The Phonetic Structures of Kalanguya

Paul Julian Santiago

University of the Philippines Diliman [email protected]

ABSTRACT
This paper examines the phonetic structures of Kalanguya, an Austronesian language spoken in Northern Philippines. This study uses static palatography to determine the articulatory positions of four coronal consonants and acoustic analysis to describe some characteristics of certain salient sounds in Kalanguya. Some remarkable findings of this study are as follows: (a) the places of articulation of coronal consonants /t/, /d/, /l/ and /n/ are at dental and alveolar regions. (b) The glottal consonant /h/ can be considered as a true fricative and compensates for the lack of /s/ in Kalanguya. (c) The voiced bilabial stop /b/ is labialized [b]. (d) The voiceless plosive /k/ in Kalanguya is uvular [q], which is quite rare in Austronesian languages. (e) Finally, there are four vowel sounds in Kalanguya: one front vowel [], two central [a] and [], and one back vowel []. Moreover, Kalanguya vowel inventory has no internal symmetry in terms of F1 and F2 measurements and gender differences and syllable stress have no effects on vowel height and frontness. Keywords: acoustics and articulatory phonetics, palatography, Kalanguya, Austronesian, uvular consonants static

means people from the forest and older Kalanguya people claim that Ikalahan is usually used as part of a derogatory statement: Ikalahan(they) have big poop an idiomatic reference to the prevalence of sweet potatoes in the Kalanguya diet.. In 2000, the population of this indigenous group is approximately 70,000 excluding the group living in Pangasinan and Nueva Ecija [3]. Figure 1 is a map of the area where Kalanguya is spoken obtained from Arsenio and Stallsmith [3]. According to my Kalanguya consultant, there are three dialects of Kalanguya: (1) the northern dialect which is the dialect spoken in Tinoc, Ifugao and Benguet, (2) the central dialect which is spoken in Ambaguio, Kayapa, and other parts of western Nueva Vizcaya, and (3) the southern dialect which is spoken in Aritao and Sta. Fe, Nueva Vizcaya and in some parts of Pangasinan and Nueva Ecija. In the ethnologue of SIL, three languages were listed with the name Kallahan and these are Kallahan Kayapa (ISO 639-3: kak), Kallahan Keley-i (ISO 6393:ify), and Kallahan Tinoc (639-3:tne). The language analyzed in this study is only Kallahan Kayapa.

1. INTRODUCTION 1.1 Kalanguya Language

Kalanguya is a Philippine-type language spoken by the Kalanguya ethnic group living in the municipalities of Kayapa, Ambaguio, Aritao and Sta. Fe in Nueva Vizcaya, Tinoc in Ifugao and some parts of Benguet, Pangasinan and Nueva Ecija. Kalanguya is listed in the Ethnologue of SIL International as an alternate name of Kallahan. Other names of this language are Ikalahan, Kalangoya, Kalangoya-Ikalahan, Kalkali, Kayapa and Akab1. However, groups living in Kayapa and other parts of Nueva Vizcaya prefer the name Kalanguya for their language, which is an autonym derived from the phrase Kelay ngo iya?, which means Why/What is this?. However, their group was also called Ikalahan, an exonym which means from the forest (I- is a prefix that means from and kalahan means forest) [22]. The people who belong to this ethnic group call their group and their language Kalanguya. Three conferences were held from 1993-1999 to reach a consensus about the group and their languages name and after several heated debates, majority of the leaders belonging to the group decided that Kalanguya is the proper name of the group. Most Kalanguya people object to the use of Ikalahan because it
1

Figure 1: Kalanguya language area map Kalanguya is not a national or official language. The only published text is the Bible translated into Kalanguya. There are efforts to publish a newspaper in Kalanguya but this never happened due to lack of funds. The language is used in informal conversation in elementary and high school but this is not the official medium of instruction. There are songs, poems, myths and stories but these are

