Understanding context in segment tasks

What is a segment task?

A segment task is a task composed entirely of translation memories (TMs): paired source-target segments that can be reused in multiple translations. In a segment task, segments do not form a single document.  They are independent of each other and are not meant to form a narrative. 

 

When you’re working on segments, you’re doing the same as you normally would do with paid tasks: fixing issues. However, it’s really important that you do as much as you can to ensure that each individual segment will work in whatever context you find it in. This will allow it to be used for different emails.

While working on segments, there are a few important aspects you need to consider: 

• As with other tasks, the glossary and register requirements must be respected. 
• Context issues only relate to the target text that you produce after you’ve finished working on it. 
• If a translation doesn’t contain any errors and would work in any context, do not check the ‘needs context’ box. If, on the other hand, the translation doesn’t contain errors on its own, but placing it in context could cause some errors, you do need to check the ‘needs context’ box.

 

What do you mean by context?

One way of translating an email or any other document is to split it into sentences, translate each sentence separately, and then put the sentence translations all together. This approach, however, is likely to produce a final document that is inconsistent and may well not make sense.

 

By context we mean all the information that precedes and follows a given segment/sentence and is necessary for the text to be understood.. 

 

The most common context-related issues that you will encounter are due to gender. English is an extreme case in that it has no grammatical gender. However, languages like Czech or Russian have three (feminine, masculine and neutral), whereas Spanish or Danish have two. In languages with gender markers, associated words must show gender agreement.

 

The following are examples of translations that, while not having any error per se, could cause  problems in context because of gender markers:

 

#1

EN: Thank you for contacting us. 

PT:Obrigado, por entrar em contacto connosco.

 

#2

EN: Hello Lucy, I’ll be happy to help you.

FR: Bonjour Lucy, je suis heureux de vous aider.

 

#3

EN: I will create it for you

ES: Se lo crearé. 

 

In the above translations, the words in bold have masculine gender markers. Examples 1 and 2 are  valid if the sender of the message is a male, but they would not work if the sender was a woman. Example 3 is valid only if the thing being created has a masculine gender.

 

All the above translations are context-dependent. Before marking the translations as context-dependent, try the strategy of gender neutralization, which consists of rephrasing the translation so no gender markers are needed. For example:

 

#1

EN: Thank you for contacting us.

PT: Agradeço por entrar em contacto connosco. 

 

#2

EN: Hello Lucy, I’ll be happy to help you.

FR: Bonjour Lucy, c’est avec plaisir que je vais vous aider. 

In the context-independent translations above, neutral forms were used. This way, there are no gender-related constraints and the sentences can work in all contexts, regardless of the gender of the sender.

 

However, this doesn’t work in all examples. In example 3 below, the gender of the Spanish translation could not be fully neutralized (see word in bold). Since it’s not possible to do it in a natural way, you need to submit a correct translation that could create issues in context, therefore you need to mark the ‘needs context’ checkbox: 

 

#3

EN: I will create it for you

ES: Se lo crearé.  [needs context]

The distinction between natural and unnatural translations is really up to your expertise as a translator. It’s okay to switch tacks and change the segment a little a bit, but don’t go overboard or create a translation that doesn’t match the source that well in the pursuit of creating a segment that is not context dependent. If, for example, we take another look at the French example above:

 

#4

EN: Hello Lucy, I’ll be happy to help you.

FR (i): Bonjour Lucy, c’est avec plaisir que je vais vous aider. 

FR (ii): Bonjour Lucy, c’est avec un grand plaisir que je peux vous aider.

 

Here, the second translation is roughly the same thing, but it’s way over the top and distorts the sentence more than necessary.

 

Context dependent vs. Context independent 

In this section, you will find examples of translations with issues that can be identified and adjusted based on the information you already have and are therefore not context-dependent. All you have to do in these cases is correct them, as normal.

 

• Segments with agreement problems within the sentence:

 

EN: My name is Antonio and I'll be happy to assist you today.

ES: Me llamo Antonio y estaré encantada de ayudarle.

 

EN: We’re sorry to let you know that the hotel is unable to fulfill your request.

PT: Lamentamos informar que ahotel não consegue satisfazer o seu pedido.

 

In the ES example above, the male name Antonio does not agree with encantada, and in the PT example the feminine determiner ‘a’ does not agree with hotel. However, all the necessary information to solve the agreement problem is contained within the sentence, so it’s not context-dependent. The only thing that you need to do is correct the issue(s) in the translation.

 

• Segments with the wrong register

If you encounter a translation where the register is not compliant with the client instructions, it should be corrected without marking it as context-dependent. This is because this is not a context issue, it is a problem related to a text not following the client’s register instructions (which you can find on the top right corner of the editing interface). In the PT example below, you only need to correct the issue in the translation, and in the Italian example, due to the unavoidable gender marker, you are also expected to mark the corrected translation as “needs context”.

 

EN: Would you please send us the following documents? [formal register]

PT: Poderias enviar-nos os seguintes documentos?

Expected correction: Poderia enviar-nos os seguintes documentos? 

 

EN: When we phoned you, you didn’t reply. [formal register]

IT: Quando ti abbiamo telefonato, non siamo riusciti a trovarti.

Expected correction: Quando le abbiamo telefonato, speravamo di trovarla. [needs context]

 

• Segments with incorrect glossary translations 

As in the previous example, if a glossary term does not match the client instructions, it should not be marked as context-dependent,  but should always be corrected. 

 

EN: Thank you for contacting our customer support. [expected translation: Atención al cliente]

ES: Gracias por ponerse en contacto con nuestro servicio de Soporte al Cliente.

Expected correction: Gracias por ponerse en contacto con nuestro servicio de Atención al Cliente.

More examples

Here you will find some more examples as well as brief comments on what is context and what is not context, and also proposals of how to turn them into gender neutral segments.

 

• EN-ES: