An analysis of translation errors in 5 literary genres based on American Translation Association (ATA) framework

Abstract The analysis of translation errors holds a great deal of importance in the realm of translation studies. Accordingly, many pieces of research have been done on the topic of translation errors analysis in different text types including religious, journalistic, tourism, commercial, and legal texts. Nevertheless, no research has been conducted within literary genres with regard to translation errors from English to Persian. Thus, the present study aimed to analyze translation errors in five literary genres including prose, poetry, non-fiction, play, and media based on the American Translation Association (ATA) taxonomy. Three objectives were set for this study. The first aim was to distinguish the most frequent error types within each literary genre. The second objective was to determine whether there is any relationship between the text types and the errors. The third and last objective was to extract the possible pattern of committed errors within each genre. To conduct the research, 28 students of English Translation Studies were selected as the participants of the study. Translations’ errors were identified and categorized based on the framework of the study. After analyzing the data, the findings revealed that the most frequent error types in the prose text were: verb tense (44.44%), cohesion (15.55%), and omission (11.11%); the most recurrent error types in the poetry text were punctuation (19.22%), terminology (16.16%), and omission (16.16%); mostly visited error types in the non-fiction text were terminology (33.58%), omission (16.29%), and punctuation (8.64%); omission (22.64%), verb tense (16.71%), and literalness (15.63%) were the most frequent error types committed in the play text; and misunderstanding (23.77%), omission (20.08%), and terminology (13.93%) were the most recurrent error types in the media text. The results obtained from the ANOVA test indicated that text type is significant except for non-fiction and play. Moreover, there was a similar pattern among errors in the five genres. The findings of the present study can be of help to translators, teachers, students, and researchers in the field of Translation Studies.


ABOUT THE AUTHORS
The first author has an M.A. in Translation Studies. She is a freelance translator. Her area of interest is assessing translation quality and she is also interested in applied linguistic in translation.
Dr Azadeh Nemati is an Assistant Professor in Iran, majoring in ELT. She is the member of Network of Women Scientists of the Islamic world and also the editor in chief of some international journals. She has already published +15 books and +40 articles nationally and internationally. Her main area of interest includes: TESL, TEFL, Sociolinguistics, Language Learning and Teaching, Vocabulary Learning and Teaching, Strategies, Gender Studies, Translation studies.
Mortaza Yamini, assistant professor emeritus of Shiraz University, is now an academic member of Zand Institute of Higher Education. His research interests include language teaching methodology, testing and phonology. He has extensively published in local and international journals and has authored several books.

PUBLIC INTEREST STATEMENT
Analysis of translation errors is one of the areas in translation studies that is of importance because producing a high-quality piece of translation has always been a matter of concern. The more errors a piece of translation has, the lower quality it possesses. Thus, a systematic study of errors which is conducted based on a standard framework can be of help to know our weaknesses and strengths and this can be a good way to improve our works. The focus of this study is on literary texts.

Introduction
Many pieces of research have been conducted on the topic of error analysis on the areas such as newspaper headlines, commercial labels, legal documents, Islamic texts, journalistic texts, and tourist guidebooks (Delforooz, 2010;Khodabandeh, 2007;Mahmoodi, 2007;Ilani & Barati, 2016), but no research has been conducted within the five literary genres with regard to translation errors. From that point of view, it seems necessary to study the issue of translation errors in literary genres in a more specialized way because of the popularity and the practicality of literary works in the field of translation. Because literary works include a wide range of texts from poetry to media, it is necessary to draw a distinction between genres.
Many scholars tried to define the term translation. "Translation is the replacement of material in one language (Source Language) by equivalent material in another language (Target Language)." (Catford, 1965, p. 20). Nida and Taber (2004) contended that "translating consists in reproducing in the receptor language the closest natural equivalent of the source-language message, first in terms of meaning and secondly in terms of style" (p. 12). Before presenting definitions of Translation Errors given by different scholars, it is better to distinguish between the terms "error" and "mistake". Goff-Kfouri (2004) differentiated them as two individual concepts by explaining that: "scientifically speaking, a mistake is generally considered as a fault in performance; it does not occur systematically. An error reflects a gap in student's knowledge and is systematic" (para. 6).
According to Neubert and Shreve (1992) "What rightly appears to be linguistically equivalent may very frequently qualify as 'translationally' nonequivalent". (p. 5) Sigrid Kupsch-Losereit, a functionalist who introduced the functional perspective on errors, defined translation errors as "an offense against: (1) the function of the translation, (2) the coherence of the text, (3) the text type or text form, (4) linguistic conventions, (5) culture-and situation-specific conventions and conditions, (6) the language system" (as cited in Nord, 1997, p. 73). Nord (1997) defined the concept of translation error with regard to functionalistic approach as anything which blocks the fulfillment of the specific function of a translation for the target readers. He classified translation errors into four categories: pragmatic, cultural, linguistic, and text-specific. Pragmatic errors are committed as a result of translator's failure to solve problems which are related to latent message in the source text. Cultural errors occur because of mismatch between cultural issues in the source text and target text. Linguistic errors are results of shortcomings in translation with regard to language structures. Text-specific errors are related to the inappropriateness of the equivalent translation to target readers.

