Transliteration refers to the method of converting the script (alphabet, syllable, characters etc.) of one writing system to another, such as writing words of a language based on an alphabetic writing system into a language based on a syllabic writing system (Haddouti &Cluzel, 1999 cited in Ayob, 2006). Simply, we can say that transliteration is the practice of changing a writing system (script) of a language into a comparable (though not the same) substitute writing system (script) of another language.

Many of the approaches and methods employed for this purpose use statistical tools and quantitative data. However, transliteration techniques, according to Sethuramalingam, are roughly divided into two major approaches, i.e. graphene-based and phoneme-based (2009). About the purpose of transliteration, Beasley (1998, cited in Humanyoun, 2006) states, “it is to write a language in its customary orthography, using the same exact orthographical conventions but using carefully substituted orthographical symbols”.

Urdu and English, being simultaneously used in Pakistan, frequently come in contact with each other in day-to-day use. According to Baumgardner (1992), lexical borrowing is an inevitable corollary of all language contact situations. However, in a contact situation between Urdu and English, it is not only lexical borrowing but also code-mixing and all sorts of mutual influences that can be witnessed simultaneously. The languages concerned in a language contact situation do not have mutual effect only at the lexical level; rather, it is a holistic phenomenon because all other linguistic systems such as graphological, morphological, phonological, grammatical, semantic etc. are equally influenced by the respective systems of either language in contact with each other. This phenomenon of borrowing/influencing at different levels by Urdu and English language has been thoroughly researched by Baumgardner in his influential work “The Urduization of English in Pakistan" (1989). This is really an exhaustive work encompassing almost all levels of Urdu and English languages, but the writing system is not covered in this work. Like all other systems of two languages in contact, the writing system of either language has significant repercussions for both languages. Baumgardner has clearly mentioned in his book “The English Language in Pakistan" (1992) that "lexical transfer is frequently accompanied by structural changes in which Urdu borrowings undergo a morphological restructuring according to the grammatical rules of the recipient language" (p. 129). He further remarks that Urdu singular nouns used in English take English inflectional morphemes for plural rather than the plural morpheme used in the source language. A very similar phenomenon can be observed in the case of English lexis transferred/borrowed into the Urdu language because the borrowed English lexis also undergoes morphological restructuring according to the rules of the target language (Urdu) rather than retaining the features of the source language (English). It is worth noting that this restructuring of Lexemes transferred or borrowed from one language to another is not limited only to the morphological level; instead, it takes place and is a necessary feature which can be evidently observed at all linguistic levels.

In this context, the present study focuses on pinpointing and ascertaining the category of words transliterated from English to Urdu, which varies in pronunciation considerably from the Standard British English pronunciation. The pronunciation of these words is obviously influenced by the orthographic conventions of the Urdu script. It is generally observed by all educated people having knowledge of English reading skills that English spelling and particularly its orthographic conventions pose serious problems and difficulties for both EFL and ESL learners. The present study is an attempt to investigate the internalisation of the transliteration and the orthographic structure of English words in Urdu by Urdu speakers (including both speakers having Urdu as L1 or L2). It also explores the ways in which this transliteration of English to Urdu influences their understanding of pronouncing words in English. In Urdu, being a phonetic language, words are spelt the same as they are pronounced. Saleem (2007) is right in his perception that everyday Urdu learners of English take it for granted. The same is also true in the case of the English language and its orthography. However, the problem is not limited only to ordinary and everyday Urdu learners of English but can be equally found and observed in educated Urdu speakers of English and most particularly in fresh undergraduate students of English Departments at various universities and degree awarding institutes. All kinds of Urdu speakers of English, i.e. ordinary, educated, and fresh students of English Departments hailing from various universities and degree awarding institutes, can be observed as lacking knowledge of basic facts about the internal linguistic systems of English and Urdu languages. For example, they have little awareness of how the two languages differ as far as their orthographic, phonological, and morphological systems are concerned. Also, they lack deep and basic know-how about the different social, political, geographic and linguistic lineage and origins of the two languages. For instance, knowledge such as that Urdu belongs to the Indo-Aryan family while English comes from the Indo-European family; that Urdu has a right-to-left writing system whereas English has left-to-right; that the rhythm of Urdu is syllable-timed whereas that of English is stressed-timed; and that Urdu uses predominantly Nastaleeq script which is described by Humanyoun (2006), as cursive, ‘context sensitive’ and complex writing system, whereas, English writing system is based on Roman script etc. as well as many other significant facts are not known to many. And writing system is no exception to the long series of differences prevalent in these two languages. That's why one can frequently observe the phenomenon of transliteration.

