Translating
from English Unit 6: Testing the Translation
Lecture Notes
B. Aron
The next step after translating the text is to test the translation.
There are many theories of translation so there are many approaches to testing translations.
Larson’s approach asks 4 main questions:
It is important to Larson to test a translation: to know whether a translation that is faithful to the source text message and that it is natural sounding in the target language.
Why Test a Translation?
According to Larson, the three elements to check when testing a translation are:
Nida and Tabor state that implicit information should be made explicit in some cases depending on the nature of the culture and language associated with the target text.
What it all boils down, “the translation must convey the intended message without extraneous information.”
Ambiguity is when a phrase or sentence could have more than one meaning in a specific context so the target audience is not sure of the intended meaning.
Who Tests the Translation and When?
The translator should check the translation section by section as its developed. For example, it is best to check each paragraph as you create it rather than wait until you have completed the entire translation, especially if it’s long.
Checking sections as you go may reveal problems you can correct early in the translation process.
Then check the translation in its entirety again after its completed.
It is ideal for a second person to review the translation. It is suggested that the person be unfamiliar with the source language, so the person can see if the target language is natural and clear but not for accuracy.
In order to check for accuracy, the tester must be familiar with both languages. It is best to have a tester who can test for accuracy.
Sometimes the translator must be responsible for accuracy and sometimes a second translator can check for accuracy.
How do you Test the Translation?
Larson suggests five ways to test a translation. There is some overlap between these approaches.
Keep in mind that these tests examine the product of the translation, not the process used to create the translation.
1. Comparison with the Source Language: This test by comparing the translation with the source language. Comparison should be done periodically during the translation process. The purpose of comparison is to check the equivalence of the content between source and target texts. Comparison can be done by the translator or a second translator.
Nida and Tabor says that of all possible criteria you can use to test the translation, dynamic equivalence is the most important. This means that the content of the source language message is well preserved and conveyed and has approximately the same impact as the target language.
2. Back Translation: The second way to test a translation is by using back translation. Someone other than the translator should translate the text back into the source language by a person unfamiliar with the source message.
This type of test focuses on the product.
The back translation is usually done in a literal rather than idiomatic fashion.
Back translation shows whether the translation has faithfully preserved the content of the source text.
3. Comprehension: This is a test of the product and is not done by the translator but by the person who is a speaker of the target language. The purpose is if speakers of the target language understand the translation correctly. This test helps the translator determine whether the translation accurately communicates the message to the intended audience. The tester retells the content of the translation and answers questions about the content. The drawback of this approach is that the retelling depends on the tester’s level of comprehension and ability to retell accurately.
4. Naturalness: One of the reviewers who read the translation should perform the naturalness test. A test of naturalness asks if the translation sounds as if it were a text originally created in the target language. Does it use the constructs of the target language? It is important to have reviewers with strong competence in the source language as well. Test of naturalness does not include testing equivalence which means it does not include accuracy and clarity.
5. Readability: Readability testing applies only to languages that are written. This type of test involves having the translation read out loud by another person. If the reader stumbles or looks puzzled at a specific area of the translation, then the translator should re-examine that portion of the translation for appropriate language use.
6. Overall Length: Nida and Tabor stress that it is usual for a written translation to be slightly longer than the original written text. It is due to the need to state everything that was included in the original and to make explicit some things that may be implicit in the original text. We do not have enough data to know if this generalization applies to signed language translations. Other key factors to watch for in the translation are that the overall organization of the target text retains the same main points and details in the same order whenever possible. The impact of the total message should not be distorted by translation. The translator should not explicitly interject his or her own opinions into the translation; however, translators do occasionally add notes to explain or clarify a point.
How Do You Use the Results of Testing?
Gile emphasizes that there are some pitfalls in evaluating a translation solely on the basis of the product. One problem with focusing on the product is that the student may not understand the real reason for the error, but think they do or they know that the translation is faulty but not know why or how to correct it.
A combined approach of examining both product and process in testing the translation is most effective. You should look at the product of your translation and also work to examine the processes you used.
The main value in looking for the errors in methods is that making changes at the method level can improve product and provide a deeper understanding of translation processes.
Awareness of where in the process errors can occur can lead to improved translations in a systematic fashion.
For grammatical errors: did not perform acceptability test properly
For an illogical translation error, did not perform hypothesis testing completely during the comprehension phase.
You must comprehend the message before you can translate it. You must determine how the major points of the text relate to each other before translating.
If you recognize ways to improve your methods will help you self-analyze your own work.
Taxonomy of Errors:
Not Serious: an error that does not skew the message; usually reformulation
Somewhat Serious: an error that does not skew the message; omission of detail or production error
Quite Serious: total skew of the message due to following English syntax instead of target language syntax
Very Serious: total skew of the message due to comprehension problems
Determining the Seriousness of Errors:
If an error occurs early in the translation process, at the analysis stage, it is likely that there will be more serious errors in the remaining stages of the process.
If you do not correctly understand the message you will not be able to transfer or reformulate it.
If you can understand the message but can not transfer it, you will have difficulty reformulating it.
By knowing exactly where the error is in the translation process, it becomes clear that if the error occurs near the beginning, the more serious the effect of the error.
If the error occurs in the comprehension stage, the error would be very serious.
If the error occurs at later stages in the translation process, it would be less serious.
Discussion Questions for Unit 6:
1. Why is it important to test a translation?
2. Who should test the translation?
3. What steps would you use to test a translation?
4. How can you determine the accuracy of a translation?
5. What does “naturalness” mean with regard to translation?