Posted on May 7, 2018 12:39 pm

Home

logo for adelaidelanguages.com

logo for adelaidelanguages.com

The Odd Language of WordPress: Lost in Translation

Part 1 : An Overview of Automatic Translation (Google)

This is an independent study on the wordPress semantic’s language used by wp experts to create, develop technical texts and write their documentation for their Themes and/or Plugins.

This is initial study that takes into consideration one main important area of work:

  • The Creation of the ORIGINAL: Source Text for WPMU DEV

Stressing from the beginning here that the idea for this exercise is not intended to criticize or evaluate wpmu dev systems in place, or any of the translations works performed by members in the form of translations.

It is a study conceived and designed from what is perceived by me personally while working as a linguist on the wpmu dev translations. In this I like to present my views on the logical aspects of writing in a tech language such as wordpress, and demonstrate with simple examples that more needs to be done in the form of editing the Original documentation, having in mind a larger audience the wp serves, a worldwide audience.

By the same token I feel that I do not have a need to go looking for any signs of errors in performed translations in the languages I know by other members (as unfortunately I am sure there are)

I like to start to demonstrate this with a simple example of a screenshot taken at random, while translating it from wpmu dev plugin.  

The Original- The WPMU DEV Documentation in a nutshell at random

If one looks closely at the source language in the screenshot above, it reveals four highlighted areas of text in red.

1 – The Expression Heads up

2 – The wp tech term Plugin

3 – the verb enable

4 – the noun features

5 – The sentence: User gets read and makes comments of posts

Therefore, we have five strings of text that to a normal English WordPress trained mind, looks simple and clear enough; that is if the person reading these strings is a wordPress developer, designer, trained or working in a similar capacity and his or her first/second language(s) happened to be English. To a lot of other people, and particularly to translators with limitations of English and who might use Gooogle Translate this is something else.

Professional translators-linguists with a grasp and knowledge for finding meaning in WordPress terminology, any good English reader will understand that  Plugin, enable and features, are words that are not easy to translate, by using automatic translation or a dictionary, without the risk of semantic and grammar errors. Plugin, enable, features are just some of these words that have too many derivations of meanings.

Therefore very clear to these using wp plugins and similar very difficult to the generalist translator and these looking up for semantic definition of these terms.

Meaning: The Original Analyses 1 to 5

Every mistake in the Original document will help, exponentially, to populate bad translations across the entire world community of WP translators.

Starting with the sentence N.5.

There is no doubt that there is a spelling error in the sentence (User gets read and makes comments of posts. The question though is: How will these hundreds of translators will document, translate/transcribe their own languages, in the language(s) they know? How will they take care of the ambiguities and errors in that sentence?  And if using AT (Automatic Translation, as it seems to be recommended and suitable for wordPress as translation tool), what will become or derive from this and other strings?

 

This is the point that I will try to bring on, and hope to be able to express and prove with simple examples, (where I still don’t know the results, and if I will succeed in doing this), butI am confident that we will arrive at the point that this language exercise is aimed at: discover things against leaving spelling and other common mistakes unchecked or not edited properly in the original text, and then later relying on ‘’Google Translate’’ for instance, to translate it in hundreds or more other languages.

Just as a quick exercise lets imagine that the sentence to be translated is/was: Enable plugin and features;

I am going to use Google Translate, (as it seems to be recommended and the tool of choice for wpmu dev), to translate from English to ES, FR, IT, PT, to arrive at result for this translation; (Just now trying for the first time, as I wrote this; and you the reader, can as well check this to see what results you getting Google, Microsoft,Amazon,Matecat to name some ).

In all probability I am hoping to find enough linguistic evidence that proves what I am thinking that will happen. That for this kind of tech language, wp terminology Google will come up with various solutions.

And this might be the boring bit, so I avoid going into the language correctness other than pointing out the discrepancies highlighted in bold text.

