Language and Translations

Although many journal articles are in a single language (often English), it is common for journal articles to have some metadata, such as the article title, abstract, and/or keywords, translated into more than one language. It is also possible for the the entire content of an article, or substantial portions thereof, to be published in more than one language. In addition, it is very common for journal articles to contain words or passages of text that are in a different language than the primary language of the article. It is important to correctly identify the language of text using the @xml:lang attribute in order to support a range of functionality, including correct pronunciation by screen readers (text-to-speech software), searchability, and particular display requirements.

The @xml:lang attribute identifies the language of text contained within the element on which it appears. The @xml:lang attribute is required on the <article> root element. An @xml:lang attribute should also be placed on elements at each place where the language of the text is different than the previous @xml:lang attribute. When an element does not have an @xml:lang attribute the language identification for that element’s content is effectively inherited from the nearest ancestor element that has an @xml:lang attribute.

The value of the @xml:lang attribute should identify the language of the text using a valid ISO-639 language code in lower-case letters. Preferably, the value of @xml:lang should use a two-letter language code from ISO-639-1. If a language does not have a two-letter language code then a three-letter language code from ISO-639-2 may be used. Optionally, the value of @xml:lang may also include a region code and/or a script code for additional specificity.

JATS allows many elements to be repeated for capturing translations or alternate language versions of specific parts of an article’s metadata and content. For example, if an article’s keywords are presented in more than one language the <kwd-group> element can be repeated with different @xml:lang attribute values for each set of keywords.

It is important that these features are used correctly to identify the language of content.

When an article title appears in more than one language each alternative language version of the title should be identified by an @xml:lang attribute. There are three different ways that alternative language versions of the article title (and optional subtitle) may be structured:

1. Use <title-group> elements for each alternative language version of the title. Place @xml:lang on each <title-group>. This method is preferred starting with JATS version 1.4.
2. Use a <title-group> element for the primary language version of the article title, and use <trans-title-group> elements for each additional language version of the article title. Place @xml:lang on each <trans-title-group>. (The <title-group> inherits its language from the nearest ancestor element that has an @xml:lang attribute, e.g. the root <article> element.) This method works in all versions of JATS.
3. Use a <title-group> element and include the alternative language version of the title as italicized text appended to the primary language version of the title, possibly after a “/” character or enclosed in parenthesis, within the same <article-title> element. Place @xml:lang on an <italic> or <named-content> element. When using this method it can be unclear whether the title is provided in more than one language or the title contains some text in another language. This method works in all versions of JATS.

Starting with JATS version 1.4, the <title-group> element can be repeated in <article-meta> to hold alternative language versions of an article title. In older versions of JATS <title-group> can only appear once in <article-meta> and the <trans-title-group> element was intended to hold translations of the article title and subtitle. The <trans-title-group> element is now deprecated although it is still available for compatibility.

In practice, the primary language version of an article title might actually be a translation and the original language version of an article title might be misidentified as a translation by using the <trans-title-group> element. The @lang-variant attribute, new in JATS 1.4 and available on the <title-group> element, provides a more flexible way to identify which alternate language versions are translations or original.

This example from https://doi.org/10.1080/25729861.2018.1532779 uses <title-group> elements for each alternative language version.

<title-group xml:lang="en">
<article-title>Disentangling war and disease in <strike>post</strike>-conflict Colombia beyond technoscientific peacemaking</article-title> 
</title-group> 

<title-group xml:lang="pt"> 
<article-title>Desvendando guerra e doen&ccedil;a na Col&ocirc;mbia p&oacute;s-conflito, al&eacute;m da pacifica&ccedil;&atilde;o t&eacute;cnico-cient&iacute;fica</trans-title> 
</title-group> 

<title-group xml:lang="es"> 
<article-title>Desenredando la guerra y la enfermedad en la Colombia postconflicto m&aacute;s all&aacute; de la pacificaci&oacute;n tecnocient&iacute;fica</trans-title> 
</title-group> 

This example from https://doi.org/10.1080/25729861.2018.1532779 uses <title-group> for the primary language version of the article title and <trans-title-group> elements for each additional language version.