http://www.ethnologue.com/show_language.asp?code=kak

unpublished. 1.2 Kalanguya

Phonology and Orthography

2. METHOD 2.1 Subject

There are only few studies done on Kalanguya language and most of them focus on morphosyntax. Hohulin and Kenstowiczs [15] study on Keley-i phonology and morphophonemics gives a brief description of Keley-i phonemes and some phonological changes that occur in this language. According to them, Keley-i has 18 consonant (/p/, /t/, /b/, /d/, /s/, /m/, /n/, /w/, /l/, /c/, /k/, //, /d/, /j/, /k/, /g/, //, /h/, // and five vowel phonemes (/a/, /e/, /i/, /o/, /u/). So far, there are no studies which deal with the phonetic structures of this language. This differs from the phonemic structure of Kalanguya of Kayapa. There are 14 consonantal sounds in Kalanguya: /p/, /t/, /d/, /k/, /g/, /b/, /m/, /n/, //, /h/, /w/, /j/, /l/ and //. A single glottal stop is not represented before vowel letters in word-initial position and in word-final position. Hyphen is used as a representation of the glottal stop when in between a consonant and a vowel and when geminated (e.g. ok-ok [k.k], i-ogip [i.gip]). The consonant /h/ will be considered as a true fricative for the purposes of this study because it can occur both at the beginning and at the end of syllables and before consonants. This compensates for the lack of the fricative /s/ which is present in other Philippine-type languages (PLs). Kalanguya also lacks the alveolar flap // which is present in other PLs such as Tagalog, Cebuano and Ilocano. Kalanguya has four vowels: /a/, /e/, /i/ and /o/. In the bible translated by the New Tribes Mission [21], the vowel /e/ is written as to represent the two dialectal variations when pronouncing the vowel. People who speak the northern dialect read this as /e/ while those who speak the southern dialect pronounce it as /a/. The word list used in acoustic recordings and static palatography followed this orthography.

Two native speakers of Kalanguya participated in palatography and still photography session; one male and one female. The palatographic techniques performed were exactly the techniques described in Ladefoged [17] and Anderson [1]. Palatograms were obtained by painting the speakers tongue with a non-toxic marking material which is a mixture of powdered digestive charcoal and olive oil. The speaker then uttered the word containing the target consonant, transferring the marking material from the tongue to the palate. Then a mirror was inserted into the mouth to reflect the contact area on the palate. This resulting image was photographed and videotaped. On the other hand, linguograms were obtained by painting the speakers palate, which transferred the marking material to the tongue when the word containing the target consonant was articulated. The speaker then put the tongue out to show the contact pattern on the tongue which was photographed and videotaped.

Figure 2: Set-up for palatography Six participants participated in digital recording; three males and three females. All the participants were native speakers of Kalanguya and all came from Kayapa, Nueva Vizcaya. Thus only the Central Kalanguya dialect will be analyzed in this study. They were recruited by a Consultant who was also a native speaker of the language to make sure that the participants (1) were native speakers of Kalanguya; (2) did not show any signs of speech and voice disorder; and (3) belonged to the Central Kalanguya group. This last condition is vital because other dialects or variety of Kalanguya have a different vowel system. The participants ages range from 17 to 40 years old. The recordings were made in a quiet room using PRAAT 5.1.122 mono sound recorder at a sampling rate of 44,100 Hz.

1.3 Current study

In this study, certain salient phonetic properties Kalanguya will be explored. Static palatography and acoustic analysis were performed to determine the articulatory characteristics of Kalanguya consonants. The formant structure of Kalanguya vowels were also examined to describe the vowel space of Kalanguya and to find out if gender differences and syllable stress affect vowel quality. A complete consonant and vowel inventory based on the results of the acoustic and articulatory analysis will be designed at the end of this paper. The study will be useful not only in teaching the correct articulation of speech sounds to the younger generation of Kalanguya ethnic group but also in encouraging other Filipino linguists and language researchers to use the methods performed in this study when documenting Philippine-type languages. This will also help Kalanguya teachers determine the similarities and differences of their language from other languages such as English and Tagalog and create new strategies in teaching the children how to make pronunciation adjustments when switching to these languages.

2.2 Stimuli
The stimuli used in this study were carefully selected by two consultants who were native speakers of Kalanguya. There were two sets of stimulus used; one for palatography and still photography and the other for acoustic recording. In palatography, the participant uttered a word containing the target consonant. Only words containing non-high vowels were used to
2

Praat is a program for speech analysis and synthesis that can be downloaded for free at http://www.praat.org

avoid multiple tongue-palate contacts. It is also important that the words used contain no other coronal or dorsal segments except the target segment. The consonants investigated were the coronals /t/, /d/, /n/ and /l/. Table 1 below shows a sample word list for these consonants.