 

According to Sapir (1921), the contact of two or more different languages has one inevitable result; the languages in contact will necessarily have a mutual effect on each other in a variety of ways, the most simple of which is the phenomenon of "borrowing" of words. Borrowed expressions happen to become lexicalised both phonetically and graphologically. Some linguists simply delimit loanwords to transliterations and graphic loans. But the major difference between transliterations and "borrowed/loan words" lies in the fact that loan words cannot go back to their original form in the source language from which they have been borrowed, whereas Transliterated words can be. This is because a borrowed word undergoes grammatical and phonemic changes so that it can be used at par in the target language a great deal like its very original words. In case the borrowed item has got the same use, same appearance, and same meanings in the target language, it is known as a cognate. However, if it has a different function, appearance and meaning, it is known as a false friend.

Ahmad (2009) argues that Urdu as a language has borrowed quite extensive vocabulary from the glossary of Arabic and Persian languages. Similarly, the Urdu language has taken a considerable number of lexical items from the English language. Importantly, these words have now become part of the Urdu language and are considered the rightful possession of the Urdu language in all its aspects. Furthermore, Ahmad (2009) believes that the borrowed words are faithfully reproduced (written) in the Urdu language as those written in the original target language. It is worth noting that in Urdu, the status of borrowed words from English is both “Loan words” as well as “transliterated words” because the words have got features of both simultaneously. 

 

A series of scholars cite Taylor’s (1883) classifications as one of the earliest attempts in this regard. DeFrancis (1989: 59), for instance, defines the classification proposed by Taylor as “an evolutionary model consisting of a succession of five stages, i.e. pictures, pictorial symbols, verbal signs, syllabic signs, and alphabetic signs of which the first three are known as ideograms and the remaining two as phonograms”. This initial categorisation by Taylor (ibid) has influenced various subsequent writing typologies that try to capture the core difference between ‘phonographic’ and ‘non-phonographic’ systems of writing. However, the large number of already existing proposed alternative terms is indicative of the elusive and difficult nature of this distinction.

Undeniably, Gelb’s (1952) seminal work on writing systems paved the way for establishing the scientific study of writing. Though, his categorisation excludes pictures as one of the proper writing forms and declares both “pictorial representation” as well as “mnemonic devices” merely as harbingers of writing. The categories of “pictorial representation” and “mnemonic devices” are not part of the category known as “full writing”, which comprises sub-categories of “word-syllabic”, “syllabic”, and “alphabetic”. Notwithstanding this significant difference, many researchers and scholars consider Gelb’s classification indisputably defective due to its undue emphasis on presenting a teleological account of writing in which he declares the alphabet as the ultimate stage in the evolution of writing systems through logography and syllabary.

One more significant typology is by Diringer (1962), which, according to Hill’s (1967) observation, seems quite alike to that of Taylor’s (1883). Like Gelb (1952), Diringer too is seriously influenced in his classification by his understanding that a system of alphabet epitomises the "most flexible and useful method of writing ever invented” (Diringer 1962: 24). Furthermore, very much like Gelb, he, distinguishes between “embryo-writing and “full writing” but unlike Gelb,  Diringer’s conception of “full writing” seems more comprehensive compared to that of Gelb’s as it includes “pictography”, “ideography”, “transitional analytic scripts”, “phonetic scripts”, and “alphabetic writing”. In the outline of his own classification, Hill (1967) criticised Diringer’s classification on the three grounds and made a tripartite classification involving the “discourse systems”, “morphemic systems”, and “phonetic systems”.

