I have recently again been asked by an agency for my rate, which they require expressed “per source word”. For those of you who can read German there was a good discussion of this matter at the site of agency Intra. It is no longer hosted there, but I got a copy from the wayback machine, and put it here.
Let me say that I do understand that agencies need to know, in advance, what a freelancer is likely to charge for a given translation. We freelancers are, I think, sensible if we reserve the right to modify this on site of the actual text, just as we really must reserve the right to decline any job when we have actually had an opportunity to see the relevant files. But it is a commercial reality that agencies need to be able to estimate what a particular translator is almost certain to charge for a given job.
There was a time when the publishing world thought of the “word length” of, let’s say, a newspaper article as a purely nominal estimate, based on a double-spaced typed page, with an average of 10 (English) words per line, 30 lines per page. The typewriter wrote at 10 characters per inch, and that was that. Because the actual, precise word count is constantly available from your word processor, people have adopted the idea of using this as a measure of the “size” of a given text. But, as any agency with any experience of German should know, this is a very bad idea when it comes to measuring German. The problems are easy to see.
Firstly, I leave aside the issue of quoting for work simply on the basis of the quantity of text, ignoring questions of complexity, subtlety or style. There are things to be considered there, but for this article I will assume that we are dealing entirely with run-of-the-mill texts, typical for what a particular translator or particular agency handles.
The most obvious point is that German is positively famous for its long collocations, written without spaces to form a single word. People who don’t know the language sometimes think that it is a particular difficulty of German: “oh, all those long, long words.” In fact, of course, it is just something you quickly get used to. Let’s take a couple of examples.
We begin with something very simple: the source text says “Ich bin Alex”, and our translator summons their skill to convert this to English: “I am Alex”. Three words, in each case, and no obvious problem.
Next, we are translating a patent, written in German, for a new and inventive hairbrush. As is typically the case with patents, the document will probably not call it a “hairbrush”, to exclude the possibility that a competitor will come along with something very similar and say: “oh, but ours isn’t a hairbrush, it’s a multi-row comb, so your patent doesn’t cover it.” For this reason they will be likely to have an expression such as, “a hair alignment device, such as a hairbrush or comb”. It sounds comical, but there are good reasons for it. So our German text refers to a “Haarausrichtungsgerät”. As I say, not as strange as it might look, because “Haar” refers to our hair “Ausrichtung” is alignment, and a “Gerät” is a device. (This is, by the way, a joke – as far as I know “Haarausrichtungsgerät” is not a real German word. But it could be!)
Now if we measure these texts by the “target word”, things are bad enough. If we do that, we are saying that the amount of work, skill, and so on involved in translating “hair alignment device” is exactly the same as that involved in translating “I am Alex”. Clearly not true. If, however, we base our charging on the number of characters, we have 21 characters as against 9 characters, and we would be saying that the translation of the long phrase is worth a bit more than twice as much as the short phrase. You might argue that the factories too high, you might argue that it is too low, but at least it is a factor, and it recognises the fact that the first phrase is longer and more complex than the second.
If, however, you want to measure by the “source word”, things are even worse. For the reasons described above, agencies will naturally want to work on the basis of the source text, because they want to know in advance what you are likely to charge. Commercial reality again – in its way, it’s fair enough. But what happens here is that we end up saying that “Haarausrichtungsgerät” is only one word, and therefore that the translation of “I am Alex” is three times as difficult, and justifies a payment three times higher than the translation of “hair alignment device”. This is no longer merely “bad enough” – it is downright ludicrous.
The solution is simple. With one click of the mouse I can see the number of characters in my text. When working in Word it’s at the bottom left-hand corner of the window. “Characters (with spaces)” is what it says. You can use this directly, but the numbers tend to be large. More conveniently, and conventionally, you can divide it by 55 and get the number of “standard lines”. Asking for advanced pricing based on counts of the number of German words is highly unsatisfactory. To my mind it suggests that an agency who wants to do this may be lacking in experience.
Taking about 1 million recently translated words from my database, I see that on average I write almost exactly five English words for every four German words, but the length in characters tends to be usually within 1% or so of the source. In fact my translation is more often marginally shorter than the original than marginally longer. The problem, of course, is not that the average German word is longer than the average English word, but that the spread of lengths is much higher, and that the really long words tend to occur in texts that are, in any case, more complex and difficult to understand and require a more specialised vocabulary to express in English.