<title-group> 
<article-title>Disentangling war and disease in <strike>post</strike>-conflict Colombia beyond technoscientific peacemaking</article-title> 

<trans-title-group xml:lang="pt"> 
<trans-title>Desvendando guerra e doen&ccedil;a na Col&ocirc;mbia p&oacute;s-conflito, al&eacute;m da pacifica&ccedil;&atilde;o t&eacute;cnico-cient&iacute;fica</trans-title> 
</trans-title-group> 

<trans-title-group xml:lang="es"> 
<trans-title>Desenredando la guerra y la enfermedad en la Colombia postconflicto m&aacute;s all&aacute; de la pacificaci&oacute;n tecnocient&iacute;fica</trans-title> 
</trans-title-group> 
</title-group> 

This example from https://doi.org/10.1080/11356405.2023.2249706 shows an alternative language version of the article title as italicized text appended to the primary language title within the same <article-title> element. The parenthesis and italicization provide only a visual distinction.

<title-group>
    <article-title>Development of vocal exercises for training musical skills (<italic xml:lang="es">Desarrollo de ejercicios vocales para entrenar las habilidades musicales</italic>)</article-title>
</title-group>

In JATS versions prior to version 1.4 the <italic> element is not able to hold an @xml:lang attribute so the <named-content> element can be used to hold the @xml:lang attribute.

<title-group>
    <article-title>Development of vocal exercises for training musical skills (<italic><named-content content-type="language" xml:lang="es">Desarrollo de ejercicios vocales para entrenar las habilidades musicales</named-content></italic>)</article-title>
</title-group>

When an abstract appears in more than one language, each alternate langauge version of the abstract should be tagged in its own <abstract> element with an @xml:lang attribute to identify the language.

This example shows an abstract with alternate language versions in English, Portuquese, and Spanish.

<abstract xml:lang="en"> 
<title>ABSTRACT</title> 
<p>In November 2016, the Colombian government and the Revolutionary Armed Forces of Colombia (FARC) signed a peace agreement to end a 52-year war.</p> 
</abstract> 

<abstract xml:lang="pt"> 
<title>RESUMO</title> 
<p>Em novembro de 2016, o governo colombiano e as For&ccedil;as Armadas Revolucion&aacute;rias da Col&ocirc;mbia (FARC) assinaram um acordo de paz para acabar com uma guerra de 52 anos.</p> 
</trans-abstract> 

<abstract xml:lang="es"> 
<title>RESUMEN</title> 
<p>En noviembre de 2016, el gobierno colombiano y las Fuerzas Armadas Revolucionarias de Colombia (FARC) firmaron un acuerdo de paz para poner fin a una guerra de 52 a&ntilde;os. </p> 
</trans-abstract> 

If it is known which versions of the abstract are translations and which version is the original language version, the @lang-variant attribute can be used to make this distinction. Attribute lang-variant="original" can be tagged on the <abstract> element that contains the original language version of the abstract, and attribute lang-variant="translation" can be tagged on the <abstract> element(s) that contain the translationed version(s) of the abstract.

The <trans-abstract> element for translated versions of an article abstract is deprecated starting with JATS version 1.4, although this element is still available for compatibility. Previous guidance was that the primary-language version of the abstract should be tagged in <abstract> and the <trans-abstract> element should be used to tag each additional language version of the abstract.

In practice, the primary-language verion of an abstract might actually be a translation and the original-language version of an abstract might be misidentified as a translation by using the <trans-abstract> element (for example, if primary language of the article is English and the original langauge of the abstract is not English). The @lang-variant attribute, new in JATS 1.4, provides a more flexible way to identify which alternate langauge versions are translations or original.

This example shows an abstract in English with translations in Portuquese and Spanish using the <trans-abstract> element. The <abstract> element in this example inherits its language identification from the root <article> element which has @xml:lang=“en”.

<abstract> 
<title>ABSTRACT</title> 
<p>In November 2016, the Colombian government and the Revolutionary Armed Forces of Colombia (FARC) signed a peace agreement to end a 52-year war.</p> 
</abstract> 

<trans-abstract xml:lang="pt"> 
<title>RESUMO</title> 
<p>Em novembro de 2016, o governo colombiano e as For&ccedil;as Armadas Revolucion&aacute;rias da Col&ocirc;mbia (FARC) assinaram um acordo de paz para acabar com uma guerra de 52 anos.</p> 
</trans-abstract> 

<trans-abstract xml:lang="es"> 
<title>RESUMEN</title> 
<p>En noviembre de 2016, el gobierno colombiano y las Fuerzas Armadas Revolucionarias de Colombia (FARC) firmaron un acuerdo de paz para poner fin a una guerra de 52 a&ntilde;os. </p> 
</trans-abstract> 

When an article’s keyword list is presented in more than one language, a separate <kwd-group> element should be used to tag each language version of the keywords list. The @xml:lang attribute is used to identify the language for each <kwd-group>.

The primary-language <kwd-group> does not require an @xml:lang attribute because the language is inherited from the xml:lang attribute on the <article> element. However, it is preferable to also add an @xml:lang attribute on the primary-language <kwd-group> when the keywords list is presented in more than one language.

The <title> element is used to provide the translation of “keywords” in each language.

<kwd-group kwd-group-type="author" xml:lang="en"> 
<title>KEYWORDS</title><kwd>Temporality</kwd><kwd>technoscience</kwd><kwd>peace</kwd><kwd>war</kwd><kwd>pharmaceuticalization</kwd> 
</kwd-group> 

<kwd-group kwd-group-type="author" xml:lang="pt"> 
<title>PALAVRAS-CHAVE</title><kwd>temporalidade</kwd><kwd>tecnoci&ecirc;ncia</kwd><kwd>paz</kwd><kwd>guerra</kwd><kwd>farmaceuticaliza&ccedil;&atilde;o</kwd> 
</kwd-group> 

<kwd-group kwd-group-type="author" xml:lang="es"> 
<title>PALABRAS CLAVE</title><kwd>Temporalidad</kwd><kwd>tecnociencia</kwd><kwd>paz</kwd><kwd>guerra</kwd><kwd>farmaceuticalizaci&oacute;n</kwd> 
</kwd-group> 

In addition to capturing the language of an author’s name, the @name-style attribute can be used to denote the style of an author’s name and ensure that the given and surname are displayed in the correct order. A list of values and how they are ordered/displayed is below:

• eastern - The name will be displayed and sorted with the family name preceding the given-name.
• western - The name will be displayed with the given-name preceding the family name but will be sorted with the family name preceding the given-name. This is the default.

In this example, the Eastern name will display as 蘊智 黃 (surname first):

<contrib contrib-type="author"><name name-style="western"><surname>WAN-CHI</surname><given-names initials="W.">WONG</given-names></name></contrib> 

<contrib contrib-type="author"><name name-style="eastern" xml:lang="zh"><surname>&#x860a;&#x667a;</surname><given-names initials="&#x9ec3;">&#x9ec3;</given-names></name></contrib> 

When multiple languages are used within the body of an article, an @xml:lang attribute should be tagged at each level where the language changes. The @xml:lang attribute identifies the language for all text within the element until there is another @xml:lang attribute on a descendant element. An element with an @xml:lang attribute should be used to identify the language for all text that is in a language that is different than the language of the surrounding text or different than the language declared by the @xml:lang attribute on the root <article> element.

The @xml:lang attribute can be placed on many different elements, including: <p>, <sec>, <disp-quote>, <italic>, <bold>, <monospace>, <overline>, <roman>, <sans-serif>, <sc>, <strike>, <sub>, <sup>, <underline>, <styled-content>, and <named-content>.

In JATS versions prior to version 1.4 the @xml:lang attribute is not available on elements <italic>, <bold>, <monospace>, <overline>, <roman>, <sans-serif>, <sc>, <strike>, <sub>, <sup>, and <underline>. In older versions of JATS the <named-content> element can be used to hold the @xml:lang attribute. For example:

<italic><named-content content-type="language" xml:lang="es">jalape&#x00F1;o</named-content></italic>

This example shows a paragraph that is in a different language (fr, French) than the surrounding text. In this example, the primary language identified on the article element is English (<article xml:lang=“en”>).

<p>In August 2016, a month before going to Colombia to start my ethnographic field research, my partner and I decided to go for a 10-day vacation to Cuba. In Havana, we walked to the Palco Hotel early one morning to meet a tour bus.</p> 

<p xml:lang="fr">Nerval est ici plus fid&egrave;le &agrave; Fourier que ne le sont ses disciples patent&eacute;s car il reprend &agrave;</p>  

This example shows a disclosure statement that is in both English (en) and Spanish (es) in an article where the primary language identified on the article element is English (<article xml:lang=“en”>). The alternative language versions of the text are presented inline, and the “/” character and italicization provide only a visual distinction.

<sec id="s0017" sec-type="coi-statement">
    <title>Disclosure statement / <italic xml:lang="es">Conflicto de intereses</italic></title>
    <p>No potential conflict of interest was reported by the authors. / <italic xml:lang="es">Los autores no han referido ning&#x00FA;n potencial conflicto de inter&#x00E9;s en relaci&#x00F3;n con este art&#x00ED;culo.</italic></p>
</sec>

This example shows a paragraph that contains italicized words in different languages than the surrounding text. This paragraph would be problematic for screen readers if the language codes are not correctly tagged.

• jalapeño (Spanish) - Without xml:lang=“es”, English screen readers would mispronounce the “ñ” as a regular “n” and stress the wrong syllable (ja-la-PEE-no instead of ha-la-PEH-nyo)
• bruschetta (Italian) - Without xml:lang=“it”, the “ch” would be pronounced as in English “church” rather than the Italian hard “k” sound, and the double “t” wouldn’t be properly articulated
• борщ (Russian/Cyrillic) - Without xml:lang=“ru”, English screen readers might attempt to read Cyrillic characters phonetically using English rules, producing unintelligible sounds instead of “borsch”
• dim sum (Chinese) - Without xml:lang=“zh”, the tonal aspects would be lost, and English pronunciation patterns would apply instead of the proper Mandarin sounds
• flan (Spanish) - Without xml:lang=“es”, English screen readers would mispronounce the ‘an’ as rhyming with ran instead of the Spanish “ahn” sound

<p>The restaurant's <italic xml:lang="es">jalape&#x00F1;o</italic> salsa complemented the <italic xml:lang="it">bruschetta</italic>, while guests enjoyed <italic xml:lang="ru">&#x0431;&#x043E;&#x0440;&#x0449;</italic> and traditional <italic xml:lang="zh">dim sum</italic>, finishing with <italic xml:lang="es">flan</italic> for dessert.</p>

The @lang-translate attribute should be used in places where it is important that text is not translated by automatic translation software.

This example shows an italicized phrase and a quote where meaning would be lost if these parts of the text were to be translated by automatic translation software.

<p>The phrase <italic xml:lang="fr" lang-translate="no">je ne sais quoi</italic> is a French expression that has been adopted as a loan phrase because it carries nuance that is not conveyed by the direct literal translation "<named-content content-type="language" xml:lang="en" lang-translate="no">I don't know what</named-content>" in English.</p>

In JATS versions prior to version 1.4 the @lang-translate attribute is not available. There is no equivalent in older versions of JATS.

When a figure or table caption appears in more than one language, each additional language version should be captured in a <p> element with an @xml:lang attribute.

<fig id="F0001"> 
    <label>Figure 1.</label>  
    <caption>  
        <p>HS-GC-MS extracted ion chromatograms and mass spectra of (a) furan and (b) furan-d<sub>4</sub>. </p>  
        <p xml:lang="es">Figura 1.&emsp;Cromatogramas HS-GC-MS de iones extra&iacute;dos y espectros de masas de (a) furano y (b) furano-d<sub>4</sub>.</p>  
    </caption>  
    <graphic xlink:href="TCYT_A_669797_F0001_B.gif" content-type="black-white"/>  
</fig> 

There are three options for tagging translations or alternate language versions of an entire article.