De Francis’ (1989) widely-cited book is the next key work in the 1980s on writing systems with a distinctive classification scheme of the writing of his own. DeFrancis’ classification contrasts between “partial” and “full” writing systems, but a very significant contribution lies in his faith in the phonetic (phonic) basis of all “full” writing systems. In line with his firm belief that writing is nothing more than a simple 'visual representation’ of speech, he proposed “pure syllabic systems” such as “Linear B, kana, and Cherokee”; “morpho-syllabic systems” such as “Sumerian, Chinese, and Mayan”; “morpho-consonantal systems” such as “Egyptian”; “pure consonantal systems” such as “Phoenician, Hebrew, and Arabic”; “pure phonemic systems” such as “Greek, Latin, and Finnish”; and “morpho-phonemic systems” such as“English, French, and Korean”.

Finally, Coulmas’ classification (1992), primarily drawing on Haas’ (1976) “pleremic”/“cenemic” dichotomy of writing systems, finally comes up with an arrangement consisting of seven general categories with the “pleremic” system, further divided by Coulmas in “logograms” + “phonograms" such as “Hittite hieroglyphs”; “logograms” + “phonograms” + “determinatives” such as “cuneiform”; and “morphosyllabic signs” such as “Chinese” and the “cenemic” systems classified into “syllabary” such as “Japanese kana”; “consonantal alphabet” such as “Phoenician”; “alphabet” such as “Roman”; and “the alphabet with independent vowel letters and integrated consonant-vowel letters” such as “Ethiopic”.

Despite the fact that it is not possible to encompass all the possible viewpoints and arguments, it certainly helps in discerning a few significant and recurring issues, such as the continuing challenge and concern with meaningful characterisation of the difference between “phonographic”, or “cenemic”, from “non-phonographic”, or “pleremic”, writing systems. Another issue that has inspired some of the typologies/classifications is to do with the challenge of identifying and characterising meaningful categories amongst the different graphemic conventions witnessed and employed across all writing systems of the world languages. The recognition and acknowledgement conferred upon “abjads” and “abugidas”, though comparatively recently, can be viewed as an advance on earlier typologies comprising fewer categories.

 

The study has significance in the sense that it explores the ways and processes indicating how the pronunciation of undergraduate L1 & L2 Urdu speakers at university is deviant or, in some cases, completely wrong when compared to the Standard British English pronunciation. It also investigates the negative role of English-to-Urdu Transliteration in the wrong pronunciation of the English language by L1 & L2 Urdu users. The study helps pinpoint that the deviant/wrong pronunciation of English words/language by L1 & L2 Urdu users is not always the result merely of the mutual negative influence of phonological features; rather, it is also triggered by the process of English to Urdu Transliteration playing an alarming negative role in along with mutual negative effects induced by language contact situation. The words listed in this study help understand the patterns of deviant/wrong pronunciation by L1 & L2 Urdu speakers and may be helpful from a pedagogical perspective while teaching English to such learners.

 

The study is based on an exploratory-qualitative approach in which transliterated words having wrong or deviant pronunciation from the standard British English necessarily (but not necessarily from Standard American English) have been carefully gleaned by means of making a long list which is based on long-term observation of the fresh undergraduate students of English department having Urdu as L1 or L2. The study is essentially exploratory in its nature and approach because any significant study on the same topic was not found. Furthermore, it endeavours to explore a list of words pronounced in a different way from Standard British English. The findings and conclusion of the study have been analysed, tabulated and elaborated on in detail. It is worth noting that not all words with a different or deviant pronunciation could be listed; however, significant attempts have been made to register the most frequently occurring words.

The population generally comprises the fresh undergraduate students of English departments having Urdu as L1 or L2. A comprehensive set comprising 197 frequently occurring words is listed and analysed in detail qualitatively. Oxford Advanced Learners Dictionary and phonetic symbols of IPA chart are used for the purpose of transcription of the selected words for both the standard British English and the deviant transcriptions.

