2. Format and transcription.

I have numbered the texts consecutively from 1 to 171. All texts are divided into paragraph-type subdivisions (though some have just one such subdivision), numbered 1.5, 171.1, and so forth. Such subdivisions are purely for convenience of reference and do not necessarily correspond to inherent structural divisions in the narrations. Occasionally a cross-reference like 1.5.2 is used; the last number indicates which line of the subdivision is intended.

A paragraph translation is provided after each subdivision. All texts have interlinear word translations directly under the Nunggubuyu text. In addition, the first six texts only have a morpheme-by-morpheme gloss (with abbreviations for most affixes) between the Nunggubuyu text and the word translations. Readers should be cautioned from the outset that the morpheme-by-morpheme glosses will be difficult to understand for readers who have not yet seen the forthcoming grammar. However, a reasonable understanding of the flow of discourse and the nature of word meanings can be gotten from the interlinear word glosses, and broader patterns of syntax, discourse, and narrative structure can usually be obtained from the somewhat freer paragraph translations. A list of abbreviations, most of them used in the morpheme-by?morpheme glosses, is provided separately.

In the Nunggubuyu transcriptions, morpheme boundaries are shown by the hyphen -. In addition, a boundary = is given in verb forms (and predicate adjective forms with verbal rather than nominal prefixes) directly before the main verb (or adjective) stem. I do not suggest that = differs from - in any significant way (e.g. phonologically); instead, the use of = is simply designed to make it easier for readers to identify the main stem. I hope this will help readers analyse the word structure, but more importantly it will facilitate identification of stems so that readers can find them in the dictionary and obtain more information about the range of forms and meanings which they (the stems) have. I do not do this with (nonpredicative) nouns, demonstratives, etc., mainly because in most cases these words have relatively transparent morphological structures. In a verb form like n^ga-n^gu=bura-n^ga-n^g 'I will sit down', the location of the stem would not be immediately apparent without the = boundary (especially since =n^ga-n^g also happens to be a verb form meaning 'will eat' and could be confused with the suffix complex -n^ga-n^g in n^ga-n^gu=bura-n^ga-n^g).

For similar reasons, the apostrophe ' has been used at morpheme boundaries to indicate that a morpheme-initial vowel has been lost (by a phonological rule combining two adjacent vowels into a single vowel, usually long). In n^ga-n^gu=bura-n^ga-n^g we do not find this since all morphemes begin in consonants. On the other hand, consider an underlying form /n^gawu=ala-i/ 'I am scratching it'. On the surface this occurs as n^gawa:li:, where the colons indicate long vowels. I write this as n^gawa:='li:-', with apostrophes indicating that morpheme-initial vowels have been swallowed up by the preceding vowels. Without the ' symbol, readers might think that the stem here was /=li-/ or /=la-/, but the ' tells them that the stem must be something like /=vlV-/ with an initial vowel. With a little experience, readers will learn to guess that /=ala-/ is the most probable base form of the stem, since the disappearing vowel at the beginning of a verb stem is usually /a/ (not /i/ or /u/), and since the long i: ending is typical of a large verb class whose stems basically end in /a/.

Vowels are these:

Of these, all but o: are reasonably productive. o: occurs as a contraction of /a-u/ in certain demonstrative forms like o:-' ba-ni /a-uba-ni/ 'that one (ANA class)', and need not be indicated in dictionary entries. It is also used as a trail-off vowel in interjections with final stylistic lengthening, hence baba 'my father', vocative (calling) form babao! or babao:::!.

Consonants:

A fuller description of these sounds will be given in the grammar; a few short remarks will suffice here. b and m are bilabial; the dh-nh-lh series is interdental and strongly laminal (nh is rare); the next column (d, etc.) is apicoalveolar and similar to English sounds (except that r is a tap); the next column (d, etc.) is retroflexed and these consonants are often perceived as having an r-colored onglide (r is an approximant like American English r); j and n^y are laminals articulated in the alveolar or postalveolar region (j is usually not released when syllable-final, unlike English ch as in lurch); g and n^g are velars. Glides (semivowels) are listed in the columns shown for convenience only; morphophonemically there are two distinct underlying w sounds, w₁ (becoming g after stop or nasal) and w₂ (becoming b), but this need not concern us here.

Some phonological rules which may disguise the identity of stems are the vowel-contraction rule illustrated above, and a series of consonantal alternations: b/w (underlying w₂), dh/lh (underlying lh), d/r (underlying r ), j/y (underlying y), and g/w (underlying w₁). In a few stems the alternations work differently (and b, g, dh, and j must be taken as underlying or basic).

Abstract morphophonemes are /G/ and /N/ at the end of certain prefixes. /G/ always disappears but behaves like /g/ (e.g. in causing hardening of following continuants to stops with the formulae just given); /N/ disappears before another nasal consonant and assimilates to the point of articulation of certain other consonants. Readers of the texts need not worry about these symbols unless they are interested in morphological structure.

Nunggubuyu is not especially hard to transcribe, except that vowel length is unstable, often merged in normal speech (at least in some positions), and subject to surface stylistic and intonational skewing. I have normalised vowel-length transcriptions so that, except for extreme stylistic lengthening (shown as:::), a given word is transcribed with the same vowel length pattern in all its occurrences even though the actual pronunciation is variable.

I have chosen to use an orthography which is both linguistically accurate and pedagogically appropriate. All phonemic contrasts are indicated, but the symbols used are all found on ordinary typewriters. In particular, I have chosen n^g for the velar nasal, avoiding both the usual symbol n with its diacritic and the usual "practical" digraph ng, which invites confusion with the cluster consisting of an n followed by a g. Similarly, n^y represents the laminoalveolar nasal, and underlining is used for the retroflexed consonants. In the case of interdentals like dh, the h (really a diacritic) is not raised since there is no true h phoneme in the language; hence dh can only be the interdental form of d and there is no need to raise the h.

This orthographic system is rather similar to those which have been experimented with by literacy people at Numbulwar (Rev. Earl Hughes, Kathie Warren, and Michael Hore). As of last report, the orthographies in use in the local literacy and bilingual programs were still in experimental stages and no final practical orthography had been agreed on.

Slight discrepancies between my orthography and whatever practical orthography is eventually decided on should not cause major problems. It is probable that the Nunggubuyu material in my publications (including this one) will reach the Nunggubuyu people through modified versions prepared by the local literacy people. In the texts presented here I have not "weeded out" false starts, intrusive English words, or grammatical errors by the narrators. In primers and other works used in the Nunggubuyu schools, it will be necessary to eliminate such elements and produce highly edited and normalised texts, and any necessary orthographic modifications can be introduced as appropriate. On the other hand, any Nunggubuyu who has attained literacy in the language should have little trouble working through my transcriptions.