Translation of literary genres
"Literary translation represents a distinctive kind of translation because it is concerned with a distinctive kind of text" (Hermans, as cited in Kuhiwczak and Littau (2007, p. 77). Newmark (1988) found translation of literature as the most problematic type of translation because the first meaningful unit of a text, which is the word, is as important as the second one, which is the sentence in prose or the line in poetry. Translation of literary texts can be a challenging work for translators because of the unique features that this type of text has. In this regards, Huang (2011) stated that "In literary translation, the typical features of the source literary text not only need to be taken into account, but also the influential elements from the target perspective, such as the linguistic and cultural differences and the target readers" (p. 14). Then, he went on introducing four specific challenges of literary translation: a) Literary language (poetic language), b) The dominant structures of literary texts, c) Translation as an individual translator's selections of the target language words, d) Target perspective: target language and culture, and target readers (Huang, 2011: 16). Similarly, Kazakova (2015) introduced three features that make the process of translation of literary texts complex. a) The personal nature of texts under translation (authorship), b) Unspecified target audience, c) Inter-lingual and/or intercultural inequality.
Although there is no fix category for literary genres, the researcher will use Drucker's category. According to Drucker (2019) the five main literary genres include: poetry, prose, drama, non-fiction, and media. According to Drucker (2019), when a text includes "meter or rhythm, and when it focuses on the way the syllables, words, and phrases sound when put together" (para 4), it is called poem. Then she went on adding "Poems are heavy in imagery and metaphor and are often made up of fragments and phrases rather than complete, grammatically correct sentences" (para 4). Newmark (1988) as a translation study scholar defined poetry as "the most personal and concentrated of the four forms [poetry, short story, novel, drama], no redundancy, no phatic language, where, as a unit, the word has greater importance than in any other type of text". Newmark (1988) suggested a three-step framework for translation of a piece of poetry: first, the translator should select a TL poetic form as similar as that of the SL; second, s/he should create the same figurative meaning and the concrete images of the poem; third, the setting and the thought-words should be rendered properly. Aforementioned, second genre of literature is prose which is described by Drucker (2019) as a text which "is written in complete sentences and organized in paragraphs. Instead of focusing on sound, which is what poetry does, prose tends to focus on plot and characters" (para 6). According to Abrams and Harpham (2010), "prose is an inclusive term for all discourse, spoken or written, which is not patterned into the lines either of metric verse or of free verse." (p. 318). Haque (2012) presumed that what makes the translation of a literary-prose difficult is its arty function besides a communicative, societal, and connective purpose. He presented three methods to deal with problems of translating a piece of prose: 1. Back Translation 2. Conference with other people (especially bilinguals) 3. Interviews or questionnaires or any kind of tests. (Haque, 2012) Previously stated, the third literary genre introduced by Drucker is drama. Drucker (2019) defined drama as "Any text meant to be performed rather than read can be considered drama (unless it's a poem meant to be performed, of course). In layman's terms, dramas are usually called plays" (para 8). Abrams and Harpham (2010) defined drama as "the form of composition designed for performance in the theater, in which actors take the roles of the characters, perform the indicated actions, and utter the written dialogue" (p. 93).
In the sixties, Hamberg's (1969) proposed certain principles for the drama translator: Drama is action […] and in translating for the radio, television and the theatre it is important to realize what the dramatic theoreticians above all demand from the spoken line. It must characterize the speaker and thus seem genuine; it must characterize time and place as well as social class; it must not be ambiguous; and it should have been given or one should be able to give it the right emphasis so that it leads the attention of the audience in the desired direction.
[…] It goes without saying that an easy and natural dialogue is of paramount importance in a dramatic translation, otherwise the actors have to struggle with lines which sound unnatural and stilted.
[…] Even where the author does not indicate in brackets how a line is to be spoken, the translator as well as the stage manager must be able to know how.
[…] A translator must be especially careful with entrance lines and exits." (p. 91-94) Non-fiction is the fourth literary genre offered by Drucker. Drucker (2019) defined non-fiction as: a vast category that is a type of prose and includes many different sub-genres. Non-fiction can be creative, such as the personal essay, or factual, such as the scientific paper. Sometimes the purpose of non-fiction is to tell a story (hence the autobiography), but most of the time the purpose is to pass on information and educate the reader about certain facts, ideas, and/or issues ….Some genres of non-fiction include histories, textbooks, travel books, newspapers, self-help books, and literary criticism. (Para 10) Bertić (2015) stated that non-fiction is of an equal importance of other literary genres which include: fiction, drama, and poetry. Losman (1998) maintained that translating of fiction is simpler than translating of non-fiction because in translating of the first one, there is a dual relationship. That means that a translator is alone with the text, but in translating the later, s/he is responsible to the author, the text, and the reader. Similarly, Bertić (2015) believed that translating non-fiction is twice as challenging because a translator of non-fiction is dealing with facts, so s/he requires a wide range of knowledge in different fields such as religion, science, music and art. Some people find translating of non-fiction a "double challenge". This idea supported by some reasons including: first, its being time-consuming when the translator does not enough knowledge in the field, s/he is translating. Second, the text's being difficult to be understood because of terminologies it contains. Third, its being full of historical names, names of places, and quotations which are translations of translations. These make a text difficult for a translator to render properly into the target language. (Bertić, 2015). Because what have been presented to the participants as a sample of nonfiction genre was a piece of a psychological self-help book, definitions and some features of psychological texts within the realm of translation will be discussed as well. Shabanipoor and Moinzadeh (2013) found translation of specialized texts problematic due to the difficulties in finding appropriate equivalents in the TL for the jargons that exist in the SL. Although there are some English-Persian dictionaries for terminologies of psychology, Shabanipoor and Moinzadeh (2013) found two major problems with these dictionaries. First, the disagreement between Persian Because what have been presented as the media to the participants of this study was a part of a movie, the matter of subtitling will be discussed. One of the new fields in the Translation Studies is audiovisual translation (AVT). The two main forms of AVT are subtitling and dubbing. Subtitling is a newer phenomenon than dubbing. O'Connell (2007) defined subtitling as "supplementing the original voice soundtrack by adding written text on screen" and dubbing as "replacing the original voice soundtrack with another voice in another language" (p. 169). Shuttleworth and Cowie (1997) defined subtitling as "the process of providing synchronized captions for film and television dialogue".

