A QUANTITATIVE AND STYLO-STATISTICAL APPROACH TO THE PHONO-SEMANTIC STRUCTURE OF TWO COMPANION- PRAYERS OF THE QURʾĀN: AL-FALAQ AND AL-NĀS

Transcription

1 A QUANTITATIVE AND STYLO-STATISTICAL APPROACH TO THE PHONO-SEMANTIC STRUCTURE OF TWO COMPANION- PRAYERS OF THE QURʾĀN: AL-FALAQ AND AL-NĀS King Abdulaziz University, Jeddah-Saudi Arabia Abstract This paper is a linguistic investigation of al-falaq and al-nās, two companion-prayers in the Qurʾān. Although both prayers exhibit a marked symmetry on both lexical and phonological levels, this symmetry has not been extensively studied by scholars due to the highly familiar nature of these two short prayers. Immediately noticeable on the phonological scale is the highly cacophonous and staccato rhythms of al-falaq, which appear to be produced by a profusion of fricatives (/f/, /kh/) and plosives (/q/, /b/, /d/), combined with a scarcity of nasals and glides. In contrast, al-nās has a much smoother sound patterning as a result of the profusion of nasals and sibilants. This striking numerical difference in the phonological configuration of these two companion-prayers leads us to suspect the presence of phonosymbolism or a correlation between phonological patterning and subject matter. In other words, this variation can be accounted for by shifts in subject matter from the dynamic process of splitting in al-falaq to the movements of the Hisser (Satan) in al-nās. Subjecting the phonological data in both texts to simple statistical checks will allow us to be sure that these observed patterns are indeed statistically significant and not attributed simply to chance variability. The perspective that I am adopting here is stylo-statistical, where the main purpose is to devise a measure that is not only statistically satisfactory but stylistically interesting as well. Ilahiyat Studies Copyright Bursa İlahiyat Foundation Volume 2 Number 1 Winter/Spring 2011 p-issn: e-issn:

2 20 Key Words: Arabic phonosymbolism, phonological symmetry, stylostatistics, Arabic Qurʾān, iconicity, functions of fricatives and plosives 1. Introduction This paper is part of a larger linguistic investigation of al-falaq and al-nās, two companion-prayers in the Qurʾān that constitute chapters 113 and 114, respectively. Collectively, they are known as al-muʿawwidhatān (the two givers of refuge) and are considered to be highly familiar prayers, commonly recited by all Muslims and the first to be memorized by school children at a very early age. It is this very familiarity that somehow hinders us from perceiving the highly symmetrical patterning that is present in these companion prayers on both lexical and phonological levels. Immediately noticeable on the phonological scale is the highly cacophonous and staccato rhythms of al-falaq, which appear to be produced by a profusion of fricatives (/f/, /kh/, /gh/) and plosives (/q/, /b/, /d/), combined with a scarcity of nasals and glides. In contrast, al-nās has a much smoother sound patterning as a result of the profusion of nasals /n/, glides /w/ and sibilants /s/. My previous studies of these prayers focused on highlighting the componential process needed to translate non-core lexemes such as falaq, waqab and waswās and on revealing the phonetic iconicity operative in the texts through the intricate use of fricatives and plosives (Fatani, 2006; 2004; 2002a; 2002b). By mobilizing the results of these previous studies and by conducting an exhaustive quantitative analysis, this present contrastive investigation hopes to reveal the two-way phonological symmetry that binds both prayers together and to contribute to current research in language universals by providing new data on phonosymbolism from a non-indo- European language, i.e., from the Arabic language system. (A rudimentary outline of this quantitative approach to the phonosymbolism of the prayers was first presented in 2005 at the Fifth Conference of Iconicity in Language & Literature in Krakow, Poland, where the substantiation of this concept in medieval Arabic was received with enthusiasm.) It must also be noted at the outset that the concept of phonetic iconicity (dalālāt al-ṣawt or muḥākāt al-ṣawt li-l-maʿnā) has been studied by many medieval scholars of the Qurʾān, such as Ibn Jinnī (d. 393/1002), al-zamakhsharī (d. 538/1143) and Abū Ḥayyān (d. 745/1355). The concept can also be found scattered in

