Methodology Ontological Nominalization .1 The Texts

Burmese forms and that superordinate forms do not process with the same structural assumptions. The texts that were selected are from different social situations, one is a text that has survived through generations of policy changes. It has remained as the third standard grade reading primer since the early 1950s. The story is about how National Day got its name. As a text, it has many purposes but in overall structure it is an expository text. The second text is a brief newspaper article about a boy who was bitten by a snake. The structure of this text overall belongs to the narrative genre. These are referred to here as National Day ND and as Snake Bite SB. The full texts can be found in the appendices, in English, an interlinear version, and also the full analytical display for each part and portion of each text. National Day consists of 480 words in 18 sentences, and 7 paragraphs. Snake Bite consists of 415 words in 6 sentences, and 5 paragraphs. National Day was written in the 1950s and Snake Bite was written in 1995. Both are in standard Formal Burmese. The intended audiences are quite different, one is for schoolchildren learning to read but also learning about the history of their own country. The other is written for adults and tells about something that happened to a boy. In both texts the author is neither a participant in the events explained nor an immediate observer of them.

5.2.2 Methodology

The methodology employed was simply to start at the beginning of the text, the title or byline, identify the form class, role, and structural relations of each word. As noted in chapter 1, the structure of Burmese is bifurcating, with a strong role for zero derivational juxtaposition in organizing structure. Higher level nominals resulting from the structural descriptions were then adjoined on the basis of two different criteria: 1 ontological structural assumptions and 2 logical-rhetorical roles of information. A machine could not precisely do this process, but it could come quite close, given a detailed lexicon and some sense about what the sentence was meaning as it proceeded in identifying constituents. Particles are usually boundary markers for higher level ontological nominals. In fact, anything that is farther to the right, moving toward the verb follows a pathway of being a constituent of an increasingly higher level. This produced a prominent structure with left branching trees. See appendix D. The simplicity of the rules and the ordering results in two kinds of text structures—ontological and logical rhetorical—that make sense of post-verbal particle roles. These structures, most of all, make some ordered clarity from the complexity described in chapter 1. It is assumed that anyone with a moderate knowledge of Burmese does something like this cognitively, particularly as a reading strategy. Since the approach suggested here for specifying grammatical forms of a Tibeto- Burman language is fairly innovative, it may be helpful to provide more detail on methodology and where the system faltered in the application of the above rules to written Formal Burmese texts. The first place it faltered was where I, a non-native speaker, was unfamiliar with a word. The method as currently defined requires that there be a lexical meaning or a grammatical function for every syllable. The particular circumstances where this is not the case were described in chapter 1—loan words or accretion and depletion for which an independent sense is now lost. Native speakers of Burmese would object to their language being ripped asunder into small pieces. Each syllable having a function or lexical meaning does not make sense for a non-analyst. Observing this method of segmentation, a native speaker might vaguely sense the meaning elements in a lexical compound, but the process of identifying them could be difficult. This may be due to conceptual blending, or formal semantic blending, or mere conventional use that reinforces the conceptual boundary of the compound. To force it apart is to act as an ignorant outlander. This inoperability may be one reason why Minn Latt 1959 classified word forms as “minimal primary” and “primary,” the minimal primary being the level used in this methodology, but the primary level that which is most perceptible and manipulable. Another troublesome aspect was occasionally encountered when determining the basic lexical category. The question for Burmese was whether to list the primary category for the form or to take a Burman point of view which would say, for instance, of the verb in \rdhkopf mrou. sac city + new = ‘new part of the city’ referential or again, in a form like ausmif:uav: kyaung: ka.le: school + little = ‘pupils’ that the verbs are nouns because the whole is an inseparable noun. Another ambiguity was in constituency. Recall that in chapter 1, a rather unusual statement was made about directionality of constituency—grammatical constituency moves toward the right, toward the verbal, semantic constituency flows back the other direction much like a tide. This particular