2.3 Analysis
Palatography was performed to determine the place of articulation of the four coronal consonants of Kalanguya. Digital photographs of the palatograms and linguograms will be presented in section III. Measurements were not necessary because the photographs alone were enough to show the palate and lingual contacts and determine the correct articulatory description of each coronal. The glottal consonant /h/ of Kalanguya seems different from the /h/ of other language such as English and Tagalog. Visual inspection of the spectograms and waveforms of /h/ was performed to determine these differences. The duration of the glottal consonants when occurring as a single sound and when geminated was also measured. The articulation of the consonant /b/ was captured by a video camera to show the unique articulatory characteristic of the sound. To inspect and verify the secondary articulation involved when uttering the consonant /b/, formant transition and spectographic analysis of the vowel following the consonant /b/ were performed. Formant measurement of vowels following the consonants /k/ and /g/ was done to determine the differences between the two velar consonants and find out if the phonetic description of /k/ is appropriate or not. The F1 and F2 of each vowel following the consonants /k/ and /g/ were measured at the stable region of the vowel and the mean values were plotted. Leftward effects of the /k/ were also explored through spectograms and formant transition analysis.

/t/ [pa.t]

/d/ [pa.dah]

/n/ [pa.na]

/l/ [pa.la]

Table 1: Static Palatography word list for Kalanguya coronal consonants Still photography was also used to investigate the consonant /b/. This method was necessary because the consonant /b/ is articulated in a unique manner by Kalanguya speakers. A mirror was placed against the side of the face to get the front view and the side view of the lips simultaneously.

Figure 3: Photograph of a Kalanguya speaker. A mirror is placed at the side of the face to see the full and side view of the speaker. Acoustic recordings of consonants /b/, /k/, /h/, and // were done by prompting the speakers to read the target words containing these consonants. These words were embedded in a frame sentence in Kalanguya Ehel moy ______ ni mamintalo which means Say three times and were written in a piece of paper approximately one meter away from them. The consonants were placed before the four vowels in Kalanguya and two words were recorded for each CV pair, giving a total of 48 tokens for each consonant (six speakers four vowels two productions). In addition to this, two productions of geminated /h/ and // before and after the vowel /a/ were also recorded in order to examine the properties of these glottal consonants when geminated. Gemination of these glottal consonants is one of the unique features of Kalanguya. In vowel formant analysis, the words used were all disyllabic. The words were embedded in the same frame sentence used for consonants. Three female speakers and three male speakers served as subjects, giving a total of 576 tokens (six speakers four vowels 24 productions). In each vowel, four productions were made; two were located in stressed syllables while the other two were in unstressed syllables. This was done to distinguish any difference in vowel height and frontness in terms of syllable stress position. Gender-related differences were also examined.

Figure 4: Edit window of Praat For each of the 560 vowel tokens, the F1 and F2 were determined using Praat and plotted using the UCLA PlotFormants 4.0 program. Before getting the formant frequencies, each tokens wide-band spectogram was visually inspected to check if there was a steady state period of the vowel. The midpoint of the steady state was measured afterwards. The following settings of Praat were used in all acoustic studies: 5500 Hz maximum frequency, 5 formants, 0.005s window length, and 40dB dynamic range.

3. RESULTS 3.1 Kalanguya Coronal Consonants

Kalanguya has four coronal consonants: /t/, /d/, /n/ and /l/. Coronal consonants are consonants articulated using front part of the tongue (tip, blade or underside part). These sounds can be divided into apical (if the tip of the tongue is the articulator), laminal (if the blade of the tongue is the articulator), or subapical (if the underside part of the tongue is the articulator). Figure 5 to Figure 8 show the palatograms and linguograms of the four coronal consonants of Kalanguya: the stops /t/ and /d/ (Figure 5 and 6), the nasal /n/ (Figure 7) and the lateral /l/ (Figure 8).

Figure 7: /n/

Figure8: /l/

The palatogram of the nasal /n/ (Figure 7) shows that the lower part of the incisors was touched and that the rearmost point was the front of the alveolar ridge. The linguogram above shows that the most active articulator was the apex of the tongue. Thus, the nasal /n/ may be classified as apical denti-alveolar. For the lateral approximant /l/ (Figure 8), no contact was made on the back of the incisors and the most active articulator was the apex of the tongue. The lateral approximant /l/ of Kalanguya can be classified as apical alveolar. The articulatory data of Kalanguya coronal consonants are summarized in Table 2. /t/ apico-laminal denti-alveolar /d/ apical alveolar /n/ apical denti-alveolar /l/ apical alveolar Table 2: Contact locations of Kalanguya coronal consonants

Figure 5: /t/

Figure 6: /d/

3.2 Kalanguya Glottal Consonants

As mentioned in the introduction of this paper, the glottal sound /h/ will be considered as a true fricative for the purposes of this paper. This is due to the fact that the /h/ of Kalanguya can appear at places where the fricative /s/ of other PLs can. True fricatives can occur at the beginning and at the end of syllables and can co-occur in these positions with other consonants. Moreover, the frication of /h/ of Kalanguya is stronger than the frication of the /h/ of other languages such as English and Tagalog. Spectographic analysis was performed to prove these claims. Figure 9 below shows the spectograms (with the dynamic range set at 40dB) of /h/ of Kalanguya at the beginning and at the end of syllables and before a consonant.

