Using Expansion Strategies in Making Untranslatable Areas of Poetry Translatable : Sa ’ di ’ s Bustan as a Case in Point

The purpose of this study was to explore poetry translatability and seek to see what the translators do to compensate those untranslatable areas of poetry. In doing so, the researchers chose a literary work, i.e., Bustan, by one of the well-known Iranian poets, that is, Sa’di (Wickens, 1990) and one of its translations, “The Bustan by Shaikh Muslihu-D-Din Sa’di Shirazi,” by Clarke (1985). They analyzed one hundred verses of his poetry, which were chosen randomly, to see what the translator did to overcome the untranslatable parts of that poetry, i.e., Sa’di’s Bustan. The study showed that the translator used expansion wherein he made some semantic and structural adjustments in his translation so as to make those untranslatable parts translatable and make the translation natural, intelligible and understandable in the receptor language. As a whole, in the selected corpus of the study, ninety two cases of expansion were identified.


Introduction
The poetry translation is commonly believed to be the most complex and perhaps satisfying translation form (Connolly, 1998(Connolly, /2001)).It has been a controversial issue, specifically in literary translation domain.The probability of translation of poetry is under discussion; however, it is a wide-spread act from at least 2000 years ago.Translated poetry has been done since then pervasively (Connolly).As an instance, one can refer to Fitzgerald's Rubaiyat of Omar Khayyam (1859) or Eastwick's Rose Garden of Sa'di (1979).There are many different views regarding the subject; and some of them may be subjective.Usually for emphasizing this task complexity, the poetry definition proposed by Robert Frost is given: poetry is "that which is lost in translation" (Connolly, 1998(Connolly, /2001, p. 170), p. 170).However, Connolly believes that studies over the real translation process and identifying the related problems existed and the solutions for overcoming them are fairly few.Accordingly, Jones (2011b) states that the "knowledge about poetry translating is, perhaps surprisingly, still fragmentary" (p.182).He further mentions that "no book-length surveys of poetry translation as a whole at least in English" cannot be found "since 1970s" (p.182).
The crucial matter that the present research has explored is that when the translators start to render some poetry verses from the source language, they might come across some words that are untranslatable thanks to the present dissimilarities among languages, chiefly from cultural point of view.The source texts include some traits which a translator may not be able to bring any equivalents for them.Accordingly, once the translator is translating, he should remind that he is rendering for the readers whose historical and cultural backgrounds are different from those of the source readers (Klaudy, 2009).To solve the aforesaid problem, some translators apply a number of semantic and structural adjustments in their translations to have intelligible, natural and comprehensible translation in the target language (Nida & Taber, 1982).

Kinds of Poetry Translation
Jones (2014, pp. 226-227) proposes four dissimilar sorts of poetry translation.They are literal translation, approximation, adaptation, and imitation.Accordingly, a literal translation continues to be "faithful" to the original "meaning of the poem" while preserving its entire "poetic integrity" (p.226).Obviously, this is hardly ever possible, because of "differences in word sounds" (p.226).Literal translation is like Nida's (1964, p. 159) formal equivalence.He mentions that in formal equivalence, "the message itself, in both form and content" is important.Here, the message in the target language should be as close as possible to the various parts of the source language.Being faithful to both form and content of the original text is the attempt of the translator in literal translation.Accordingly, Newmark (1981, p. 39) states that "in communicative as in semantic translation," on condition that "equivalent-effect" is available, "the literal word-for-word translation is not only the best, it is the only valid" translation method.
Approximation is one other sort of poetry translation wherein the translator is faithful to some facets of the original poem like "sense, structure" and "rhyme scheme" (Jones, 2014, p. 226).He, however, has some liberties in order that a rational translation might be created.Adaptation has a lesser amount of faithfulness to the original text comparing to the approximation.In view of that, Munday (2001, p. 58) states that adaptation concerns with altering the cultural reference once a condition in the source culture is not present in the target culture.Adaptation, as Vinay and Darbelnet (2000) mention, is applied in certain situations in which the kind of condition that the source language message is referring to is unidentified in the target language culture.Creating a novel poem in the receptor language with the source language theme is imitation (Jones, 2014).This kind of translation of poetry is typically done by poet-translators.Accordingly, Johnson (cited by Jackson, 2011) views imitation as only a poem loosely rooted in another poem.Jackson (2011) cites from Dryden that the translation needs freedom, both to change sense and words, and to relinquish them regarding the coming circumstance.The translator should consider merely some common clues from the source text to produce whatever he wants on their basis.Imitation, as Dryden said, is a type of free translation and is relatively like adaptation (cited by Jackson, 2011).For imitating a poem, the imitator is affected by it in creating a novel poem of his own, preserving the original theme.As a result, while the two poems have dissimilar wordings, they have the same theme.Folkart (cited by Nouss, 2011) states that she has "no intention of bogging down in a futile attempt to draw a hard and fast line between translation and imitation" and it does not make any difference "as long as the results are esthetically viable" (p.356).Folkart (2007), considering imitation, mentions that "if you can't make the poem yours, you won't be able to make a poem.…When it comes to poeming, the end justifies just about every conceivable means" (pp.377-378).Accrodingly, Boase-Beier (2009) states that a different method of coping with the hardness of translating poetry is to retreat from the original poem and try to create an imitation or a "version" as labeled by Paterson (2006, p. 73). However, Hamburger (1989, p. 51) believes that this retreating from the original poem is "an admission of defeat".It is the case while a number of poetry translators think it is the mere method for creating translations of poems which are poems themselves (Boase-Beier, 2009).
As mentioned by Jones (2014, pp. 226-227), approximation, adaptation and imitation in poetry translation all specify that there are some elements untranslatable in poetic discourse.According to Nida and Taber (1982), anything that could be mentioned in one language could be mentioned in another, if the form is not an indispensable part of the message.In view of that, Nabokov (cited by Hatim & Mason, 1990) believes it is merely a literal translation which is able to communicate the true essence of the poem.
Regarding the English translation of Persian poetry, Browne (1927) mentions that the Persian poetry is outstandingly dependent for its efficiency on the twist of phrase, the words or rhyme choice.Besides, he states that in this case a literal translation is frequently tedious and incomprehensible if no clarification is given.It is opposite to the aforementioned idea believed by Nabokov that it is merely a literal translation which could express the poem essence; however, Browne declares that the beauty of Persian poetry depends on its language and its music instead of its meaning.

Poetry Translation Methods
For rendering a text, there are numerous ways; however, for poetry translation, some of them are proper to be applied.Accordingly, Lefevere (cited by Bassnett, 2002, p. 87) mentions seven methods for poetry translating: phonemic translation, literal translation, metrical translation, poetry-to-prose translation, rhymed translation, blank verse translation, and interpretation.
1) Phonemic translation remakes the source language sounds in the target language.And, simultaneously, the translator seeks to convey the meaning.According to Lefevere (cited by Bassnett, 2002), on the whole, the outcome appears awkward and every so often does not express some parts of the source meaning.
2) Literal translation is word-for-word translation.In this method, "the sense and the syntax of the original" are distorted in translation (p.87).
3) The metrical translation focuses on recreating the original meter in the receptor language.And, "like literal translation" it focuses on only one facet of the original text "at the expense of the text as a whole" (p.87).Since every language has its own particular stressing and pronunciation system, this method brings about an unsuitable translation regarding structure and meaning.4) Poetry-to-prose translation has also some weaknesses.In this method, as Lefevere mentions, the source text "sense, communicative value and syntax" are distorted (p.87).The other vital weakness of this method is that the source poem beauty cannot be achieved.5) Rhymed translation focuses on conveying the rhyme of the original poem to the target text.The outcome is semantically unsuitable and according to Lefevere it is "merely a 'caricature' of Catullus" (p.87).6) In blank verse translation method, the translator might bring the proper equivalents in the target language.However, the rhyme and meter are overlooked.Therefore, the outcome is not physically similar to the source one, but it semantically looks identical.7) Interpretation is the last method mentioned by Lefevere.In line with him, there are two kinds: version and imitation.A version of a poem in the target text is semantically similar to the source one though it is physically dissimilar.And, an imitation is precisely a dissimilar poem, but the topic and opening spot are identical with the source poem.
Lefevere's clarification of the aforementioned methods appears to re-accentuate that the weaknesses of the methods of translating poetry are owing to just considering one or some of the poetic elements in the process of translation (Bassnett, 2002).The literal, metrical, and rhymed translation methods appear to focus on the form or poetic structure of the poem; whereas the rest give emphasis to the conveying of the exact meaning to the receptor text.It looks none of the mentioned methods are able to aptly provide the needs of poem translators.However, according to Nida and Taber (1982), translators, while translating, should compensate as much as they can via applying some syntactic and lexical adjustments in their translation.

