YT Academy Class:

Obtaining an accurate meaning

Accuracy may be called the most important part of translation ‒it is essential as it conveys the intended meaning from the source text. However, some of these errors might be so subtle that do not stand out when reading a text, so they need to be spotted by carefully comparing the translation to the source text in another language. To put it in figures, accuracy errors represent 22.4% of our total made errors due to multiple reasons. The good news is that we have multiple solutions available to prevent us from making accuracy mistakes.

'' Translators have the moral and personal obligation to produce an accurate translation of every written text they work on. Quality control is important to ensure that there is consistency in the delivery of accurate translations. ''

~ Bernadine Racoma

In collaboration with: Ignachio Urrutia

Ignacio Urrutia holds a BA and a MA in Translation and Interpreting Studies. Starting his career as a game tester in the UK, he then worked as an in-house project management and translator for several firms until he finally started his own business specialising in technical, legal, IT and marketing contents. For over a decade, he has been working full time in the translation/interpreting industry, while cooperating with several universities in their final year internship programmes through his agency based in Spain.

The importance of accuracy

In order to assess accuracy, we need to understand the purpose of the translation as the source text intends to produce an effect in its audience, hence serves a purpose. The translated text must keep the original meaning while serving the same purpose ‒conveying the same meaning in an analogous way. ​

The problem may be obvious when we spot a clear mistranslation or when an idea or fact is missing in the translation. However, there is a more subtle range of translation errors that ranges from choosing a translation that does not have the same semantic scope as the source to other solutions that produce ambiguity in the meaning or even add information that was not present in the source text. ​

As stated before, accuracy is paramount in order to produce a faithful translation but isn’t easy to always achieve. Moreover, as the range of possible accuracy errors range from subtle to obvious, a high level of attention to detail is needed in order to spot them all. ​
However, there are a lot of things we can do to prevent loss of accuracy. The most important part is to understand the meaning so we’re able to convey it in the right way. If you cannot understand the source text, the readers of the target text will end up reading something else than what the author intended to say.

Our classification

Following our proposition, we may classify accuracy errors as follows:
• Addition
• Mistranslation
• Omission
• Untranslated


EN: Santander Bank will issue 15,000 shares by the end of Q3.
ES: El Banco Santander emitirá unas 15.000 acciones a finales del T3.
FR: Banco Santander se lance dans une emission de quelques 15 000 actions pour la fin du 3ème trimestre.
PT: O Banco Santander anuncia a emissão de umas 15.000 ações no fim do 3º trimestre.
As we can see here in bold, each example of the translations in Spanish, Portuguese and French add an item before the figure, introducing a nuance of meaning that was not in the source. This may lead to a slightly different meaning or understanding for the reader.


EN: Santander Bank will issue 15,000 shares by the end of Q3.

ES: El Banco Santander emitió 15.000 acciones a finales del T3.
FR: Banco Santander a fermé une emission de 15 000 actions pour la fin du 3ème trimestre. PT: O Banco Santander emitiu 15.000 ações no fim do 3º trimestre. ​

Here we see a change in verb tense, which leads us to understand the action has already occurred in the past instead of being expected in the future. The reader is misled and the meaning does not correspond to the source, so this is a major type of error as the idea(s) conveyed are wrong to certain degree.

In this category, we can find different subtypes like errors in numbers and figures or false friends - which are words similar in their signifiers but with a whole different meaning. So we have the following list of mistranslation subtypes:

Ambiguous translation

Languages in general do not have 1:1 correspondents on the lexical, syntax and grammar levels, hence some of the most common -and even easy- items to translate like the possessive pronoun “su” in Spanish has 4 possible translations in English such as “his”, “her”, “its” and “their”, making a simple sentence like “su ordenador” -without any more context,- ambiguous as the resulting translation may be “his PC”, “her PC”, “its PC” or “their PC” depending on the subject.

