Exploring the Textual Metafunction in a Biblical-Fictional Parable in Ngũgĩ wa Thiong’o’s Devil on the Cross

This paper explores the textual metafunctional patterns in two fictional versions of a Biblical Parable in Ngũgĩ wa Thiong’o’s novel Devil on the Cross (1982) in order to find the similarities and differences between their compositional features. It critically draws on the theory of systemic functional linguistics expounded by specialists like Halliday (1994), Halliday and Matthiessen (2004) and Eggins (2004) to overview the theoretical background to the study, with focus on the grammar of textual meaning, proposes a new Theme classification and uses it to analyze the Thematic structure and taxis system of the two extracts. The researcher has come to the conclusion that, though they are initially spoken, these texts show such interesting textual-meaning properties as the density of ellipsis, of circumstantial and interpersonal thematization, of taxis and rank shift that they should be qualified to belong to both spoken and written mode of discourse.


Introduction
It is well-known that the Kenyan writer Ngũgĩ wa Thiong'o heavily draws on the Bible for his literary composition, namely in terms of characterization by description and speech (Ngara, 1985;Amoussou, 2011Amoussou, , 2015. Indeed, in most of his novels, the writer enrolls his characters after figures from the Bible and Kenyan politics by making them talk or think the Bible without any forewarning to the readers. This is exactly what happens in his fifth novel Devil on the Cross (1982) when two fictional characters, a socialist-inclined one (text1) and a capitalist-inclined one (text2), draw on Matthew 25: 14-46 to speak, first to a group taxi-boarders and then to an assembly of 'robbers' or 'capitalists'. While text1 is a near word-for-word reproduction of the source text, text2 is a fictional adaptation of it.
This article analyses the Theme patterns and logical relations in the two texts in order to reveal their major characteristics. Such an endeavor falls within the realm of the grammar of textual meaning/THEME. Following the tradition of the Prague School linguists, the word 'Theme/Thematic' is written with a capital initial as a label for 'the textual functional constituent' (Halliday & Matthiessen, 2004: 64) but it is fully capitalized when it refers to the whole grammar or when used to display the 'Thematic structure' of clauses (Eggins, 1994).
In addition, there is some confusion from one scholar to another regarding the 'Theme-Given' hierarchy and the 'Rheme-New' one. For instance, Halliday and Matthiessen (2004: 93) argue that 'the Theme falls within the Given and the New within the Rheme" while Eggins (2004: 229, 300) contends that 'the Theme typically contains familiar, or 'given' information and that "the Rheme typically contains unfamiliar, or 'new', information" (my emphasis). Even the conceptual definitions cannot at times help to clear the fog. Such definitions of 'the Theme' as 'the point of departure of the sentence', 'the left-most constituent of the clause or sentence', 'the starting point of the utterance' and 'the constituent which is 'the concern of the message' (Halliday, 1967: 212;Brown & Yule, 1988: 126, 135;Halliday & Matthiessen 2004: 55;Eggins, 2004: 136 ) fail to define the boundary of the Theme, especially when the concept of 'left-most constituent' becomes difficult to apply to such functional morphemes as conjunctions which often occur in initial positions to play a more cohesive/non-structural function between clauses than a structural/grammatical one within a single clause (Eggins, 2004: 47-51;Halliday & Matthiessen, 2004: 538-59). There also is no way to clearly cut the Theme from the Rheme in this definition by Halliday and Matthiessen (2004: 64): "the Theme is the element which serves as the point of departure of the message; it is that which locates and orients the message within its context....The Rheme is the remainder of the message, the part in which the Theme is developed".
It thus emerges that determining what stands for the 'Theme' is not an easy task. Fortunately, Halliday (1994: 53) offers a definition that can serve as the linchpin for Theme analysis: "the Theme extends from the beginning of the clause up to (and including) the first element that has a function in transitivity". In other words, Halliday and Matthiessen (2004: 79) specify that: "the Theme of a clause ends with the first constituent that is participant, circumstance or process. We refer to this constituent in the textual function as topical theme". This means that a transitivity-labeled item is the obligatory constituent in the structure/composition of Theme: "every clause must contain one and only one topical Theme" (Eggins, 1994:277;2004: 302).