On the day(  ) I got the following results for the sentence: Enable plugin and features

In Spanish – Habilitar el complemento y las funciones (habilitar = setup; capacitate; someone or something) 

In French – Activer le plugin et les fonctionnalités (activer = activate; accelerate; something or some control)

In Italian – Abilita plugin e funzionalità (abilitare (v )= same as above for Spanish)

In Portuguese – Ativar plugin e recursos (ativar (v)= same as above for French)

These are the results I got and I am amazed at what I got. I have to agree that Google Translate has come a long way, when comes to recognizing gender, masculine from feminine, singular from plural, actions, tenses, verb endings and so on. This is the main reason why to the unaccustomed eye these translations look acceptable and true, when in fact these are only half right. On the surface ”Google Translate”, surpasses my expectations. When it comes to avoid spelling mistakes there aren’t any.

Are these translations suitably correct or good enough to be used by a reputable, any business and proud individuals, selling WP merchandise and or top shelf WPMU DEV Tools and Plugins? It is questionable as I do not have an answer.

Proceeding with this previous language exercise:

Firstly we have a look at the initial actions that are going to take place and expressed by the verb – Habilitar(ES) Habilita (re) (IT) Activer (FR) Ativar (PT) (habilitar (ES) says little about ”enabling or making a plugin function”.

From the above one can automatically realize two things in particular:

1 – The verbs are different for the action they performing ”Enable” as we see above

2 – The translations in the four languages are incorrect as ”plugin” remains untranslated with the exception of the text in Spanish where it gets translated as ”complemento” and is also wrong used here.

3 – And more importantly Google could not ”see between the lines”; the sentence Enable plugin and features, is incomplete and Google can’t see that but a human brain can. We will want to know which plugin and features we need to enable; So we read through the text, ask questions, research the subject and so on.

Next we analyse the idiomatic expression ”Heads up

In Spanish: Aviso

In French: La tête haute

In Italian: Dritta

In Portuguese: Atenção

Of course it makes no sense whatsoever; this serves to show that Google translation as well many other automatic translation tools and resources should be used with great care, in all documentation, subjects and at all levels of language. Hands up, from these who agree with this statement?

”Hands up” . I can’t wait to see what our Google translator guess will be.

As expected and a good BAD guess for Hands up:

In Spanish: Manos arriba

In French: Les mains en l’air

In Italian: Mani in alto

In Portuguese: Mãos ao ar

Even though ”Hands up” is shown without an exclamation mark (!) it makes no difference to Google; it reads, translates this basically as ”Hands up (hands up in the air!)”, instead of ”Raise your hand” …if you know’/if you can answer this/that; and not something like  (‘’Hands up! This is a holdup! The’ holdup’ is my creation, sorry.)

This Brings us to the last full sentence in the picture above – User gets read and makes comments of posts. What results will Google come up with in the translation we can only imagine since the sentence is clearly missing words and some context or at least concision.

It can be at least three things.

  • User gets read and makes comments of posts (original) He/she /user is read (slang language)
  • User gets toread and makes comments of posts  (guess 1) He/she/user reads
  • User gets to be read and makes comments of posts (guess 2)He/she /user is read (correct)

These three possibilities above will all bring a different result, be it derived from automatic or human translation contexts. 

Concentrating first on the original WPMU DEV text and the meaning extracted from Google:

1 – User gets read and makes comments of posts.

In Spanish: El usuario es leído y hace comentarios de publicaciones. = User is read…

In French: L’utilisateur est lu et commente les messages. = User is read…

In Italian: L’utente viene letto e rende commenti dei post. = User is read…

In Portuguese: Usuário é lido e faz comentários de posts. = User is read…

User gets read, is what Google give us back in the four languages above and, is pretty much correct in the interpretation of the original English sentence. It may not be entirely correct but contrary to ‘’hands up’’ meaning we get the message without being held up!