3 A Quantitative and Stylo-Statistical Approach to fragmented form in various exegeses of individual chapters of the Qurʾān. Modern Arab commentators (al-ʿabd, 1984; Bū ʿUmāma, 2002; Abū Mūsā, 1987; al-rāfiʿī, 1990; Ḥassān, 1993; al-ṣaghīr, 2002) have also contributed extensively to the domain. However, most studies focus on the phonosymbolical structures of individual iconic words. To date, no attempt has been made to reveal phonosymbolism in a whole sūra or to compare the phonosymbolism operative in two sūras or companion texts. It is taken as axiomatic that semantic significance of various kinds underlies such variation in language use. Hence, this striking numerical difference in the phonological configuration of these two companion-prayers leads us to suspect a correlation between phonological patterning and subject matter. In other words, this variation in the way consonant types are used can be accounted for by shifts in subject matter from the dynamic process of splitting (falaq) in al-falaq to the movements of the Hisser (al-waswās) in al-nās. I contend that these two dynamic processes are mapped onto the phonological structures of the prayers, hence producing two texts that are quite similar in syntax but strikingly different in terms of phonological construction. In this paper, my main objective is to subject the observed phonological data in both texts to thorough quantitative analyses and simple statistical checks that will allow us to be sure that the observed phonological patterns prevalent in both texts are indeed statistically significant and not attributed simply to chance variability. In more technical terms, these statistical checks will allow us to reject the null hypothesis, which states that any variation between the observed numbers in groups and what one would expect is due to chance. If there is a significant difference, the variation is more than is expected by chance, which suggests that some other factor is involved. The perspective that I am adopting here is stylo-statistical (Crystal, 1997), where the main purpose is to devise a measure that is not only statistically satisfactory but stylistically interesting as well. 2. Primary Observations Primary quantitative results reveal a significant difference between the data of obstruents (stops, fricatives and affricates) and sonorants (nasals, liquids and glides), reflecting a shift away from fricatives and stops in al-falaq (Text 1) in favor of nasals, glides and liquids in al-

4 22 Nās (Text 2). By making use of statistics, I have attempted in this paper to substantiate the initial impression of word-position choices and to provide quantitative evidence to prove the presence of a pervasive fricative-plosive word patterning in Text 1 and a nasal-sibilant word patterning in Text 2. It is my hypothesis that the sequential ordering of phonemes in these two adjacent texts are meant to mimic and reenact the external processes and experiences referred to on the lexical or semantic levels. Both prayers should therefore be perceived as iconic diagrams (Fischer & Nänny, 2001) in which the form iconically mirrors the content in the same way that charts, maps and graphs are motivated and governed by their real-world data. To test the validity of this phono-iconic structuring, a statistical comparison will be made between both adjacent texts. The striking numerical differences in their semantic and phonological configuration should sensitize us to the pivotal role that stylo-statistical studies can play in revealing layers of hidden meanings in Qurʾānic texts. Although this paper seeks to suggest that there are semantic constraints governing the use of obstruents and sonorants, no attempt will be made to provide an extensive analysis of the subject matter or the lexico-semantic structure of the texts; for that, the reader must look at the numerous exegeses of these highly familiar prayers. 3. Methodology As Crystal (1997: 67) observes, stylo-statistics investigates matters of frequency and distribution in three main areas: 1. Formal characteristics that do not relate directly to the meaning of a text, such as parts of speech, and the length of words, sentences or lines. 2. Characteristics that relate directly to meaning, such as the size and diversity of an author s vocabulary. 3. The detailed study of single words, or small sets of words, such as and, or the use of on vs. upon. This paper deals with the second area of investigation, that is, with characteristics that relate directly to meaning. However, the main focus is not on vocabulary but phonology. It is my contention that the consonantal structures of the content words (i.e., verbs, nouns and adjectives) in both texts, in particular the positional distribution of

5 A Quantitative and Stylo-Statistical Approach to plosives, fricatives and nasals in word-initial and word-final positions, function as an important instrument of disambiguation. Four basic steps are followed: (1) identifying the most-frequently occurring consonant types in content words, (2) devising a frequency and positional preference by counting all the instances of plosives, fricatives and sonorants and identifying the preferred position of each consonantal class, (3) placing these consonant classes in descending rank order of frequency and (4) comparing and contrasting the phonological structures of al-falaq and al-nās in terms of frequency, rank and preferred word-position of consonant types. To check the reliability of the data, two simple statistical checks of significance known as the preference test and the distinctiveness ratio (DR) are used (Kenny, 1982: 69-72). These are extremely versatile techniques applied to many different kinds of quantitative work where the basic problem is to compare the scatter of scores. They are employed here to look at differences in the way the phoneme classes are distributed among the content words in both texts. The results show that there is in fact a real and reliable relationship between the patterning of obstruents and sonorants and the variation in subject matter. With this limited amount of linguistic features and with such a small language sample, the analysis should corroborate the motivated use of the fricative-plosive pattern in al-falaq and the nasalfricative pattern in al-nās. Before presenting the texts, it is important to note from the outset that highlighting the interrelationship between form and subject matter necessitated a new translation of both prayers. Reliance on an existing English translation was not possible because no single published translation can answer the purposes of an article such as this. This is basically because most translators aim at rendering the communicative meaning of the text rather than the denotational meaning of individual lexemes. Crucial words to our analysis, such as splitting, sputterers and Hisser, even though they represent a corresponding one-to-one equivalency, have not been incorporated into any existing translation to date. 4. Stylo-statistical Analysis of al-falaq (T1) 4.1 The Text 1. Say I seek refuge with the Lord of the splitting qul aʿūdh u bi-rabb al-falaq

