Consonants

^{Footnote 1,Footnote 2}

The examples show the consonants in intervocalic position, which allows the largest number of consonantal contrasts in this language (see section ‘Stress and word structure’ below). The examples provide a phonemic representation, followed by a representation in practical orthography, and the abbreviation for the relevant speaker’s name (see above). Given the nature of the corpus, I do not have sound clips of sufficient quality by one and the same speaker for all of the examples. The same applies to allophonic variation (discussed in the second half of this section and the next), which cannot be systematically demonstrated with alternating tokens from one and the same speaker, and is illustrated with examples from different speakers instead. Whenever a specific variant is idiolectal, however, this is mentioned explicitly; all other variation is to be regarded as general for the different speakers in the corpus.

I first describe the inventory in the context of Australian languages more generally, and then focus on individual (sets of) phonemes, including patterns of allophony. The consonant inventory of Umbuygamu is unusual for an Australian language, for a number of reasons. First, its general architecture deviates markedly from the fairly uniform inventories found in most Australian languages. Australian consonant inventories have generally been described as ‘long and thin’ (Butcher Reference Butcher, Harrington and Tabain2006), combining relatively few contrastive manners of articulation with relatively many places of articulation (see further Butcher Reference Butcher, Harrington and Tabain2006, Fletcher & Butcher Reference Fletcher, Butcher, Koch and Nordlinger2014, Gasser & Bowern Reference Gasser, Bowern, Kingston, Moore-Cantwell, Pater and Staubs2014). Umbuygamu has the ‘Standard Australian’ manner contrasts of paired plosives and nasals for most places of articulation, plus trills, laterals and approximants for a smaller set of places of articulation. In addition, however, it also has a contrastive fricative series, and a voicing contrast for plosives (see further below). Both features are rare in Australia. Contrastive fricative series are found in a number of subgroups of Paman languages in Cape York Peninsula, including Lamalamic (Sommer Reference Sommer1976, Rigsby Reference Rigsby, McConvell and Evans1997) and Northern Paman (Hale Reference Hale1964, Reference Hale1976a), as well as in the Western Torres Strait language just north of the Peninsula (e.g. Hunter, Bowern & Round Reference Hunter, Bowern and Round2011), and in the Daly River area in the continent’s northwest (Evans Reference Evans and Goldsmith1995; Dixon Reference Dixon2002: 605–616). VOT-related contrasts for plosives are found, again, in some subgroups of Paman languages in Cape York Peninsula, including Lamalamic (Sommer Reference Sommer1976, Rigsby Reference Rigsby, McConvell and Evans1997), Alaya-Athima (Alpher Reference Alpher, Verstraete and Hafner2016) and Southwest Paman (Alpher Reference Alpher1972), as well as some other regions of Australia (see Austin Reference Austin1988 and Dixon (Reference Dixon2002: 605–616) for overviews, and Butcher & Reid Reference Butcher and Reid1989 on the nature of the VOT contrasts involved). A second noteworthy feature in Umbuygamu is the presence of a contrastive glottal stop, both in initial position, as in /ˈʔada/ ˈada ‘from the south’ (FB), and intervocalically, as in /ɛˈʔa/ eˈanh ‘plains turkey’ (RL). Contrastive glottal stops are generally rare in Australia, but found in many languages of Cape York Peninsula, as well as Arnhem Land (Evans Reference Evans and Goldsmith1995; Dixon Reference Dixon2002: 615–618). A third unusual feature is the presence of a voicing contrast for alveolar trills, found in no other Australian languages apart from the Lamalamic languages and Yaygir of northern New South Wales (Dixon Reference Dixon2002: 578). A final feature that is typologically noteworthy is the presence of a dental approximant //. This is again rare in Australian languages: it is found as a contrastive element in some languages of Western Australia – Yindyibarndi (Tabain & Butcher Reference Tabain and Butcher1999: 334–335), Bunuba (Rumsey Reference Rumsey, Dixon and Blake2000: 42) and some less well-documented cases (Olson et al. Reference Olson, Mielke, Sanicas-Daguman, Pebley and Paterson2010) – and in Cape York Peninsula as an allophone of a dental plosive in the Southwest Paman language Kuuk Thaayorre (Gaby Reference Gaby2006: 28). This extends the series of approximants found in Umbuygamu to almost the same set of places of articulation found for plosives and nasals.