As humans one can assume that improper ‘slang’ language has been used in the original; that the preposition (to), and verb form (to be) is or  could be missing; more importantly, as humans we can, and should be asking questions about any clarification of these sentences before transcribing/translating from one, to exponentially hundreds of  languages, as happens nowadays.

2 – Assuming that ”User gets (to) read and makes comments of posts…, is the answer will bring totally different results from GoogleTranslate:

In Spanish: El usuario lee y hace comentarios de publicaciones = User reads…

In French: L’utilisateur peut lire et commenter les messages = User can/is able to read…

In Italian: L’utente legge e fa commenti sui post = User reads…

In Portuguese: O usuário começa a ler e faz comentários de postagens = User starts to read…

The discrepancy in time of action starts to be visible as verb tenses for French and Portuguese are incorrect while in Italian and Spanish offer the right gist.

3 – The other assumption: ”User gets (to be) read and makes comments of posts’’, will show what follows:

Spanish – El usuario se lee y hace comentarios de publicaciones = User reads itself/ one reads…

French – L’utilisateur peut être lu et commente les messages = User can/may be read…

Italian – L’utente deve essere letto e fa commenti sui post = User must be/should be read…

Portuguese – Usuário começa a ser lido e faz comentários de posts = User starts to be read…

Results from ”Google Translate” became now dubious and erroneous informing us of a different action taking place in the four different languages in question and as seen by the difference in the verb tenses and grammar results.

The conclusion one can take from this small translation exercise, is that no matter what Machine Translation Engine – ”Google Translate” or any other form of Machine Translation one uses, be it well written or ill prepared source/original texts, media and any other documentation, the results can be disastrous or even catastrophic.

 

Again this shows totally different and totally wrongly results, even though as far as I am concerned the original sentence is kind of correct in English and Google didn’t recognize this.

This is one very important factor that I believe has to be taken into consideration would be the quality of the ORINALITY expressed in English and factored in the translation aspect of a hundred plus languages

This is the kind of danger that can be observed on how misinterpretation can be derived from any original text, not just about wp, intended to be translated in more than an hundred languages plus; THIS is the case of WordPress and WP MUDEV’s and / or any other official documentation.

The Odd Language of WordPress: Lost in Translation

  • Original Text Relating to WP in general: Themes/Plugins
  • The Tasks Considered In Writing All This Documentation In English
  • This Documentation Being Translated in Many Languages, (100’s) Across the World.

 

Part 2 – Who Is Responsible To Write/Create This English Documentation for WordPress?

An Overview. Editing of Texts by Linguists/Other Before Going to Print.

In the case of wpmu dev documentation it appears that more has to be done towards this end. These are, should be the solid foundations for all the construction that follows.  

Going back to Google translate automatic translation is unpredictable. One thing is for sure it beats many of us at remembering things. It sure can read but it can’t think!

if we go back to previous examples ( Part 1 Link Here) strangely enough, we seem to be getting worse results from an original correct(ed) sentence – User gets to be read and makes comments of posts…, than from the previous one – User gets read and makes comments of posts, that could be considered to be written in ”slang style”.

The point is: It seems wrong and highly impractical to avoid proofing the original documentation as much as possible, after developers and other responsible write it. This is documentation for plugins, themes, and other documentation to go over the world to millions of peple from all walks of life. If one considers the great number of these and the very many sectors of the economy these document texts will serve, there will be expectations for improvements. (WordPress is used …)governments and legislation departments all over the world use wpCMS’s and in the USA the list of these department include the White House. Many are the the international Institutions,  major enterprises, entertainment, film and television, newspapers, universities and the list goes on  https://themeisle.com/blog/wordpress-website-examples/ (http://www.honeytechblog.com/top-10-government-sites-powered-by-wordpress/ ) and is forever widening and getting more specialized (https://www.quora.com/

that surely work with WordPress on their daily bases and WPMU DEV tools and plugins.

 

Above all these same original WordPress documents, are going to be creating thousands of translations. If a translation is riddled with typos and spelling mistakes, then this instantly renders the work poor quality.

 

Leaving these text documents, without being properly translated, transcribed, revised and so on, before goes to print, can contribute to serious complications as many different translation outcomes can be derived from ‘’GoogleTranslate’’ and similar search engines.

It seems wrong and highly impractical to avoid proofing the original documentation if one considers sectors of the economy as whole: governments and legislation, educational establishments, businesses and the private sector in general, that surely work with WordPress on their daily bases and WPMU DEV tools and plugins.

Above all these same original WordPress documents, are going to be creating thousands of translations. If a translation is full with typos and spelling mistakes, then this instantly renders this work as of poor quality.

 

Starting with what seems easy in English and hard to translate in other languages as an example in these few strings for wpmu dev. Starting from the top down the word

‘’Easily’’

  • Matching length of word: as an adverb will create an extension of the word and longer text in FR<IT<PT<ES etc.
  • It isn’t written in a appropriate careful style; Is more related to talking language
  • Lacks of proper context and meaning
  • Easily, (pardon the pun,)It is not necessary to be there, as only adds to confusion

‘’Her’’

  • Assumptions and presumptions lead to guess work
  • As a typographical spelling mistake for ‘’here’’ will leave doubts in somelinguist minds
  • These with limited knowledge of English will just use results from Automatic Translation
  • Could easily be found and corrected if proofread.

 

Problems with translation of any form is often underestimated and overlooked as a normal activity. No more than in the case of wp.org for instance where anyone with an account can login and start translating. (translate.wordpress.org.

However this is understandable as wordpress.org is completely free to everyone and there are controls in place and a chain of linguists appointed to take over acceptance of translations before these are published to the web.

I believe that such system, as the wp.org above, would not be so easier to implement; in the case of wpmu dev and similar wp companies, considering that these are payable systems, members feel that they have the right and expect these localized translations to be available from day one full stop.

Developers, designers, customer, other stakeholders have to accept what is given to them.

I equally believe that, in the case of wp  paid membership is questionable leaving this to every member to do; it also would pay off, (as more members would get involved in volunteering to translate), if documentation for plugins etc, all written in English – original – was a bit more cared for; having in mind that, number one, it has to make sense first and foremost, to Google translate and then to the translator.

 

In doing this exercise I decided on two things important to me and I believe to readers:

1 – Presenting and using examples taken from wpmu dev as I work/translate same. Examples that in the original – English seem short and clear and that to these translating in hundreds of local languages is a nightmare. And we do not even mention what smorgasbord Google Translate can make of it!

Below are some of these examples that affect translator’s potential to help more than they could do:

 

The original In no particular order; images taken at random from wpmu dev Translation_Plugins.

2 – Divorcing myself from the quality of translations that might exist because this is all about the original documents taken at random. I believe that my notes above explain it a bit and need no further comments from my side.

 

 

Besides presenting original documents for translation as much as possible with clear meaning, and proper linguistic guidelines there are other established important factors to take care of; that is that text can grow or shrink when it is moved between localized languages.

To come to an acceptable result for example abbreviating text, sentences in the English original is not always the ideal situation.

A sentence that is 15 words long in English for a wp website for instance may require 25-30 words in another language, as:

  • more contest is need to achieve the same result for obtaining the same meaning.
  • Is dependent on the knowledge and the mastering skills of each linguist/translators.

 

These economy of words will be discussed later in Part 3 Linguists – The Revision Process and Discussions that Follows Creation of text documentation.

 

website’s homepage and a translator does not take into account the need to be economical with their words, the translation may be great in terms of accuracy and style, but it simply won’t work for its purpose. In this way, it is vital that a translator pays due attention to formatting requirements as well.

they say, if a job is worth doing, it’s worth doing well.