Translatability and Untranslatability
Jae (2005) believes that translation is practically probable, since it is theoretically feasible.This outlook emphasizes on the point that it is true that translation actually happens among languages and inside one language; however, it should be considered as well that language per se is a translation of another thing.In view of that, Ramat (1987, p. 10) believes in translatability and mentions that due to the presence of language universals, it is "possible to understand why languages, in principle, can be translated into other languages".Wong (2006) states that translatability is "something slightly different: the degree of ease with which one language lends itself to translation into another" (p.124).According to Pym andTurk (1998/2001, p. 273) ''the question of translatability is also sometimes used to illustrate general methodological or philosophical concepts".They further mention that the issue of translatability might emphasize the source text or the target text; it might be about translating of different kinds of texts like religious, literary, or scientific texts and so forth.
Thus, talking about the issue of translatability depends on the text types, and the barriers involved in the process of their translations.Accordingly, Bassnett (2002) mentions that recreating the form and the content of the original language by the translator is affected both by the receptor language system and the original language system; besides, it relies on the translation function as well.In view of that, Gutt (1991), cited from House (1981), states that functional equivalence success is not feasible in overt translation and it is also hard to accomplish this equivalence through covert translation due to the socio-cultural dissimilarities.Covert translation is a kind of rendition that does not look like a translation; however, it is like an original one, better to say that, it is so natural that it does not seem to be a translation.On the other hand, everything in overt translation shows that it is obviously a translation not an original work (Gutt).According to Larson (1984), some notions of the source text might be like some recognized notions in the target culture; however, if one analyzes them again, he might realize that they have dissimilar purposes and implications.Here, on the occasion that the target text readers get them with regards to their own culture, the real meaning might change.
However, considering the issue of equivalence, one should bear in mind that there is a significant dissimilarity between linguistic and translation equivalence.According to Toury (1979), inter-linguistic or intra-linguistic equivalence is a two-sided association.In other words, X is the equivalent of Y to the extent that Y is the equivalent of X.In opposition to it, translation equivalence is a one-sided association of target text to source text or target language to source language.Consequently, linguistic equivalents are exchangeable in a particular condition technically, while a translation equivalent substitutes a unit or structure of a new text or language under study.In other words, linguistic equivalents preserve matching places in the linguistic system, while, translation equivalents keep equivalent places in a text, regardless of the fact that whether there is such a place in the whole linguistic systems.
Untranslatability appears wherever there is no equivalent in translation for a word or expression in the source text.Accordingly, Nida and Taber (1982) believe that here, a translator might use a suitable translation kind so as to have an appropriate translation via applying structural and semantic adjustments.They further talk about one specific kind of meaning and mention that the facet of the meaning that concerns with one's "emotional reactions to words" is labeled as connotative meaning (p.91).This facet of meaning refuses to be translated and stays untranslatable, since it belongs to the word implicit part which has a translatable denotative part.Therefore, connotations are indescribable and carry sensational emotions with themselves that are hardly translatable.Sun (2012) states that cultural connotation is a challenge in translation.The issue of untranslatability fortifies a nonstop need for readability, accompanied by the need for relevant cultural and cross-cultural concerns in translation.It has high significance in literary translation.Literary translation is predominantly involving in rendering the untranslatable factors or the hardly translatable factors.Accordingly, it is crucial to explore diverse kinds of untranslatability that are confining and forming translation.Moreover, readability is plausible with applying some adjustment techniques in an attempt to generate some type of synthetic naturalness and transparency in translation.
In poetry translation wherein form is a vital fraction of the message, untranslatability is an inevitable issue, as, accordingly, Nida and Taber (1982, p. 4) state that whatever is told in a language can be rendered in another language, "unless the form is an essential element of the message".In view of them, Jakobson (1959Jakobson ( /1966) ) resolutely believes that poetry is untranslatable.In line with him, Bonnefoy (1992) believes that poetry is not translatable, since translators face with many challenges which they cannot overcome; and, they have to "make too many sacrifices" (p.186).He further mentions that translating poetry via compensation does not, nonetheless, bring about "the poetry of other languages" (p.188).Discussing the issue of literary texts translatability, Boase-Beier (2011) states that cultural and linguistic differences between two languages, namely, SL and TL, might be regarded as one of the reasons of literary translation impossibility.He also mentions that the close link between meaning and form in literature is another reason for its untranslatability.Because, if translators detach form and meaning, keeping merely the latter, consequently translating literary texts becomes unfeasible.Accordingly, Salah Salim Ali (2005) believes that literary works and specifically poetry have to be studied as a whole meaningful construction which involves the semantics and pragmatics of the text.He states that there are some stylistic traits that entail an unusual language usage, which considerably change the text content.They are the untranslatable or hard to be translated poetry parts and need to be considered carefully by the translators.
One of the issues that has been claimed to be untranslatable is literary style.Though, Shiyab and Lynch (2006) believe that style is translatable if the translator has the theoretical information, sensible abilities, and the capability to cautiously understand the "tone" and "spirit" of the original text (p.264).For rendering literary style, comprehending paralinguistic characteristics is significant, and translators must have the ability of skillfully understanding and examining the artistic symbols and images at hand.One of the most significant concerns of literary translation, chiefly poetry, is the poetry nature and its specific language.Comprehending poetry entails realizing the characteristics of poetic language.The capability to do so is one of the requirements the translator of literature must have (Savory, 1975).In view of that, Yang (2010) states that poetic languages of different cultures, even if somewhat dissimilar from daily languages, are similar in being special.Moreover, the history has shown that "poems cannot only be translated, but also translated well" (p.168).Pedro (1999, p. 546) explains two outlooks about translatability, universalist outlook and monadist one.Accordingly, those who believe in universalist view argue that translatability is guaranteed by the linguistic universals.And the followers of the monadist view declare that speakers of any language "interprets reality" in their individual specific way, and it makes translatability vulnerable.Hermans (2009) mentions that the everyday translators' performances clearly prove that translation is feasible.Rejecting translatability does not typically hypothesize total untranslatability; however, it wonders whether it is feasible to attain completely adequate translation.
Accordingly, Wilss (1982, p. 71) states that one can evaluate the text translatability by looking to the point that to what extent the text might be "re-contextualized" in the receptor language, while considering the whole linguistic and non-linguistic parameters.He has gone on to say that the text translatability is therefore possible, as there are universal classifications in syntax and semantics.Hermans (2009), cited from Derrida, states that untranslatable elements challenge the translator to deal with the unfeasible.As the degree of untranslatability of a text goes up, its request to be translated increases as well.In view of that, Xiumei and Qinyan (2012), discussing the issue of translatability vs. untranslatability, take the relevance theory view in which the equivalence is rejected and instead the term resemblance is proposed.They have believed that whether translation is possible or not is not the question anymore; however, the relevance theory now looks for the degree of translatability.

