Separate meaning from language

6 February 2004

9 comments

A Blue Perspective: Separate meaning from language

The virtues of separating document structure from style are well documented, and in fact is the cornerstone of the XML/CSS pairing. The aim is to keep data entirely isolated from the way it is presented, allowing the data to be visually reformatted for different purposes without having to affect the data itself – as the CSS Zen Garden demonstrates well.

While trying to find out exactly what an Estonian web site was saying about me, it struck me that a similar division exists between meaning and language. Essentially when you speak/write you are communicating meaning using a semantic construct, such as the English language. "I'm hungry" has the same meaning whether you say it in English or Chinese, it is merely the way that the writer represents it that differs.

Just as style sheets allows us to alternately view XML as a graphical web page or as text on a palmtop, if we encoded our communications as pure meaning then we should be able to write "language sheets" that display that meaning in a particular language. Granted, such language sheets would be incredibly more complex than any existing style sheets, but it would give rise to a truly accessible Internet, with no boundaries to content whatsoever, irrespective of whether you speak only Ancient Sumerian or l33t. Could you imagine opening up Opera and visiting a web page with your custom language sheet, being oblivious to what native language the author might arbitrarily communicate in?

I know, it begins to get into Star Trek Universal Translator territory, and way too far into linguistics for me to actually appreciate the difficulty, but I see it as the last true barrier to creating a unified World Web (and Google's language tools just don't cut it), so energise, Geordi!

(If you're using your Original Star Trek Language Sheet that should read as "beam me up, Scotty!", but this page defaults to Star Trek: The Next Generation)

Categories

, ,

Comments

  1. 1/9

    Kaimar commented on 6 February 2004 @ 08:52

    Hi there

    The translation would be roughly as follows:

    10 tips on how to alienate people from other people. Tip no 1 - program a web-based Scrabble. But, actually, it's pretty cool. The game, that is.

    Regarding your thoughts on thought and language :) - underlying thought structures do differ in case of different language speakers, and some thoughts cannot be neither represented nor understood in some languages. It has been suggested that translation is impossible altogether, and I pretty much endorse that view - one's thoughts and writings can be adapted, but not translated to other languages. An adaption can be good or bad - but never precise. The reason I'm playing a smartass here :D is that I wrote my Bachelor's Thesis on the topic of thought and language and I'm studying it further at the moment for my Master's Thesis as well.

    And the Scrabble was cool indeed :)

  2. 2/9

    Keith Bell commented on 7 February 2004 @ 00:24

    I am Scottish, and currently I am living in Sweden. Having been born and raised in Edinburgh, my native tongues are English and Lallans, or "Lowland Scots" ( see http://en.wikipedia.org/wiki/Lallansand http://www.lallans.co.uk/). I have reasonable fluency in French, I speak rudimentary German, I am stumbling along in Swedish, and I have varying levels of understanding of Italian, Spanish and Dutch.

    The "language sheet" is a wonderful idea, but a dream I fear that will never be realised, whether you prefer the Star Trek technical type of solution, or Douglas Adams' more organic approach with the Babelfish...

    I can only agree with Kaimar's observations on thought structures, which holds true even within Europe, where we are not so very different from each other. And even without these subtle complexities, speakers of one language are often bound together by culture and tradition, as well as mother tongue. Consequently speech is highly seasoned with cultural references that cannot be translated easily, or would make no sense to a speaker of another language. Here in Sweden, television programmes imported from the UK and USA are broadcast in the original English, with Swedish subtitles. It amuses me to watch how the subtitlers struggle sometimes to convey the sense of what is being said, when the original speech is coloured with references to people, places, events and traditions that are simply unknown to the majority of Swedes.

    And literature... aah, literature! What would the "language sheet" make of James Joyce's "Finnegans Wake", I wonder? If you're not familiar with it, you can get a flavour of the book by reading The Tale of the Prankquean at http://www.english.swt.edu/cohen_p/avant-garde/Literature/Joyce%20folder/Prankquean.html
    For frequent thoughts on the challenges of translation from someone who does it for a living, I recommend Gail Armstrong's Open Brackets: http://www.openbrackets.com/

  3. 3/9

    The Man in Blue commented on 7 February 2004 @ 01:42

    Hmmm ... even I have trouble reading Joyce, and apparently it's in English :o]

    I can understand the impossibility of translating such creative prose as that, but surely more objective information must be able to be translated, otherwise human translators would be also rendered useless.

