DURATION FORMANT PATTERNS AND ACOUSTIC SPACE OF PAHARI MONOPHTHONGS.
The primary aim of the study is to present the duration formant patterns and acoustic space of Pahari Monophthongs. To achieve this aim an experimental study is conducted. Ten Pahari native speakers participated in the study and were given a list of 12 Pahari monophthongs in CVC context for recording where V is the target monophthong. The results show that the monophthongs are described as occurring in the form of long-short pairs that differ significantly quantitatively (vowel duration). Long vowels are over two times longer (e.g. / o:/and / o/ are 221 ms and 88 ms respectively) in terms of duration than that of their short-counterparts. Spectral differences show that Pahari has four close vowels six mid vowels and two open vowels. Spectral differences also occur among long-short pairs i.e. all the short front vowels show low F2 while all the short back vowels show higher F2 than their corresponding long vowels.
The study further shows that long vowels are peripheral in quadrilateral while short vowels are a bit centralized as compared to their corresponding long vowels. Long vowels enclose more vowel space than that of short vowels.
Vowel inventories in the world's languages differ considerably in size. Maddieson's (1984) survey of the corpus of 317 languages shows that 5.7% languages have only three vowel phonemes while at the other extreme 4.1% have 17 or more contrastive vowel categories. The most common vowels are: Ia! /i/ Iu/. On the basis of number of vowels in a language Maddieson (1984) divides the vowel inventories into: small vowel inventory (2-4) average vowel inventory (5-6) and large vowel inventory (7-14).
In the small amount of literature previously written about Pahari only a short description is found (Karnai 2003). It addresses the topic of vowels. In terms of vowel inventory Pahari is an interesting case to study as it has nineteen vowels both oral and nasal (Karnai 2003). But he did not use IPA symbols for vowels; rather he uses Urdu alphabets.
Khan (2012) presents the descriptive study of Pahari vowels based on minimal pairs and distribution of sounds in words. He established that Pahari has twelve vowel system. This study aims to analyze the Pahari vowels acoustically to establish the vowel inventory of the language in the form of quadrilateral.
2. Literature Review
2.1 Temporal Characteristics
Cochrane (1970) states that vowels may be distinguished in terms of duration for those languages and dialects that employ phonemic vowel length. This durational contrast may or may not be the only distinctive feature between two vowels. For example English short vowels / i I /o/ and /o/ are qualitatively distinct from their corresponding long vowels. According to Tsukada (2002) vowel duration is used contrastively in some languages but not in others. He states that English and Hindi differ from Japanese and Thai since both these languages use vowel duration as an acoustic cue for the length distinction in addition to qualitative differences to maintain the contrast between lu and /i:/ /u/ and lu:/ Ia! and /a:/. It shows that in Japanese and Thai length contrast is a prominent cue whereas English uses other acoustic cues such as spectral differences.
Watson and Harrington (1999) claim that vowel classification experiments show increased accuracy when frequency and durational information are combined. According to Cochrane (1970) the major difference between long and short vowels is simply one of total vowel duration. However the difference is relative rather than absolute as contextual and prosodic factors affect the ultimate length of the vowel. Peterson and Lehiste (1960) describe short vowels as showing a short target and slow relaxation whereas for long vowels the target is maintained for longer followed by a more rapid off glide. Lindblom (1967) documents that openness is positively correlated with length and therefore open vowels tend to be longer than close vowels. He suggests that this is universally the result of the increased biomechanical effort required to produce low vowels.
Erickson (2000) presents summaries of the factors that affect segmental duration including speaking rate phonological/phonetic influences such as inherent segmental duration and the effect of a postvocalic consonant.
2.2 Vowel Formant Characteristics
In articulatory phonetics vowels are mainly described in terms of three features: (1) height of the tongue (2) backness of the tongue and (3) lip rounding. Acoustic studies approach the description of vowel differently. "An acoustic analysis of vowel stresses the different formant configurations that are characteristic of each vowel.
The relationship among the vowels can be examined by comparing their formant values" (Olive Greenwood and Coleman 1993). Vowels are frequently described with reference to their formant structure which provides an indication of vocal tract resonance and therefore articulatory shape (Fant 1960).