Models of error analysis
Many scholars tried to offer taxonomies for translation errors to establish a framework for assessing the quality of a translated work. Liao (2010) proposed a classification of error types with three categories of rendition, language, and miscellaneous. Rendition errors refer to the failure of translator in conveying the exact meaning of the source text. Language errors refer to problematic expression in the target text. And miscellaneous errors refer to omissions. Hurtado suggested a model for the translation quality assessment that is based on error analysis. The model is categorized under three headings: (as cited in Waddington, 2001) 1. Translation mistakes: Inappropriate renderings which affect the understanding of the source text; these are divided into eight categories: contresens (misinterpretation), faux sens (incorrect meaning), nonsense (absence of meaning), addition, omission, unresolved extra-linguistic references, loss of meaning, and inappropriate linguistic variation (register, style, dialect, etc.).

Translation major errors:
Inappropriate renderings which affect expression in the target language; these are divided into five categories: spelling, grammar, lexical items, text and style.
3. Translation minor errors: Inadequate renderings which affect the transmission of either the main function or secondary functions of the source text.
House (1997) offered a framework to evaluate the quality of a translation which is divided into three main categories: 1-Pre-linguistic studies, in which subjective and not-so-much clear statements, regarding the quality of a translated work, are the major trend.
2-Psycholinguistic studies, in which translation qualify is judged in terms of the effect a translated piece of work should have on the readership.
3-Source-text based studies, which attempt to build linguistic criteria in order to account for both the source text and the target text.
The researcher used American Translation Association (version 2017) structure retrieved from http://atanet.org as the framework of the study which includes 26 types of errors which are presented below: 1) Addition 2) Ambiguity, 3) Capitalization, 4) Cohesion, 5) Diacritical marks/ Accents, 6) Faithfulness, 7) Faux ami, 8) Grammar, 9) Illegibility, 10) Indecision, 11) Literalness, 12) Mistranslation, 13) Misunderstanding, 14) Omission, 15) Punctuation, 16) Register, 17) spelling/ character, 18) Style, 19) Syntax, 20) Terminology, 21) Text Type, 22) Unfinished, 23) Usage, 24) Verb Tense, 25) Word form/Part of speech 26) Other Errors. Khodabandeh (2007) analyzed errors of students in translating of headlines. Fifty-eight graduate students of English were participants of this study. They asked to translate thirty Persian and thirty English headlines. She used Keshavarz' (1993) model as the framework of her study. The results of the study revealed that the students had committed grammatical and lexical errors in their translations from Persian to English. The findings also indicated that they had grammatical, discoursal, and lexical problems in translating from English to Persian. Grammatical errors resulted from the translation of tense forms and the wrong use of declarative sentences. Lexical errors are committed because of the inappropriate rendering of vocabularies, proper nouns, acronyms, and abbreviation. Mahmoodi (2007) conducted a research titled "An Error Analysis Approach to Translating Persian Commercial Product Labels into English". She intended to discover the nature of these translation and identify the most frequent error types which had been taken place by participants. She used a group of senior students who majored in English Translation Studies. The researcher gave a test to the participants and asked them to translate the test which included some commercial labels. In order to analyze the errors committed by the participants, she used Keshavarz's (1997) taxonomy of errors as the framework of her study. The results revealed that 72.44% of translations were not acceptable. Most frequent errors were due to semantic, then grammar, and the least frequent errors were due to pragmatic. At the end, the researcher came up with the result that the cause of such errors is the lack of target language competency of participants.

Empirical studies
Delforooz (2010) investigated translation problems of tourism industry books which were translated from Persian to English. The objective of the study was to discover the syntactic, pragmatic, and semantic errors in the process of translation from source language to target language. He extracted 300 sentences from three tourist books randomly. The frameworks for this study were Keshavarz (1997) and ATA's taxonomy of errors. At the end he came up with the result that most of errors, 60 cases were due to semantic; 44 cases, were due to syntactic; 18 cases belonged to pragmatic; and 10 cases appertained to translation-specific errors.

Popescu (2013) conducted a research titled "A corpus-based Approach to Translation Error
Analysis: A case-study of Romanian EFL Learners". Poescu embarked on finding out error patterns produced by the participants of the study in the journalistic texts. The researcher used 30 students at an intermediate level of proficiency, aged from 20 to 24. The texts consisted of 15,555 words. Three types of error had been identified: linguistic, comprehension, and translation. Translation errors included: distorted meaning, addition, omission, and inaccurate rendition of lexical items. As Popescu (2013) asserted "Most of the errors in this case are due to misunderstanding of lexis, distorted meaning and to some extent, limited linguistic (morphological) competence" (p. 245).
Ilani and Barati (2016) analyzed errors of translation of journalistic texts. The aims of the study were to determine and categorize the errors and to determine the most frequent ones. The third objective was to see whether there is a pattern among the errors committed by the students in their translations. The researchers used Keshavarz's Model and ATA as a hybrid model. Participants included forty English Translation students who were at the final semester of their undergraduate studies. Four texts were selected to be given to the participants to be translated. Two experts helped the researchers in this process. The results revealed that the most frequent errors belong to the categories a) grammar, b) terminology, c) misunderstanding. They found that there is not a pattern among errors committed by participants.

Aim and methodology
Three objectives are set for this study. One of the aims of this study is to distinguish the most frequent errors within each literary genre. Another aim of the study is to determine whether there is any relationship between the types of literary genres and the errors which occur in the process of translation. The third objective is to extract the possible pattern of committed errors within each genre. Thus, three questions are posed in this study which are as follows: 1-What are the most frequent errors in each literary genre? 2-Is there any relationship between the types of literary genres and the errors which occur in the process of translation?
3-Is there any pattern among errors committed by participants within each literary genre?

Design
The design of the study was mixed. Both quantitative and descriptive research designs were implemented. Both descriptive and inferential statistics were used to analyze the data. To answer the first and the second questions, the method was descriptive. In order to answer the third question, the method was quantitative because the aim was to find the frequency of errors in each genre.