The palatogram of the voiceless stop /t/ (Figure 5) shows that only part of the incisors was touched and that the rearmost point was the front of the alveolar ridge. The linguogram above shows that the most active articulator was the apex and front blade part of the tongue. Thus, the voiceless stop /t/ of Kalanguya may be classified as apico-laminal denti-alveolar. Figure 6 shows the palatogram and linguogram of the voiced stop /d/. As can be seen, the back of the incisors was not touched and the rearmost point of contact was the front part of the alveolar ridge. The active articulator was the apex of the tongue. /d/ may be classified as apical alveolar. We can also notice that the tip of the tongue barely touched the alveolar region. This suggests that the contact of the tongue to the alveolar ridge is very fast.

/ h

a / / b a

/ l a h t i / Figure 9: Kalanguya /h/ can be considered as a true fricative because it can occur at the beginning and at the end of syllables and even before consonants. When compared with the /h/ of Tagalog and English, it is clear that the frication of the /h/ of Kalanguya is stronger than the frication of the /h/ of the two other languages. Figure 10 shows the spectogram of /h/ of the three languages. Kalanguya /h/ English /h/ Tagalog /h/ /a h

/a

a /

a /

Figure 11: Wide-band spectogram and waveform of single and geminated /h/ In Figure 11, the duration of the single /h/ is 0.148 milliseconds whereas the duration of the geminated /h/ is 0.315 milliseconds. Moreover, the spectogram shows that the frication of the first /h/ is stronger than the second /h/. This means that when /h/ is geminated, the frication becomes weaker at the second half of the utterance. Figure 12 shows the wide-band spectograms and waveforms of the single glottal stop // and the geminated glottal stop //. The length of the single // is 0.051 milliseconds whereas the length of the geminated // is 0.223 milliseconds or almost two and a half longer than the single //.

Figure 10: Waveforms and spectograms of /h/ in Kalanguya, English and Tagalog Looking at the waveforms and spectograms of the /h/ in three languages, the /h/ of Kalanguya has a strong frication that it may be considered as a breathy-voiced glottal // but the voicing of Kalanguya /h/ is beyond the scope of this study. Further studies such as aerodynamic investigation, spectral analysis and acoustic analysis of phonation types are recommended. Another unique characteristic of the glottal sounds in Kalanguya is that they can be geminated. Other Philippine-type languages do not allow this kind of gemination. When geminated, the utterance of the glottal fricative /h/ and the obstruction of the airway in glottal stop are prolonged. This means that when the glottal stop // is geminated, the release is delayed. Figure 11 below shows the wideband spectograms and waveforms of the single /h/ and geminated /h/, both occurring in between two low vowels /a/.

/ a

/a a / Figure 12: Wide-band spectogram and waveform of single and geminated //

/ e

Figure 14: Wide-band spectogram of /b/ with the formant values indicated by the red dots. The F1-F2 was measured before the release of the burst and start of the stable region of the vowel. In Figure 14, the F1-F2 value after the burst (start of the vowel region) is 480Hz-1004Hz and when it reached the stable region, the F1-F2 value increased up to 633Hz-2266Hz. The F1-F2 values from the release of the burst to the stable region are plotted and shown in Figure 15.

3.3 Voiced Labialized Bilabial Plosive [ b ]

Several measurements and visual investigations were performed in order to describe the features of the voiced bilabial plosive of Kalanguya. Video photograph of a Kalanguya speaker uttering the sound /b/ intervocally is shown in Figure 13.

