Translation memory (TM)
Translation memory (TM) is a specialized database that supports human translators in the process of translating documents from one language to another. It serves to enhance efficiency and accuracy by storing previously translated phrases and sentences, allowing translators to reuse them in new translations. This tool is particularly beneficial for individuals and organizations that frequently work with similar documents, such as legal texts, government forms, and surveys. Unlike machine translation, which operates independently of human input, TM requires a translator to review and finalize the outputs, thus ensuring the subtleties of language are respected.
The process of translation is complex, often involving the breakdown of text into manageable segments for precise rendering. TM software streamlines this process by automatically identifying and suggesting translations from its database while flagging any partially matching phrases for human assessment. As the database expands, the accuracy and consistency of translations improve, which is crucial in maintaining clarity across different texts. Despite its advantages, some critics raise concerns regarding the software's reliability, particularly if errors exist in the database, as well as the potentially high costs associated with obtaining TM software. This duality makes TM a valuable resource for larger translation operations, while smaller entities may find it financially challenging.
Translation memory (TM)
A translation memory (TM) is a database that helps in the routine translation of documents. Translation is an important, difficult job that requires a translator to be fluent in multiple written languages. When translations must be completed at a rapid pace, translators sometimes make errors. Tools such as translation memory software were designed to aid this process.
TMs are used by businesses, governments, and linguists. They are most useful to people who transcribe similar documents on a regular basis. Translation memory is distinct from machine translation software. While translation memory is a tool utilized by a human translator, machine translation is completely carried out by a computer.
Background
Translating is the process of transcribing text from one language to another. Interpreters, people who verbally translate one language into another, have existed for as long as spoken word. However, the earliest anthropological records of translators have been dated to roughly 1000 years after the earliest records of writing. For this reason, historians believe that translators were rare, highly educated individuals.
The earliest evidence of translators is the existence of ancient clay and stone tablets. Several of these tablets, including the famous Rosetta Stone, contain the same text repeated in multiple written languages. They could be used as reference material for multilingual areas and cultures, or as translation keys for future generations.
During the Middle Ages, translators were often members of the clergy or nobility. The average person did not have the money or the leisure time required to study multiple languages and become fluent in each. Many English scholars and monarchs were fluent in Latin, and spent time translating their favorite works into English. Some of these works exist as different English versions based upon different translations. For example, author Geoffrey Chaucer translated his favorite works from French to English.
One of the most famous translation projects of all time, the Bible, has been translated into more languages than almost any other text. The bible was first translated in the third century, when a community of Jews living in Greece translated the Old Testament into their native language for easier reading. Several hundred years later, as Christianity spread throughout Europe, a Latin translation of the text was needed. Many scholars attempted the complex translation, resulting in many different Latin variations of the Bible, until Pope Damasus I ordered the creation of a single, definitive Latin bible in 382 CE.
Once the Western Roman Empire fell, Christians realized that people in many of the areas once controlled by the Romans spoke their own regional dialects. These included varieties of French, Spanish, and Anglo-Saxon. For this reason, only particularly educated individuals, such as the clergy or the nobility, could read the Latin Bible.
The Church discouraged further translations until the sixteenth century, preferring to keep the power of interpreting Scripture for educated individuals. The reformer Martin Luther, a German professor who disagreed with many of the Catholic Church's attitudes, thought that Christianity would spread more easily if the common man had access to Scripture. He took the task upon himself, studying Greek and translating a Greek Bible into German. Several other scholars followed Luther's example, translating the Bible into English, Spanish, and many other languages.
Overview
Translating a text through traditional methods is a long and difficult process. To begin, the translator must be educated in both the native language of the text, and the language into which the text will be translated. The process of learning a second language can take many years of work.
After learning the languages involved in the translation, the translator may begin to tackle the text. Translations usually are completed in small chunks, translating smaller sentences a few at a time. Because long sentences in a second language can be complicated and difficult to comprehend, most translators prefer to work with five- or ten-word chunks. Splitting the sentence up allows translators to focus on correctly translating individual phrases and clauses.
Once the entire document has been translated in such a manner, the translator goes back to polish the work. They make sure no content was lost, and may alter the document to make sure that it flows well in its new language. Some phrases that fail to literally translate into the new language, such as idioms or plays on words, may be completely changed so that the document makes sense to its new readers.
Translation memory software is designed to make this complicated process quicker and easier for translators. Unlike translation machines, translation memory is designed to work with a human translator. It cannot translate an entire document on its own. Instead, it organizes a database of commonly translated phrases. It then scans the document, replacing any chunks that perfectly match those in its database. Additionally, most translation memory software will attempt to translate phrases that partially match those in its database. However, it will flag these phrases for review by the human translator.
The end result of using translation memory software is a partially translated document ready for review by a professional translator. While the translator must still do a substantial amount of work to complete the document, the workload should be lessened. As the software's database grows, its translations will become more accurate, further reducing the work of the translator. For this reason, translation memory software works best in situations in which similar documents need to be routinely translated. These may include government forms, legal documents, surveys, and many other situations in which repetitive text must be repeatedly translated.
Because the translator is pulling common phrases from the database instead of their mind, translation memory makes translators more consistent. Important terminology will always be translated identically, reducing confusion for the reader. Many translation memories also reformat the document for the translator, keeping the new document's appearance as close as possible to the older document.
Critics of translation memory software question its effectiveness in day-to-day use. They argue that if the software's language database contains an error or if it learns a phrase incorrectly, it will translate sections incorrectly while flagging them as completely correct. Additionally, the cost of translation memory software tends to be prohibitively expensive. Businesses that routinely require the translation of similar documents, such as international law firms, will save money from utilizing translation memory software in the long run. Smaller services, however, may not be able to afford the initial license.
Bibliography
"The 5 Step Language Translation Process." Pacific International Translations, 2014, www.pactranz.com/blog/language-translation-process/. Accessed 7 Nov. 2024.
"History of Bible Translations." HistoryWorld, 2017, www.historyworld.net/wrldhis/PlainTextHistories.asp?historyid=ac66. Accessed 7 Nov. 2024.
"Machine Translation vs. Translation Memory." QuickSilver Translate, 2015, www.quicksilvertranslate.com/2798/machine-translation-vs-translation-memory. Accessed 7 Nov. 2024.
Park, Jiyoung. "Ethical Approach to Translation Memory Reuse: Discussions from Copyright and Business Ethics Perspectives." Translation Studies, vol. 17, no. 1, pp. 37-52, 2023, doi.org/10.1080/14781700.2023.2183248. Accessed 15 Nov. 2024.
"Translation Memory: Potential Pitfalls." QuicksSilver Translate, 2011, www.quicksilvertranslate.com/2219/translation-memory-potential-pitfalls. Accessed 7 Nov. 2024.
"What's the Difference Between Translation Memory (TM) and A Glossary?" Language Scientific, 2017, www.languagescientific.com/whats-the-difference-between-translation-memory-tm-and-a-glossary/. Accessed 7 Nov. 2024.