• Separate articles
• Within the article using <sec>
• Within the article using <sub-article>

A translation or alternate language version of an article may be published as a separate article with its own DOI.

Each article should include a <related-article> linking to the other form of the article. The <related-article> element should have these attributes:

• @related-article-type=“translated-article”
• @xlink:href should hold the DOI of the linked article or a URL if no DOI is available.
• @ext-link-type should hold “doi” or “uri” to indicate the link type of @xlink:href
• @hreflang should identify the language of the linked article.

In JATS versions prior to version 1.3 the @hreflang attribute is not available. In older versions of JATS the @xml:lang attribute on <related-article> can be used instead.

For example, article 10.1080/13696815.2013.789248 was originally published in Gĩgĩkũyũ (ki) and then an translated into English (en) and published as article 10.1080/13696815.2013.789251.

The article that is the translation should have a <related-article> link pointing to the article that is the original:

<related-article related-article-type="translated-article" ext-link-type="doi" xlink:href="10.1080/13696815.2013.789248" hreflang="ki"/>

The article that is the original should also have a <related-article> link pointing to the article that is the translation:

<related-article related-article-type="translated-article" ext-link-type="doi" xlink:href="10.1080/13696815.2013.789251" hreflang="en"/>

For more more information about <relate-article> see Related Articles.

A translation or alternate language version of an article may be published within the article itself as a section within the article. If the translation is located after the main text of the article and before the article’s backmatter the translation should be contained within a <sec> element that has an @xml:lang attribute to identify the language.

The entire text of the translation or alternate language version should be contained within a <sec> element with an @xml:lang attribute. If the entire text of the translation or alternate language version cannot be contained within a single <sec> element, the language of the text must still be identified using @xml:lang on available elements.

A translation or alternate language version of an article may be published within the article itself following the main article. If the translation is located after the article’s backmatter the translation should be contained within a <sub-article> element that has an @xml:lang attribute to identify the language.

When an article has one primary language the @xml:lang attribute on the root <article> element identifies the primary language of the article, however this is not sufficient when an article has more than one primary language. An article that contains alternate language versions or translations of the article’s main content should have the <content-language> element in <article-meta> to identify each language.

For example, if an article’s main text is in English (en) and a translation in Spanish (es) is included within the article (using <sec> or <sub-article>), each language should be identified by a <content-language> element located in /article/front/article-meta:

<content-language>en</content-langauge>
<content-language>es</content-langauge>

In JATS versions prior to version 1.4 the <content-langauge> element is not available. There is no equivalent in older versions of JATS.

Instructions will be provided if the additional language metadata attributes described in this section are needed for any specific journals or articles.

The @lang-group attribute can be used to connect alternative language versions of the same content. When using @lang-group, each element that contains an alternative language version of the same content should have the same @lang-group attribute. The @lang-group attribute links the elements using an IDREF, and an @id attribute with the same value must be placed on one of the alternative language versions to create a basis for the link.

<disp-quote>
    <p xml:lang="en" lang-group="lang0001" id="lang0001">Cure sometimes, treat often, comfort always.</p>
    <p xml:lang="es" lang-group="lang0001">Curar a veces, tratar a menudo, consolar siempre.</p>
    <attrib>&#x2013; Hippocrates</attrib>
</disp-quote>

The @lang-variant attribute can be used to identify a member of a language group as an original, a translation, a transcription, etc.

The @lang-focus attribute can be used to indicate how members of a language group relate to each other, for example by identifying which are primary or secondary.

The @lang-source attribute can be used to describe the source of an alternative language version, for example “author” or “translator”.

In JATS versions prior to version 1.4 the @lang-group, @lang-variant, @lang-focus, and @lang-source attributes are not available. There is no equivalent in older versions of JATS.

Non-keyboard characters, such as Chinese symbols and acute accents, should be tagged using Unicode characters where possible. See Character Encoding.