TYPES OF TRANSLATION ERRORS IN THE TRANSLATION FEATURE OF LOCAL GUIDE REVIEWS ON GOOGLE MAP

: Translation errors can occur due to the delivery of a translated source language that the target language cannot accept. Moreover, the need for more attention to new terms adopted by a language means that translation errors often occur. Therefore, this study aims to describe and correct errors in the translation results in the see translation feature in local guide reviews on Google Maps. Sources of data in this study were taken from the Google Maps application and the results of interviews. Researchers used a qualitative descriptive approach with thematic analysis techniques to describe the types of errors. Documentation and interview techniques were used for data collection. From the research results, 194 errors were found in the results of the translation. Based on orthography taxonomy, 19 types of errors were found, with a percentage of 9.79%. Then, in the lexis taxonomy, 36 data were found with a percentage of 18.54%. In the grammar taxonomy, 23 data were found with a percentage of 11.85%. In the semantic taxonomy, 91 data were found with a percentage of 46.88%. And the last one, namely discourse taxonomy, found 25 data with a percentage of 12.88%. From the research findings, the most common type of error found is in the semantic taxonomy. This research implies that by knowing these errors, the creator of the translation application can produce a better translation and know the techniques for improving the translation results in accordance with the context of reviews made by local guides on the Google Maps application, which in general discussing the service, atmosphere and conditions of a place


INTRODUCTION
One of the linguistic studies that has a very important role at this time is translation study.Translation study is a field of study related to translation theory, description, and application.Koman (2019) argued that translation is necessary because it is an important task in the process of exchanging information.He also stated that translation is a challenging task because it requires good mastery of both the source language and the target language.Translation, according to (House, 2014) can be interpreted as the result of linguistic-textual surgery in which a text in one language is recontextualized into another language.
Meanwhile, Machali (2009) believes that translation is an effort to replace the source language text with equivalent text in the target language.What is translated is the meaning as intended by the author.Translation has an important role in conveying knowledge, culture and language.This is because translation is an activity to convey messages or intentions contained in one language into another language accurately.
Nowadays, the development of technology is increasingly sophisticated, and it has an automatic translation feature that is used in various applications such as social media, online transportation, ecommerce, and so on.However, the translation results of these various features certainly have translation errors that need to be in accordance with the intended meaning.Translation errors occur due to the formation of incorrect words, phrases or sentences in the target language, which is caused by insufficient knowledge of the rules contained in the target language.Errors in the translation are not only due to a lack of knowledge about the target language but also a lack of knowledge in both the source and target languages , such as structure, vocabulary and punctuation (Yousofi, 2014).
In this case, this study used the Google Maps application as the object platform for this study.Google Maps provides a review feature to express opinions or comments on a place that has been visited.Both in positive or negative comments, so that someone who wants to visit a place can consider it first for their visit.In the reviews made, various kinds of comments or opinions were conveyed by the public.Some of them used Indonesian language, and some used English.Google Maps provides a translation feature that can help someone understand the intent of the reviews that other people have written.
According to the people who have used the translation feature, they said that the feature is quite helpful even though the translation from English to Indonesian still needs to be improved and has several errors in the translation of the reviews.One example of a problem or error found in the translation results from a machine translation, namely regarding the translation of idioms or language styles.This problem is generally found in the translation of novels that took more work to translate.The problem is the translation of these idioms, or figurative languages, requires a level of equivalence that must be accurate with a level of reasonableness that must be acceptable in the target language and the cultural context of the user community and a fairly high level of readability by readers of the target text.However, the use of idioms is also sometimes used in reviews given by a local guide on Google Maps, so that in the application's translation feature, the translation machine has not found a meaning that is commensurate with the intended target language.
Therefore, considering how important translation is as a medium of information between worlds, the field of translation has finally begun to develop and enrich theory so that it can solve various problems in translation.In addition to theory, translation studies are also experiencing applicative development, namely, with the presence of machine translation.However, many aspects must be considered in translating.One of them is to look for equivalent words from the source language to the target language, as revealed by (Halimah, 2018) in her article.This research is intended to reveal the types of translation errors found in the see translation feature in the Google Maps application, especially from English to Indonesian.Apart from that, this research is also useful in adding information regarding translation errors in the Google Maps application from English as the source language to Indonesian as the target language, as well as increasing knowledge regarding the level of accuracy of the see translation feature in the Google Maps application in translating a written review.