Part of the confusion is due to the fact that constituents with different lexico-grammatical labels (experiential, interpersonal, and textual) can function as Theme, giving rise to three major types of Theme: topical, interpersonal and textual Theme. The topical Theme is the constituent which serves as the departure point and to which a transitivity label (participant, process, or circumstance) can be assigned (Halliday, 1994;Eggins, 1994Eggins, /2004. The interpersonal theme is the one to which a mood label (subject, finite, vocative…etc) can be assigned while the textual theme is the one bearing neither transitivity nor mood label, but which does a rather cohesive work by relating a whole clause to its context. It must also be pointed out that, in addition to these three general categories of 'Theme', other specific types can be distinguished. For example, a topical Theme can be 'unmarked' or 'marked'. A topical Theme is said to be 'unmarked' when it is used in its normal or expected subject position. However, when such an item occurs in a position that is not normally its, then it becomes used as 'a marked topical theme'. Even the other two types of Theme (interpersonal and textual) can be marked. In (1b), (2b)and (3b) in Table1 below, the item 'in this country' is a circumstantial adjunct used as a 'marked topical Theme', 'never again shall..' are modal locutions used as 'marked interpersonal Themes' and 'thus' is a conjunction used as a 'marked textual Theme'. (6) (a) I shall never bow down to the lifeless god of money again.
(b) Never again shall I bow down to the lifeless god of money.
(7) (a) I was thus left the only child of my mother.
(b) Thus was I left the only child of my mother.
There also is what is known as a 'multiple Theme'. Indeed, Eggins (2004: 307) argues that when several textual Themes or/ and interpersonal Themes occur before the obligatory topical Theme, the combination is known as a 'multiple Theme'. Distinction should also be made between 'a textual Theme' and 'a structural one'. While the former is generally a coordinating conjunction (and, then, but, yet, however, etc.), a subordinating conjunction (when, before, after, how, etc) or a continuity adjunct (oh, yes/yea, no, well, etc.), the latter has to do with the use of relative pronouns such as 'who, which, that, etc.' to introduce embedded clauses.
The foregoing entails that exploring the textual metafunction should involve, in addition to the analysis of Thematic and Information Structures, that of taxis and rankshift, as Matthiessen, Teruya & Lam (2015: 221) argue that the textual metafunction includes such systems as "THEME, INFORMATION, CONJUNCTION, SUBSTITUTION-ELLIPSIS, REFERENCE and LEXICAL COHESION" (authors' capitals). In fact, the tactic system describes the kind of interdependency relationship between clauses linked together into complexes. This involves parataxis/coordination, when clauses are related as equal, independent entities, and hypotaxis/subordination, when clauses relate to a main clause through a dependency relationship (Eggins, 2004: 258). Interestingly, Halliday (1994: 224) describes parataxis and hypotaxis as "the two basic forms taken by logical relations in natural language". The tactic system is somehow opposed to rankshift/embedding. While taxis, through coordination and subordination (addition, contrast, variation, temporality, causality, etc.), involves expansion, rankshift implies compression, packing more meanings into units by bundling a whole clause into a unit of a lower rank (Eggins, 2004: 269).
Finally, it is worth noting that Thematic Structure analysis is particularly interested in such deviations from the normal patterning that convey particular messages. A key step in the exploration of the textual metafunction is thus to observe such deviations in theme-patterns as the inversion of word-order, logical connectors and the construction involved in the hierarchies of clauses, all of which must serve some aesthetic function such as emphasis or explicitness, or their opposites -the aesthetically justified blurring of distinctions or obscurity. As Martin (1992: 12) puts it, "the different patterns and meanings made by the choice of Theme can be manipulated and exploited, consciously or subconsciously, by the writer in order to convey their 'angle' or viewpoint'".

