which is for much of the psalm a prayer to God; and the “lower” or accessed level, dealing with a particular situation the psalmist is describing, involving an enemy sometimes plural enemies. The indicated paragraphs follow the NASB, which for present purposes I will follow. The first three paragraphs vv 1–2, 3–5, 6–7 have prominent vocative and second person references to God, whereas in the remainder of the psalm references to God are in the third person, with the exception of a brief intercalated prayer vv 8b–9a. Whereas in vv 1–7 the encoding situation prayer is highly prominent, in vv 8–17 it is much less so, but it still remains close. Since God is a participant in both the access space as addressee and in the accessed space as the righteous judge, he is potentially a second or a third person referent at any point. This in itself is not remarkable, but the apparent ease with which the Hebrew writer switches between second and third person—which may seem strange to us—is a linguistic expression of the easy switching of attention and conceptual interpenetration between the encoding situation and the accessed situation. 115 The easy interpenetration between accessor and accessed spaces is itself just one manifestation of the more general phenomenon of seemingly unannounced changes involving the accessor space. • In Psalm 14 the topic of the accessed space is the wicked except for the last verse of the psalm, v 7. In v 5, the psalmist invites his addressees to join him as virtual accessors in viewing the topic: ‘There they are in great dread….’ In the following verse v 6, for a single time in the psalm, the psalmist addresses the wicked in the second person: ‘You would put to shame the counsel of the afflicted….’ This modifies the accessor space by “pulling up” into it the accessed topic, to function as addressee. • In the 9 verses of Psalm 20 the accessor space has three distinct configurations: vv 1–5 addressed to the king, vv 6–8 addressed neither to the king nor to God, v 9 addressed to God. • In Isaiah the first person singular is often used in a formally ambiguous way: it is not clear whether it refers to the prophet in the accessor space or to another speaker being cited as having his own accessor space within the prophet’s accessed space. In 61:1, for example: ‘The Spirit of the Lord GOD is upon me, because the LORD has anointed me to bring good news to the afflicted….’ That is, there are seemingly unannounced switches between accessor spaces on the two levels. Because of modern literary techniques, the Western reader is accustomed to unannounced changes in the accessed space, although he can still sympathize with the Ethiopian eunuch’s question in Acts 8:34: “Please tell me, of whom does the prophet say this? Of himself, or of someone else?” But unannounced changes in the accessor space are predictably more difficult, since that is the locus of the epistemics of the text.

3.7 Universals of topicality