Machine Translation
Machine translation (MT) is a tool used in the process of translating from one language to another language automatically.According to (Irfan, 2017), machine translation is a machine that works automatically.The translation machine is the result of artificial intelligence (AI).Machine translation is a computer program designed to translate a text from one source language into another with human assistance.According to (Hutchins, 1982), machine translation is an attempt to automate all or some of the translation activities.
Meanwhile, according to (Levesque, 1986) , machine translation is a computer application that can translate text from one natural language to another.According to him, a translation machine should be seen as a tool that can help translators in terms of costs and expenses.In translation studies, machine translation is classified into applied translation studies, as described by (Venuti, 2021)in his paper entitled "The name and nature of translation studies" (Munday, 2001) as follows:

Translation Error
Error, according to Hansen in the book of (Gambier & Van Doorslaer, 2010), is that there is something wrong.In written texts, both original texts and translated texts, errors can be classified as pragmatic, semantic, idiomatic errors, linguistic errors or stylistic errors.So, translation errors arise from the relationship between the two texts.A misunderstanding of the TL translation can cause translation errors because it cannot find a more accurate or equivalent TL meaning.In theory, based on the concept of equivalence between TL and SL, translation errors are considered a result of a mismatch of meaning between SL and TL.According to (Fitria, 2020), errors can also caused by a lack of language skills, such as a lack of knowledge about the correct rules of the TL or errors in translating when using certain translation tools.
Translation errors can be categorized into two categories, namely scientific errors, which consist of referential errors and linguistic errors, according to (Munday, 2001).Referential errors are errors related to facts.Meanwhile, linguistic errors are errors that show the translator's ignorance of the language being translated, which includes grammatical errors.In analyzing grammatical errors, four steps must be carried out by researchers, namely identifying the problems found, representing or describing the grammatical errors found, and identifying and evaluating language errors made (Ellis, 2003).Classification of errors for machine translation includes orthography, lexis, grammar, , and discourse (Costa et al., 2015).As shown in the following diagram:  et al., 2015).

Orthography
Errors at the orthographic level include all errors related to misuse of punctuation and spelling errors.Orthographic errors are divided into three types, namely punctuation, capitalization and spelling errors.They are using punctuation which is not suitable, which means having punctuation errors.The second error is in capitalization, which refers to the use of capital letters-for example, using the first letter at the beginning of a sentence.Spelling errors involve replacing, adding, or deleting one or more letters to the word orthography.

Lexis
Lexis errors contain the following types of errors: omission, addition, and untranslation.Translation errors in omissions and additions that are analyzed are regarding the types of words, such as content words, namely words that carry sentence meanings, such as nouns and adjectives.Then, function words are words that have a function to express grammatical relationships with other words in a sentence, for example, prepositions and pronouns.

Grammar
Grammar errors come in two types such as misselection and misordering.Misselection errors appear at the word class level; if, for example, the source text is an ad or adjective, then the target text changes to a noun or verb (tense and person), agreement errors (gender, number, person and in contractions between prepositions and articles.Misordering errors are errors related to syntactic problems that the sentence may indicate.Misordering errors are errors in placing words in an incorrect order (Ellis, 2003).A good translation is not only about using the forms in the right context but also about arranging them in the right order.

Semantic
Semantics is the study of the meaning of words, phrases and sentences (Kreidler, 2013;Wongranu, 2017).(Faqih, 2018) added that in transferring the message of a text from SL to TL, translators must also pay attention to semantic aspects and equivalence in SL and TL.In this case, semantic errors are classified into confusion of sense, wrong choice, collocational errors, and idioms.Sense confusion occurs when a word in the source text is translated into something that represents one of its possible meanings in the target text, but that needs to be corrected.Then, the wrong choice means that the source word is translated into the wrong target word without a clear relationship.Collocational errors mean wrong combinations of natural words.Idiom errors are related to errors in idiomatic expressions made by translators who do not know and translate the source text as plain text.Keraf (1985) stated that idioms are structural patterns that deviate from general language conventions, usually in the form of phrases, while their meanings cannot be explained logically or grammatically, relying on the meanings of the words that make them up.These expressions cannot be translated literally because their meaning is not literal, and in some cases, the equivalent expressions in the target language are very different.

Discourse
Errors at the discourse level are classified into style variations and should not be translated.Stylistic errors are a poor choice of word style when translating sentences, such as repetition of words in a close context where using synonyms from the translation is a better choice.Variation errors occur when the translation results do not represent the target language.Then, it should not be translated, and it means that the source text does not use any translation in the target text, for example, the title of a book, film, places, and many others.

