Vowels

There are three vowels, /i a u/, used in Tena Quichua and in words adopted from other languages, such as Spanish. All three appear in both stressed and unstressed positions. Example (1) shows each of the vowels in stressed position in combination with each vowel in unstressed position.

1. (1)

   

Note that the mid vowels /e o/ are not found in words with Quichua origin, but may appear in Spanish words that have not been fully incorporated into the TQ sound system. For example, a loan word into Quichua such as /paˈbo-ɾa/ ‘duck-acc’ maintains the mid back vowel /o/ from Spanish /ˈpabo/ ‘duck’, while a word such as /ˈɡaʎu/ ‘rooster’ shows a more fully incorporated word with a high back vowel /u/ instead of the original /o/ from Spanish /ˈɡaʎo/ ‘rooster’.

In order to examine the quality of the vowels, stressed vowels in open syllables were extracted from recordings of semi-spontaneous speech and measured in Praat (Boersma & Weenink Reference Boersma and Weenink1992–2011). For the five speakers, approximately 10 tokens of each vowel quality were measured. The data points labeled i a u in Figure 1 represent the mean for each speaker (minimum = 5 vowel measurements per data point, average = 9.7 vowels; total = 146 vowels). The measurements, taken in Hertz, were then converted to the Bark scale and normalized using the procedure described in Guion (Reference Guion2003) to account for individual differences in vocal tract length. Figure 2 represents the average vowel quality in Bark for the five speakers combined. Error bars show the average standard deviation for the five speakers. Note that the height (F1) of the back vowel is lower than the high front vowel. Orr (Reference Orr1991) observes that the back vowel is between the /u/ and /o/ sounds. In comparison to the Cardinal Vowels (Delattre et al. Reference Delattre, Liberman, Cooper and Gerstman1952: 200), the high front vowel may be considered to be closer in realization to [y] in terms of backness (F2), although it is also slightly lower, and the high back vowel may be considered closer to between [o] and [ɤ] in terms of backness (F2), although it is also lower than these vowels. Since Quichua vowels are generally transcribed as /i a u/, they will be transcribed in this way here also, even though, as noted in Guion (Reference Guion2003), their actual quality is considered to be closer to lower and more centralized variants [ɪ] and [ʊ]. In the case of the present data, the near-back vowel is even lower, and appears between the Cardinal Vowels [o] and [ɔ] in terms of height. The realization of the low vowel [a] is also more centralized (higher in F1) and is also less back in F2 than the corresponding cardinal vowel.

Figure 1 Data points represent average vowel quality of five Tena Quichua speakers (in Hz).

Figure 2 Vowel quality averaged over five Tena Quichua speakers (converted to Bark scale and normalized).

Diphthongs

The presence of diphthongs is described with respect to the overall syllable structure found in TQ. A coda consonant is possible in TQ, such that the addition of a progressive marker /-u/ forms a falling diphthong with the preceding nucleus, as in (2). Since falling diphthongs are possible, other instances of a glide (or semivowel) following a vowel in (3) are analyzed as diphthongs as well, although a coda consonant is not present in these cases.Footnote ⁶

1. (2)

   

2. (3)

   

In the case of rising diphthongs, the evidence is less clear with respect to syllable structure. That is, complex onsets are found, as in the second syllable in (2). However, there are no instances of a complex onset in combination with a glide. Therefore, in the following examples in (4), a central approximant is considered to be part of the complex onset.

1. (4)

   

In cases where there are two adjacent identical vowels, they are reduced to one long vowel rather than forming a diphthong, as in (5).Footnote ⁷

1. (5)

   

Under this analysis, examples such as (6) represent a falling diphthong preceded by a complex onset (rather than a triphthong).

1. (6)

   

Stress

Stress in Tena Quichua generally falls on the penultimate syllable (as shown in the examples in (1) above). However, a subset of suffixes causes stress to fall on the last syllable of the word, including: the locative (loc) suffix when added to polysyllabic words ending in a vowel (see example (5) above), third person singular /-n/ and plural /-nun/ present tense markers (see examples (6) above and (7)), and the suffix /-nɡax/ used to express future obligation (see example (8)).

1. (7)

   

2. (8)

   

There are also cases in which the addition of a suffix does not cause stress shift to the new penultimate syllable, as in the topic marker (top) /-ɡa/ (Orr Reference Orr1991). In the present data set, this phenomenon was also observed in some cases with the accusative marker (acc) /-ɾa/ (as in (9), from /ˈkwika-ɾa m iˈku-n/ earthworm-acc eat-3s ‘it eats/is eating (an) earthworm’). However, in other examples, this phenomenon is not as clear (as in (10)), while other cases do show penultimate stress (as in (11), from /a ˈtʃa-ɾa m iˈku-n/ meat-acc eat-3s ‘she eats/is eating meat’) (see Figure 3). Whether or not penultimate stress is observed with /-ɾa/ may be due to several factors, including word order and placement within the discourse, although this matter lies beyond the scope of the present work.