problem arose quite often where, without the subsequent string, constituency was clear until introduction of the following noun or verb. The added element would trigger a shifting weight that shook the tree structure. This reverse flow is due to the headedness factor. It is due to semantics and not grammar. It is the “pull” of the head toward semantic constituency, probably the result of the propensity for blending of some type when likely nominals are juxtaposed. An example of this process from the National Day text ND8.2: 103 \refrm w rsdK : mranma ta. myui: Burma 1 kind ‘Burma one kind’ would mean “Burmese or Burma, they’re something else” The constituency rule fits perfectly. The first noun mranma is followed by a numeral and its classifer. The compound numeral word ‘one-kind’ is conjoined as the head with modifier mranma. The result is ‘one kind of Burma’. This fits the normal grammatical pattern for numerals: ‘two houses’, ‘four books’, and so forth will have this type of construction [N [Numeral + Classifier] ]. Just as particles are generally heads, so are classifiers in these constructions. All was well until the next word was added om: sa: ‘son’ or ‘offspring’. The semantics began to realign the constituency structure because myui: sa: kind + son is a compound that means ‘national’, ‘citizen’, ‘one who belongs to the race’. So, the problem was not just the local semantic pull of a semantic head, but to the nature of the text itself. The text is about National Day, which in Burmese is called rsdK:om: myui: sa: ‘kind + son day’. This compound is tightly blended semantically and pragmatically. Thus, what was a regular, left-branching structure grammatically was reconfigured semantically into right-branching. Another case is in the Snake Bite text, SB6, which was the example used in figure 19 and figure 22. Figure 45. Grammatical versus semantic organization In figure 45, the verb ‘die’ ao se is followed by a nominalizing particle r‘ hmu., establishing a noun ‘death’, which is then followed by a repeat of the same verb ‘die’ ao se. This establishes the beginning of a doubling pattern so common and widely used in Burmese, of V 1 X V 1 X with an identically repeated first member and a different but related second member set. The second pair forms a compound verb ao se cif: hkang:, while the first pair forms a noun, and together N + V a predication forms a nominal which means something similar to a clause ‘death arrange’. This entire nominal is then adjoined to the following nominal derived by t- a- prefixation. This derived nominal becomes the head of the compounded Word. The sense of this is regular and meaningful. The last head means a ‘matter’ t|m a-ra, so that the meaning is ‘a matter of arranging for death, or about a death’. All is well. Then the following nominal pulls in the last derived noun ‘a matter’ to its sphere of semantics changing the constituency. The pull is strong because the meaning of the following noun contiguous to ‘a matter’ is a form that has grammatical, lexical and phonological similarity to ‘a matter’. That word ta|: a-re: means ‘a case’ or ‘ a legal matter’. It is also a derived noun with the t- a- prefix. Phonologically, the initial consonant is identical with the previous nominal ‘a matter’. The only difference between the second member phonologically is the vowel and tone. The pairing of these two forms together creates a preferred phonological group a four-syllable rocking sound with the structure A 1 Y A 1 Y so that the result would be a series of rocking syllables establishing a rhythm of two pairs of four syllables in a row V 1 X V 1 X + A 1 Y A 1 Y which is the doubling pattern, and doubling of doubles, so attractive in Burmese. In the end, the conspiracy of the semantic and phonological draw is to take immediate constituents from their normal grammatical structure and move them into patterns of semantic and phonological enablement, in particular semantic compounding and phonological doubling. Now, regarding the methodology, it is clear and easy to apply. After the grammatical ordering is established, other subsystems operate to alter the initial grammatical constituency relations. The grammatical form yields to the semantic and phonological reanalysis of constituents, which is a further example of why categoriality has been such a problem in this language see the discussion in 3.4. Although it is difficult to construct a grammar that excludes semantic roles, this is what has been attempted here. Simple though it may be, it is helpful in appreciating grammatical and ontological structuring processes in Burmese and keeping separate the grammatical form from the semantic operations and roles. Nominalization is the principle key to understanding grammatical formation as a separate process from the semantic. Now to address the second question: Does grammatical form, as shaped by nominalization at the Word and Expression levels, extend insightfully to the text as a whole? Can Longacre’s 1996 model of text structure be correlated with the ontological forms generated by nominalization processes?

5.2.3 Text Analysis