See Translation Feature
The translation feature, or what is usually called auto-translation, is a feature found in various applications, especially Google Maps.This feature gives users access to see translations in several languages, one of them from English to Indonesian and vice versa.The results of the see translation feature match the intended intent and meaning from the source language (SL) to the target language (TL), according to Supriyadi (2016).However, auto-translation also sometimes has some errors in translation where the meaning could be more accurate and acceptable.This is because the autotranslation feature still uses the word-forword translation method so that the meaning does not match what is meant, according to (Susanti & Ekasani, 2021).The use of the see translation feature on the Google Maps application is very easy, and people only need to press the see translation button on the local guide's review that uses English.The translation of the review appears immediately.

RESEACH METHOD
This study used descriptive qualitative research in the form of thematic analysis.Thematic analysis is one way to analyze data with the aim of identifying patterns or finding themes through data that has been collected by researchers (Tay et al., 2021).The approach used in this research is the applied linguistics approach, specifically in the field of translation error analysis.The data sources used in this paper were taken from the Google Maps application and limited to 150 reviews from local guides regarding tourist attractions and restaurants in Sumatra Utara.Furthermore, the data collection method used in this research is documentation.
The data analysis technique used in this research is thematic analysis.According to (Boyatzis, 1998), thematic analysis is a method for identifying, analyzing and reporting patterns contained in data and being able to interpret various aspects of research topics.

RESULT AND DISCUSSION
This study found several translation errors in the translation results of the see translation or auto-translation features, which were reviewed based on the types of translation errors according to (Costa et al., 2015).There are 150 reviews of the translation results in the see translation feature with 3145 words and 323 sentences.The number of errors from all categories is 194.The following is an explanation of the results of the research that has been done.

Translation Errors in Orthography Taxonomy
The findings of this research prove that the translation results of the see translation feature have errors in the grammatical structure of the Indonesian language orthographic aspect.The errors found in a table: The data above is quoted from a local guide review written by Dennis Mitchel which reviews Gundaling Farmstead.Upon further review, there were punctuation errors in the review.In this case, the SL does not place a question mark (?) in the written review, but when it is translated into Indonesian, the translation feature system raises a question mark which should not need to be used, because the review itself does not contain a question sentence.Therefore, the results of this translation are classified as punctuation errors.

Capitalization Error Data Source
The place is very nice natural atmosphere is beautiful and cool, we have to maintain the nature of Indonesia.
Translation Result on the Translation Feature Tempatnya sangat bagus suasana alamnya asri dan sejuk, kita harus menjaga alam indonesia The data above is quoted from a local guide review written by one of the local guides who reviewed Mount Sibayak.After reviewing, there was an error in the use of capital letters in the review.In this case, SL is correct and appropriate in placing capital letters where they should be.However, when translated into Indonesian, the words that should start with a capital letter are incorrect.Therefore, the results of this translation are classified as errors in the using of capital letters.

Data Source
Nice place, but today is so crowded.. my fav desert is always unavailable Translation Result on the Translation Feature Tempat yang bagus, tapi hari ini sangat ramai.. gurun favorit saya selalu tidak tersedia This data is quoted from a local guide review written by a person who does not use name in the local guide account, he reviews Gundaling Farmstead.In this case, the word "dessert" in SL is written "desert".Therefore, if there is only one letter s in the word, then the meaning will be "gurun".Therefore, it can be said that if you make a mistake in writing the SL, the TL you are referring to will also be wrong.This is in accordance with the cause and effect of errors in spelling that occur due to the deletion of one letter which causes the translation results to be inappropriate.

Translation Errors in Lexis Taxonomy
In this study, the researcher found translation errors in the see translation feature in terms of lexis taxonomy.The errors found in the lexis taxonomy are divided into three categories, namely; 1) omission; 2) addition; and 3) untranslated.Errors at the omission and addition levels are analyzed regarding word types such as content words and function words with the percentage of errors shown in the following table: Based on the table above, it can be concluded that error in the lexis taxonomy in this study was 18,54%.The omission error in the content word error with a percentage of 1,03%, while the error in the function word is not found.Then the addition function word error is with a percentage of 0,51% while the content word error is also 0,51%.The highest error in this lexis taxonomy is the untranslated error with a percentage of 16.49%.The three categories of translation errors at the lexis taxonomy level are explained as follows:

Sumber Data
Good place to spend time with friend with outdoor concept.A little bit empty space at back.this place u order foods or drinks and paid by yourself at the counter.If rain then you cannot dine in This data is quoted from a local guide review written by Hardie Chen which reviews Dominico Coffee.Based on the table above, there are translation errors in the lexis taxonomy, especially at the level of omission errors, where the translation of the word "dine in" is only translated as "makan di".Therefore, this translation error is included in the translation error of omission because the translation is not complete.