The same example goes for “you” treatment in English, which in Spanish, French and Portuguese can be translated as a plural “you” (people) or as a singular “you” (individual) depending on who we are referring or speaking to, and the same goes whether we use formal or informal style, as “you” does not reflect which one is to be used in the other 3 languages, but the context!

False friends

As mentioned above, a false friend is a word from the source language that looks “familiar” or almost the same to another in the target language, as its written form or signifier is very alike to one another. Hence, this may lead the translator to use a wrong translation instead of the right correspondent(s) in a given target language. These are highly recognisable as when reading a target text they tend to highlight among the rest of the contents as they do not seem to fit in well, they do not belong or they just feel odd for readers. As they are highly recognisable, they should be avoided above all.

Here you have shortlist of some well-known false friends.

assister á

False friend in English
assist at
attribute to
elaborate (verb)

Why is it wrong?
‘actual’ means ‘real’
ádequate’ means ‘sufficient’
‘assist’ means ‘help’
‘attribute to’ means ‘consider to be due to/characteristic of’
‘complete’ means ‘finish’
‘a delay’ means ‘a postponement or hold up’(=retard in french)
‘to elaborate’ means ‘to go into detail’
‘eventual’ means ‘predicted’
‘foreseen’ means ‘predicted’
‘important’ is right if you mean ‘significant’;but not if you mean ‘large’
‘material’ means ‘matter’, ‘information’
‘opportunity’ means ‘chance’
‘perspective’ means ‘standpoint’
‘to respect’ means ‘to value’ or ‘honour’ someone or something
‘sensible’ means ‘reasonable’

What’s the correct word?
current, topical
attend, participate in
allocate to, assign to
deadline, time limit
draft, develop, produce
provided for, planned
supplies, equipment
prospects, outlook
comply with (rules), meet (a deadline)

Overly literal

Sometimes the mistranslation come from the oddness produced by texts (translations) that, despite being “correct” in the target language, do not read like natural texts as they follow too closely the narrative, lexical choices or even the syntax of a foreign language - the one of the source language the text was written in originally. This sort of mistranslation is much more subtle and sometimes hard to realise. Frequently one might think this is a question of taste or written style, but we must compare in depth both source and target texts in order to find out where the problem comes from - lexical choices, syntax structures, punctuation and length of sentences, etc.

Numbers and figures

Numbers and figures constitute a category of its own as there are many different conventions between languages and also between variants of the same language to designate quantity, units, time, etc.
As a translator, not being aware of those can lead to errors and a lack of accuracy (mistranslation), as when using decimal dot in English for Spanish, French and Portuguese when the convention for these latter is to use decimal comma instead. The same can be said about a billion in English, which does not equal a “billón” in Spanish, but rather a thousand millions, while a Spanish “trillón” will be the equivalent of the English “billion” - a figure of 12 zeros.

Should not have been translated

This category has to do with mistranslation led by overtranslating, like items that should remain the same both in the source and in the target text, such as proper names. There is not much to say, but to be attentive to what is expected to be translated from the point of view of the target reader and what should remain as is - like a street name and the like.


EN: Santander Bank will issue over 15,000 shares by the end of Q3.

ES: El Banco Santander emitirá 15.000 acciones a finales del T3.
FR: Banco Santander se lance dans une emission de 15 000 actions pour la fin du 3ème trimestre.
PT: O Banco Santander anuncia a emissão de 15.000 ações no fim do 3º trimestre.
Here we have the opposite case to “addition”, where an item in the source has no correspondent in the target text or is missing. This makes the translation incomplete, not wrong, as an idea(s) has not been effectively conveyed. ​


EN: Santander Bank will issue 15,000 shares by the end of Q3. ​

ES: El Banco Santander emitirá 15.000 acciones a finales de Q3.
FR: Banco Santander se lance dans une emission de 15 000 actions pour la fin du Q3. PT: O Banco Santander anuncia a emissão de 15.000 ações no fim do Q3.
When we have an item, a part of a sentence or a sentence in the source with a correspondent in the target language but that remains the same (in the source language) in the translated text, this can be due to unintentional omission or because the source text was not fully understood and hence leads to a lack of meaning or ideas conveyed as the target audience may not recognise the item written in a foreign language or feels the oddness of the translates text due to it. This is not a category of those items to be kept the same both in the source and target text, as these have a correspondent translation and needs to be translated.