Figure 15: Plot of the F1-F2 values of /e/ during the transition from high back (/u/) to low mid (e) in Kalanguya. Figure 13: Kalanguya speaker uttering the syllable /ba/. Notice the rounding of the lips before the release of the sound. The rounding of the lips is one of the articulatory features of labialization [16]. Before unrounded vowels, the lips are rounded before releasing and uttering the following vowel. All the six Kalanguya participants pronounced the /b/ with the lips rounded even in word-final position. Examining the formant transition provided more evidence for the claim that the Kalanguya /b/ is labialized. Figure 14 shows a spectogram of /b/ before /e/ with red dots indicating the formant frequencies. Figure 15 shows the transition of the vowel /e/ when preceded by the bilabial /b/. We can clearly see that after the release of the burst or the start of the vowel, the F2 value is relatively low and towards the middle of the transition, the F2 value increased until it reached the stable region of the vowel. This suggests that the vowel sound right after the plosive occupies the high back region /u/ and there is a long transition from the high back region to the low mid region. The inspection of the formant transition of the following vowel proves that there is a secondary articulation involved when /b/ is articulated. Due to the rounding of the lips when articulating the bilabial /b/, the tongue is positioned at the back part of the mouth before uttering the unrounded front vowel /e/ and will take at least 30 ms to reach the front region. Labialization of consonants is marked by a diacritic []. The purpose of placing the diacritic [] to [b] is to represent the secondary articulatory feature of this consonant.

3.4 Voiceless Uvular Plosive [q]

Uvular consonants are one of the less common consonants; 80.9% of all the languages surveyed in Maddieson [19] have no uvular consonants. He also stated that uvular consonants are absent in Southern Asia including the Philippines. However, the acoustic investigation of Kalanguya suggests that the voiceless velar plosive /k/ is uvular [q]. There is very little published data on the difference between velar and uvular stops [16]. Post-velar or uvular consonants have phonetic effects on vowels and acoustic analyses of these effects have been done for languages such as Nuuchahnulth [27], Interior Salish [4], [5] and several dialects of Arabic including: Jordanian Arabic [28], Egyptian Arabic [9], and Palestinian Arabic [24]. Wilsons [27] study on Nuuchahnulth provides a very detailed description of both the leftward and rightward effects of post-velar on the neighboring vowels. He also reviewed previous researches on vowel lowering / retraction and schwa offglides and onglides. It was found in his study that the F1 of /i/ following uvular consonants is significantly higher than when it follows non post-velar consonants. However, his study did not include the effects of uvular consonants on F2 of vowels. Lowering of F2 of the vowel following the uvular stop occurs due to the retraction of the back of the tongue root when the uvular stop is articulated. Zawaydehs [28] acoustic analysis of uvularization in Ammani-Jordanian Arabic supports this idea of F2 lowering of the neighboring vowels. Figure 16 shows the wide-band spectograms of /g/ and /k/. Notice the lowering of the F2 value of the vowel when preceded by the uvular plosive.

/a/ /e/ /i/ /o/

675 613 377 436

1879 2231 2568 1111

810 723 605 472

1524 1821 2193 899

Table 3: Mean values in Hertz of the F1 and F2 for each vowel following the velar and uvular plosive.

Figure 17: F1 and F2 plot of the four Kalanguya vowels following the velar and uvular plosive. The plot of the means of F1 and F2 of the four vowels following the velar and uvular stop shows that the space of the vowels following /g/ slightly overlapped with the space of the vowels following /k/. We cannot say that the difference is due to the fact that /g/ is voiced while /k/ is voiceless. Voicing might have some effects on vowel quality but they do not have significant effects on F1 and F2. The most affected vowel is /i/ due to its height and frontness. When /i/ is articulated, the tongue body is high and forward and the tongue root advanced, whereas uvular consonants have the tongue back and the root retracted. Thus there is a gestural conflict when /i/ follows an uvular consonant [27]. We can see that the high front vowel /i/ in Kalanguya is never realized as [i] or [] but rather it is lowered to mid-front vowel [e]. Wilson [27] also stated that uvular consonants have a negligible effect on vowels /u/ and /a/ in terms of vowel height or F1. However, this is not the case in Kalanguya because all the four vowels have higher F1 when they follow /k/. One of the leftward effects of uvular consonants on vowels is the occurrence of schwa offglides on high front vowels. Carlson et al. (as cited in Wilson) [27] state that a schwa transition occurs between /i/ and /q/ in /iq/. This means that as the tongue retracts from /i/ to /q/ position, the tongue is expected to pass through schwa space in the course of its transition [11]. This is also true in Kalanguya when /i/ occurs before /k/. We can see in Figure 18 that there is a transition from /i/ to a schwa vowel // at the end of the vowel.

/k

/ /g

Figure 16: Wide-band spectograms of /ke/ and /ge/. A slight rise in F1 value and lowering of F2 are some of the effects of the uvular stop to the neighboring vowels. Looking at the results of F1 and F2 measurements of four Kalanguya vowels in Table 3, we can see that vowels following /k/ have higher F1 and lower F2 than when they follow the voiced velar stop /g/.