6 24 2. From the evil of what (He) creates min sharr i mā khalaq 3. And from the evil of a plunger when it flows wa-min sharr i ghāsiq in idhā waqab 4. And from the evil of the female-sputterers in the knots wa-min sharr al-naffāthāt fī l-ʿuqad 5. And from the evil of an envier when he envies wa-min sharr i ḥāsid in idhā ḥasad 4.2. Content Words and Function Words in al-falaq and al-nās Content Word List Function Word List qul say bi- with aʿūdh u i-seek-refuge al- the rabb lord min from falaq splitting mā what sharr evil wa and khalaq creates-3ms min from sharr evil idhā if ghāsiq plunger wa and waqab flows-3ms min from sharr evil al- the naffāthāt female-sputterers fī in ʿuqad knots al- the sharr evil wa and ḥāsid envier min from ḥasad envies-3ms idhā if 4.3. The Dynamic Process of Splitting Although this paper deals exclusively with stylo-statistics and does not offer the reader an extensive semantic interpretation of the text, one specific semantic issue should be borne in mind when examining the companion-prayers. The title-word al-falaq contains the central idea advanced by the text and is thus crucial to our analysis because it allows us to decode the prevalent iconicity of the text. A precise definition of the term is thus necessary from the outset of the study.

7 A Quantitative and Stylo-Statistical Approach to As evident from the word splitting, the English equivalent I have used, al-falaq is derived from the verb fa-la-qa the primary sense of which is to split, to cleave, or to break. Both the Arabic term and its English equivalent are commonly associated with physics and biology and are used to refer to the dynamic process of fission (Fatani, 2004: 174). The dynamic meaning of the word is all important to our analysis Frequency Distribution of Obstruents & Sonorants There are 15 content words (nouns and verbs) and 15 function words (prepositions and conjunctions) in the text. Within content words, there are a total of 40 phonemes, 30 obstruents (fricatives and plosives) and 10 sonorants (trills, liquids, nasals and glides), distributed as follows: Fig. 1. Frequency of distribution of obstruents & sonorants in al-falaq From this chart, we can see that the most popular values on the X- axis are the fricatives and the plosives, i.e., the obstruents. Together, these two high-frequency consonant classes account for 75% of all consonants in the lexical category of words (or display a 0.75 rate of occurrence). In terms of consonantal distribution, the two high points

8 26 of frequency represented by the fricatives and plosives indicate the bimodal nature of the text. In contrast, the sonorants have an extremely low rate of distribution, making up only 25% of the consonants. The ratio of obstruents to sonorants is thus 3:1. It is important to note at this stage that the sonorants almost always occur in monosyllabic words like qul, rabb, and sharr, which have less communicative value than the polysyllabic words. Within the polysyllabic category, we find that nasals and glides have a statistically significantly low rate of occurrence because they occur only once in the corpus. They are thus minus-words, which mark genuinely discriminating phonological characteristics of the text. In other words, the sonorants are strikingly less prevalent but nevertheless quite significant statistically by virtue of the fact that both nasals and glides occur only once in the whole text. In terms of percentage, the obstruent/sonorant data are as follows: Fig. 2. Distribution of obstruents & sonorants by percentage 4.5. Positional Frequency of Obstruents and Sonorants Obviously, the most crucial question to our study is what proportion of the text is made up of occurrences of the fricatives in wordinitial position and plosives in word-final position. To use Lyons

9 A Quantitative and Stylo-Statistical Approach to (1977: I, 43) terminology, what is the positional frequency of the occurrences of obstruents? The relevant data are as follows. Of the 17 fricatives, 10 are in word-initial position, 6 in word-medial and 1 in word-final position. Within the plosive category, 9 of the 13 plosives are in word-final position, 2 in word-medial and 2 in word-initial position. There are thus 10 words that have fricatives in word-initial position as opposed to 2 words that feature plosives. In word-final position, there is only one word that ends with a fricative as opposed to 9 words that end with plosives. Two words, waqab and naffāthāt, retain the final stops but substitute the fricatives with the approximates /w/ and /n/. Of the total number of content words (15), 6 start with fricatives and end in stops. Only one word reverses this phonological pattern, the matrix verb aʿūdh u, which starts with a plosive and ends with a fricative. The positional frequency of the occurrences (Lyons, 1977: I, 43) of obstruents and sonorants may be presented more clearly in graphical form: Fig. 3. Positional frequency of occurrence of obstruents & sonorants Positional Frequency of Occurrence in Polysyllabic Content Words (PCW) If we exclude the 6 monosyllabic words qul, rabb and the 4-times repeated sharr and count only the 9 polysyllabic words, an even

10 28 more striking patterning is revealed. We find that the 28 phonemes that make up the polysyllabic category of words are divided into 13 fricatives, 11 plosives, 2 liquids, 0 trills, 1 nasal and 1 glide. Hence, the obstruents (24 occurrences) now control 85% of the words. The following chart outlines the positional frequency of occurrence of obstruents and sonorants in the polysyllabic category: Fig. 4. Positional frequency of obstruents & sonorants in polysyllabic words As is evident from the chart, plosives in the final position represent the highest category in the scale (8 occurrences), followed by fricatives (6 instances) in both the initial and medial positions. The fact that sonorants are now reduced almost to a minimum and totally blocked from word-final position is evident; the trill is reduced to zero instances. Hence, we can conclude that the preferred phonological pattern in the text is that of a fricative-fricative-plosive structure. The following pie chart highlights this striking phonemic distribution in terms of percentages.