 

The Odd Language of WordPress: Lost in Translation

Part 3 – An Overview. Proofing of Texts by Linguists Before Printing

This process is usually done by professional translators but anyone with good language knowledge of the language at task and with a keen eye to spot mistakes can do this. In the case of wpmu dev it requires some knowledge and research into this type of tech terminology.

 documentation is edited as necessary usually keeping the same file format and taking care of the following:

Making sure that original is:

  • Accuracy, clarity and completeness of the content
  • Whether the purpose of the text has been fulfilled (informative, descriptive, narrative, explanatory, argumentative)
  • Appropriate word choice/terminology according to the subject field, client, target audience and context
  • Appropriate style
  • Consistent and coherent wording and spelling
  • Spelling
  • Typing errors
  • Syntax (word order and sentence construction)
  • Grammar
  • Punctuation
  • Hyphenation

 

can help you to significantly reduce upcoming costs. Interlingua is here for you; we can provide you with consulting services, and we can offer the most a

The linguist focuses on the separate words in the text, their meaning and their contextual use – in other words: syntax.

The review is treated as a final revision of the text’s integrity in comparison to the original and the provided guidelines. It includes copy editing, reevaluation and change of the whole text structure, if necessary.

 

nsure that your translation is accurate and error free?

 

 

 

 

PROOFREADING

Proofreading is also a monolingual review. In a proofread, all formal aspects of a text or a translation are checked one last time before the text is printed. Aspects of the text that are reviewed include punctuation, capitalisation, obvious word separation and the layout (formatting of headings, page numbers, spacing, insertions, the position of graphics, image descriptions or captions). Changes to the content or wording should not be required at this stage.

 

It helps if you know that wpmu dev language is highly technical and can be ambiguous sometimes

 

In Part I wrote about and gave examples, (obtained at random), of how simple it was to prove that one shouldn’t rely on translations automatically conceived or created by Internet Translators such as Google.

The same goes for Amazon, Microsoft, Matecat and many other good CATS working with algorithms and memories, and other digital systems and spit out fuzzy translations that devalue human translators as it gives the idea that now anyone can be a linguist. I intend to talk more about this topic in part 3.

But to be fair there are other advantages besides disadvantages worth mentioning in using Google Translate. Three advantages being the fact that:

Automatically takes care of a lot of spelling errors that can be missed easily by humans

Is able to employ and use grammar; it recognizes verb tenses, gender,

  • vagueness
  • ambiguities
  • doubts and assumptions
  • guess work
  • the length of text
  • unnecessary capitalization
  • level of language used and so on

 

Part 3 – The Revision Process and Discussions that Follows Creation of text documentation.

 

I must stress here that Google and all these automatic translators have come a long way and in the hands of knowledge people and linguists it can save those heaps of time doing short quick translations. Polyglots can work on as many languages as they need to.

Google Translate or any other engine, dictionary and literary sources are here to help us that with some creativity we adapt. Google as a machine is allowed to make mistakes that humans can then correct.

https://www.independent.co.uk/news/long_reads/long-read-google-translate-interpreter-misunderstanding-human-mind-a8209851.html

If sections are translated by different individuals, there will inevitably be inconsistencies in the finished product, not only in terms of style, but also word choice, grammar and structure. No two translators are alike – even if they are reproducing the same writer’s work.

Furthermore, something may be introduced in one part of the document and then referenced elsewhere, with the assumption the reader will have seen the whole piece, which the translator may not have done.

 

I like also to acknowledge that these simple examples given through images were based on the language of WordPress and in particular in relation to WPMU DEV documentation of which personally I gather these examples, as a translator-polyglot myself of this community.

It’s purpose is to evaluate systems employed without going over the good or bad quality of translations. or notor criticize the works or  systems employed, but simply to analyze things from the side of the linguist and as a result, if possible at all, arrive at a good conclusion.