/g/ vowel F1 F2 F1

/k/ F2

/l

w/

Figure 18: Wideband spectogram of /i/ before /k/ with the formants represented by the red dots. Based on the results of the formant measurements and spectographic investigation of the neighboring vowels, Kalanguya /k/ is articulated as uvular stop and will be phonetically transcribed as [q]. The results presented here remain suggestive. Further studies such as spectral, cinefluorographic and perceptual analysis may help us clearly determine the differences between uvular and velar plosives.

Figure 19: Plot of the F1 and F2 in Hertz of 576 vowel tokens of both male and female Kalanguya speakers.

3.5 Kalanguya Vowels 3.5.1 Acoustic Properties of Kalanguya Vowel Space

The F1 and F2 for each of the 576 vowel tokens were determined with the BURG algorithm as built into the Praat Program. The stable region of the vowel was visually determined by looking at the spectogram and then the formant values were measured at the center of the stable region. Table 4 gives the mean values in Hertz of the F1 and F2 for each vowel in Kalanguya from three male and three female native speakers. Then the F1 and F2 of the 576 vowel tokens are plotted in Figure 19 and the mean values plotted in Figure 20. Vowel /a/ /e/ /i/ /o/ F1 720 573 397 447 F2 1360 1624 2158 965

Figure 20: Plot of the mean values of F1 and F2 for each vowel. One remarkable characteristic of the vowel space of Kalanguya is that both the F1 and F2 of each vowel differ from other vowels significantly. Unlike other languages, Kalanguya has no internal vowel symmetry; the high front vowel has no corresponding back vowel and the mid front vowel /e/ has corresponding back vowel. Unlike other languages with small vowel inventories, the /i/-/u/ contrast does not apply to Kalanguya. A comparison F1 averages of /i/ and /o/ using student's t test provided significant evidence of difference [t(10)= 3.166, p = 0.0101]. The difference in F1 of /o/ vs /e/ is considered to be extremely statistically different [t(10)=7.212, p < 0.0001]. Finally, the average F1 of /a/ and /e/ is considered to be very statistically different [t(10) = 3.670, p = 0.0043]. Three front vowels are present in Kalanguya: /i/, /e/ and /a/. But the F2 values of /i/ and /e/ differ significantly [t(10) = 10.5345, p < 0.0001] and a comparison of the F1 value of /a/ and /e/ provided evidence of difference [t(10) = 5.8354, p = 0.0002]. This suggests that /i/ is the only front vowel in Kalanguya and both /e/ and /a/ are central. Based on the results of the F1 and F2 average for each vowel, the following phonetic transcription will be designated: (1) The

Table 4: Mean values of F1 and F2 in Hertz for each vowel in Kalanguya

vowel /i/ is phonetically realized as near close near front [] and (2) The vowel /o/ is phonetically realized as near close near back []. Although the average F1 values of /i/ and /o/ differ significantly in Kalanguya, back vowels having slightly higher F1 value than their front counterparts is universal [10]. (3) The vowel /e/ will be transcribed as open mid central [] because the average F2 is significantly lower than the F2 of the front vowel /i/ and the F1 value is slightly higher than the F1 value of [e] in Filipino (F1=487) [23], English (F1=476) [14] and Spanish (F1=458) [6]. Finally, (4) the low vowel /a/ is phonetically realized as a low central vowel [a]. The phonetic symbol [a] refers to both front low vowel and central low vowel.

3.5.2 Gender-Specific Differences and Effect of Stress on Kalanguya Vowel Space

For each of the four vowels, the following results were obtained for the mean of the F1 and F2 of both male and female speakers. Vowel /a/ /e/ /i/ /o/ F1 644 541 378 430 F2 1310 1646 2096 923 Figure 21: Male and female vowel space. Vowel height seems to be most affected by gender but the difference of F1 for /i/ is considered to be not statistically significant [t(4)= 1.6642, p = 0.1714]. For other vowels, the difference of male F1 and female F1 is statistically significant. This is also true for F2 of all the vowels. Female speakers might have higher F2 value than male speakers but the difference is not statistically significant. As for the effects of syllable stress on vowel formants, the difference of F1 and F2 of the stressed vowel and the F1 and F2 of the unstressed vowel is not significant. This is true for all four vowels of Kalanguya and for all speakers. This may suggest that the acoustic correlates of stress do not have any effect on Kalanguya vowel space. However, discussion of this point is beyond the scope of the study.