The plosive series shows a contrast between voiced plosives and voiceless unaspirated plosives (except, of course, for glottal stops). In the Australianist literature, there has been some debate over the status of voicing as a contrastive feature. In some Australian languages, voicing has been demonstrated to be a typical correlate of consonant length, such that the length of the closure phase is the basic contrastive feature, with long plosives typically being voiceless, and short ones typically voiced (Butcher & Reid Reference Butcher and Reid1989, Butcher Reference Butcher2004). In Umbuygamu, by contrast, the basis of the contrast is related to VOT, and there is no correlation with length: consonant lengthening is a general allophonic process associated with a specific consonant position (see below), rather than a contrast restricted to plosives. For instance, in the tokens illustrating plosive contrasts at the start of this section, the voiceless plosives have VOT times between approximately 10 ms and 30 ms, regardless of length (compare long /p/ in /ɔˈpa/ opa ‘palm tree’ (RL) with short /k/ in /ɔˈraŋ ˈmaka/ orrang maka ‘message stick’ (RL)). Figure 1 illustrates /p/ in /ɔˈpa/ opa ‘palm tree’ (RL), with a long closure phase, and a VOT of approximately 15 ms. The figures in this paper were created using Praat (Boersma & Weenink Reference Boersma and Weenink2015).

Figure 1 Spectrogram and waveform for /ɔˈpa/ opa ‘palm tree’ (RL), VOT highlighted.

Within the plosive series, there is one phoneme that is very rare, namely the voiced palatal plosive /ɟ/, found in eight roots, and one phoneme that is relatively rare, namely the voiceless velar plosive /k/, found in 16 roots (versus almost four times as many roots with voiced velar plosives). The rarity of the voiceless velar plosive in Umbuygamu is typologically somewhat unexpected. Cross-linguistically, voiceless velar plosives seem to be favoured over voiced ones (see Ohala Reference Ohala and MacNeilage1983), as reflected, for instance, in the fact that inventories with a voicing contrast tend to lack voiced back plosives rather than voiceless ones (see Maddieson Reference Maddieson, Dryer and Haspelmath2013). In Umbuygamu, the relative rarity of voiceless velar plosives can be attributed to historical developments, specifically the development of velar plosives to glottal stops in intervocalic contexts (e.g. aˈa ‘close’ < Proto-Pama-Nyungan *nyaka ‘here’ (Alpher Reference Alpher, Bowern and Koch2004)), and to voiced plosives in the context of homorganic nasal-plosive clusters (e.g. agarr ‘flesh’ < Proto-Paman *pangkarr ‘flesh’ (Hale Reference Hale1976b)), which worked together to almost eliminate voiceless velar plosives from the lexicon. One final note on the plosive series concerns the glottal stop. As can be seen in Figure 2, the glottal stop is reflected in creaky voice in the surrounding vowels, a phenomenon that is cross-linguistically not uncommon (see Ladefoged & Maddieson Reference Ladefoged and Maddieson1996: 74–75).

Figure 2 Spectrogram and waveform for /ɛˈʔa/ eˈanh ‘plains turkey’ (RL).

The nasal series shows allophonic variation between plain and prestopped nasals. Prestopping is found with all nasals, as illustrated in (1), but always optionally.

1. (1)

   

There is no evidence that prestopped nasals constitute a separate contrastive series, as claimed in Ogilvie (Reference Ogilvie1994: 26–27) and Sommer (Reference Sommer1998: 8, 11). All of the items they present as contrastive with prestopped nasals (e.g. okngal ‘mosquito’ in Ogilvie Reference Ogilvie1994: 26) actually show variation between plain and prestopped nasals in the present corpus.

The stop phase preceding the nasal is consistently voiceless, unlike with prestopped nasals in other Paman languages like Olkola (Hamilton Reference Hamilton1997), and it is usually longer than (or at least as long as) the nasal phase. This is illustrated in Figure 3, where in /aˈmal/ [aˈp^mɐl] amal ‘foot’ (RL), the stop phase is nearly twice as long as the nasal phase. In fact, the relative prominence of the stop phase would justify a phonetic analysis as postnasalized (or nasally released) plosives rather than prestopped nasals (also when compared with other cases of allophonic prestopping in Australia (see Round Reference Round, Pensalfini, Turpin and Guillemin2014, Harvey et al. Reference Harvey, Lin, Turpin, Davies and Demuth2015), where the stop phase is usually quite short).Footnote ³ I transcribe the sounds as such in (1) above and elsewhere in this paper, but given the systematic alternation with ‘plain’ nasals in Umbuygamu, as well as the tradition of work focused on prestopping in Australia (see Butcher Reference Butcher, Ohala, Hasegawa, Ohala, Granville and Bailey1999, Round Reference Round, Pensalfini, Turpin and Guillemin2014), I retain the label prestopped nasals in this description.