The high-low and front-back distinctions are represented by the first and second formants on the spectrogram (Olive Greenwood and Coleman 1993). First Formant (henceforth Fl) shows the high-low distinction. That is the lower the formant value the higher the vowel. Second Formant (henceforth F2) shows the front-back distinction. If the value of F2 is high the vowel is closer to the front position. The relationships between (Fl) and the height of tongue and F2 and the front/back dimension ensures that when the first two formants of a set of vowel targets are plotted on axes with appropriate scaling characteristics the result closely resembles the traditional auditory vowel map. Such vowel spaces with axes Fl and F2 rely on the concept of the vowel target. The target is the vowel component least influenced by its surrounding phonetic context and it is considered to be either a point in the time course of the vowel or else a section of time during which the vowel position remains stable.
A single point is often used to provide a representation of the target position and for most vowels this can be assumed to be approximately mid way though the nucleus.
2.3 Vowel Space
Catford (1988) talks about the concepts called "vowel space" and "vowel limit". He says that the idea of the Cardinal Vowels by Daniel Jones is based on the concept that the vowels are limited by vowel space. In the production of a vowel there is a certain fixed area within oral cavity beyond which the vowel takes space of an approximate type sound.
Thus theoretically speaking "any vowel of any language must have its tongue-position either on the vowel limit itself or within the vowel space" (Catford 1988 p. 130).
Maddieson (1984) claims that vowel inventories in his sample vary from having three to fourteen distinct vowel qualities with two-third of the languages having between five and seven distinct vowel qualities.
He further makes a claim that the specific vowels that comprise these statistically preferred vowel inventories tend to be the same. For example five-vowel systems tend to have /i e a o u/ seven-vowel systems tend to have these five vowels plus /e/ and /0/ and six-vowel systems usually have /i e o o u/. Furthermore the vowel inventories of the vast majority of the worlds languages include the three vowels that define the extremes of the general vowel space namely /i a u/. Accordingly these three vowels are known as the "point vowels" and have been awarded a special status in theories of vowel systems. Previous work has led to the development of several theoretical positions regarding the structure of vowel systems. Dispersion Theory (DT) claims that the vowels of a given language are arranged in the acoustic vowel space so as to minimize the potential for perceptual confusion between the distinct vowel categories.
Using computer programs to generate the optimal configurations for vowel systems of various sizes this approach to vowel inventories has been proved fairly successful (Liljencrants and Lindblom 1972; Lindblom 1975 1986; Disner 1984). However these investigations of DT focus exclusively on intercategory distance as the determiner of vowel system configuration in a universally defined acoustic vowel space. As a result this approach fails to account for the observation that certain languages such as Swedish with nine vowels and Danish with ten vowels crowd their vowels into a small corner of the entire vowel space rather than dispersing them throughout the available space (Disner 1984).
In more recent developments the dispersion principle has been expressed as a principle of sufficient rather than maximal contrast (Lindblom 1989 1990). Furthermore the theory has been extended to account for within-speaker variation. A speaker's vowel space will be expanded relative to his or her casual speech vowel space.
The Quantal Theory of Speech (QTS) (Stevens and Blumstein 1975) suggests an alternative approach to vowel systems. According to this theory there are certain regions of stability in the phonetic space. In particular it is claimed that there are stable regions corresponding to the point vowels lu Ia! and /u/. Thus this theory predicts that the point vowels should be in approximately the same locations across all languages regardless of vowel inventory size. Furthermore QTS predicts that since the point vowels are in phonetically stable regions they should show less within-category variability than non-point vowels.
Ten Pahari native adult speakers (five male and five female) participated in this study. The participants selected for this study were the residents born in Poonch dialectal region and had lived there in most part of their life. They all speak Pahari with their friends family and at market places as well. They speak Urdu with people who do not speak Pahari.
Their ages range from 20 to 40 years at the time these recordings were made in October 2010. All the participants were interacted with to ensure that they had no hearing or articulation problems.
3.2 Data Collection
Monosyllabic words exemplifying 12 monophthongs of Pahari were selected such that the target vowels occur between /m/ and /1/. A list containing CVC context was constructed and all the participants were asked to read each word three times in random order. The participants were instructed to read words with normal speed. Ten speakers of Pahari give a total of 360 tokens (ten speakers x three repetitions x 12 vowels). All recordings were made in Muzaffarabad.
They were recorded directly on PRAAT software by using high fidelity microphone. In addition participants were also given a list of words containing all the vowels to pronounce and they all were recorded.