Table 5: Mean Values in Hertz of the F1 and F2 for Each Vowel of Male Speakers Vowel /a/ /e/ /i/ /o/ F1 795 606 415 465 F2 1409 1601 2220 1008

Table 6: Mean Values in Hertz of the F1 and F2 for Each Vowel of Female Speakers Kalanguya exhibits the following near-universal effects of gender on vowel quality: (1) Females have higher F1 and F2 than males. (2) Females tend to have a larger vowel space than males. Males have the tendency to reduce their F1 space size because their F1 values are easier to discriminate by listeners than female F1 values (Goldstein, 1980; Ryalls and Lieberman, 1982; Diehl et a1., 1996 as cited in Escudero et al. [10]).

4. DISCUSSION 4.1 Kalanguya Phonetic Inventory

* b - voiced labialized bilabial plosive

Figure 22: Phonetic Inventory Chart of Kalanguya

The complete and precise phonetic inventory of Kalanguya based on the results of the study is presented in Figure 22. The voiced bilabial plosive /b/ is articulated with the lips rounded therefore phonetically transcribed as [b] to represent the secondary articulation involved when uttering the sound. Kalanguya lacks the alveolar flap // and the alveolar fricative /s/; both sounds can be found in other Philippine-type languages. The four coronal consonants are considered alveolar instead of labeling them as dental because the alveolar ridge is where most of the contact is made. The glottal fricative /h/ is considered as a true fricative in this paper due to the fact that it can occur both at the beginning and at the end of syllables and even before consonants. This compensates for the lack of the alveolar fricative /s/ in Kalanguya. Finally, Kalanguya /k/ is phonetically realized as voiceless uvular stop [q]. Effects of the uvular on the neighboring vowels prove that /k/ is uvular. High vowels following [q] are always lowered to a mid-vowel and an excrescent schwa occurs in between a high vowel and [q] when the high vowel precedes the uvular stop. Kalanguya has a four-vowel inventory: one front vowel [], two central vowels [] and [a], and one back vowel []. Although [] and [] are both considered as near-close vowels, their F1 differs significantly. This means that Kalanguya vowel inventory has no internal symmetry in terms of F1 and F2 measurements. Moreover, gender differences and syllable stress have no effects on vowel height and frontness. Related studies such as acoustic correlates of stress and other prosodic features are recommended to further improve our understanding of Kalanguya phonology and phonetics. Other methods such as cinefluorography, electropalatography, spectral analysis and aerodynamic investigations may enhance the reliability of the findings in this study.

[1] Anderson, Victoria. 2008. Static palatography for language fieldwork. Language Documentation & Conservation 2(1):1 27. [2] Anderson, V., I. Ko, and W. OGrady. 2003. A Palatographic Investigation of Place of Articulation in Korean Coronal Obstruents. Korean Linguistics 12.1-24 [3] Arsenio, B. and G. Stallsmith. 2008. Preserving livingm traditions in live performances: A traditional music and dance troupe of the Kalanguya of the Northern Philippines. The 2nd International Conference on Language Development, Language Revitalization, and Multilingual Education in Ethnolinguistic Communities, Bangkok, Thailand. Bessell, Nicola J. 1998a. Phonetic aspects of retraction in Interior Salish. In Salish languages and linguistics: Theoretical and descriptive perspectives, ed. Ewa Czaykowska-Higgins and M. Dale Kinkade, 125-152. Berlin: Mouton de Gruyter. Bessell, Nicola J. 1998b. Local and non-local consonantvowel interaction in Interior Salish. Phonology 15:1-40.

[4]

[5]

[6] Bradlow, A.R. 1995. A comparative acoustic study of English and Spanish vowels. Journal of the Acoustical Society of America., 97, 1916-1924. [7] Boersma, P., and D. Weenink. 2009. Praat: Doing Phonetics by Computer Version 5.0.43 Computer program, Retrieved 1 December 2009 from http://www.praat.org/. [8] Dart, Sarah. N. 1991. Articulatory and Acoustic Properties of Apical and Laminal Articulations. UCLA Working Papers in Phonetics 79, 1155. [9] Elgendy, Ahmed M. 2001. Aspects of pharyngeal coarticulation. Amsterdam: Netherlands Graduate School of Linguistics.

4.2 Documenting the Sounds of Philippine -Type Languages

This is one of the first attempts to investigate the articulatory and acoustic features of a Philippine-type language and perhaps the first to provide a complete phonetic description using static palatography and acoustic analysis. There are more or less 170 languages spoken in the Philippines and a few of them are becoming severely endangered. Majority of the sounds of the languages of the Philippines are fairly common but some languages have unique sounds that are needed to be documented and preserved before they become completely extinct. It is now the time for linguists and language researchers and members of different ethnolinguistic groups in the Philippines to start to engage in data collection and documentation of not only the sounds of Philippine-type languages but also other linguistic structures and practices of these languages.