Types of Untranslatability
Catford (1965) considers two types of untranslatability, linguistic and cultural one.Linguistic untranslatability occurs when a translator cannot "find a target language equivalent" because of the dissimilarities between the source and target languages (p.98).There are some instances such as puns, ambiguity, and polysemy.Cultural untranslatability comes up where a situational element which has a particular function in the source text does not fully exist in the culture of the target language, for example, the names of some tools and foods.
With regards to the specificity of poems, translating it needs several skills and translators have to be skilled readers of the original poem and skilled writers of the target poem (Jones, 2011;Folkart, 2007;Bassnet, 1998).They should be expert in both languages so as to be able to choose suitable equivalents for the multifaceted characteristics of the original poem; moreover, once those traits are hard to be rendered, the translators have to have the literary ability to judge what should be recreated and what should be discarded.And, when they want to recreate them, they should know how to do it.Accordingly, Nida and Taber (1982) believe that the translators need to use some expansion techniques in order to make the translation possible.And, as poetry language is usually condensed, its translation needs on the part of the translator to apply several expansions.
According to Nida and Taber (1982, p. 163), there is "a tendency for all good translations to be somewhat longer than the originals".This does not obviously imply that every long translation is inevitably good.It merely implies that in the transfer process from one cultural and linguistic structure to another, it is approximately unavoidable that the translation will become longer.Nida and Taber (1982) state that this inclination to greater length is owing basically to the reality that the translator wants to mention everything that is in the original text; however, he is also forced to explicate in the target language what can stay implicit in the source text, as the original readers of this text apparently had all the essential background to realize the message contents.
The foundation for the actuality that good translations are apt to be fairly longer than the original text can be, nonetheless, described much more specifically.Firstly, it might be supposed that every message that is given has two fundamental aspects, length and difficulty (Nida & Taber, 1982, p. 164).In the original text, any well organized message is intended to match the channel capacity of the readers.In other words, it is written in a way that can be understood by the readers.
If, nonetheless, the translator renders a message literally from the source language to the target language, and in this process uses a message with the identical aspects of length, unavoidably the difficulty aspect will be substantially greater (Nida & Taber, 1982).However, it changes to a severe problem since the normal channel capacity of the readers in the target language is much smaller than that of the original readers.It is undoubtedly the case when the languages are members of fairly dissimilar linguistic families as well as the time when the cultures are fairly dissimilar (Nida & Taber).Because in the literal translation the aspect of difficulty is greater, whereas the channel capacity is smaller, the merely feasible way out is to "draw out" the message, i.e., to add in redundancy (p.164).Accordingly, the translators should apply different kinds of expansion in their translations to make the flow of the target texts smooth for better understanding of their readers and to give them their needed information.