Tips & Tricks

When it comes to accuracy, we must prepare ourselves to deal with a text having the right tools for each scenario. We must stay focused on the type of text and contents we are facing, so our lexical choices will be comprehended in certain sematic field. I.e., when translating general contents from English into Spanish, a policy will be “una política”, while when dealing with insurances and insurers, this same word will become “una póliza”, so the semantic field will be restricted to this speciality or industry, not only when selecting our words or terms, but also the tone, treatment and way to convey meaning.

In order to get into context, we can read the text as a whole before starting to translate. If there are any words that we do not understand or that we find don't fit in that context, we should look them up and read other reference texts to gain more insight. By doing so, we will prevent accuracy errors when we start to translate. The same goes for proper names and items that need to remain as in the source text when we get to them while translating.

Once we have our translated text, we must correct any possible flaws by cross-checking source and target texts looking for inaccuracies. This spot-check must focus primarily on highlighted items such as proper names, numbers and figures, text with a different format like bold or italics, bullet points, etc., so we can make sure that these have been correctly translated and none are missing.

We will also verify that no necessary items in the source are missing in the target text and the same for any additions to translation that were not intended by the author of the source text. While doing so, meaning re-reading our translation, we will verify that ideas have been faithfully conveyed from the source into the target language in the most natural way from the point of view of syntax, lexical choices and punctuation.

If our role, is a reviewer one for texts produced by another linguist, we must also get into context but instead of reading parallel texts and contents alike, we must deep dive into the choices made by the translator and validate them.
In order to do so, we must be aware of the different items that can be wrong to a higher or lesser degree regarding the different accuracy errors explained above. Therefore, we must apply certain procedures in order to spot and correct them ‒ using a multi-prism check or screening looking into the following items:

Check for any omissions

We cross-check source and target texts in order to spot any missing parts ‒which are mistranslations by omission ‒ and if so add any missing items to the target text. It is very easy to realise omissions such as proper names, capitalised words or expressions, numbers, dates and figures, as well as other highlighted items, such as when the text has bullet points, uses interrogation and exclamation signs, contents in italics, bold or underlined, etc. So before we get into other accuracy aspects, we must make sure that all items from the source text are present in the target one.

Check for untranslated text

Apart from making sure that none of the source items have been left behind in the translated text, we should check that all that needed to be translated has been effectively translated. As we have explained in the categories and subcategories of mistranslations, there are items that need to remain the same both in the source and target text as they are proper names or parts that do not need translation, so these should not be mistaken with those items that have a correspondent in the target languages and that has been mistakenly left untranslated. Once spotted, we must make sure to correctly translate them and place them along the text in the most suitable way.

Check for any additions

While cross-checking both source and target texts, we must put close attention to detail in case we realise any “addition” to the translation that were not intended by the author(s) and therefore introduces a new meaning or a different meaning in the target text. Once spotted, these must be removed by editing the contents in the target language in a way that it mirrors the source.

Check for mistranslations

While reading through and cross-checking both source and target texts, we must pay attention to any evidence of false friends, errors in numbers and figures, text mistakenly translated or overly literal texts in the translated text. This sort of screening entails a really deep attention to detail and review skills from any linguist, making sure when reading the target text, sentence by sentence and paragraph by paragraph, that the translation reads natural as a native speaker in the target language. We achieve fluency in a text when it flows well from the point of view of the language (syntax, lexical choices and punctuation) and contains no false friends or superfluous translated items.

All in all, the aim of an accuracy review is to assure the final text will be faithful to the source in conveying all information intended by the author(s), that will read natural and appropriate for its purpose and that no other additional information or items will remain mistakenly included in the translation. As translators is our duty to review our own texts and as reviewers we must assess other’s texts with deep attention to detail in order to validate their choices.