Controlled Authoring

tumblr_lptmh1EY1E1r1sle6o1_500I recently attended a seminar on Mechanical Translation, not because this is something I’m particularly concerned with per se, but because I always  find it interesting to hear someone talk about a topic they are especially enthused by.  In case this is a concept you’ve not previously encountered it is very much what it says on the tin.  You type into your computer whatever it is you want to translate and the engine you’ve called up offers you a translation.  If you want to try it out you can experiment with the engine our students  most commonly use at Google Translate.

Of course, this is a pretty rough form of translation and probably responsible for all those sets of instructions you’ve tried to follow over the years with about as much success as trying to build a snowman in a heatwave.  However, apparently you can minimise the errors in the translation by assuring that the original text is written using controlled authoring.

Now this was a concept completely new to me, at least in this context.  What it means is that you write the original piece obeying certain rules that will make it easier for the machine translation program to translate it.  So, for example, if the original text is in English no sentence should have more than twenty-three words.  What is so magical about the number twenty-three I have no idea, but that’s the rule.  Neither should you employ the passive.  All your sentences should be in the active voice.  No the cow was milked by the farmer then, just in case it gets translated as the cow milked the farmer,  which would be fun but unlikely.

While I may not have come across controlled language in this context, however, it isn’t a concept with which I am completely unfamiliar.  In another life I was very much involved with teaching children to read and evaluating the reading schemes that were used in schools and there, let me tell you, controlled language was the order of the day.  Each scheme would have its own set of rules which dictated what could and couldn’t appear in the books at different levels of so-called difficulty.  I say so-called because a lot of the difficulty came about because of the very rules that were being imposed.

So, you might have a scheme built around controlled vocabulary,  which, in the instance I’m thinking about meant using only the words judged to be those most commonly found in written language even if that resulted in a sentence no one in their right minds would ever use and no child would ever predict.  For example, when telling a dog he can’t come on a walk we have:

No, you cannot come. You cannot come.  You must be at home.

No, you can’t use stay.  It just isn’t common enough.  I know it’s logical, but it isn’t the rule.

Many of these schemes also insist on an unnatural repetition of words, presumably to hammer the idea home.  This is what leads to such gems as:

Look, John, look.

Come John, come.

I have friends whose party trick is to read such texts out aloud as dialogue.  It’s great fun.  You can make them sound pornographic enough to warrant an X rating.

There are dozens of other schemes each obeying an entirely different set of rules and each insisting that their rules are the only ones that should be followed.  However, my own personal favourites are those that are built round spelling patterns.  So, for instance, you may have a book which is built as far as possible around the pattern -ack and therefore contain all the words the authors can think of that end in that letter combination.  This comes to mind because this is the book that has a picture of a horse in it and a word beginning with used to describe what is in that picture only the word isn’t horse, it’s hack.  In forty years teaching I never found a child who read that accurately the first time round, nor one who knew what a hack was when they did read it correctly.  Even better was the book for five year olds introducing them to words that began with p and which contained that classic line:

A quaking punter parted from his punt, met queer pond fish in Queen Pam’s park.

I can’t begin to tell you the problems that that particular sentence caused in some of my classes with a high immigrant population.  But I’m sure I don’t need to; you can imagine them for yourselves.

You will have gathered then, that I am not a great fan of controlled authoring when it comes to material used for teaching children to read.  Oh yes, you apply common sense, but hard and fast rules, no.  Although, of course, such schemes do come up with some gems that might never otherwise have reached the printed page.  Fortunately none of the children I taught ever saw the possible double meaning in this, but my students used to have hysterics.  We are concentrating on the spelling rule that says when you have a three letter consonant – vowel – consonant word you double the consonant when adding a suffix.

Dan slipped off Flash and hugging Fred

Went hopping, skipping back to bed.

Fred said as Dan bounced up and down

“Dan is the happiest man in town.”

Don’t you just wonder how any of us ever learnt to read in the first place?


7 thoughts on “Controlled Authoring

  1. I find this article very fascinating, yet I remember how many twists and turns my own reading lessons took; I am still (at the age of 55) learning the odd word here and there, and I suspect that there is no real limit to what we learn over the course of a lifetime, as long as we are willing to keep reading and revising our previous understanding(s). Thank you for the insight on Google Translate, too. Back in about 2009, there was a fun site on the Internet called “Translation Party.” On it, you went and typed in a sentence or some poetry, and it was translated back and forth from English to Japanese repeatedly until it reached something called “equilibrium.” This didn’t mean, however, that at that point it made sense, it just stopped. Sometimes the results were hilarious, and I used to do it for fun. Finally, however, the researchers who invented it had had enough and made it (if I’m not mistaken) into a part of a translation site for real, and the fun was over. Too bad; it passed many a dull moment.

    1. I must ask the students if they ever came across ‘Translation Party’. Because translation isn’t my field I missed out on it, which is a great shame; I could have had tremendous fun with it.
      Of course, you’re right about learning to read being an on-going process. One of the great problems with the controlled authoring process is that all it really does is teach children to decode, which is not the same thing at all.

  2. This was so interesting! I’ve resorted to Google translate myself at work, when I have to write a letter overseas – just to find a word or quick phrase (often the local version of “I’m sorry I can’t write this whole letter in your language”), and I used to use the Babel fish site, which seems to have disappeared.

  3. This was fun, it reminded me of the Dick and Jane books when I was a kid. There would be sentences like “See Dick run. Run, Dick, run!” Babel Fish is still alive and kicking it’s my go-to online translator when I need to look up a word or phrase.

    1. Oh yes, we had Dick and Dora over here too. Thanks for the information on Babel Fish. I’m sure it’ll be useful to those who’ve been missing it.

  4. Alex wrote: However, apparently you can minimise the errors in the translation by assuring that the original text is written using controlled authoring.

    Yes, you can, but not with online translators such as Google Translate.

    Typically, industrial-grade automated translation uses both a controlled language and a customized translation engine. For example, refer to

    To get satisfactory translations with free machine translation tools such as Google, the best option is to use the guidelines from ‘The Global English Style Guide’ (

    ‘Evaluation of international English and machine translation’ shows the quality of translations that you can expect from Google Translate if you optimize the text for machine translation (

Your thoughts are welcome

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s