    Undoubtedly there must be judgement calls made by a translator upon the certain meanings and inferences between two languages. Such judgements should theoretically be reduceable to algorithms, but that also raises questions about the state of human/computer learning.

    I'm interested whether the profession of translation has progressed alongside technology, or whether it has remained fairly static over the years and relies instead upon each individual's own connections between languages?

  4. 4/9

    Pathfinder commented on 10 February 2004 @ 06:14

    Score: 1068

    Long dictionary version.

    Beautiful Game.

    Thanks a lot.

  5. 5/9

    sadie commented on 1 July 2004 @ 01:05

    Translating from one language to another is famously difficult, but what might be more realistic is writing straight into the raw meaning - probably represented as a complex tree structure - it could probably be renderable out as almost any language (if not very elegantly).

    It's comparable to trying to produce raw HTML for a given stylesheet, by feeding the computer an image of the page supposedly rendered by that sheet. Very tricky. But write the markup yourself, in a purely structures way, and it all starts to make sense.

    One side-effect of this would be to encourage simple, clear writing - the UI would make anything else naturally difficult.

  6. 6/9

    The Man in Blue commented on 1 July 2004 @ 04:20

    That's sort of what I was proposing, but the previous guys shot me down :o]

  7. 7/9

    Steve Learmonth commented on 5 August 2004 @ 16:10

    I know this is all hypothetical but I'm wondering how we would go about "encoding communications as pure meaning"? If, as sadie suggested, we decided to represent meaning with some sort of complicated tree structure, then haven't we just defined a new and different arbitrary language (as opposed to separating out meaning from language?).

    The author of "The Language Instinct" argues (convincingly I thought), that the mind has a language of its own ("mentalese") which is entirely independent of the language that we use to communicate. Perhaps if we had access to this raw, native data format that the brain uses for thoughts and meaning then we might be able to get somewhere... but that's not really my point. Uh, what was my point?.. Oh yeah, maybe meaning cannot exist without some form of language to represent it and thus the two can never really be separated in the way document structure can be separated from style.

    BTW, nice little place you've got here Cameron!


    BTW, nice little place you've got here Cameron!

  8. 8/9

    The Man in Blue commented on 5 August 2004 @ 16:53

    True, in order to encode "pure meaning" you would have to either create an entirely new grammar/language to encode it in, or subjugate an existing language, which kind of defeats the purpose.

    I think, in essence, what we are talking about is a standardsiation of language. The ideal would be if everyone spoke the same language. Perhaps that will happen in a few millenia, maybe not. But the next best thing is to have a standard language which can act as an intermediary between all the other languages, and which allows lossless translation of meaning between all other languages.

    You would then end up with a programmer-type class of people who understand this language and how to manipulate it.

    Actually, kind of reminds me of Neal Stephenson's "Snowcrash" which delves into the origins of language and brain "hacking". Good read.

  9. 9/9

    The Man in Blue commented on 5 August 2004 @ 16:56

    BTW, good to hear from you Steve :o]

  10. Leave your own comment

    Comments have been turned off on this entry to foil the demons from the lower pits of Spamzalot.

    If you've got some vitriol that just has to be spat, then contact me.

Follow me on Twitter

To hear smaller but more regular stuff from me, follow @themaninblue.

Monthly Archives

Popular Entries

My Book: Simply JavaScript

Simply JavaScript

Simply JavaScript is an enjoyable and easy-to-follow guide for beginners as they begin their journey into JavaScript. Separated into 9 logical chapters, it will take you all the way from the basics of the JavaScript language through to DOM manipulation and Ajax.

Step-by-step examples, rich illustrations and humourous commentary will teach you the right way to code JavaScript in both an unobtrusive and an accessible manner.

RSS feed