Method
This study was carried out in a comparative mode on Bustan and its English translation extracted from "The Bustan by Shaikh Muslihu-D-Din Sa'di Shirazi" by Clarke (1985).Bustan, which is also referred to as Sa'di-name, is a moralizing verse work (Katouzian, 2006).In fact, it is one of the most famous poem in Persian literature and has many unique features (Wickens, 1990).Accordingly, this research intended to investigate poetry translatability and sought to see what the translators do to compensate those problematic areas of poetry.The present study was done in a parallel unidirectional Persian-English corpus.For this purpose, one hundred verses of Sa'di's Bustan were selected at random to be analyzed.The unit of the analysis was the verse.The researchers went through the verses and their translations to extract the expansion cases that are evident in the process of translation.According to Nida and Taber (1982), there are two kinds of expansion: syntactic and lexical.Each of them has some sub-categories.In view of that, the researchers have applied this theory on the corpus of the study and have attempted to indentify, in each case, first the expansion kind and then the subcategory.And, the due explanations have been given in each case.

Theoretical Framework
Nida and Taber's (1982) expansion theory has been followed in this study by the researchers.Here, a through explanation of it is given.

Expansion Kinds
As expansion comprises the chief constituents in exploring dynamic equivalence, it might be helpful to indicate those specific kinds of expansions which frequently take place (Nida & Taber, 1982).In this procedure, the types of expansions that are often obligatory and the types of expansions that are not essential or legitimate will become obvious.There are two kinds of expansions which are syntactic or formal expansions and lexical or semantic ones (p.166).

Syntactic expansions
The most widespread expansions needed by the syntactic structure of the target languages comprise: 1) identification of the participants in events, 2) identification of objects or events with abstracts, 3) more explicit indication of relationals, and 4) filling out of ellipses (Nida & Taber, 1982, p. 166).In the following, they are discussed.
1) The first category of syntactic expansion is identification of the participants in events.It is clear that so as to recognize explicit participants, nouns have to be introduced instead of pronouns; this can be considered as a substitution as well (Nida & Taber, 1982).In fact, the discussing subject is more multifaceted.For instance, in the following clause, "I am the resurrection and the life", the reader might entirely fail to notice the point except the subject of the "resurrection" is evidently recognized.Since "resurrection" is not a transform of an intransitive verb however of a causative one.To be precise, "I am the resurrection and the life" implies "I am the one who causes people to rise again and to live".Except the subjects of these two events of "rising" and "living" is obviously recognized, the receptor might think that this is merely an indication to the truth that Jesus himself will become alive from the dead and continue to live again (p.166).Indeed, this is exactly the case wherein most native English speakers comprehend this text.According to Nida and Taber (1982), another kind of participant, nonetheless, exists that must be evidently distinct in translation.It is the subject of direct sentences.For instance, in numerous languages a relatively literal translation of Mark, "as it is written in Isaiah the prophet, Behold I send my messenger…," recommends that Isaiah is "sending his messenger".Certainly, this misinterpretation could be corrected by clarification.This kind of clarification is, however, not needed once the subject of the direct quotation is made completely obvious in the translation, whether by bringing a phrase like, "Isaiah, speaking as a prophet on behalf of God," or "as Isaiah the prophet wrote, God says, Listen, I am sending my messenger…".
2) The identification of objects and events with abstracts might be a very uncomplicated issue, as in the example of "know the truce," that might be expanded to be "know the true word" or "know the true message" (Nida & Taber, 1982, p. 166).Conversely, these expansions might be fairly multifaceted as well, for example in Luke, in which "turn… the disobedient to the wisdom of the just" might need substantial expansion, e.g., "to change the disobedient persons so that they will act wisely as the just people do" (p.166).
3) The third category of syntactic expansion is more explicit indication of relational.Nida and Taber (1982) state that a language relationals commonly need more specification in another language.For instance, "Be angry but do not sin" could be rendered as two fairly imperatives (p.166).They are legitimate and might be somewhat linked to each other.To be fair to the translating of this kind of semantic sentence, the translator has to fairly explicate the relationships.For instance, a probable rendition of this statement could be: "Even if you do get angry, you must not sin" (p.167).
4) The fourth category of syntactic expansions is filling out ellipses (Nida & Taber, 1982).Every one of the languages uses ellipsis; however, the ellipsis patterns are typically fairly varied in various languages.For instance, in a number of languages the equivalent of "He is greater than I" is "He is greater than I am great," whereas in other languages it has to be translated as "He is great, I am not".Some problems are created by these ellipses; however, one might not notice some of them.For instance, in many languages, "the Sabbath was made for man, not man for the Sabbath" has to be rendered as two joint positive-negative statements: "The sabbath was made for the sake of helping people; people were not made for the sake of honoring sabbath".In this case, there are two types of ellipsis: (a) the lack of the verb in the second clause, and (b) the fact that the events in the two examples are fairly dissimilar.Because of this, the implied words, "helping" and "honoring" have to be inserted (p.115).Some parts of speech are sometimes ellipted in some languages and they can be filled out in the course of translation like conjunctions or verbs.