Participants
The participants of this study were a group of 28 BA students of English Translation Studies. They were all seniors, selected based on the availability sampling. They were 22 female students and 6 male students, with the age range of 21 to 30 years. The mother tongue of all the participants was Persian.

Instruments
The researcher used ATA's explanation of error categories in order to identify and categorize the extracted errors. The framework includes 26 types of error. Explanations are provided for each type in detail. The framework is presented in Appendices. In order to have a homogenous group of participants, just 28 students among 64 students who translated the materials were considered as the participants of the study. These 28 students received grades of pass according to the ATA framework for standardized error marking. According to the framework "A passage with a score of 18 or more [negative] points receives of a grade of Fail". (Retrieved from http://atanet.org). The chart is presented in appendix F. A table developed by the American Translators Association was used in order to show which errors undergo which category, meaning transfer/strategic errors, mechanical errors or errors in formal properties of the examination. The table also indicates the maximum negative points that would be given for each error type. The table is presented in appendix G.

Materials
Three paragraphs from each genre of prose, drama, and non-fiction and four stanzas from poetry were given to the participants to translate. A three-minute-extract from a movie named Doubt (Shanley, 2008), directed by John Patrick Shanley, was presented to the participants. For poetry, Yeats' Leda and the Swan was given to the participants to translate. It consisted of 113 words. Victoria Aveyard's the Red Queen, consisting of 299 words, was presented to the participants as a sample of prose. John Patrick Shanley's play named Doubt was a sample of drama which consisted of 301 words. And I'm Ok-You're Ok, a self-help book by Thomas Anthony Harris was given to the participants as a sample of non-fiction which consisted of 290 words. Drucker (2019)'s categories was used in the study to categories literary genres. She introduced poetry, prose, drama, non-fiction, and media as the five main literary genres.

Data collection procedures
The selected texts were presented to the participants and they were asked to translate them. The five types of texts were presented in five consecutive weeks. Participants were tested individually in the presence of the experimenter. They had an hour and a half to translate each text. They were allowed to use paper dictionaries. For translation of media, the experimenter played the selected part of the movie once and then it was played sentence by sentence and she asked the students to translate the sentences. The length of pauses was variable, depending on the length of the sentence which was played at the time. In order to have better translations, the selected part of the movie was played twice.
The reason for choosing the above-mentioned books and the movie is that all of them are considered as modern works. They were selected randomly among all other modern works. The decision on selecting the parts was reached after consultation with an expert who was teaching Translation of Literary Texts.
The collected papers of sixty-four students were examined and scored. Two factors were involved in scoring, first the number of errors, and second, the seriousness of errors. The more errors the students had, the more negative points they received. The more serious the errors were, again the more negative points the students received. The decision on the seriousness of the errors was made based on the flowchart for error point decisions. In the left column of the chart (mechanical errors), the more Yes answers, the rater could give to the questions of the flowchart, the more serious the errors would be, so the more negative points the students would receive. In the right column of the chart (meaning transfer or strategic errors), the more No answers, the rater could give to the questions of the flowchart, the more serious the errors would be, so the more negative points the students would receive. The negative points were deducted from 100. Passages with scores of 18 or more negative points received grades of fail and they were excluded from the study. Thus, 28 students were considered as the participants of the study. So the minimum score would be 83 and the maximum score would be 100, although no one received 100. Twenty-eight papers for each genre were considered as the data of the study. All of the papers were examined carefully. Then, the researcher identified and categorized errors, which were found in translations, based on ATA translation errors taxonomy. Each error was categorized under one or more heading(s). This was done for each genre. Then, the number of errors for each 26 types within five genres were estimated and calculated. In order to reduce the degree of error in the process of analysis and calculation, the researcher reviewed the translations once An expert helped the researcher in the process of spotting translation errors and scoring the translated texts. Thus, two raters rated translations.

Inter-rater reliability
The researcher and an expert rated translations. In order to calculate the consistency of the rating system, the researcher ran correlation in SPSS. As illustrated in Table 1, Inter-rater reliability was 0.99.

Data analysis procedures
The researcher used both descriptive and inferential statistics to analyze the data. For descriptive statistics, the researcher determined the mean and frequency percentage. The American Translation Association explanation of error categories was also used as the criterion to identify, categorize, and describe the errors. For inferential statistics, repeated measures ANOVA was used.