The present study aims to investigate the effects and influences of English to Urdu transliteration in the wrong or, in some cases, a sort of odd pronunciation of certain well-known words. The word list comprises miscellaneous and most frequently occurring lexical items that are generally transliterated from English to Urdu.

 

The set of items in the list was analysed in accordance with the rules of English and Urdu phonological, morphological, and orthographic systems in mind employing the concept of deep versus shallow orthography. It was found that most of the errors induced by transliteration can be safely accorded to the difference in various levels, such as segmental, suprasegmental, orthographic etc., of English and Urdu languages. The errors resulting from transliteration are due mainly to differences in phonetic features, phonemic inventories, and phonological processes in sounds of English and Urdu languages.

 

The phonological processes of English and Urdu are systematically different in detail which is a major source of pronunciation errors induced by the transliterated English words into the Urdu language. It naturally results in a distorted shape of SL items as well as their pronunciation in the TL. For example, both the voiced and voiceless dental or (interdental consonants as Yule 1996 call it) {/θ/, /ð/} and the voiced and voiceless alveolar phonemes {/t/, /d/} of the English phonological system are always replaced with their alveolar counterparts (nearby sounds) of Urdu phonological system. It is worth noting that the phonemes with which Urdu replaces these English phonemes are not included in the English phonemic set/inventory. In this way, all words in English with any of the above-mentioned four consonants that occur initially, medially or finally are transliterated with a different pair of consonant sounds; hence different pronunciation in the case of such words is inevitable.

The second major difference between English and Urdu is that no one-to-one mapping between scripts of the two languages is possible, as found by Ahmad (2009) in his work “Urdu- Roman transliteration”. There is no possibility of implementing English-to-Urdu transliteration by sheer one-to-one replacement of characters or letters of the alphabet in both languages. Unlike Urdu, which is a phonetic language and has no problems with sound and letter correspondence, English has got a variety of issues in the representation of sounds by letters of the alphabet. For example, in English, one sound (phoneme) can be represented by as many as five-to-eleven different combinations of letters (graphemes) such as /tʃ/ phoneme can be represented by five different combinations of letters as in words like “cheap”, “catch”, “picture”, “question”, “righteous”  (the bold and underlined combination of letters in these words give /tʃ/  sound). Similarly, the phoneme /dʒ/ is represented by eight different combinations of letters in different words as in words like “July”, “gem”, “pigeon”, “soldier”, “grandeur”, “bridge”, “sandwich”, “adjective” (the bold and underlined combination of letters in these words give /dʒ/ sound).

The more complex and more surprising result can be found in the case of the representation of vowel phonemes by such variant and unusual combinations of letters (graphemes) as can hardly be realised by Urdu script. For example, the sound /iː/ is represented in English words by as many as eleven different combination of letters as in words like “be”, “seed”, “leaf”, “field”, “receive”, “police”, “Caesar”, “Phoenix”, “people”, “key”, “quay””(the bold and underlined combination of letters in these words give /iː/sound).

In English, different combinations of letters (grapheme) in different words result in the same sound (phoneme), but it is not possible for the Urdu language to realise this in its script because, in Urdu phonology, each letter of the alphabet stands for one sounds/phoneme rather than many different combinations of letters in different words giving the same sound. This shows that, unlike the English language, there is a one-to-one correspondence between Urdu phonemes and graphemes. The problem is that a particular combination of letters in English that represents a particular sound in a word when transliterated to the Urdu language with that particular sound, Urdu readers/speakers generalise and take it for granted that the same combination will give the same sound in other words as well which is not the case in English.