Figure 3 Spectrogram and waveform for /aˈmal/ [aˈp^mɐl] amal ‘foot’ (RL).

Prestopping in Umbuygamu is found most typically at the onset of a stressed syllable, less typically at the onset of unstressed syllables, and almost never in clusters. Not all speakers in the corpus show this variation, however: in the older generation, four out of five speakers use prestopped nasals, while in the younger generation only one uses them. It is difficult to interpret this in a straightforward way, but given that the one older-generation speaker who lacks prestopping is described as less than fluent in the recording notes, the distribution suggests that prestopping could be an older feature that is lost among younger speakers.

The fricative series has two phonemes that are reasonably frequent in the lexicon, and two that are rare. The rare fricatives are the voiceless bilabial fricative /ɸ/ and the voiceless alveopalatal fricative /ɕ/, found in four and six roots, respectively. Both also show variation: /ɸ/ alternates with a labiodental fricative [f], as shown in (2) below, and /ɕ/ alternates with an alveolar fricative [s], as shown in (3).

1. (2)

   

2. (3)

   

Footnote ⁴

There are not enough tokens to determine the basis of this variation, although for /ɕ/ it may be idiolectal, as one speaker (NG) almost consistently has an alveolar realization [s]. The dental and glottal fricatives, by contrast, are not rare at all. The glottal fricative shows variation between a voiceless glottal fricative [h], available in all contexts, a voiceless velar fricative [x], available in the onset of stressed syllables, and a voiced glottal realization [ɦ], audible (mainly) in post-tonic positions, as illustrated in (4).

1. (4)

   

In addition to the fricative phonemes, two phonemes in the inventory of Umbuygamu have fricative or affricate allophones. The voiceless palatal plosive /c/ has an (infrequent) affricate allophone [tʃ], as shown in (5).

1. (5)

   

The bilabial approximant /w/ can have a voiced bilabial fricative allophone [β] before a high or mid front vowel,Footnote ⁵ illustrated in (6) (compare also with the verb in (5)).

1. (6)

   

As mentioned earlier, Umbuygamu has contrasting voiced and voiceless alveolar trills, as in /aˈram/ arram ‘wallaby sp.’ (RL) and /aˈam/ arham ‘sandridge goanna’ (FB). Voiced trills are voiced throughout, while for voiceless ones, voicing is suspended for the duration of the trill, as shown in Figures 4 and 5.

Figure 4 Spectrogram and waveform for /aˈram/ arram ‘wallaby sp.’ (RL).

Figure 5 Spectrogram and waveform for /aˈam/ arham ‘sandridge goanna’ (FB).

In the series of approximants, there is one element that deserves further comment, namely the dental approximant //. In earlier work on Umbuygamu, this has been analysed as a fricative, the voiced equivalent to the dental fricative /θ/ (Ogilvie Reference Ogilvie1994: 27–28). However, there is no evidence to analyse this as a fricative, phonetically or phonologically. Phonetically, there are two arguments: first, there is no perceptible friction in its realization, and second, in post-tonic position [] occasionally alternates with a dental lateral [], as shown in (7a–b), which is also an approximant in articulatory terms.Footnote ⁶

1. (7)

   

Phonologically, moreover, the distribution of // is that of an approximant and not that of a fricative. Fricatives (and plosives) never occur in word-final position in Umbuygamu, whereas // frequently occurs in word-final position (nine out the 30 roots with /ð/ in the lexicon), as shown in (8). Hence the present analysis of // as an approximant.