3.3 Data Analysis Procedure
Spectrograms were used to determine first two formant frequencies. Measures of the lowest two formants of vowels were made using Praat software. The vowels were segmented on the basis of visual information in a wide band spectrogram. Fl and F2 were determined and measured in Hz in the middle of the target vowel since it can be assumed that the influence of an adjacent segment is minimal and the articulatory target is maximally achieved in this position. The target is the vowel component least influenced by its surrounding phonetic context and is considered to be either a point in the time course of the vowel or else a section of time during which the vowel position remains stable. A single point is used to provide a representation of the target position and for most vowels this is assumed to be approximately mid way through the nucleus.
Figure 3.1 shows the spectrogram of a speaker's utterance `mal' with the target vowel [a] selected in three cursors. The mid cursor indicates the middle of the formants at which measurement of the formant frequencies are taken in Hertz (Hz) while the two extreme cursors measure the duration of the vowel from the beginning of the sound (left) to the end (right) in seconds.
3.4 Statistical Analysis
An inferential statistical test repeated measures ANOVA was used with alpha level set to .05. In the vowel duration studies the independent variable was vowel (phoneme) and the dependent variable was each speaker's mean vowel duration. In the vowel quality studies the independent variable was Vowel (phoneme) and the dependent variable was each speaker's mean Fl and F2.
4. Results and Discussion
As discussed in section 3.2 the parameters selected for acoustic analysis are vowel duration and the formant frequencies (F1 F2). Table 2 shows mean vowel duration and Table 4 illustrates mean F1 and F2 values with standard deviations for the data from ten Pahari speakers in CVC context. In both the tables the standard deviation represents the variance of means of the three tokens for each of the ten speakers. The measurements of three tokens within each participant are treated as repeated measures and therefore averaged.
4.1 Vowel Duration
The means for the duration values are provided in Table 2 while Figure 2 illustrates the relative nucleus durations for each vowel. It is observed that when speakers of the same geographical region produce same vowel the result is different vowel durations. This is because some speakers speak faster and some speak slower.
The bar chart above shows the average duration of each long and short vowel in the speech of selected participants. The data indicate that durational contrast between long and short vowels is very much prominent and distinctive. The average duration of long and short vowels is 224 ms and 94 ms respectively which is shown by last two bars on bar chart above. The short vowels are approximately 40 percent the length of their corresponding long vowels. The data further show the increase in length with vowel openness. Ia:! has 235 ms duration that is longer than that of two other front mid and front close vowels /e:/ and /i:/ respectively. The length of back vowels decreases with vowel closeness as the close back vowel /u:/ is shorter than mid back vowel /0:!. It is also evident that all the long vowels are over two times longer than their corresponding short vowels.
Statistical analysis also proves that vowel duration was significantly affected by length. The difference between short and long vowels was significant (F (1 764) = 4350.636 p= .000). A significant difference in terms of duration is found between long and short vowels but no significant difference is found within long vowels and same is for the short vowels.
The above results show that durational contrast is significant. It is phonemic in Pahari. Though it is the most distinctive and prominent and does help the listener to place vowels in large categories such as long and short it is not sufficient in itself to enable identification of any individual vowel. For example [e] and [i] can't be differentiated on the basis of duration as both are short and have mean duration 93 and 92 ms respectively.
In addition the data analyzed also gives evidence of six short oral vowels in Pahari corresponding to each of the long oral vowels. For example long vowels map into short vowels when grammatical categories of certain words are changed. Some examples of this process of vowel mapping are listed in Table 3.
It can be concluded from above discussion that the duration contrast in Pahari is phonemic. Only quantity (duration) is not enough to identify the individual vowels. To identify individual vowels along with quantity quality cues are also important.
4.2. Formant Frequencies
Table 4 gives the mean F1 and F2 values with standard deviations for the data collected from ten Pahari speakers in CVC context. Figures 3 and 4 display the mean F1 and F2 frequencies in Hz of the monophthongs respectively.
The bar diagram above shows that there is very little difference between the F1 values of short and long vowels. The correlation is statistically insignificant at p greater than 0 .05.
Figure 4 also shows that there is very little difference in F2 values of long vowels and their corresponding short vowels. The correlation is statistically insignificant at p greater than 0 .05.