11 A Quantitative and Stylo-Statistical Approach to Fig. 5. Percentage of obstruents & sonorants in polysyllabic content words 4.7. Phonemic Patterns of Combination To highlight the prevalence of fricative and plosives in the text further, we examine 9 phonological features, all to do with plosives, fricatives and sonorants in polysyllabic content words (PCW). The results reveal 9 possible combinations of fricatives, plosives and sonorants in word-initial and word-final positions. The following table shows these phonemic combinations, the number of occurrences of each combination in the 9 polysyllabic words, their value in percentage and their rank. Number of Percentage Rank occurrences Fricative-plosive 6 66 % 1 Fricative-sonorant 0 0 % 4 Fricative-fricative 0 0 % 4 Plosive-fricative 1 11 % 3 Plosive-sonorant 0 0 % 4 Plosive-plosive 0 0 % 4 Sonorant-fricative 0 0 % 4 Sonorant-plosive 2 22 % 2 Sonorant-sonorant 0 0 % 4 Fig. 6. Ranking of phonemic combinations

12 30 In terms of ranking, we find that the highest rank is allotted to (1) the fricative-plosive combination, followed by (2) the sonorantplosive combination, (3) the plosive-fricative combination represented by the matrix verb aʿūdh u and finally (4) a group that includes 6 combinations that were never lexicalized in the text, i.e., zero percentage. Equally important is the fact that all the polysyllabic words make use of three voiced plosives, /q/, /b/ and /d/, which can all be grouped under the Arabic qalqala group, a category of five consonants that includes all the voiced plosives, namely, /q/, /ṭ/, /b/, /d/ and /j/. Of these, the phoneme most frequently used is /q/, which appears in all three word-positions: word-initial (qul), word-final (falaq-khalaq) and word-medial (waqab-ʿuqad). In terms of distribution, [q] alternates between 1 and 2 instances per line until its final reduction to vanishing point in line 5. In other words, /q/ is repeated in various phonological combinations in each of the rhyming words, except in the final word, where it is replaced by the fricative /s/ in ḥasad. It is also significant to note that /q/ is the strongest of the voiced plosives and has a phonological structure that consists of the following articulatory features: +voiced, +back and also +emphatic; it is the most powerful when placed in final position where it immediately acquires aspiration. As such, it functions as an important marker of dynamic action in the text and can be seen as a phonetic metaphor of the sheer energy involved in the process of splitting. Because the /q/ appears to migrate in a sequentially decreasing order from final to medial to zero position, we can safely assume that there is an attempt to decrease the acoustic intensity of the text. This phonological decrease is meant to accommodate a semantic shift to a new form of supplication in al-nās, where the subject matter is not the kinetic turbulence of splitting but the continuant and slithering movements of the waswās (the Hisser) as he hisses and incites people to sin Word-position Preference & Distinctiveness Ratio (DR) As researchers in stylometry explain, to study word-positionpreferences, counting the number of occurrences of each of the word-positions is a sufficient substitute for counting the consonants in the entire text (Kenny: 68). A good place to start is by calculating the texts preference between fricatives in word-initial position A

13 A Quantitative and Stylo-Statistical Approach to (henceforth WIP) and fricatives in word-final position B (henceforth WFP). We can give numerical expression to the text s preference between word-position A and B by calculating the proportion as follows: number of occurrences of A number of occurrences of A + number of occurrences of B Obviously, if the text prefers A to B, then this proportion will be greater than 0.5; otherwise it will be equal to or less than 0.5. If we make use of this simple statistic, we find the following proportions. 10 (fricatives in WIP) 10 (fricative in WIP) + 1 (fricatives in WFP) The proportion of fricatives in initial position is 0.9, an extremely high rate of occurrence that reveals the text s preference for fricatives in WIP in comparison to WFP. If we define the proportion of fricatives from the point of view of WFP, we find that it is 1 (1 + 10) = 0.09, an extraordinary low-frequency position that further emphasizes the fact that fricatives are more common in WIP, and that this preference is a distinctive characteristic of the text. If we wish to calculate the text s preference between plosives in WFP and plosives in WIP, the proportion is as follows: 9 (plosives in WFP) = (plosives in WIP) The proportion of plosives in WFP is 0.8, again an extremely high rate of occurrence that is almost identical to the high-frequency distribution of fricatives in WIP. By contrast, the proportion of plosives in WIP is 0.1, an extraordinary low number. We use proportions again to illustrate how essential these wordpositions are for fricatives and plosives. For instance, we can compare the proportion of fricatives in initial position with the proportion of plosives in initial position by applying a simple statistic called the distinctiveness ratio (DR). If we have the rates of occurrence for both fricatives and plosives in WIP position, we can calculate the ratio:

14 (rate of occurrence of fricatives in WIP) = (rate of occurrence of plosives in WIP) The distinctiveness ratio of the fricatives in WIP is 9, a high DR that is indicative of the text s preference. (The consonants that have a DR greater than 1.05 are plus-consonants; those that have a DR less than 0.67 are minus-consonants.) To calculate the DR of plosives in WIP, we simply reverse the rate of occurrences. The calculation for this word-position is 0.1, an extremely rare occurrence for plosives: 0.1 = In terms of WFP, the ratio is as follows: 0.8 (rate of occurrence of plosives in WFP) _ = (rate of occurrence of fricatives in WFP) The DR of plosives in WFP compared to fricatives is an extraordinarily high 8.8, indicating a preferred position for plosives in the final position. If we reverse the rate of occurrences, we will find that fricatives in WFP are minus-consonants with a low DR of = 0.1. There is a significant difference between plosives in WFP and fricatives in WFP, indicating that the preferred position for positives is the final position and the preferred position for fricatives is the initial position. Subsequently, these phonological patterning figures do not represent chance variability. 5. Stylo-statistical Analysis of al-nās (T2) 5.1. The Text 1. Say I seek refuge with the Lord of the people qul aʿūdh u bi-rabb al-nās 2. the King of the people malik al-nās 3. the God of the people ilāh al-nās 4. from the evil of the retreating Hisser min sharr al-waswās al-khannās

15 A Quantitative and Stylo-Statistical Approach to who hisses in the hearts of the people alladhī yuwaswis u fī ṣudūr al-nās 6. from among the Jinn and the people min al-jinnat i wa-l-nās 5.2. Content Words & Function Words in al-nās Content Words Function words qul say bi- with aʿūdh u i-seek-refuge al- the rabb lord al- the nās people al- the malik king min from nās people al- the ilāh god al- the nās people alladhī who sharr evil fī in waswās hisser al- the khannās retreating min from yuwaswis u hisses-3ms al- the ṣudūr hearts wa and nās people al- the jinnat i jinn nās people 5.3. The Dynamic Process of Hissing In terms of semantics, al-nās follows the same formulaic pattern used in al-falaq: Say: I seek refuge with the God of X from the evil of Y. However, the supplicant here is seeking the protection of God not in His capacity as lord of the splitting but rather in His capacity as lord of the people. In contrast to T1, the supplicant is not seeking the protection of God from a diversity of evils but rather from one major evil, namely the Hisser (al-waswās), a familiar and iconic epithet for Satan. This more or less subtle change in subject matter is accompanied by a dramatic change in the phonetic configuration of the text that mimics the basic activity referred to in the text, i.e., the incessant hissing of Satan in the hearts of man. Thus, the repetitive use in word-final position of the fricative /s/, a phoneme low in fre-

16 34 quency and high in susurration, comes close to being a direct transcription of the devil s repetitive and secret incitements to evil (for a comprehensive componential analysis of the verb waswās see Fatani, 2006: ; 2002a: 51-70). There thus appears to be an attempt to introduce a new controlling phoneme, namely the /s/ versus the /q/, which figured so predominately in T1. The /s/ grapheme in Arabic object) / is significantly an acrophonic sound (i.e., a picture of an س/ that originally referred to a pillar or column. It thus corresponds closely with the text s classification of God s authority and power in ascending order, from the Lord of the people, to the King of the people and finally to the God of the people. This classification also implies a hierarchical division of people into families, nations and races. This iconic stacking image contrasts with the circular image / ق/ predominant in T1 and curiously inherent in the Arabic /q/ or grapheme, an acrophonic sound originally referring to a knot or cavity Frequency Distribution of Obstruents & Sonorants As further evidence of the semantically motivated use of plosives and fricative in al-falaq, a quick phonological analysis of its companion-prayer, al-nās, is warranted in an attempt to confirm the pivotal roles that obstruents and sonorants play in the text and to validate the hypothesis that both texts must be viewed as complementary iconic diagrams. One must keep in mind that both prayers are companionpieces or adjacent texts and that, therefore, any striking statistical deviation from the fricative-plosive pattern outlined in al-falaq should be viewed as being motivated by constraints of subject matter. If we compare both texts in word-list style, three facts become immediately clear. First, both texts are composed of 30 words, but the distribution of content words and function words in al-nās is less symmetrical; there are now 16 content words and 14 function words. Second, al-nās features the use of an adjective, al-khannās (the retreating), a word-class completely excluded from al-falaq. Third, and most striking, the end-rhyme word al-nās repeats except in line 4, where the word is embedded within the highlighted adjective alkhannās. This apparent phonological difference between both texts is highly significant to our study. Of particular importance is the difference between the data of obstruents and sonorants, which reflect a

