An Analysis Model for Multilingual Content Management
I've been doing a lot of thinking lately about agile translating and the future of proofreading and translating generally, all within the context of multilingual content strategy and publishing. It's clear that the industry is changing fast, but I think a lot of the innovation in globalizing written content is going in the wrong direction. Today, I came up with a graphic representation of why I think that is.
Anyone who's ever come across Integral Theory will recognize this four-quadrant graphic (which I threw together on my phone, so…), but it's easy enough to understand even if you've never heard of Ken Wilber or don't care for the guy. Essentially, the y-axis represents “individual” and “collective”. Here, I've labeled x-axis “connection” (inner, subjective, human connection) and "information" (outer, objective data).
In terms of the creation of written content, the light blue upper-left (UL) quadrant represents the inner world of the individual writer. The darker, blue-gray lower-left (LL) quadrant represents the human connection of the entire community engaged by a given piece of writing. The dark red upper-right (UR) quadrant represents information conveyed to an individual reader, and the pink lower-right (LR) quadrant represents systems whereby objective information is conveyed to a community of readers.
So what does all of this mean in more practical terms? Pretty much all writing will have aspects of each of these quadrants. Some will have been written primarily for the writer himself, or the connection with the writer will be of fundamental importance (UL). Other writing may be more about uniting a community around some thing or idea (LL), whereas other texts will be more about conveying information to an individual (UR) or group (LR).
Deciding what a piece of writing is all about, though, can help us to determine how that content can best be conveyed in a multilingual setting. Things that might fall primarily in the upper-left quadrant may include literature, poetry and journalism, because they tell a story or paint a picture that could only be created by that specific author, poet or journalist. To bring this sort of text to another culture in another language will necessarily call for a human translator as the intermediary between creator and reader.
In the upper-right quadrant, we have any sort of text that is purely intended to convey objective information from one person to another. The personal, cultural or otherwise subjective content is minimal, non-existent or not important. This sort of information can be easily converted from one language to another by way of machine translation, since getting the gist is all that's needed.
Moving on to the lower-left quadrant, we find things like marketing copy and anything else that is more about connecting a community around an idea, a cause, an organization, or a product. Here, the actual writer of the content will be less important than the idea being conveyed. I would argue that texts like these shouldn't be translated at all, but rather written in each language directly by a representative of that language's culture who has a vested interest of some sort in the community concerned.
Finally, in the lower-right quadrant, we might have things like company reports, financial statements, and other documents intended to convey objective information to a broad audience. Human connection is minimal since the primary goal is to convey information, not to nurture a community. These sorts of texts really don't require a human translator either and can be more efficiently handled by automated content-generating systems and artificial intelligence.
Where does this leave the human translator? Primarily working with texts in the upper-left quadrant, where the connection with the content creator is key. In these situations, a human is required to understand and intuit what the author or poet is trying to say, and the translator must do their best to convey this in another language while remaining an invisible conduit between creator and reader.
Of course, not all texts will fall neatly into one specific quadrant. Financial reports, for example, tend to serve a twofold purpose of conveying data and promoting the company's image. In this case, the vast majority of the report can be computer generated, but some parts may require either a human translator - such as for the letter from the chairman - or a human content creator or editor.
A great deal of time and resources are being poured into machine translation in the industry these days, but it seems to me that more money could be saved while keeping content quality high by taking the right approach for the specific content at hand. For example, an enormous amount of time and money is being wasted on humans writing and translating documents that could be done directly by a content generator with minimal or no human involvement. At the same time, trying to apply either machine or human translation to content intended to build human bonds of some sort is only short-circuiting or unnecessarily stretching this connection and is really not the most effective or efficient way to go.