Figures 5 and 6 show that for front vowels F1 becomes lower when the constriction in the oral cavity increases. As /i:/ is the most constricted front vowel it has the lowest F1. It means that F1 increases as the tongue position gets lower for front vowels i.e. F1 of high front vowel /i:/ is 310 Hz while F1 of mid front vowel /:/ is 567 Hz. In case of back vowels F1 decreases with the height of the tongue i.e. mid back long vowel /o:/ has 460 Hz while high back long vowel lu:/ has 370 Hz. In contrast to F1 /i:/ has the highest F2 and lu:/ has the lowest F2. This suggests that high vowels have low F1 and low vowels have high F1. Pahari has only one low vowel /a:/ and it has the highest F1 value (625 Hz). Figure 6 also displays that the maximum separation between F1 and F2 occurs with the close front vowels and it is the smallest with the low vowels. For back vowels F2 is much lower and closer to F1.
4.3. Vowel Space
This section shows the vowel plots that were generated with Plot Formants. The graphs show F2 on the horizontal axis and F1 on the vertical axis.
Figure 7 shows the acoustic space enclosed by long and short vowels. As it is seen all the short vowels are a bit centralized and do form the inner circle of vowel space as compared to their corresponding long vowels which are distributed peripherally in vowel space. This shows that the acoustic space enclosed by long vowels is more than that for the short vowels.
Figure 8: Quadrilateral of Pahari vowels
Figure 8 shows that only 2 vowels one long and one short are central while other 10 are peripheral in Pahari. The long and short vowels are located close to each other in the quadrilateral. This shows that there is very little spectral/qualitative difference between long and short vowels.
This suggests that duration is the primary and distinctive cue in Pahari. Figure 8 further exhibits that all the short vowels are a bit centralized as compared to their long counterparts. Another trend appears that F1 of 1mid short vowels is lower than that of their corresponding long vowels.
In contrast all the close central vowels have higher F1 than that of their long counterparts.
4.4 Pahari Vowels Description
4.4.1 Front Vowels
Pahari has three sets of front vowels: a close front and two mid front. It has two high front vowels: long [i:] as in [pi:p]pus' and short [i] as in [lit]' a piece of wood'. These two vowels differ in length as well as in quality. As the data show F1 and F2 of [i:] are 310 and 2234 Hz while F 1 and F2 of [i] are 332 and 2140 Hz respectively. Lower F2 value of lu than /i:/ suggests that /i/ is inclined toward center. [i:] has the lowest F1 value and the highest F2 value among the entire vowel system which makes this vowel a high front vowel. In the articulation of [i:] the body of the tongue moves forward and upward in the direction of the hard palate. The lips are widely spread. The vowel /i:/ has shown the maximum separation between F1 and F2 among the twelve vowels of Pahari. Vowel duration is 218 ms for /i: /.This can be classified as high-front or close-front long vowel Both are unrounded vowels.
On the other hand /i/ has the 2nd lowest F1 value and the 2nd highest F2 value among the close front vowels. In articulatory terms the body of the tongue is displaced upward and forward in the direction of the hard palate. The lips are spread. Vowel duration is 93 ms for / i I. This vowel can be classified as high-front or close-front short vowel of Pahari.
The first set of the two mid front vowels is: long [e:] as in [k"e:l] game' and short [e] as in [beln] `rolling'. As the data show F1 and F2 of [e:] are 412 and 1909 Hz while F1 and F2 of [e] are 401 and 1850 Hz respectively. These are found just below cardinal vowel 2. Duration of [e:] is 222 ms. In articulatory terms the body of the tongue moves forward and upward in the direction of the hard palate for this vowel of Pahari. The lips remain neutral and the mouth remains half closed for this vowel. Since the tongue body is displaced forward and upward in the direction of the hard palate this vowel can be classified as mid-close front long vowel of Pahari. Like [e:] the production of [e] involves the same articulators (the body of the tongue moves forward and upward in the direction of the hard palate). The difference is of duration. [e] has 92 ms duration and it can be classified as mid-close front short vowel Both are unrounded vowels. This is mid-close short vowel in Pahari.
The second set of two mid front vowels is: [a:] and . It is found in words like [g:v]'cow' and [ml] wb'. These are found just above the cardinal vowel 3. Their F1 are 578 and 567 Hz and F2 are 1730 and 1660 Hz respectively. In articulatory terms the body of the tongue moves forward and upward in the direction of the hard palate from neutral position to the mid high position in mouth for both the vowels.