17 A Quantitative and Stylo-Statistical Approach to shift away from the fricative-plosive combination in al-falaq in favor of nasal-fricatives or glide-fricative combinations in al-nās. There are a total number of 43 consonants within the category of content words. These are distributed as follows: 18 sonorants, 17 fricatives and 8 plosives. (Note that the affricate /j/ in jinnat i is classified as both a plosive and a fricative because the production of this sound involves both consonant types.) Collectively, the sonorants and the fricatives now make up 81% of the content words rather than the fricatives and plosives. The sonorants, which represented only 26 percent of the content words in T1, are now the highest ranking phoneme in T2 representing 41%. (Note that the phoneme /j/ in yuwaswis u is also classified as a glide.) By contrast, the plosives made up 32% of the content words in T1 but now ranks lowest among the consonants with only 8 occurrences (18%) placed most often in word-initial rather than in the emphatic word-final position. The /q/ featured prominently in T1 but now occurs only once in the monosyllabic verb qul that is shared by both texts. We also find that within the sonorants, the nasals that had only been used once in T1 in the word naffāthāt now total 7 in number and are placed in word-initial position in all of the rhyme words represented by the repetitive al-nās. This number is actually much larger because many of the nasals are doubled or duplicated in pronunciation (e.g., khannās and al-nās). T2 also has a high percentage of fricatives in the final position, whereas the reverse is the case in T1 with fricatives never occurring in word-final position, except in the shared matrix verb aʿūdh u. The following graph makes clear this variation between T1 and T2 in term of distribution of obstruents and sonorants.

18 36 Fig. 7. Contrastive distribution of obstruents & sonorants in T1 & T2. The above chart shows that the fricatives and the liquids represent constant variables in both language types; their numbers are fixed at 17 and 3, respectively. The plosives, nasals and glides succeed in bringing about this phonological variation between both texts. The above chart reveals the sudden rise of nasals and glides in T2 accompanied by the marked decrease in plosives. To further highlight the constant and variable elements in values, here is a XY chart of all 6 phonemic variables in both texts: Fig. 8. XY chart of obstruents & sonorants in T1 & T2

19 A Quantitative and Stylo-Statistical Approach to As the preceding table reveals, the most striking variation occurs in the nasal category. Indeed, one of the defining criteria of T2 is its high frequency of nasals that occur 5 times more frequently than in T2. In turn, this nasal prominence succeeds in placing strong emphasis on the word naffāthāt in T1 and also sensitizes us to the crucial role that the /n/ phoneme plays in both texts. It is significant to note that nasals are acoustically the opposite of obstruents because they represent a radical rerouting of airflow from the oral cavity to the nasal cavity. The fact that this rerouting occurs as soon as the feminine principle is introduced could be seen as an iconic means of foregrounding a contrastive male/female principle operative in the text. Does the text use nasality as an index of femininity? An additional interesting observation is that, morphologically, naffāthāt represents a turning-point because it is the only word in T1 that is both structurally feminine (through the use of feminine plural morpheme āt) and semantically feminine in terms of gender (i.e., female-sorcerers). For one thing, the nasals have universally been acknowledged as representing the maternal principle lexicalized by the word mother as opposed to the plosives /p/ or /b/, which are commonly used to lexicalize the paternal element. The use of the nasal might thus metaphorically be seen as an acoustic representation of the female gender. As for the variation in plosive values in T1 and T2, the following chart of the number of plosives per verse-line better reveals this dramatic difference in frequency distribution:

20 38 Fig. 9. Frequency distribution of plosives per line in T1 & T2 Line 4 stands out as representing both a relatively high score or peak in terms of plosives in T1 and the lowest point in T2, which means that the plosives in T1 correlate negatively with plosives in T2. We can conclude, therefore, that the plosives are meant to be seen as diametrically opposed values in T1 and T Phonemic Combinations in al-nās A comparison of al-falaq and al-nās in terms of the distribution of fricatives, plosives and sonorants further augments the striking difference in the phonological configuration of both texts. There are a total of 37 consonants divided among 13 polysyllabic words in al-nās compared to the 28 consonants and 9 polysyllabic words in al-falaq. The following table provides a contrastive distribution of obstruents and sonorants in polysyllabic content words (PCW) in T1 and T2:

21 A Quantitative and Stylo-Statistical Approach to Fig. 10. Contrastive distribution of obstruents & sonorants in PCW in T1 & T2 As the preceding table reveals, both texts are bimodal because they are dominated by two high peaks represented by the fricatives and plosives in T1 and by fricatives and nasals in T2. In T2, the plosives now exhibit a low rate of occurrence, dropping from 10 occurrences to 5 occurrences, but there is a sudden rise in nasals from 1 single occurrence to 10 occurrences. The table below provides all possible consonantal combinations within the polysyllabic category in T2. The word nās is included within this category not only because of its semantic significance as a rhyming word but also due to its use of the long vowel /aa/, which compensates in length for any lack in consonants.

22 40 Number of occurrences Percentage Rank Fricative-plosive 0 0 % 4 Fricative-sonorant 1 7 % 3 Fricative-fricative 1 7 % 3 Plosive-fricative 2 15 % 2 Plosive-sonorant 0 0 % 4 Plosive-plosive 1 7 % 3 Sonorant-fricative 7 53 % 1 Sonorant-plosive 1 7 % 3 Sonorant-sonorant 0 0 % 4 Fig. 11. Consonantal combinations within polysyllabic category in T2 If we compare the above table with the corresponding one of al- Falaq, we can calculate how far the rank order of scores for each variable is similar to the rank order of scores in T2. The aim here is to measure how closely the rank order of these phonemic combinations match. A quick analysis reveals that the ranking is now completely reversed, with the highest ranking now allocated to the sonorantfricatives that ranked last in al-falaq. In addition, the plosive-fricative combination, which was number 1 in al-falaq, is now relegated to final rank with zero percentage. Pattern Ranking T1 Ranking T2 Fricative-plosive 1-66% 4-0% Fricative-sonorant 4-0% 3-7% Fricative-fricative 4-0% 3-7% Plosive-fricative 3-11% 2-15% Plosive-sonorant 4-0% 4-0% Plosive-plosive 4-0% 3-7% Sonorant-fricative 4-0% 1-53% Sonorant-plosive 2-22% 3-7% Sonorant-sonorant 4-0% 4-0% Fig. 12. Comparison of ranking order of phonemic combinations in T1 & T2