1. (8)

   

In fact, the articulatory features of // in Umbuygamu correspond quite well to the description of dental approximants provided in Olson et al. (Reference Olson, Mielke, Sanicas-Daguman, Pebley and Paterson2010), the only general study available so far, which is mainly based on a detailed analysis of a number of Austronesian languages of the Philippines. The authors describe the sounds as voiced approximants, with a dental or interdental place of articulation, and showing perceptual (as well as articulatory, see Mielke et al. Reference Mielke, Olson, Baker and Archangeli2011) links with laterals. Most of these features are found in Umbuygamu, although with just a few good tokens and no chance to do visual checking, it is difficult to determine whether the precise place of articulation is dental (in the narrow sense) or interdental. Acoustically, Mielke et al. (2010) report a clearly visible formant structure, with smooth transitions from and to the surrounding vowels. The same applies to // in Umbuygamu, as shown in Figure 6, with F1 and F2 roughly within the ranges reported for Kagayanen // in Mielke et al. (Reference Mielke, Olson, Baker and Archangeli2011), though a separate study is required to verify this.

Figure 6 Spectrogram for /ˈhaar/ haðarr ‘good’ (NG).

Finally, Umbuygamu exhibits two general allophonic processes that apply across the different classes of consonants. One concerns length: intervocalic consonants in the onset of stressed syllables allow lengthening, as illustrated for two different consonants in (9).

1. (9)

   

Consonant lengthening is found in the same context in other languages in the region, for instance in several Middle Paman languages to the north of Umbuygamu – see Verstraete & Rigsby (Reference Verstraete and Rigsby2015: 73–74) for an overview. The other process concerns labialization of consonants following the high back vowel /u/, as illustrated in (10a–b).

1. (10)

   

Labialization is obligatory, with two exceptions: it is not found when the consonant precedes a high front vowel, as shown in (11), and rarely found when the relevant syllable is unstressed.

1. (11)

   

Labialization is audible as a [w]-offglide following the consonant, which can also affect rounding of a following back vowel (see further in next section).

Vowels

The examples show vowel contrasts in two positions. The first position is the (unstressed) initial syllable of a vowel-initial bisyllabic root (for the canonical root type in this language, see next section). In this context the examples form (near-)minimal pairs, except for the front mid vowel, which tends to show vowel harmony within the root (see below).

The second position for which contrasts are illustrated is the (stressed) second syllable of a bisyllabic root. Given that back vowels are relatively rare in this context, it is difficult to find one set of (near-)minimal pairs illustrating all vowel contrasts at once. However, all height contrasts can be illustrated pairwise, for front vowels and for back vowels, as well as backness contrasts for high vowels and for non-high vowels. The contrast between high and mid back vowels relies on consonant-initial forms, as no (near-)minimal pairs are available in other contexts in the corpus.

Unlike its large consonant inventory, Umbuygamu’s vowel inventory is typical of the relatively small vowel inventories commonly found in Australian languages (see Fletcher & Butcher Reference Fletcher, Butcher, Koch and Nordlinger2014, Gasser & Bowern Reference Gasser, Bowern, Kingston, Moore-Cantwell, Pater and Staubs2014). There are three features in the architecture of the inventory that deserve some further comment. One concerns vowel length: some Australian languages have phonemic vowel length in at least the stressed syllable, but this is not found in Umbuygamu, which only shows length as an allophone in specific contexts (see the discussion of allophonic processes below). The second feature is the notable presence of a low back vowel [ɑ] in a number of lexemes, as /aˈpɔr/ [aˈpɑːr] aporr ‘turtle sp.’ (RL). Given that the low back vowel never contrasts with the mid back vowel, however, the two are analysed here as allophones of the same phoneme (see below on the relevant context), and the inventory as counting five rather than six vowel qualities. A third noticeable feature is the presence of vowel harmony for front mid vowels. In the lexicon, front mid vowels almost exclusively co-occur with a mid or low front vowel, and rarely with high or back vowels. Similar patterns of vowel harmony are found in other languages in the region, specifically vowel height harmony for mid vowels in some Middle Paman languages, including Umbuygamu’s northern neighbour Umpithamu – see Verstraete & Rigsby (Reference Verstraete and Rigsby2015: 84–86) for an overview.

There are two general allophonic processes that apply to all vowel qualities. One concerns lengthening: vowel length is not phonemic, as mentioned earlier, but vowels can be lengthened in three contexts. First, they are often lengthened preceding voiced trills, as illustrated in (12), except in post-tonic syllables, as illustrated in (13).

1. (12)

   

2. (13)

   

Second, vowels are occasionally lengthened in stressed syllables, as illustrated in (14).