Part 1 and 2 is basically about the difficulties and issues that can occur when there are errors in the original texts that have to be translated in hundreds of languages by a community of wpmu dev members with various levels of linguistic knowledge and capacities.

No two translators are alike and this is the main reason why important texts media and official documentation need more than on translator and proofreader; if there are errors, missing words and so on in the original the chances are that in the case of technical terminology they’ll take guesses instead of asking questions.

 

This is nothing new to the community of professional translators. In fact is quite possible the number one problem or cause to so many being translations being done and accepted worldwide.

In Part 1, I briefly mentioned how important it would be to take into consideration spending more time and resources on professionally linguistics semantically and gramatically in proofing all the documentation before going to print; as this would have some advantages to translators.

Capitalization of certain words/leaves many doubts in wp?

Carrying on with these principles in mind part 2 of this study is to analyze closer the effects of language in its original form and as used for the creation of terms, titles and various programs plugins for wpmu dev.

 

Then follow on…

5 Are Professional Linguistics Discussed Or Taken Into Account In Proofing and Decision Making?

Language used…. Slang, jargon, spoken words

In Part 1, I briefly mentioned how important it would be to take into consideration spending more time and resources on professionally linguistics evaluation; by that I mean semantically and grammatically and in particular in proofing all the documentation before going to print. This, without going into the linguistically technicalities, could clarify and do away with many original issues for translators, such as:

  • spelling errors
  • vagueness
  • ambiguities
  • doubts and assumptions
  • guess work
  • the length of text
  • unnecessary capitalization
  • level of language used and so on

In Part 3 we will discuss this further from the viewpoint of the professional translator

 

 

If the original test in a document doesn’t make sense, is ambiguous or contains errors it shouldn’t be left to translators to guess the intended meaning(s), which in turn means that authors of these written texts and documentation etc, should also be made available for questioning and referencing, when in doubt.

In the case of technical texts as for WordPress, full of English jargon, chances are that these texts for themes, plugins and so many other documentation that allows and can be translated in hundred plus languages by the community of its members will vary considerably when comes to meanings and expected results.

Great advances in MT no doubt, but Not ”Google” perfect as one can see. More about MT:

There are still questions about the different usage of the verbs and the the timing that  the actions relate to. Some of the tenses used in the English (original language) lack of some clarity and cause us to question the validity of MT to the non linguist standard user. Moreover translation’s context has changed and text is now closer to the truth. User gets to read and makes comments of posts. Instead of user gets (to) be read and makes comments of posts.

In the next posts I will take some time to question, research and examine the ways (wordPress) and wpmu dev words are created and try to figure out, with real examples, how these words come to be accepted and translated by the international community.

 

Part 2 – Who Is Responsible To Write/Create This English Documentation for WordPress?

2 – What Sort Of Background And Knowledge Do these Professional Have In Linguistics?

3 – Do These Professionals Think Of The Language Complications That Can Occur When Odd Words need To be Translated?

4 – Do These Professionals Think Before Naming These Themes And Plugins i.e. ‘’Smush’’ ‘’Hummingbird’’ ’’Appointments’’ etc, etc.

5 – Are Professional Linguistics Discussed Or Taken Into Account In Decision Making

Part 3 – The Revision Process and Discussions that Follow Creation

1 – Who Makes The Final Decision For The Creation And Naming of Themes plugins

2 – Can Linguists Collaborate In The Naming Of These Themes And/Or Plugins What Is The Best Way To Translate The Untranslated

5 – What Is The Best Way To Translate ‘’shortcodes’’Errors In The Source Language

 

3 – The Role Of Translators/Proofreaders and How They View It All

One must keep in mind that many wpmu dev plugins, technically and linguistically can be quite complex to understand how they perform at these various levels. These are then translated by different individuals and agencies as members associated with the community free of charge.

There will inevitably be inconsistencies in these finished translations, not only in terms of style, but also word choice, grammar and structure. No two