Answer to the first research question
3.1.1. What are the most frequent errors in each literary genre?

Is there any relationship between the types of literary genres and the errors which occur in the process of translation?
The results obtained from repeated measures ANOVA test indicated that text type is significant except for non-fiction and play ( Figure 1).
Scores were provided for each student based on ATA framework for standardized error marking Version 2017. Means of scores were obtained for each type of text. Repeated measures ANOVA was run.
Means of scores were obtained for each type of text. Mean of scores in poetry was higher than the other four text types. Then, media, play, and non-fiction had the higher means, respectively. The lower mean was for prose text type.

Is there any pattern among errors committed by participants within each literary genre?
The same pattern is true for all the five genres. The most frequent error type was devoted to the transfer errors. The second frequent visited error type in all text types was mechanical errors. And the least frequently committed error type was errors of exam form.
The researcher used ATA framework for standardized error marking and categorized the twentysix error types into three schematic category to identify the possible extractable pattern of committed errors by participants in each genre. Koby (2014) defined the framework "A table developed by the American Translators Association that lists categories of transfer errors, language mechanics errors, and errors in formal properties of the examination, with columns for 1, 2, 4, 8, or 16 points per error" (Chapter 13). The framework is presented below in Table 5: The mean frequency of each category of error types was calculated for each genre. In Table 6 the frequency of each three types is shown in all five genres: In all genres, the most frequent error type was devoted to the transfer errors. The second frequent visited error type was mechanical errors. And the least frequently committed error type  was errors of exam form. According to Table 6, the same pattern is true for all the five genres. Thus, the patterns of committing errors in the five genres would be ( Figure 2):

Discussion
As was already mentioned, the most frequent error type in the translations of the prose text appertained to verb tense category (44.44%). The analysis revealed the fact that there is a tendency for most of the participants to translate the simple present verbs of the text into simple past tense in TL (Persian). The same is true for translations of the play text. The second frequent error type occurred in the play text is verb tense (16.11%). While there is no such a tendency in the translations of verbs in the non-fiction and media texts. The source of such an error type would be negligence, not lack of language competence. The second frequent error type is cohesion. In this case, by increasing the number of verb tense errors, the number of cohesion errors was increased as well for each person. Based on the observations, the logical reason behind this relationship is due to students' constantly shifting between simple present and simple past tenses, so the texts were hard to follow. Thus, the consistency of texts was lost. The third frequent error type appertained to the omission category. The omission error is one of the three highfrequency error types in all five genres. Students most committed omission errors, when they were unable to understand one or more elements of the source text. In these cases, students did not transfer a word, phrase, or a whole sentence into target text (TT). Missing parts were marked one or more errors of omission, depending on how much was omitted. According to analyses, the sources of such an error type were due to students' lack of foreign language competence, lack of knowledge, and carelessness while reading the text. Previously stated, the most frequent error type found in the poetry text is punctuation (19.92%). Most of the students did not observe paragraphing and wording. They failed to follow poetic format. Most of them translated the whole poetry in a single stanza, while the presented poetry contains three stanzas. In some cases, they violated the conventions of the TL regarding punctuation. The findings put emphasis on the importance of first principal of Newmark's (1988) three-step framework for translation of a piece of poetry which draws attention of translators to poetic forms and proposed translators to select a TL poetic form as similar as that of the SL. Terminology and omission errors are the second frequent error types. Terminology errors occurred because of wrong choices of content words or phrases with an incorrect or less appropriate meaning. The