23 A Quantitative and Stylo-Statistical Approach to A contrastive analysis of T1 and T2 in terms of consonantal wordpositions (WIP and WFP) in polysyllabic content words provides us with the following results for the 4 major consonantal classes, fricatives (WIP and WFP), plosives (WIP and WFP), nasals (WIP) and glides (WIP): Fig. 13. Contrastive analysis of T1 and T2 in terms of consonantal WIP & WFP in PCW The preceding phonemic combinations and word-position chart points to a significant degree of phonological modification taking place in T2. Foremost is the shift away from the plosive-fricative pattern that is prevalent in T1 to a nasal-fricative pattern. Additionally, we notice the following phonological restructuring: A decrease in plosives in word-final position (WFP), The devoicing of plosives in WFP (malik - jinnat i ), The introduction of /j/ as a phonological unit representing the merging of fricatives and plosives, The absence of /q/, A rise in the number of occurrences of /s/ in all positions, especially WFP, A shift in fricatives to word-final position, The prevalence of nasals and sonorants as opposed to obstruents, and

24 42 The mobilization of the words /waqab/ and /naffāthāt/, which represent the only occurrences of sonorants in word-initial position in T1, to foreshadow the strikingly new phonological configuration in T Word-position Preference & Distinctiveness Ratio (DR) in al-nās If we look at the word-position preferences of content words in T2, we find that the fricatives have a low proportion in WIP (0.2) and a high proportion in WFP (0.7), suggesting that T2 prefers fricatives in word-final position. The plosives in WIP and WFP have nearly equal proportions because there are 4 instances in WIP (0.57) and 3 instances in WFP (0.42). The nasals only occur in word-initial position (6 occurrences) with zero instances in word-final position. They thus have a high 1.0 proportion in WIP and a zero proportion in WFP. Using the above proportions, we can calculate the DR of fricatives, plosives and nasals, keeping in mind that a plus ratio is greater than DR 1.05 and a minus ratio is less than DR For fricatives, the DR in WIP is a minus 0.4 : 0.2 rate of occurrence of fricatives in WIP 0.5 rate of occurrence of plosives in WIP In WFP, the DR of fricatives is a plus 1.4 ( = 1.4). For plosives, because of the low number of occurrences of fricatives in the initial position, we need to calculate using the ratio of both fricatives and nasals. Thus, the DR of plosives in WIP is 0.6 ( (rate of fricatives) (rate of nasals) = 0.6). In WFP, the DR of plosives is a low 0.6 ( (rate of fricatives) = 0.6). For nasals, the DR in WIP is a plus rate of DR 2.0 ( (rate of fricatives) = 3.0. The following table provides a comparison of the DR for T1 and T2: DR of T1 DR of T2 Fricatives in WIP Fricatives in WFP Plosives in WIP Plosives in WFP Nasals in WIP Fig. 14. Comparison of DR for T1 and T2:

25 A Quantitative and Stylo-Statistical Approach to Conclusion By making use of statistics, I have attempted to substantiate the initial impression of word-position choices and to provide quantitative evidence to prove the presence of a pervasive fricative-plosive patterning in T1 and a nasal-fricative patterning in T2. The fact remains significant that, of the 9 polysyllabic words in al-falaq that constitute the so-called materials of the argument, 6 follow the fricative-plosive pattern (a high proportion rate of 6 8 = 0.75). The two occasions of unfulfilled patterning, waqab and naffāthāt (both of which retain final plosives but substitute sonorants for fricatives), are highly significant variations because they are statistically mobilized to foreshadow the strikingly new phonological configuration in al-nās, where there is a marked increase in nasals and glides. The comparison of scores would suggest that there is a motivated attempt to alter the phonological configuration of al-nās so as to correspond to the new subject matter. To summarize, when we compare the frequency of obstruents and sonorants in T1 with that of T2, four facts become immediately clear. First, the variables show a shift away from plosives in end-position, a patterning that appears to be obligated by the subject matter of T2. Second, this shift also includes an absence of the /q/ phoneme that was an integral part of the phonetic configuration of T1. Third, the fricative scores show a striking uniformity in terms of frequency of occurrence; they seem equally at home in T1 and T2. This result suggests that it is the plosives and sonorant classes that feature instances of variability. However, these fricative scores also point to a shift to word-initial position combined with a prevalence of the /s/ phoneme. Fourth, the distribution of nasals rises significantly in T2, again reflecting the shift away from the fricative-plosive pattern in favor of a nasal-fricative pattern. The differences in the scores of T1 and T2 for the plosives and the nasals lead us to conclude that plosives function as important markers of the splitting process in T1, just as nasals and the strident /s/ function as markers of the movements of the Hisser. The application of the preference and DR tests to both texts indicates the extremely low probability of the fricative-plosive pattern and the nasal-strident pattern occurring as a result of chance. There seems therefore to be a real and reliable relationship between these phonological scores and the subject matter of each text. In addition, be-