1. (14)

   

Finally, root-initial vowels can also be lengthened as part of a sandhi process repairing underlying vowel hiatus, illustrated in the compound structure in (15) (see further in next section on such structures).

1. (15)

   

The second general allophonic process concerns vowel reduction: front vowels are occasionally centralized in post-tonic position, as illustrated in (16). This type of reduction is not found for back vowels but, as already mentioned, these are rare in post-tonic position anyway, so this may be an effect of the relatively small number of tokens in the corpus.

1. (16)

   

A second set of allophonic processes concern specific vowels in the inventory. There are two notable processes for the low vowel. First, the low vowel varies between a front [a] and central [ɐ] realization. The front realization is more frequent in pre-tonic position, the central realization in post-tonic position; in tonic position both occur in what looks like free variation. Examples (17) illustrate these patterns: front realization in pre-tonic position in (17a, b), central realization in post-tonic position in (17c), and variation between front and central realization in tonic position.

1. (17)

   

I choose /a/ as the phonemic label because the pre-tonic position (initial V in V-initial roots, the canonical root type in the language) has a higher functional load for vowel contrasts than the other positions, which, as mentioned earlier, show some historical levelling of vowel contrasts, especially for back vowels.

A second process affecting the low vowel is raising: low vowels can be raised following palatal elements, as illustrated in (18), and raised and rounded following labialized consonants, as illustrated in (19).

1. (18)

   

2. (19)

   

Raising following palatals is relatively rare in the corpus, but raising following labialized consonants is very frequent, effectively leading to neutralization of the contrast with mid back vowels in contexts of labialization.

For front vowels, there are two patterns to be discussed. One concerns the high front vowel /i/, which in the speech of Nellie Salt can be realized as lax [ɪ] in stressed contexts, as illustrated in /aˈriwir/ [aːˈrɪβɪr] arriwirr ‘owl sp.’ (NS). This is probably an idiolectal feature, as in the speech of all other speakers, vowel reduction is only found in unstressed syllables (see the discussion of vowel reduction above). The second pattern concerns the front mid vowel /ɛ/, which in root-initial position (i.e. pre-tonically) is frequently diphthongized to [ɛɘ] or even [ɛa] when preceding a stressed syllable with a low vowel. This process is found for roots in isolation, as shown in (20), but also in the resolution of vowel hiatus between roots, as in (21).

1. (20)

   

2. (21)

   

Given that this is the same context that can trigger lengthening for all vowels, as noted, there is free variation between [ɛ], [ɛː] and [ɛə] in this context.

The back vowels also show allophonic variation. For two speakers in the corpus (Bob Bassani and Nellie Salt, both from the older generation of speakers), the high back vowel /u/ in post-tonic position alternates with -Vw following nasals, trills and laterals preceded by high or mid front vowels. As shown in (22), V is usually a centralized vowel in such forms.

1. (22)

   

Lastly, the other back vowel varies between a mid [ɔ] and low [ɑ] realization. The low realization is restricted to one context, namely in a stressed syllable before word-final /r/, illustrated in (23).

1. (23)

   

In this context, [ɑ] contrasts with all other vowel qualities except [ɔ], which is why [ɑ] is treated here as an allophone of /ɔ/.

Stress and word structure

The analysis of stress presented in this section is inherently limited because it is based on the judgement of a non-native speaker: given the nature of the corpus, I cannot rely on native speaker judgements of stress perception. The same applies to determining the relevant acoustic cues: following work on other languages of the region (Verstraete & Rigsby Reference Verstraete and Rigsby2015: 86–94), my working assumption is that stress perception is associated with the primary pitch movement. The analysis presented here is partly corroborated by patterns of vowel centralization, which, as mentioned in the previous section, are mainly found in post-tonic contexts, but again this is an issue for further research.

To discuss stress and word structure in Umbuygamu, it is necessary to distinguish between vowel-initial and consonant-initial roots, and between roots and compound structures. Umbuygamu is unusual among Pama-Nyungan languages in that a large proportion (about two-thirds) of its roots is vowel-initial, deriving from a historical process of the loss of initial consonants (see Hale Reference Hale1964, Reference Hale1966 and Alpher Reference Alpher1976 on initial-loss in other subgroups of Paman). Lexicons with relatively large proportions of vowel-initial roots are also found in other languages of the region, e.g. the other Lamalamic languages (Sommer Reference Sommer1976), at least one Middle Paman language (the present author’s fieldnotes on Umpithamu), Northern Paman languages (Hale Reference Hale1964, Reference Hale1966) and Alaya-Athima (Alpher Reference Alpher, Verstraete and Hafner2016).