1. (9)

   

2. (10)

   

3. (11)

   

Figure 3 Word-initial stress in [ˈkwika-ɾa] ‘earthworm-acc’, penultimate stress in [a ˈtʃa-ɾa] ‘meat-acc’; pitch shown with solid line.

Last, there are some words where secondary stress may be perceived, either on the first or sometimes second syllable, as in (12) and (13). Hintz (Reference Hintz2006: 482–483) provides a summary of stress in varieties of Quechua, including cases in which primary or secondary stress is found on the initial syllable. Cerrón-Palomino (Reference Cerrón-Palomino1987: 128–129) notes that which syllables attract stress may be due in part to syllable structure and the types of suffixes involved in word formation. In order to further characterize stress in TQ, a detailed study of the acoustic correlates of stress, including duration, intensity, and fundamental frequency (f0) is needed as well as a comparison of vowel quality of stressed and unstressed vowels.Footnote ⁸ Preliminary observation suggests that changes in intensity may participate in signaling tonic and pretonic syllables, as in Figure 4: the greatest degree of intensity appears during the penultimate syllable, whereas the antepenultimate syllable shows the least. In this example, unstressed vowels are not noticeably reduced in quality or duration. Rather, the stressed syllable appears longer than the other atonic syllables (due in part to the presence of a diphthong) and is further highlighted with an f0 peak.

1. (12)

   

2. (13)

   

Figure 4 Secondary stress on initial syllable, penultimate primary stress in [ˌrimanuˈna ka] ‘they were talking to each other’; pitch shown with solid line, intensity with dashed line.

Intonation

A final note on prosodic features is related to intonation. Since more work on stress (and stress shift) in TQ is needed, the suprasegmental feature of intonation is mentioned here as an area for future research. In terms of Quichua varieties, Cole (Reference Cole1982) provides a description of intonation in Imbabura Quechua, which shows a schematic representation of the peak occurring during the stressed syllable. Words spoken in isolation follow this general description (see Figure 5). However, the representation in Cole (Reference Cole1982) shows only contour movement at the end of the utterance, such that the alignment and scaling of peaks in non-final position within an utterance needs to be determined.

Figure 5 Words produced in isolation with peak during the stressed syllable [ˈʃimi] [ˈɾiŋɾi] ‘mouth’, ‘ear’.

In the declaratives analyzed for TQ, tonic-alignment of peaks is found in both final and non-final positions within an utterance (e.g. Figure 6).Footnote ⁹ The final pitch accent or tonal movement related to the stressed syllable may be lower than previous pitch accents, as in Figure 6; in other cases, the final pitch accent is not lower, as in Figure 7. The extent to which downstep operates within TQ is another area to be explored further, that is, whether or not pitch is successively lower at certain tonal events along an utterance, as described in Liberman & Pierrehumbert (Reference Liberman, Pierrehumbert, Aronoff and Oehrle1984). Longer utterances with successive peaks are needed to observe this phenomenon (or the lack thereof) in more detail.

Figure 6 Tonic-aligned final and non-final peaks, lower final peak:

• 

Figure 7 Tonic-aligned final and non-final peaks, non-lowered final peak (OV word order):

• 

In the previous examples of declaratives with broad focus (neutral object–verb word order), the utterance ends in a final fall. In Figure 8, the utterance with verb–object word order shows a wider pitch range and slightly lower final pitch accent. Longer utterances demonstrate a similar pattern, with peaks during the stressed syllable and lack of downstep, as in Figure 9 and Figure 10. For the present data, using AM notation for peaks (H) and valleys (L), tonal association (*), and boundary tones (%), the nuclear and prenuclear pitch accents appear to be tonic-aligned or H* (as in Figure 8), in some cases preceded by a low tone L+H* (as in Figure 9); declaratives end in a low boundary tone L% (Figures 6–10). Several other areas to be explored within TQ intonation include examination of other basic utterance types, such as yes–no and wh-questions, exclamatives, and imperatives, as well as the shape of final and non-final contours.

Figure 8 Tonic-aligned final and non-final peaks, lowered final peak, expanded pitch range (VO word order):

• 

Figure 9 Tonic-aligned final and non-final peaks, non-lowered final peak:

• 

Figure 10 Tonic-aligned final and non-final peaks, non-lowered final peak:

• 

Stops

In intervocalic position, voiced stops undergo spirantization, such that fricatives are found, as in /ˈka -bi/ [ˈka -βi] ‘this-loc’ (Figure 11) and /aˌʃanɡa-ˈɡuna/ [aˌʃaŋɡa-ˈɣuna] ‘basket-pl’. In addition, voiced stops that originated in loans from Spanish may undergo further spirantization intervocalically, producing approximants, as in /paɡaˈɾatʃu/ [pa aˈɾatʃu] ‘Thank you’ < Sp. /paˈɡaɾ/ pagar ‘to pay’ (Figure 12). More investigation is needed to determine if voiced stops undergo spirantization word-initially when following another vowel.