Methodological Perspective
For the purpose of this analysis, each of the two texts is divided in its constituent clauses. Each clause is numbered and each Theme-type is underlined, labeled and quantified (see the appendix). Plain numbers -i.e. (1) (2)-show ranking clauses while these numbers followed by a dot and other numbers -i.e. (1.1) (2.1.2)-indicate rankshifted clauses, a rankshifted clause being a complete clause carrying out the function of a noun phrase or of just a word acting either as complementiser or modifier within the noun phrase. In addition, as different lexico-grammatical labels (experiential, interpersonal, and textual) can coexist within the Theme, I propose the structure-oriented classification summed up in table 2: This classification offers a big advantage: it makes Theme-identification within the clause less cumbersome, as it helps to avoid using many theme-labels in the same clause. In addition, it helps to clearly show the hierarchy between the clauses: classes (b) (c), and (d) reveal clause coordination and subordination while classes (e), and (f)) shed light on rankshift in terms of complementization and relative definition. It must be noted that some letters are followed by (+); this signals markedness. In addition, while Eggins (2004: 311) considers the particle 'let's/let us' as 'a topical Theme', it is taken as 'an interpersonal One' in this research not only because the real 'verb/process' is the base-verb after the particle, but also, especially as it is used as a synonym for 'shall we....?' Indeed, the combination 'let/may +topical theme/subject +base verb' is used to express theoretical assumptions in mathematics or hope/wish in Biblical invocations (Genesis 1: 3,6,9,14,20,24,26,etc.;Psalms 25: 21;31: 17,18,21;32: 8;33: 22;34: 3;35: 4,5,8,25,26,27,etc.). Moreover, Halliday and Matthiessen (2004: 513) consider the use of 'let' in such a causative-like structure as a form of verbal modulation. The same applies for such praise-laden phrases like 'long live' (i.e. may...live long) in text2, which are thus treated as class (c) Theme: 'interpersonal +topical items'.
The first section of the analysis looks into the Thematic structure in terms of Theme classes and structures while the second deals with Taxis and Rankshift.
While this significant occurrence of elliptical topical and interpersonal Themes can well give the impression of a carefully written mode, the rate gap between the two texts in the density of these also implies that text1 is in a more carefully written more than text2. This means that even though the writer draws on Bible discourse, he somehow adjusts the mode character-wise or register-wise to reflect linguistic differences. Indeed, both the imaginary and real speaker/Lord in text 1 is Christ while real speaker in text 2 is a robber, speaking from the point of view of an imaginary Lord, a colonial capitalist, a master robber.
The foregrounding of circumstantial elements in the texts has three possible interpretations among others. First, it stresses not only the importance of the situational context in the description of participants and processes but also the speaker's or writer's psychological priority to this (Halliday & Matthiessen, 2004: 55;Eggins, 2004: 136). This thematization of circumstances is perhaps the best illustration of the view of the Theme as 'the psychological subject' (Halliday & Matthiessen, 2004: 55;Eggins, 2004: 136). Secondly, this circumstantial markedness appears as a realization of a carefully-written mode as the writer must have planned the rhetorical development of the text to achieve this level of foregrounding (Eggins, 1994: 319). Finally, Eggins (2004: 339) argues that significant thematization of circumstantial adjuncts shows that the information expressed in both texts is presented as non-arguable, a strategy to express the speaker's authority on the addressee. Eggins (2004: 315) suggests two levels of analysis for subordinate clauses in frontal position: first a clause-by-clause analysis in which each is taken to have its own Thematic structure; then a second level of analysis in which the firstplaced dependent clause serves as the Theme for the whole complex, the main clause serving as the Rheme. The first level of analysis is thus already carried out in treating such foregrounded subordinate clauses as '51, 89, 114' (text1) and '7, 13/14, 25, 27, 43, 93, and 164' (text2) as any other clause in terms of their Thematic structure. At the second level, clause '51' serves as 'the Theme' for '52' which is 'the Rheme', just as '89' does for '90' and '114' for '115' (text1). In text2, the same applies for '7-8', '13/14-15', '25-26', '27-28', '43-44', '93-94', '164-165' and '202-203'. These clauses must thus be seen to play the same function as marked circumstantial elements in simple clauses (Eggins, 2004: 315). Table 4.1 shows the two levels of analysis for '89-90' (text1) and table 4.2 does for same for '43-44' (text 2), the same thing being applicable for any of the pairs listed above. textual Topical topical THEME RHEME THEME RHEME THEME RHEME textual Topical topical THEME RHEME THEME RHEME THEME RHEME Themes of classes(c) and (d) prove to be very revealing as to tenor in both texts, namely in text1. Indeed, most Themes of class (c) have a vocative adjunct, i.e., the item 'Lord' as the interpersonal Theme (19, 27, 35, 42, 76, 78, 80, 82, 84, 85, 106, 107, 108, 109, 110, and 111). Here, it is useful to point out the combination 'vocative +topical theme' 'Lord, when' used in (76) and (106) covers respectively '78, 80, 82, 84, 85' and '107, 108, 109, 110, 111', which is why they are taken as belonging to the same class, despite the initial conjunction 'or' in the last five cases. The other Themes of this class have a mood adjunct as the interpersonal part of the Theme: 20, 22, 28, 30, 39, 89, and 114, even though there is a combination of a mood adjunct and a vocative in (30). All class (d) Themes are in the form of 'then+ shall... ' (54, 60, 75, 105, 113, and 114). The foregrounding of this strong modulator not only reinforces the third interpretation earlier provided for circumstantial markedness, but also the master-servant relationship between the speaker and his listeners, on the one hand, and the imperativeness of the judgment given to both the lawful and unlawful servants, on the other.