Data Source
Very packed during weekends too.Needs more people on the floor maintaining cleanlines.One of the most go to hotspring in doulu.

Translation Result on the Translation Feature
Sangat dikemas selama hari kerja juga.Membutuhkan lebih banyak orang di lantai untuk menjaga kebersihan.Salah satu yang paling banyak dikunjungi di pemandian air panas di doulu.This data is quoted from a local guide review written by Gerard Sembiring which reviews Pariban Hot Springs.Based on the table above, there are translation errors in the lexis taxonomy, especially at the level of addition errors, where the SL itself does not say "in", but in the TL there is a translation of "in".Therefore, this translation error is included in the additional translation error, because the translation contains additional translations that are not in the SL.

Data Source
Menu is not special.Place is okay and warm.Building is Woody.Many tress.But lot of ants.
This data was taken from local guide Selvester Cia's review of the Rumah Pohon cafe.In this translation, the source language "woody" is not translated into the target language.Therefore, this translation error is included in the untranslated translation error, because the word "woody" in the translation is not translated.

Translation Errors in Grammar Taxonomy
The findings of this study prove that there are errors in the results of the translation of the see translation feature in terms of grammar taxonomy.The errors found in the grammar taxonomy are divided into two categories, namely misselection and misordering with the percentage of errors shown in the following table: Based on the table above, it can be concluded that the translation error in the grammar taxonomy in this study was 11,85%.The misselection error in the word class type is with a percentage of 3,09% and in the tense type is 0,51%.Meanwhile, misordering errors are at a percentage of 8,25%.In this case, misordering errors are more common than misselection errors.The two categories of translation errors at the grammar taxonomy level are explained as follows:

Data Source
It was good to learn about different culture.Highly recommended place to visit!!

Translation Result on the Translation Feature
Bagus untuk belajar tentang budaya yang berbeda.Sangat dianjurkan tempat untuk kunjungan!!This data is quoted from a local guide review written by I-Anne which reviews Lake Toba.In this translation, the word "visit" is included in a type of error at the word class level where the translation of the word "visit" is a noun and the expected translation should be "visited" which is a verb.So this error is included in the word class error.

Data Source Cheap and fast service Translation Result on the Translation Feature Pelayanan murah dan cepat
The data above is quoted from a local guide review written by Hendra Darma which reviews Agem Senyum, a restaurant located in the city of Medan.The translation carried out by the translation feature is correct, but the word is not quite right.So this is included in the misordering error.

Translation Errors in Semantic Taxonomy
Researchers prove that there are translation errors in the see translation feature in terms of semantic taxonomy.The errors in the semantic taxonomy that have been found are divided into four categories, namely; 1) confusion of senses; 2) wrong choice; 3) collocational errors; and 4) idioms with the percentage of errors shown in the following Based on the table above, it can be concluded that the translation error in the semantic taxonomy in this study was 46,88%.In this taxonomy, the confusion of senses error is 1,54%.Then the highest error lies in the wrong choice error with a percentage of 37,62%.Furthermore, collocational errors are 4,12%.The last error is an idiom error with an error percentage of 3,60%.The four categories of translation errors in the semantic taxonomy are explained as follows: This data is quoted from a local guide review written by Djunaidy Tehnika which reviews Agem Senyum.Based on this translation, the source language of "hot & spices local" has errors in the confusion of senses error type.In this sentence, one of the meanings is not clear, resulting in a confusion of senses error.

Sumber Data
Great place to hangout and eat.Good vibes, similar vibes like ubud bali.

Translation Result on the Translation Feature
Tempat yang bagus untuk hangout dan makan.Getaran yang bagus, getaran serupa seperti ubud bali.This data is quoted from a local guide review written by Nevin Lojie which reviews Morning Glory Coffee.In this translation, there is a wrong choice error in the words "Good vibes, similar vibes", where the translation results do not have a clear relationship with the SL or can be said to have no continuity with the meaning of the SL itself.