They are found above cardinal vowel 3. Both are unrounded vowels. Their duration is 228 ms and 99 ms respectively. They are classified as mid-low front long and mid-low front short vowels respectively in Pahari.
4.4.2 Central Vowels
Pahari has two central vowels [a:] and [a] which contrast in length. They are found in words like [ba:l] `hair' and [bol] `alright' respectively. Their FT and F2 values are 625 andll43 and 582 and 1158 Hz respectively. has a high F1 frequency value and a middle range F2 value which makes it a central vowel. In articulatory terms the body of the tongue is raised up in the direction of hard palate from the neutral position to the middle position in the mouth. The lips remain neutral and the mouth remains opened. Mean duration of  is 96 ms and is classified as mid central short vowel. [a:] has the highest F1 frequency value among the twelve vowels of Pahari and a mid range F2.
In articulatory terms the tongue body is raised up towards hard palate from the neutral position. The lips remain neutral and the mouth remains widely opened during the production of this vowel. [a:] is open- central long vowel with mean duration 235 ms.
4.4.3 Back Vowels
Pahari has two sets of back vowels: Set 1 [o: and o] is articulated from the back mid position as in [kho:a:] `donkeji' and [khotta:] `mean' while set 2 [u: and o] is articulated from the high back position as in words [bu:ta:] plant' and [mol]/e This set of vowels has middle range F1 value (460 Hz) and a small F2 value (955 Hz) which make these vowels back series vowels. The tongue body moves upward and backward in the direction of the soft palate. The lips are rounded and the mouth remains half closed during the production of these vowels. Their mean duration is 221 ms and 88 ms respectively. They can be classified as mid-back long and mid-back short vowels respectively.
F1 and F2 of [u:] are 364 and 854 Hz while F1 and F2 of [u] are 370 and 950 Hz respectively. Table 3.16 shows that [u:] has the lowest F1 and F2 values among the back series vowels which make this vowel the high back or close back vowel of Pahari. In the production of both [u:] and [o] the tongue body moves backward and upward in the direction of the soft palate. The lips are rounded and the mouth is simultaneously narrowed down during the production of these vowels. Their mean duration is 218 ms and 96 ms respectively.
Both the vowels are rounded. They can be described as close-back long and close back short vowels respectively.
Based on the acoustic analysis it is concluded that Pahari operates on 12 vowel system. These vowels are presented in Table 6.
With respect to Pahari monophthongs the analysis shows that Pahari has four close vowels six mid vowels two open vowels. The data suggest that its monophthongs are described as occurring in the form of long-short pairs that differ significantly quantitatively (vowel duration) and less in terms of quality (spectral characteristics). This study shows that long vowels are over two times longer (e.g. /i:/ and /i/ are 218 ms and 93 ms respectively) in terms of duration than their short- counterparts. Spectral differences also occur among long-short pairs i.e. all the short front vowels show low F2 while all the short back vowels show higher F2 than that of their corresponding long vowels.
This shows that long vowels are peripheral in quadrilateral while short vowels are a bit centralized as compared to their corresponding long vowels. Long vowels enclose more vowel space than short vowels. Temporal cue is the most prominent as the language clearly exhibits long-short distinction and duration is phonemic in Pahari vowel system.
Amina L. (2002). Phonemic inventory of Siriaki and acoustic analysis of voiced implosives. CRULP 43-48.
Baart J. (1999). Acoustic phonetics. Dallas: Summer Institute of Linguistics.
Campbell W. and Isard S. (1991). Segment durations in a syllable frame. Journal of Phonetics 19 37-47.
Cochrane G. R. (1970). Some vowel durations in Australian English. Phonetica 22 240-250.
Cox F. (1999). Vowel change in Australian English. Phonetica 56 1-27.
Clark J. and Yallop C. (1992). An introduction to phonetics and phonology. Oxford: Blackwell.
Clements G.N. (1980). Vowel harmony in nonlinear generative phonology: an auto segmental model. Bloomington: Indiana University Linguistics Club.
Clements G. N. and Sezer E. (1982). Vowel and consonant disharmony in Turkish. In H.
van der Hulst and N. Smith eds The structure of phonological representations (pp. 213-55). Dordrecht: Foris Publications.