omission errors have been already discussed. Aforementioned, the most recurrent error type in the translations of this text type is terminology. The findings are in line with statement of Bertić (2015) about non-fiction texts' being difficult to be understood because of terminologies it contains. In parallel, the findings support Shabanipoor and Moinzadeh (2013) who found translation of specialized texts problematic due to the difficulties in finding appropriate equivalents in the TL for the jargons that exist in the SL. Students' being unfamiliar with psychological jargons and terminologies was the source of such an error type. Omission is the second frequent error type found in the translations of students in the non-fiction text. Punctuation errors are the third most visited error type on this text type. Approximately all of the students received negative points for the already mentioned error type because of incorrect or unclear paragraphing. Three paragraphs were presented in English to the students, though they translate the whole text in a single paragraph. In some cases, problems in the use of commas, semicolons, and quotation marks were observed as well. The sources of such an error type might refer to students' carelessness and lack of knowledge of the conventions of Persian language regarding punctuation. As noted earlier, the most frequent error type found in the translations of the play text is omission. Referring to the previous argument, this type of error is one of the three high-frequency error types in all five genres. Referring to analyses and earlier discussion, as the students had difficulty in understanding elements of the source text (ST), they omitted them. In some cases, carelessness of the translator caused one or more missing words, sentences, or lines. Aforementioned, similar to the translations of the prose text, verb tense errors frequently occurred in the translations of this text type. Most of the students translated simple present tense verbs into simple past tense verbs. Conversely to translations of the prose text, cohesion of texts in play was not that much lost. Because the students who committed verb tense errors were not shifting between simple present and simple past tenses. Most of them either translated all verbs in simple past wrongly or translated all verbs in simple present. The third most visited error type in the translations of the play text is literalness. In view of short sentences that play texts contain in the form of dialogues and descriptive sentences of acts and states of actresses and actors, the students keep the subject pronouns in the TT, though the use of subject pronouns in Persian does not seem natural. The findings put emphasis on Hamberg's (1969) principals for the drama translator "It goes without saying that an easy and natural dialogue is of paramount importance in a dramatic translation" (P. 91-94). The source of such an error type was students' lack of translation practice. Earlier indicated through Table 2, misunderstanding is the most recurrent error type in the translations of this genre. According to ATA frame work "A misunderstanding error occurs when the grader can see the error arises from misreading a word, for example, or misinterpreting the syntax of a sentence" (p. 4). Considering the obvious differences of audiovisual translation (AVT) and written translation, in order to conduct this study in media genre, "misreading" was considered as "mishearing" a word. So mishearing errors were categorized as misunderstanding category. This type of errors arose from students' lack of language competence regarding listening skill and their lack of practice in the realm of AVT translation. The second and third frequent error types which are terminology and omission have been already discussed.
Four error types had not been observed in the translations of all five genres, including: capitalization, diacritical marks/accents, text type, and unfinished. In view of TL in this study which was Persian, thus no capitalization errors occurred because there is no such a system in Persian language system. The same is true for diacritical marks/accents errors. Diacritical marks are not common and practical in Persian language. According to ATA framework, text type and unfinished categories merged with register or style and omission categories, respectively.
It is worth adding that referring to Appendices A-E, some error types, especially errors of exam form, and faithfulness errors are somehow person-based. It means that most of the indecision errors were committed by the participants number twenty (P20), P28, and P10. Likewise, illegibility errors were frequently committed by P9. Faithfulness errors were mostly visited by P10.