Figure 11 Intervocalic fricative [ˈka -βi] ‘this-loc’.

Figure 12 Approximant [pa aˈɾatʃu] ‘thank you’ < Sp. [paˈ aɾ] ‘to pay’.

Affricates

After nasals a voiced affricate [dʒ] is found. Since the distribution is limited to this context, [dʒ] is considered an allophone of /tʃ/, as in /ˈpuntʃa/ [ˈpundʒa] ‘day’ (Orr Reference Orr1991). However, the affricate /ts/ is not voiced after the nasal (e.g. /ˈtsuntsu/ [ˈtsuntsu] ‘poor’). This lack of voicing after nasals may be due in part to the development of /ts/ from /tʃ’/ (see Cerrón-Palomino (Reference Cerrón-Palomino1987: 188) for a discussion of the evolution of /ts/ from glottalized, aspirated, and simple voiceless affricates). Note that a distinct release of the stop closure appears in the onset of the second syllable in /ˈtsatsa/ [ˈtsatsa] ‘sand’ and /ˈtsuntsu/ [ˈtsuntsu] ‘poor’ (also apparent in a spectrogram of these realizations). This realization may be due to careful pronunciation but should be investigated further according to the evolution of the affricate word-internally.

Nasals

Nasals in coda position share the place of articulation with the following stop (e.g. /ˈtʃimba/ [ˈtʃimba] ‘across’, /ˈtʃinda/ [ˈtʃinda] ‘brush pile’, /ˈtʃunɡa/ [ˈtʃuŋɡa] ‘ten’); note in this last case that [ŋ] is considered to be an allophonic variant of /n/ since its distribution is limited to preceding other velars. Orr (Reference Orr1991) also reports a similar (velar) pronunciation preceding /m/ or word-finally (e.g. /ˈatun/ [ˈatuŋ] ‘large’). However, the extent to which this velarization process occurs in these contexts consistently needs to be investigated further, including formant transitions in the preceding vowel.

Fricatives

After nasals, the alveolar fricative /s/ is not voiced, as in /ˈkinsa/ [ˈkinsa] ‘three’. In Cordero (Reference Cordero2006 [1892]), Caimi ñucanchic shimiyuc-panca (1982), a voiced /z/ is shown after a nasal (e.g. /ˈpunzu/ [ˈpunzu] ‘fluffy, frizzy’); there are also very few tokens of /z/ cited word-initially and intervocalically. The voiced alveolar fricative may appear in early loan words from Spanish (also noted in Cerrón-Palomino (Reference Cerrón-Palomino1987: 186) and references therein). The Quichua word /ˈduzi/ [ˈduzi] ‘zenith’ shows an early Spanish loan from the Old Spanish /ˈdoze/ [ˈdoze] doce, which became a voiced dental in Medieval Andalusian and American Spanish; later developments in the devoicing of sibilants resulted in modern Latin American Spanish [ˈdose] doce (see Penny Reference Penny2002 for a description of the evolution of Spanish sibilants). However, in elicitation of this token for the present data set, only a voiceless alveolar fricative was observed ([ˈdusi] ‘zenith’). A similar result was found for [kuɾˈzana] ‘place in a crossed fashion’ < Old Sp. [kɾuˈzaɾ] ‘to crossʼ; that is, [kuɾˈsana] ‘place in a crossed fashion’ was observed). Since in modern Spanish orthography 〈z〉 represents a voiceless fricative [s] in Latin American Spanish, the production of a voiceless variant for Quichua duzi and kurzana as opposed to the voiced fricative predicted for Quichua may be due to influence from Spanish orthography. More investigation is needed in this area to determine the possible sources of this variation. Also, the fricative /z/ may be realized with an affricate pronunciation word-initially, as in /ˈzas-ʎa/ [ˈdzas-ʎa] ‘quick-lim’, or in some cases after a nasal, as in /ˈaswa a nzi/ [ˈaswa a ndzi] ‘solid residue from corn beer’, compare /ˈpunzu/ [ˈpunzu] ‘fluffy, frizzy’. While the tokens above were observed in the present data set, further investigation is needed to determine how consistently this epenthesis takes place, including examination of other environments such as /z/ in intervocalic position. Variants of another fricative are found in the place of articulation of /x/, which may range from velar to glottal in native Quichua words and in loans (e.g. /ˈpaxu/ [ˈpaxu] ‘illness’, /ˈxutʃa/ [ˈhutʃa] ‘debt, sin’).