The presence of vowel-initial roots in Umbuygamu is relevant because the stress pattern of roots differs depending on whether they are vowel- or consonant-initial: vowel-initial roots have stress on the second syllable, as in (24), while consonant-initial roots have stress on the first syllable, as in (25).

1. (24)

   

2. (25)

   

This difference is consistent with the hypothesis that historical loss of initial consonants may have gone hand in hand with a shift in stress pattern (Blevins & Marmion Reference Blevins and Marmion1994). The majority of roots are bisyllabic, which implies that a bisyllabic vowel-initial form, as in (24) above, is the most typical root. There are a handful of monosyllabic roots, mainly verbs and cardinal directions, illustrated in (26), as well as a larger number of trisyllabic roots, illustrated in (27), although some of these may be complex diachronically.

1. (26)

   

2. (27)

   

The distinction between roots and compounds is relevant because compound structures are defined by a distinct stress pattern, with stress falling on the stressable syllable of the second root. Furthermore, the first member of a compound can also undergo reduction if it is vowel-initial and/or vowel-final. When a vowel-initial root serves as the first element of a compound, it can drop the initial vowel, as in (28) and (29), in a process that is specific to compounds.

1. (28)

   

2. (29)

   

When a vowel-final root serves as the first element and is followed by a vowel-initial root, the first element loses its final vowel, as in (29) and (30), in a general sandhi process to resolve vowel hiatus (which can also be observed in the transcribed passage in next section).

1. (30)

   

As mentioned earlier, the second process can be accompanied by lengthening of the first vowel of the second member (or diphthongization in the case of a front mid vowel).

I use the label ‘compound’ for any structure that has these formal features, i.e. a distinct pattern of stress and vowel elision. These features are most typically associated with forms that also look like compounds semantically, like (28)–(30) above, but not necessarily or exclusively. On the one hand, combinations of lexemes that look like compounds semantically also allow regular construal, as is shown in (31a), which is construed with stress on both lexemes. Conversely, lexeme combinations with other semantic relations also allow compound construal, as is shown in (31b), which is semantically a generic-specific combination.

1. (31)

   

The vowel elision processes associated with compound structures create consonant-initial forms with second-syllable stress. Given that roots only allow second-syllable stress for vowel-initial forms, any consonant-initial forms with stress beyond the first syllable should be analysable as compounds, as in /θɛˈɟɛr/ aθa edyerr ‘ashes’ (BB). There are a few such forms in the lexicon for which a synchronic compound analysis is difficult because not all elements are attested as independent roots, as is the case in the examples in (32).

1. (32)

   

In such cases, however, there is often comparative evidence to show that they derive from compounds and retain the typical compound stress pattern. The initial morpheme wa- in (32) is only attested in complex forms, but may be related to a Proto-Paman etymon *caawa ‘mouth’ (Hale Reference Hale1976b), which suggests that they derive from compounds with an element meaning ‘mouth’ as the first element. The semantic transparence of the second element varies in these structures. In (32a), the second element is attested independently as othanan ‘skin’, which suggests that the structure derives from a compound like ‘mouth skin’. In (32b), the second element is not attested independently, but is probably related to an etymon *calparr ‘chin’ (Alpher Reference Alpher, Bowern and Koch2004). In (32c), the second element is not attested independently and not obviously reconstructible, which means that wa-, as well as the paradigm of forms in (32), are the only evidence for an origin in a compound structure.

Transcribed passage

The following text is the first episode (about one-third) of a longer text describing a hunting trip: it was recorded in 1974 by Bruce Sommer; the speaker is Nancy Gunnawarra (NG). The text is elicited, with the fieldworker prompting the speaker for every new utterance. Although this is not the most representative type of text, it is the only longer text in the corpus that can be reliably glossed and translated (partly based on glossing provided in Bruce Sommer’s fieldnotes, as well as prompts in the recording). I provide a broad transcription of the text, followed by a morphemic breakdown in practical orthography, interlinear glosses, and a free translation.