Conclusion
The present study aimed to analyze translation errors in different literary genres including: prose, poetry, non-fiction, play, and media. The analyses revealed that the most frequent error types in the prose text were: verb tense (44.44%), cohesion (15.55%), and omission (11.11%); the most recurrent error types in the poetry text were: punctuation (19.22%), terminology (16.16%), and omission (16.16%); mostly visited error types in the non-fiction text were: terminology (33.58%), omission (16.29%), and punctuation (8.64%); omission (22.64%), verb tense (16.71%), and literalness (15.63%) were the most frequent error types committed in the play text; and misunderstanding (23.77%), omission (20.08%), and terminology (13.93%) were the most observed error types in the media text. Repeated measures ANOVA was ran in order to find out whether there is any relationship between the text type and the errors. The obtained result indicated that text type is significant except for non-fiction and play. Based on the analyses the pattern is the same for all genres. The obtained pattern is based on the mean frequency of the three types of errors (strategic, mechanical errors, and errors of exam form) found in the translations.

Funding
The authors received no direct funding for this research.

Limitations of the study
There were a number of limitations in the process of conducting this study. First, due to the limitation of time regarding an extremely time-consuming process of assessing translations, presenting longer sample texts was not possible. Second, in view of the lack of equipment, considering the difficulty of providing a PC for each student, participants were asked to write down their translations of media text. Furthermore, only BA students majoring at translation studies participated in this study. Lastly, due to the large number of text types supposed to present to participants, presenting all the five genres' texts at the same section was not possible. Thus, the five types of texts were presented in five consecutive weeks.

Recommendations for further research
Some suggestions are provided for further research. First, it is recommended to replicate the study with other models of analysis of translation errors. It is also recommended to conduct the same research from Persian to English. Other genres with models of analysis of translation errors would be studied. Last, the same studies can be done with any text type in other languages.