In addition to problems with consonants, the vowel segments were found to have more serious issues in the process of transliteration. For example, many of the words, when transliterated from English to Urdu, have a changed vowel from the original English words and their standard British English pronunciation. Surprisingly, the vowel sounds in some words do partially match with Standard American English, such as in words like 'student' /ˈstjuː.d ə nt/  (standard British English) and /ˈstuː.d ə nt/  (standard American English), the pronunciation of this word in Standard American English is same to Urdu transliterated version in terms of consonant segment whereas it differs from Standard British English in the sense that the former has two consonants whereas the latter has three consonants at onset position in the same word. The same is also true of many other words also.

Another major issue in Urdu transliteration is the substitution of English diphthong or triphthong with monophthongs or single vowels. Many of the English words have a glide from one vowel to another and, in some cases to a third one also, are transliterated and pronounced with a single vowel sound in Urdu, which leads to wrong or at least deviant pronunciation of English words on the part of learners of English having Urdu as L1 or L2. For example, words like /meɪ/, /steɪ/, /teɪk/, /həʊm/ etc. are all transliterated and pronounced with single vowel (monophthong) rather than a diphthong.

Besides these segmental issues, more serious and drastic are errors and problems of suprasegmental features such as stress, pitch, intonation, and problems related to aspects of casual, rapid and connected speech such as assimilation, elision, gemination, and juncture. Since rhythm of both the languages are different from each other in the sense that English is a stress-timed language in which stress syllables tend to recur at equal intervals of time (Roach, 1983), whereas Urdu is a syllable-timed language in which all syllables tend to recur at equal interval of time. All English words of more than one syllable, i.e. disyllabic or multiple syllable words, have a distinctive stress pattern. That is, the stress will either be on the first, second, third, penultimate or ultimate syllable of multisyllabic words. As for as Urdu transliteration is concerned, it neither has any stress pattern nor has any provision for stress marking, i.e. no diacritics or any other symbols or signification is there in Urdu script to indicate the stress pattern of a word. So all multisyllabic words of English are transliterated without any stress marking, and as a result, a faulty pronunciation is bound to take place.

Finally, English is a non-phonetic language (shallow orthography), whereas Urdu predominantly tends to be a phonetic language (deep orthography). In a phonetic language, there is a strong correspondence between phonemes and letters of the alphabet and all phonemes are realised in its script, whereas in a non-phonetic language, it is not necessary that a word shall necessarily be spelt the same as said. There is no one-to-one correspondence between spelling and pronunciation of words. In English, the processes of assimilation, dissimilation, insertion, intrusion and elision etc., are always at work both in isolated words and in connected speech, but in Urdu, it is not that frequent. In Urdu transliterated version, English words are treated as phonetic i.e. they are spelled/written the same way as they exist in their original English script. There is a huge difference between English spellings and their pronunciation, but transliterated lexical items in Urdu treat it as if there were one-to-one correspondence between the written and the corresponding spoken forms of English words. For example, the word "talked”/tɔːkt/ has a /d/ sound in its spelling but a /t/ sound in its pronunciation due to the process of voicing assimilation but in Urdu transliteration, the word would appear with a /d/ sound instead of /t/ sound because Urdu transliteration does not account for the processes or aspects of connected or casual and informal speech in English. Similarly, pronunciation of words such as “pizza” /ˈpiːt.sə/ seems quite contrary to its spelling. The written form of this word is not a guide to its pronunciation at all. It appears from the spelling as if the word would be pronounced as /ˈpiːzə/, but it is very different from its written version in that there is an insertion of the/t/ phoneme, which is nowhere visible in the written form, and there is a change of /z/ to /s/. Urdu transliteration would treat the word as phonetic and will transliterate it in a straightforward manner as /ˈpiːzə/ (considering the spelling/orthographic form as a good guide) instead of /ˈpiːt.sə/.

 

To make the matter worse, Urdu transliteration has no idea of how to treat and pronounce foreign or non-native words in the English language (that are imported from other foreign languages such as French, Spanish, Italian, Dutch etc.) that retain the pronunciation of their respective native language. Again, Urdu transliteration would treat these words as phonetic in nature. i.e. it assumes that there will be strong affinities between written and spoken forms of such words but the reality in such cases is very different from the assumption. For example, there seems to be no affinity at all between the written and spoken forms of the word “rendezvous”/ˈrɒn.deɪ.vuː/ taken from French and Urdu transliteration would produce the word as/ˈren.de.vʌs/, which is not only wrong pronunciation but quite ridiculous as well when compared to the original pronunciation.