Rhotics, laterals, and approximants

The tap /ɾ/ is realized in general as a vibrant with a single occlusion. In the present set of recordings, in addition to a simple tap, a trilled variant [r] was observed word-initially (e.g. /ˈɾuna/ [ˈɾuna], compare [ˈruna] ‘person’). Note that both cases are realized with a prothetic vowel; see Figure 13. The presence of a rhotic trill may be due in part to influence from the multiple trilled variant in Spanish, especially in more careful speech styles, such as in reading. In addition, the rhotic may be naturally strengthened to a trill at the onset of a stressed syllable. There appears to be a degree of variation in this aspect since both the simple tap and trill are found after a pause, such that the trill is not obligatory in this context. Also, a retroflex approximant [ɹ] was found (e.g. [tʃuɹa-ˈɾi-ʃka] ‘put-refl-past.part’; Figure 14). In some cases, although transcribed as a tap, the actual realization was longer than a prototypical tap and/or was between a tap and an approximant.

Figure 13 Word-initial tap [ˈɾuna] and trill [ˈruna] ‘person’.

Figure 14 Approximant and elongated tap in [tʃuɹa-ˈɾːi-ʃka] ‘put-refl-past.part ‘put on, worn’.

According to Catta Q. (Reference Catta1994: 10), in Eastern varieties (including Tena) the palatal lateral approximant /ʎ/ may be realized as [ʎ] word-initially and intervocalically, and as [l] in syllable-final position (e.g. /ˈʎakta/ [ˈʎakta]). In TQ, this depalatalization occurs in particular before the high front vowel /i/ (Carpenter Reference Carpenter1984: 13), as in /ˈali/ [ˈali] ‘good’; since the /l/ phoneme is already posited, the underlying form of [l] is considered to be the alveolar approximant /l/ rather than an allophonic variant of /ʎ/ based on comparison with other Quechuan dialects. A more detailed acoustic analysis is needed of the formants for both /ʎ/ and /l/ in order to determine the extent to which this depalatalization process is complete in syllable-final and intervocalic positions.

Semi-narrow phonetic transcriptionFootnote ^{10 ,} Footnote ¹¹

tʃi w a ɾa i skinamanda i ndiwa ɾimanuˈnun / ma kanda tʃi iʃkipuɾamanda m as ʃindʒi a ka // tʃi ɾju r una p asaka ʃu k ilpaɾina tʃuɹaɾiʃka // pa na tʃuɾanuka p aɾehumanda aʎi ɲawpapunda p uɾiŋɡawa / ma kambas tʃi tʃuɾana uɾmakta ɲa ɾikuŋɡawa p a ʃindʒi aʃkaɾa // ʃinakpi tʃi iʃkimanda ʃamuk w a ɾa p ukuka p unkuŋˈɡa k aʎaɾika w a ɾawa // tʃi m as ʃindʒi w a ɾa ʃitaka tʃi ɾju r una ʃindʒiɾa a pika p a wa k ilpaɾinaɹa // tʃi iʃkiˈuna ɾimanuʃa w a ɾa s ambajaŋɡama // ʃinakpi tʃi i ndi ʃindʒi r a taka p a wa ɾikutʃina p undʒawa // ʃinakpi d zasʎa tʃi ɾjux r una a paʃitaka p a wa tʃuɾanaɾa // ʃinakpi w a ɾa i skinamanda w a ɾa ɲa j atʃaka aʎiɾa tʃi i ndi aʃka ʃindʒi aʃkaɾa //

English translation

This wind from the (North) corner, with the sun, they talk to each other (about) which of the two was stronger. This traveler passed wearing a cloak. They agreed that of the two (whoever) is first which also makes the clothing fall will be recognized as stronger. So this wind coming from the corner blew, began to blow with wind. The stronger the wind blew, the (more) strongly the traveler held his cloak, the two talking to each other until the wind became tired. Then the sun shone strongly with his visible day(light). Then suddenly this traveler took off his clothing. Then the wind, the (North) corner wind already well knew that the sun had been much stronger.

Spanish text used in recording of Quichua translation

El Viento Norte y el Sol discutían cuál de los dos era el más fuerte, cuando pasó un viajero envuelto en una capa. Se pusieron de acuerdo en que aquel que primero pudiera obligar al viajero a que se le cayera la capa sería considerado el más fuerte. Entonces el Viento Norte comenzó a soplar con mucha furia. Pero, cuanto más soplaba más el viajero agarraba a su capa hasta que el Viento Norte desistió. Entonces el Sol brilló con todo su esplendor e inmediatamente el viajero se quitó la capa. Así que el Viento Norte tuvo que reconocer la superioridad del Sol.