Davenport M. and Hannahs S. J. (2005). Introducing Phonetics and Phonology. London: Hodder Arnold.
Disner S. F. (1984). Insights on vowel spacing. In I. Maddieson Patterns of sounds 136-55. Cambridge: Cambridge University Press.
Erickson M. (2000). Simultaneous effects on vowel duration in American English: A covariance structure modelling approach. Journal of the Acoustical Society of America 108 29 80-2995.
Fant G. (1960). Acoustic theory of speech production. The Hague: Mouton and Co.
Jha S. K. (1985). Acoustic analysis of the Maithili diphthongs. Journal of Phonetics 13 107-115.
Jones G. E. (1984). The distinctive vowels and consonants of Welsh. In Ball and Jones (eds.) 40-64.
Karnai M. K. (2007). Pahari aor Urdu: ik Taqabali Jaiza. Islamabad: National Language Authority.
Kenstowicz M. (1970). On the notation of vowel length in Lithuanian. Papers in Linguistics 3 73-113.
Kent R. D. and Charles R. (2002). The acoustic analysis of speech. San Diego: Singular Publishing Group.
Khan A. Q. (2012). Phonolgy of Pahari: A Study of Segmental and Suprasegmental Features of Poonch Dialect. (Unpublished doctoral dissertation) University of Azad Jammu and Kashmir Muzaffarabad.
Khan A. Q. Sarwar N. and Bukhari N. H. (2011). Syllable Onset Clusters and Phonotactics in Pahari. Accessed from: www.languageinindia.com/march201/quadirkhanpaharifinal.
Klatt D. (1976). Linguistic uses of segmental duration in English: acoustic and perceptual evidence. Journal of the Acoustical Society of America 59 1208-122 1.
Ladefoged P. (2001). A course in phonetics. USA: Harcourt College Publishers.
Ladefoged P. and Maddieson I. (1996). Sounds of the world's languages. Oxford:Blackwell.
Ladefoged P. (1996). Elements of acoustic phonetics. Chicago: University of Chicago Press.
Lewis M. P. (ed.) (2009). Ethnologue: Languages of the World Sixteenth edition. Dallas Tex.: SIL International. Online version: http://www.ethnologue.com/
Liljencrants J. and Lindblom B. (1972). Numerical simulation of vowel quality systems:the role of perceptual contrast. Language 48 839- 62.
Lindblom B. (1967). Vowel duration and a model of lip-mandible coordination. speech Transmission Laboratories Progress Status Report 4 1-29.
Lindblom B. (1975). Experiments in sound structure. Paper presented at Eighth International Congress of Phonetic Sciences Leeds.
Lothers M. and Lothers L. (2010). Pahari and Pothwari: A sociolinguistics survey. Islamabad: Summer Institute of Linguistics.
Maddieson I. (1984). Patterns of sounds. Cambridge: Cambridge University Press.
Masica C. P. (1991). The Indo-Ayan languages. Cambridge: Cambridge University Press.
Olive J. Greenwood A. and Coleman J. (1993). Acoustic of American English speech. New York: Springer-Verlag.
Peterson G. E. and Lehiste I. (1960).Duration of syllable nuclei in English. Joumal of the Acoustical Society of America 3 693-703.
Pickett J. M. (1999). The Acoustic of speech communication fundamentals speech perception theory and technology. Boston: Allan and Bacon.
Stevens K. and Blumstein 5. (1975). Quantal aspects of consonant production and perception: a study of retroflex stop consonants. Journal of Phonetics 3 215-233.
Tsukada K. (2002). An acoustic comparison between American English and Australian English vowels. 7th International Conference on Spoken Language Processing ('ICSLP 2002) Denver Colorado USA. 2257- 2260.
Watson C. and Harrington J. (1999).Acoustic evidence for dynamic formant trajectories in Australian English vowels. Journal of the Acoustical Society of America 106 458-468.
|Printer friendly Cite/link Email Feedback|
|Publication:||Kashmir Journal of Language Research|
|Date:||Dec 31, 2013|
|Previous Article:||THE NATURE OF ERGATIVE CASE MARKING IN PAHARI.|
|Next Article:||GENDER CONSTRUCTION IN OBITUARIES: AN ANALYSIS OF PAKISTANI NEWSPAPERS.|