Data Source
Fresh original cow milk industry.You can also find a restaurant here.Some friend snack also here.
This data was written by Robin Huang who reviewed Gundaling Farmstead.In this case, the meaning of "friend's snacks" is not appropriate to describe the purpose of the review.This is because "friend's snacks" will mean that the snacks belong to one of your friends.Therefore, this type of error is included in the type of collocation error, which means that the natural combination of words in the translation is incorrect.

Sumber Data
Best place for relax and quality time

Translation Result on the Translation Feature Tempat terbaik untuk bersantai dan waktu berkualitas
The data above is quoted from a local guide review written by Dion Silalahi which reviews Denatural Resto.Based on the table above, the translation of the phrase "quality time" is not quite correct if it is translated as "waktu berkualitas".This is because the translation of the phrase "quality time" involves errors in idiomatic expressions made by translation machines that translate a term literally.The word "quality time" is often used by people to express activities that involve spending time together in one place.

Translation Errors in Discourse Taxonomy
The findings of this research prove that there are errors in the translation results of the see translation feature from a discourse perspective.The errors found in the discourse taxonomy are divided into three categories, namely; 1) style; 2) variety; and 3) should not be translated with the error percentage shown in the following table: Based on the table above, it can be concluded that the translation errors in the discourse taxonomy in this study were 12,88% with a stylistic error of 7,73%.However, in this taxonomy no variation errors were found.Furthermore, the should not be translated error is with a percentage of 5,15%.In this discourse taxonomy, more errors are found at the style level than should not be translated errors.The three categories of translation errors at the discourse taxonomy level are explained as follows:

Data Source
Outdoor, nature bath.A nice place to get fresh air, unwind and to chill should not need to be translated into Indonesian because the words "53 steps" are the name of a place.
After describing the research results on the types of translation errors in the see translation feature in the local guide review on Google Maps, the following is a table of the overall data on the types of errors obtained along with the error categories.Based on the table above, it was found that there were 194 types of errors found in the translation results of local guides' reviews using the see translation feature.In orthography taxonomy, there were 19 data (9,79%), lexis taxonomy 36 data (18,54%), grammar with 23 data (11,85%), semantics amounting to 91 data (46,88%), and the last one is discourse with 25 data (12,88%).In this case, the most common errors found were errors in semantic taxonomy, namely 91, with a percentage of 46,88% of the 194 errors found.

CONCLUSIONS
From the results of the data analysis that has been carried out, the type of grammatical errors that are most often found are semantic errors, the type of wrong choice errors.Wrong choice errors are characterized by the presence of various types of word choices that are incorrect or do not match the context of the source language.The choice of inappropriate words is caused by the translation machine not being able to determine the context of the review, so that it only translates a word in general terms.However, in this study, several types of errors were not found, such as errors in grammar taxonomy in the misselection section.Only errors were found at the word class and tense levels.Then, discourse errors at the variety level were also not found in this research.
This research implies that by knowing these errors, the translation machine can produce better translations and know techniques for improving translation results that are appropriate to the context of reviews made by local guides on the Google Maps application, which generally discuss the service, atmosphere and conditions of a place.Based on the research findings, several practical implications can be identified, including that the results of this research can help develop an understanding of the phenomenon under study, both for application developers, application users and machine translation application researchers.Then, provide guidance and recommendations to application makers in developing corrections for errors made by machine translation.This research still has limitations in terms of the number of objects studied and the methods used.

Figure 2 .
Figure 2. The Taxonomy of Errors (Costaet al., 2015).OrthographyErrors at the orthographic level include all errors related to misuse of punctuation and spelling errors.Orthographic errors are divided into three types, namely punctuation, capitalization and spelling errors.They are using punctuation which is not suitable, which means having punctuation errors.The second error is in capitalization, which refers to the use of capital letters-for example, using the first letter at the beginning of a sentence.Spelling errors involve replacing, adding, or deleting one or more letters to the word orthography.LexisLexis errors contain the following types of errors: omission, addition, and untranslation.Translation errors in omissions and additions that are analyzed are regarding the types of words, such as content words, namely words that carry sentence meanings, such as nouns and adjectives.Then, function words are words that have a function to express grammatical relationships with other words in a sentence, for example, prepositions and pronouns.
Error Data Source Well known for the instant noodle cooked, added hot & spices local for seasoning Translation Result on the Translation Feature Terkenal dengan mie instan yang dimasak, ditambahkan panas & bumbu lokal untuk bumbu

table :
Types of Translation Errors in The Translation Feature of Local Guide Reviews on Google Map Page 1375Asian Journal of Engineering, Social and Health Volume 2, No. 11 November 2023