 

Another noteworthy yet not accounted-for fact is that the pronunciation of English words does change with the change in its grammatical category, but Urdu transliteration is found quite negligent in this aspect in the sense that it treats English words in all their different grammatical roles/categories remaining same whereas English words are often found with some sort of change in some aspect of its pronunciation either at segmental or supra segmental level. For example, the word "close” as a verb has the pronunciation /kləʊz/ but the same word used either as an adjective or noun has a changed pronunciation as /kləʊs/ with a change in a final segment from /s/ to /z/. A similar process at work can be witnessed in the case of the word "use" and many other lexical items where the orthographic form remains the same, whereas its phonological form is changed.

 

The change in the pronunciation of functional words (grammatical words) when used in their weak forms is another issue that Urdu transliteration is unable to tackle or account for. The closed class of words such as article, pronouns, conjunctions, prepositions, and determiners have two different forms in the spoken version, known as 'strong' and 'weak' forms, which is dependent upon the fact of how and which way they are used in a sentence. For example, a generally weak form of prepositions is used if they are followed by a complement (noun or pronoun), e.g. the phrase "for him” would be transcribed as /fə. ɪm/, weak forms of both the preposition and the complement are used. On the other hand, if the sentence begins with a pronoun, e.g. “he," or ends with a preposition, e.g. “for",; in either situation, the strong form will be used, such as /fɔː/ and /hiː/. Urdu transliteration neither accounts for nor can afford to take into account such niceties of English pronunciation which can be viewed as providing a wrong guide for English pronunciation.

 

It can be concluded that the pronunciation of both educated and uneducated L1 & L2 Urdu speakers varies considerably from the standard British due to the fact that English words are written in different orthographic conventions compared to the Urdu language. For example, in certain cases, additional sounds, especially schwa /ə/, are inserted in different places, such as in the beginning, medially or finally. Another noteworthy difference that can be evidently noticed is that both consonant and vowel segments (sounds) but vowel sounds, in particular, are generally supplanted with other similar homorganic vowel sounds, which gives a different or deviant and, in some cases, a wrong effect or impression of the original transliterated English word. The study lays emphasis on the fact that this different transcription and hence pronunciation of transliterated words from English to Urdu has an enormous influence on the wrong pronunciation of English words, and their correction will definitely improve the pronunciation of Urdu learners.

The study has attempted to explore the problems and issues related to and caused by Urdu transliteration of English words in the pronunciation of the English language, especially for educated Urdu L1 or L2 users. These issues need to be resolved and necessary changes be made which can significantly improve English pronunciation of educated L1 or L2 Urdu users well before they study rules of English phonology and have a good working knowledge of English orthographic conventions as well as knowledge of the fact that there is no one-to-one mapping exist between English phonological and graphological levels/systems. It is worth noting that since the orthographic systems of English and Urdu largely differ from each other, the use of Urdu orthography is not appropriate without careful and customised use of diacritical marks.

 

The following recommendations are hereby suggested.

1.         The pronunciation and representation of transliterated items from English to Urdu need to be guided by the conventions and rules of the source language: its phonology, orthographic conventions, and prosodic features rather than keeping in view the target language system of orthography and phonology.

2.        Appropriate strategies or rules should be devised to mark prosodic features such as stress and intonation patterns of the source language in transliterated versions of the target language words.

3.        There is a strong need for necessary readjustments in the standardised orthography of the target language in order to represent the source language words in their real spirit and shape rather than reproduced in a distorted form and with deforming tendencies.

4.       An analysis of properties mutually shared by bilinguals (English-Urdu bilinguals or bilinguals from any two languages) corpora that causes problems in the process of accurate transliteration is also recommended.