[10] Escudero, P., P. Boersma, A. Schurt-Rauber and R. Bion. 2009. A cross-dialect acoustic description of vowels: Brazilian and European Portuguese. to appear in Journal of the Acoustical Society of America. [11] Gick, Bryan and Ian Wilson. In press. Excrescent schwa and vowel laxing: Cross-linguistic responses to conflicting articulatory targets. In Laboratory Phonology 8, ed. Louis Goldstein, Douglas Whalen, and Catherine T. Best. Berlin: Mouton de Gruyter. [12] Gordon Matthew. 2003. Collecting Phonetic Data on Endangered Languages. 15th International Congress of Phonetic Sciences, 207-210 [13] Gordon, M. and I. Maddieson. 1999. The Phonetics of Ndumbea. Oceanic Linguistics, 38, pp. 66-90. [14] Hillenbrand J., L. A.Getty, M. J. Clark, and K. Wheeler. 1995. Acoustic Characteristics of American English Vowels. Journal of the Acoustical Society of America 97, pp. 3099 3111.

5. REFERENCES

[15] Hohulin, L and M. Kenstowicz. 1979. Keley-I Phonology and Morphophonemics. In Nguyen Dang Liem (ed.) South-East Asian Linguistic Studies Volume 4, Pacific Linguistic Series C-No. 49 the Australian National University, 241-254. [16] Ladefoged, P. and I. Maddieson. 1996. The Sounds of the Worlds Languages. Oxford: Blackwell. [17] Ladefoged, Peter. 2003. Phonetic Data Analysis. Oxford: Blackwell. [18] Lewis, M. Paul. 2009. Ethnologue: Languages of the World, Sixteenth edition. Dallas, Texas. SIL International. Online version: http://www.ethnologue.com/. [19] Maddieson, Ian. 2008. Uvular Consonants. In: Haspelmath, Martin & Dryer, Matthew S. & Gil, David & Comrie, Bernard (eds.) The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, chapter 6. Available online at http://wals.info/feature/39. Accessed on November 15, 2009 [20] Mayes, Sharon. 1979. An Acoustic Analysis of Thompson Velar /k/ and Uvular/q/. University of Hawaii Working Papers in Linguistics 11, 11-22. [21] New Tribes Mission. 2001. Hota Bllon Tolagan Diyoh (The New Testament). 2nd Edition. Bambang, Nueva Vizcaya: Philippine Bible Society [22] Rice D. 2008. The Search for Abundant Life. Available online at http://anthropologywatch.blogspot.com/2008/04/search-forabundant-life.html. Accessed on November 25, 2009 [23] Santiago, P.J., Delos Reyes, J. Tadena, D., Zubiri, L.A. 2009. Acoustic Characteristics of the Filipino Vowel Space. In Proceedings of the 6th National Natural Language Processing Research Symposium. De la Salle University, Manila, Philippines. [24] Shahin, Kimary N. 2002. Postvelar harmony. Amsterdam: John Benjamins Publishing Company. [25] Strange, W., A. Weber, E. Levy, V. Shafiro, M. Hisagi, and K. Nishi. 2007. Acoustic variability within and across German, French, and American English vowels: Phonetic context effect, Journal of the Acoustical Society of America, 122 (2), 1111-1129. [26] Taff A., L. Rozelle, T. Cho, P.Ladefoged, M. Dirks and J. Wegelin. 2001. Phonetic structures of Aleut. Journal of Phonetics 29, 231271 [27] Wilson, Ian. 2007. The Effects of Post-Velar Consonants on Vowels in Nuu-chah-nulth: Auditory, Acoustic, and Articulatory Evidence. Canadian Journal of Linguistics 52:1/2. 43-70. [28] Zawaydeh, Bushra Adnan. 1997. An Acoustic Analysis of Uvularization Spread in Ammani Jordanian Arabic. Studies in the Linguistic Sciences, 27, 185-200.

[29] Zhobov, Vladimir. 2004. Uvulars in the Erkec Dialect Location: Global, Area, and International Archive. Retrieved from: http://www.escholarship.org/uc/item/153487vt. January 10, 2010 [30] Zygis, M., D. Recasens, and A. Espinosa. 2008. Acoustic Characteristics for Velar Stops and Velar Softening in German, Polish and Catalan. In R.Sock, S. Fuchs, Y. Laprie Eds., Proceedings of the Eight International Seminar on Speech Production. pp. 97100, Strasbourg.