Lexical expansions
According to Nida and Taber (1982, p. 167), there are three kinds of lexical expansions: (1) classifiers, (2) descriptive substitutes, and (3) semantic restructuring.In the following, they are discussed.1) Classifiers are fairly ordinary and could be employed once a borrowed word requires some semantic redundancy, in order that the receptor would get some information about its function and form, for instance, Mashhad as the city Mashhad, Zayanderud as the river Zayanderud, and Orumiyeh as Lake Orumiyeh.
2) Descriptive substitutes have approximately constantly more length than the equivalent originals (Nida & Taber, 1982); since they need a number of dissimilar lexical elements to explain the function and form of the event or object under consideration.For instance, "a synagogue" might be described as "the worship-house of the Jews" (p.167).
3) Several expressions, nevertheless, frequently need significant expansion in the target language, as they are semantically very condensed in the original text (Nida & Taber, 1982).For instance, "I am a jealous God" could be misinterpreted if rendered literally; since it might merely recommend that God behaves similar to some jealous lovers or that he has a low, selfish nature.In view of that, in several languages, this statement has to be semantically reformed through expansion to be, "I am a God who demands that my people love no one else other than me" (p.167).In the following figure, the whole strategies can be observed.

Data Analysis
As a whole, one hundred verses were analyzed, but, here in this section, only some examples of expansion adjustments made by the English translator of Sa'di's Bustan are discussed.It has to be mentioned that more explicit indication of relationals, the third sub-category of syntactic expansion and classifiers, the first sub-category of lexical expansion were not observed in this study.

Syntactic Expansion
In the following, some instances for the observed types of syntactic expansion have been given.

Identification of the Participants in Events
Example 1
The translator added "angles" in his translation.It was not directly mentioned in the source text.It is an identification of the participants in events; and it is a type of syntactic expansion.By using this expansion, the translator said to his readers that the subjective pronoun they refers actually to angles and the fact that angles are the doers of the action of seizing the hands.Therefore, this explicitation of the implicit information of the source text will add to the knowledge of the target text readers in a way that they will not cope with the ambiguity anymore.Otherwise, there is a question that keeps appearing in the minds of the readers, that is, who seized our hands?But, the translator has duly helped the readers in this case by expanding his rendition.
Example 2 TT At this open table, whether enemy (infidel), or friend (the faithful) -what matter?
In the translation, he has added "infidel" and "the faithful" to make the meaning of "enemy" and "friend" clear to the target text readers.These two words are expansions.They are cases of identification of the participants in events.Here, the translator has truly gained the source text meaning and expanded his translation in order to convey it properly.In fact, the original text author by using enemy has meant those who do not believe in God; and, by using friend, he has meant those who have faith in God.Therefore, the readers will understand the target text better.
Example 3 TT And, if a traveller (a pious one) becomes acquainted with the secret of God.
They (the angles) will shut on him the door of returning (to the world).
In the translation, the translator has used two syntactic expansions in his translation.They are "a pious one" and "the angles" which are implicit in the source text.Both of them are identifications of the participants in events.
Here, the translator has explicated his translation by making the participants clear.The first instance explicates what the author meant by a traveler.The translator has expanded his rendition by mentioning that a traveler should be a pious one not every traveler.This being a pious traveler is implied by the source text.Thus, the readers of the target text will have better understanding of the translation.The translator has made the participant clear by using the angles.As the subjective pronoun they is not clear.Thus, the translator has properly expanded his translation so that his readers will not have difficulty in comprehending the text.
Example 4 TT Joy began to be afraid of me, at that time, when my evening (black hair) began to blossom as the down (white hair) In this case, the translator has used the phrase black hair to explicate the meaning of the subjective phrase my evening.In fact, he has attempted to convey that the original author has meant the pass of time and its effect on human in a way that the black hair changes to white hair.This is the identification of the participants in events where the subject of the sentence has been made clear for the readers.