Analysis of Taxis and Rankshift in the Texts
-temporality (when, until, after, before)

60
It must be noted that 103 logical relations have been identified in Text1 while Text 2 contains 181 such relations. As can be seen in table6, the most frequent relation between the clauses and sentences in both texts is that of coordination or parataxis -[text1: 65.05%); text2: 44.20%] -, this entails that sequences of clauses of similar status or equal importance linked either through an additive relationship (and, then), or a contrastive one 'but, however, yet' or variative one 'or, instead' or a causal one (so, therefore, thus, then'. Additive conjunctions appear as the most dominant in this paratactic system, accounting for 51.46% of use in text1 and 37.02% in text 2. This involves that little contradiction is involved in the line of argument, even though all these conjunctions, be they (additive, contrastive, variative or causal) contribute to the logico-semantic expression of expansion (Halliday & Matthiessen, 2004). In addition, the pervasive use of the conjunction 'and' to link main clauses in both texts gives the impression of a recitation of events in the structure of a list, as much as a child might (do) in telling a story (Cummings & Simmons, 1983: 93), even though this use is more intense in text1 than in text 2, given the rate gap between them. Interestingly, there is a case of Theme-reiteration (10a+b) in text1, which is considered as a feature of rapid conversational speech (Eggins, 1994: 289;2004: 312).
The use of hypotaxis is higher in text 2 than in text 1 -[text1: 13. 59%; text2:22. 65%] -This means that the second text attaches twice more dependent clauses to main ones in a relation of unequal status than text1. However, even though the dependent clauses appear to be structurally of lower status to the main ones, they help to expand their meanings by adding dimensions of time, place, causality, conditions, etc. to them. In fact, in most cases, the removal of the subordinating conjunction and the placement of a comma or period would make a main clause and its subordinate one appear as two independent clauses, even though the logico-semantic relationship between them would become fuzzy in the absence of the conjunction. Adding parataxis to hypotaxis, it can be said that text1, with a tactic density of 78.64%, is tactically denser than text 2 which has 66. 85 % of this density.
In addition to taxis, there is a relatively significant occurrence of rankshift in both texts, which indicates a greater sense of dependency, hierarchy and value within some of the clauses. Just as for hypotaxis, text2 has a higher rate of rankshift than text1: [text1: 21.36%; text2: 33.15%]. It must be noted that the rates if rankshift in both texts should normally be equal to the sum 'e+ f' (table3) if the total number of clauses (t1:140; t2: 277) were considered instead of that of logical relations (t1: 103: t2: 181) considered in table6.
In general, the combination of parataxis, hypotaxis and rankshift leads to complex grammatical structures resulting into the creation of depth, as Cummings & Simmons (1983: 140) claim: The number of nodes in a sentence is the measure of its depth, i.e., the complexity of its articulation … the more divisions in a unit complex, and the more rankshift; the more nodes occur and the greater is the depth. If the situation is a serious one, the fact that depth is intellectually demanding, and hence slows the progress of communication helps to reinforce the considered and weighty nature of what is being said … it helps to convey a serious and thoughtful tone.