26 44 cause the phoneme /q/ is used at higher levels than other plosives, it can be seen as the most dominant phoneme in T1. The clear tendency for the level of /s/ in T2 to increase with the introduction of waswās, the Hisser, is equally important. The variable /s/, occurring 11 times in all, emerges as the dominate phoneme of T2 correlating with the hissing of Satan. In the final analysis, it is important to note that these statistical tests confirm the validity of our impressions; they do not tell us why the pattern should occur. To find motivation or reasons behind these contrasting phonetic prominences, we need to link these results to the controlling image of both texts. As linguists have indicated, sound prominence alone is not sufficient, which poses the interesting question if T1 would be just as iconic and powerful if the fricative-plosive sequence appeared in the reverse order or if it would lose much of its force and mimetic quality. I think we may have concluded that splitting was a less turbulent and violent activity if it had started with a sudden eruption of plosives that dwindled into friction or a smoother continuant flow. Similarly, T2 would have lost much of its iconic dynamism had it not been saturated with the hissing phoneme in wordfinal position and the smooth gliding movements of the Hisser represented by the sonorants /n/ and /w/ strategically placed in wordinitial position. REFERENCES al-ʿabd, Muḥammad (1984), al-mufāraqat al-qurʾāniyya (Cairo: Dār al-fikr al-ʿarabī). Abū Mūsā, Muḥammad (1987), Dalālāt al-tarākīb (Cairo: Maktabat Wahba). al-andalusī, Abū Ḥayyān Muḥammad ibn Yūsuf al-gharnāṭī (1983), Tafsīr albaḥr al-muḥīṭ, 8 vols., (Beirut: Dār al-fikr). Bū ʿUmāma, Muḥammad (2002), al-ṣawt wa-l-dalāla: Dirāsa fī dawʾ alturāth wa-ʿilm al-lugha al-ḥadīth, Majallat al-turāth al-ʿarabī 85, Crystal, David (1997), The Cambridge Encyclopedia of Language (2 nd ed., Cambridge: Cambridge University Press). Fatani, Afnan H. (2006), Translation and the Qurʾan, in Oliver Leaman (ed.), The Qurʾan: An Encyclopedia (London & New York: Routledge),

27 A Quantitative and Stylo-Statistical Approach to (2005), The iconic-cognitive role of fricatives and plosives: A phono-semantic analysis of a classical Arabic prayer Al-falaq, in Constantino Maeder, Olga Fischer and William J. Herlofsky (eds.), Outside-In Inside-Out [Iconicity in Language and Literature 4] (Amsterdam: John Benjamins Publishing Company), (2002a), A Diachronic Semantic Error Analysis (DSEA) of Arabic Non-core Lexicon: Chapter 114 of the Quran/An-naas ( The People ), Language Forum 28/1, (2002b), The Lexical Transfer of Arabic Non-core Lexicon: Sura 113 of the Qurʾan al-falaq (The Splitting), Journal of Qurʾanic Studies 4/2, Fischer, Olga and Nänny, Max (2001), Introduction: Iconicity and Nature, European Journal of English Studies 5/1, Ḥassān, Tammām (1993), Al-bayān fī rawāʾiʿ al-qurʾān (Cairo: ʿĀlam al- Kutub). Ibn Jinnī, Abū l-fatḥ ʿUthmān al-mawṣilī (1974), al-khaṣāʾiṣ, 3 vols., (ed. Muḥammad ʿAlī al-najjār; Beirut: Dār al-hudā). Kenny, Anthony (1982), The Computation of Style: An Introduction to Statistics for Students of Literature and Humanities (Oxford: Pergamon Press). Lyons, John (1977), Semantics, 2 vols., (Cambridge: Cambridge University Press). al-rāfiʿī, Muṣṭafā Ṣādiq (1990), Iʿjāz al-qurʾān wa-l-balāgha al-nabawiyya (Beirut: Dār al-kitāb al-ʿarabī). al-ṣaghīr, Muḥammad Ḥasan ʿAlī (2002), al-ṣawt al-lughawī fī l-qurʾān (Beirut: Dār al-muʾarrikh al-ʿarabī). al-zamakhsharī, Abū l-qāsim Jār Allāh Maḥmūd ibn ʿUmar (1947), al- Kashshāf ʿan ḥaqāʿiq ghawāmiḍ al-tanzīl wa-ʿuyūn al-aqāwīl fī wujūh al-taʾwīl, 4 vols., (ed. ʿAbd al-razzāq al-mahdī; Beirut: Dār Iḥyāʾ al-turāth al-ʿarabī).