TT
The Messenger of good news and the Observer (Muhammad) sent an army; They took captive a multitude of them.
To identify the subject of the sentence, the translator has added the real intended subject by the original author, i.e., Muhammd the Muslims' prophet.Thus, he has omitted the ambiguity that it might have for his readers.This is the identification of the participants in events.

Identification of Objects or Events with Abstracts
Example 1

TT
The pious ones gave not their hearts to the covering (external beauty).
In this example, the translator has used the strategy of identification of objects or events with abstracts by adding the phrase external beauty to his translation.Indeed, he has clarified what the original author has meant by covering and has mentioned that it refers to the appearance; therefore, his target text readers will have better understanding of the meaning.
Example 2 TT By power, the Guardian of high (sky) and low (earth), The Lord of the Court of the day of reckoning (Judgment-day).
In the translation, there are three expansions.In the first line Clarke has added two words, that is, "sky" and "earth" to clarify the words high and low respectively.In the second line, he has used "Judgment-day" to explain the "day of reckoning".They are identifications of objects or events with abstracts.As can be seen, if these expansions have not been used by the translator, the readers might keep asking what the author means by high or low or the day of reckoning.But, the translator, by expanding his translation and clarifying them, has provided intelligibility for his readers.Therefore, the translator has successfully conveyed the message of the original text into the target text.And, accordingly, he has suitably expanded his translation.

TT
The lamp of the world-consumer (tyrant) extinguished-is best.
In this case, the translator has clarified what the original author has meant by the world-consumer, i.e., tyrant; therefore, he has used the identification of objects or events with abstracts.Accordingly, he has removed the possible ambiguity that the target text readers might have.
Example 4 TT Those who are yet invisible (unborn) will come, and pass over our dust.
In this instance, the translator has used the identification of objects or events with abstracts strategy wherein he has added unborn to his translation so as to clarify the word invisible for his readers.He has explicated the implicit meaning behind the word invisible which refers to those who are not born yet, that is, the next generations.

Filling Out of Ellipses
Example 1
In the above mentioned instances, the translator has used the strategy of filling out of ellipses.As can be seen, the to be verbs have been ellipted in the source text; however, the translator has filled out the ellipted verbs in the target text.
Example 2 TT That the infidel sits secure from contest with him; "(But) a Muslim escapes not from the violence of his tongue." In this example, the translator has added the ellipted conjunction to his translation so as to show to his readers that there is a contrastive relation between these two sentences.Using this expansion, he has tightened the two sentences together semantically.Accordingly, he has used the strategy of the filling out of ellipses.
Example 3 TT (To be) a man of black condition (in distress)-in it, foot-bound, Is better than to carry strife from place to place.
As can be seen, the translator has used the strategy of the filling out of ellipses.In fact, due to different syntactic rules of English and Persian languages, he has had to fill out the ellipted verbs of the source text in his translation.

Lexical Expansion
In the following, some instances for the observed types of lexical expansion have been given.

TT
When the teacher of truth lets fall clay on the corpse, His heart will burn, not for it, but for himself.
In the translation, Clarke has used "the teacher of truth" for the Farsi word [mohaeghegh] in his translation.It is a descriptive substitute strategy.He had to expand his rendition so as to express the meaning of the word in the source text.Therefore, several words have to be used instead of that condensed word of the original text so as to convey the meaning properly.

TT
The embroidered leather surface of the earth is His common table.
In the above example, the translator has used a descriptive substitute, i.e., The embroidered leather surface for the original word, i.e., [aedime].In fact, he has attempted to convey the meaning of the word by describing it in the receptor language.

Semantic Restructuring
Example 1 TT My body trembles, when I bring to recollection, the prayers of one distraught in the sacred enclosure at Makka.
In the translation, the translator has used "the sacred enclosure at Makka" for [haeraem].In this case, while expanding his translation, Clarke has attempted to convey the meaning into the target text and make the word meaningful for the readers.It is a semantic restructuring of lexical expansion.The translator has restructured the Persian word in his rendition.This act increased the number of the words of the target text.In his translation, he has explicitly made use of the word Makka which is not explicit in the source text.Indeed, he has used "the sacred enclosure at Makka" for conveying the intended meaning; therefore, he has fully understood that this word has implicit meaning.In other words, sacredness is implied in this Persian word.Besides, this sacred place is a special place; it is not any sacred place like a mosque or a church; in fact, it is Kaaba, Muslims' God's house which is located in Makka, a holy city for Muslims.It is the case that all the source text readers know these facts; therefore, they are implicit in the source text.But, the translator has to make them clear for his readers.And, in this case, the translator has expanded his translation so as to convey the meaning clearly to the target text so that the readers will have better intelligibility.
Example 2 TT In memory of God, they have fled from the world; So intoxicated with the splendour of the Cup-bearer (God) that they have spilled the wine!
The translator has used expansion strategy for translating the Persian word säghi.Indeed, he has used the semantic restructuring strategy by using the splendour of the Cup-bearer (God).He has mentioned here that the beauty and the splendor of God or cup-bearer is the reason for being intoxicated.He semantically restructured the very condensed original word.