This means that sentence depth is equivalent to the level of its internal grammatical complexity/articulation and is associated with the seriousness and intellectual respectability of the topic or situation. As a result, text 2 has more depth than text1. In addition, Eggins (2004) contends that taxis is considered to be more dynamic -it involves little forward planning as the speaker can simply chain on another unit of the same type -and rankshift is more static as it requires some forethought in the construction of the clauses. That is why taxis characterizes spontaneous, spoken language or informal written texts while rankshift is associated with formal, careful written texts. In this regard, Halliday (1994: 224) writes: The clause complex is of particular interest in spoken text, because it represents the dynamic potential of the system -the ability to 'choreograph' very long and intricate patterns of semantic movement while maintaining a continuous flow of discourse that is coherent without being constructional. This kind of flow is very uncharacteristic of written language".
As it appears, though both texts are spoken, they exhibit features of both spoken and written mode. However, the higher density of hypotaxis and rankshift in text2 entails that, though the writer has drawn on a more spoken mode, he has twisted it toward a more written-like one by strengthening this feature of written-ness (Cummings & Simmons, 1983;Halliday, 1994).

Recapitulation and Conclusion
This paper helps to reveal a few important things about the grammar of textual metafunction. First, it appears the most difficult to apply as such features as ellipsis can make Theme identification a hard labor. This is probably why less research is carried out in this area at the Maîtrise-level in our university (Benin, West Africa). However, it is a worthwhile enterprise as it helps to gauge/challenge the researcher's knowledge of structural intricacies. Secondly, the concern of 'Information Structure Analysis' with intonation patterns makes it difficult to fully apply this grammar to such written-down-to-be-read texts as the present ones.
Likewise, while both texts exhibit significant uses of taxis and rankshift (table6), there is a significant difference in the use of parataxis (65.05% vs. 44.20%), hypotaxis (13. 59% vs. 22. 65% and rankshift (21.36% vs. 33.15%) (table6). While the first gap is earlier interpreted as an orientation of text1 toward more spoken-ness than text2 (Cummings & Simmons, 1983;140;Halliday, 1994: 224), the second one can be in terms of the 'main/dominant' versus 'dependent/subordinate' hierarchy. As a fact, the lesser use of hypotaxis in text1 may be due to the fact that Christ holds a near 'equal-to-equal' tenor with his disciples while the capitalist stresses more the 'master-slave' one to his servants. As for the gap in the use of rankshift, it may be interpreted, on the one hand, as the capitalist master's stronger emphasis on the God-ordained perfect system of unequal master-slave tenor, and on the other, as the writer's inclination toward written-ness.
On the whole, by showing that both texts blend features of both spoken and written modes, this study confirms Bahktin's (1981: 262-3) view of the novel as kind of dialogism -a reproduction of the intrinsic cultural and ideological heteroglossia of language, as he writes: The novel can be defined as a diversity of speech types (sometimes even a diversity of languages) and a diversity of individual voices, artistically organized…..These distinctive links and interrelationships between utterances and languages, this movement of the theme through different languages and speech forms, its dispersion into rivulets and droplets of social heteroglossia, its dialogization -this is the basic distinguishing feature of the stylistics of the novel.
Finally, this exploration into the textual metafunction has yielded quite interesting insights into the mode variable encoded in the texts, as one of the interpersonal metafunction is likely to yield into the tenor variable, an option for future research. At any rate, the writer's adoption of Bible topics and speech has not prevented him from manipulating Thematic structure and Taxis to convey the different ideologies expressed by the two characters or imitators of Bible discourse in the articulation of their political beliefs (Martin, 1992: 12).  [and they (b) seize the key of the country], (15) [I (a) shall never be able to repossess these plantations and industries]. (16) [This tea, this rice, this cotton, this coffee, these precious stones, these hotels, these shops, these factories, these fruits of their precious sweat -these and more ( [and I (b) shall be able to plant seeds] (21.1) [that (e) will take root more firmly] (21.2) [than the ones (f) I planted before].  [and especially (he (c) taught them) the trick of sprinkling theft and robbery with the sweetest-smelling perfumes, and the trick of wrapping poison in sugar-coated leaves, and many tricks for dividing the country's workers and peasants through bribery and appeals to tribe and religion]. (56) [And it (b)    [who (e) has the audacity to reject slavery]. (194) [Throw (a) him into jail or into everlasting darkness], (195) [so that his family (b) will harvest only tears and the gnashing of teeth!] (196) ['Good! Good!You people (c) have done a fine job]. (197) [Mete out (a) the same treatment to all such rebels], (198) [so that the other workers (b) will be too scared to strike for higher wages or to take up arms to smash the chains of slavery.]'