Conclusion
In the area of literary translation, most of the researches have been about the problems of translating poetry instead of other literary styles (Bassnett, 2002).One of these problems that a translator is faced with in translating poetry is the untranslatable elements as a result of linguistic and cultural limitations, and literary and poetic traits.These parts might cause limitations in translation.Here, nonetheless, the translator might use an appropriate kind of translation.He also considers target language linguistic and cultural standards for creating suitable, intelligible and natural translation.In doing so, he may apply some structural and semantic adjustments (Nida & Taber, 1982).
Accordingly, Nida and Taber (1982) state that expansion is sharing of the semantic constituents over several dissimilar words.Once a word of the source language is translated by some words in the target language, expansion adjustment is used.In view of that, Larson (1984, p. 14) mentions that there is rarely a total "match" among languages.Therefore, it is frequently essential to render one word of the original text to some words in the target language for expressing similar meaning.Indeed, in a translation, expansion adjustment is an analytical translating of the source expression.
One hundred verses from Sa`di's Bustan were analyzed to realize what the translator did to overcome the untranslatability problem in his translation.One of the findings of this study is the number of expansions the translator used in his translation.Accordingly, Clarke has used ninety two cases of expansion in his translation.
The number of descriptive substitutes in translation was more than the other kinds of expansion.The frequency and percentage of lexical and syntactic expansions are shown in tables 1 and 2. As far as types of syntactic expansions are concerned, in Clarke's translation, identification of the participants in events was used 16 times; identification of objects or events with abstracts was identified 26 times; and filling out of ellipses in his translation was observed 8 times.But, more explicit indication of relationals was not used by the translator.In Figure 2, Clarke's overall usage of expansion types can be seen.In the current research, Sa'di's Bustan verses were so condensed wherein the language is principally connotational.This is why the translator has used expansion for expressing the meaning of the Persian source text to the English target one.In this study, the researchers came to the point that the translator used the strategy of blank verse for his translation; as a result, the source rhythm and music were not achieved in the translation.
The other noteworthy result is that the studied translation is chiefly source based.Clarke has attempted to express in so far as feasible of the source text content even to the extent of removing the formal beauties and traits, the harmonious form, and the internal delicacies of the source poem.Persian poetry is generally musical and rhymed.Accordingly, Browne (1927, p. 59) states that "the variety of Persian poetry meters is infinite; and while some make a ready appeal to the western ears, others sound heavy and halting; though to the Persian themselves they are equally musical".It is hardly feasible to translate these poetic features to another language without fundamental alterations.
According to the mentioned issues, it can be concluded that the lexical and syntactic adjustments are proper for making the translation of poem semantically intelligible and natural to the receptors in the target language.The poetry translation feasibility does not, nonetheless, imply that the entire poetic features are translatable.Every language has its individual linguistic traits of poetry which is merely for that specific language; besides, it is not feasible to recreate them in other languages.Consequently, due to those untranslatable parts which are not achieved in the translation process, translation of a poem does not have precisely the similar impact on the receptor readers, and it is not able to create the identical feeling, comparing to the source poem.Thus, owing to the differences among linguistic systems, sociocultural standards, and poetical characteristics of all languages, it is approximately unfeasible to render a poem so that the rendition entirely matches the source one.Among languages, thanks to the existed dissimilarities, there are no complete matches; besides, there have been always untranslatable elements in translation.In poetry translation, specifically, all elements of the source poem are not always achievable.But, translators have to do their best to lessen the untranslatable factors as in the present study, the translator, Clarke has attempted to render Sa'di's Bustan as semantically accurate as possible via using expansion strategies.

Figure 2 .
Figure 2. Expansion types used by Clarke

Table 1 .
Kinds of lexical expansionIn Clarke's translation, descriptive substitutes were used 31 times and semantic restructuring was identified 11 times.No use of classifiers was detected by the researchers in the selected corpus of the study.