What’s your word-to-segment ratio?

Ever translated a legal document?  Did you find it harder than texts in other fields? Specialist terminology and background knowledge aside, there’s one fundamental aspect of legal texts that do make them harder to translate: a higher word count per sentence.

I ran a quick analysis of some recent translation projects I did, and I found that the word-to-segment ratio is significantly higher in legal texts than in other type of text, whether it be a technical manual, a stock exchange analysis or a marketing copy.  In all legal texts the average number of words per sentence was above 13. The average count was between 14 and 15. And if discount the fact that every legal text includes some very short segments (section titles such as “Scope” or “Miscellaneous” or addresses of signatory parties), the average number of words per actual sentence is even higher.

By contrast, the same ratio in technical manuals was somewhere around 7 or 8; whereas in GPS navigation software it was under 3.

Just why is it more difficult to translate texts that consist of longer sentences?

For starters, legal texts tend to have rather complex sentence structures. It is not rare to see several layers of embedded  subordinate clauses. This may be hard to translate into any foreign language, but when the target language has a fundamentally different inherent logic than the source language (think English to Hungarian), the building blocks of the sentence have to carefully selected, and often twisted around to create a grammatically correct sentence that also happens to convey the same meaning as the source sentence.

Even if an English source sentence does not appear to include any subordinate clauses, the corresponding Hungarian sentence may have to be constructed with 2 or 3 clauses to make sense and to be grammatically correct. (More on this in a separate future blog post.)

Legal texts frequently include long lists within a sentence. Such list may include nouns or nominal structures only. These are usually easier to translate than list of verbs or verbal structures. (More on this in a separate future blog post.)

A unique difficulty in translating legal texts from English is the difference between the British and American legal systems on one hand, and continental European legal system on the other.  Often there’s no direct correspondence between an English legal term and a continental one, which means some imperfect solution has to be found to circumvent the problem. (More on this in a separate future blog post.)

All this leads me to the conclusion that it would be fair to charge higher-than-the-average rates for legal texts to make up for the additional time required to process complex sentences.

Here’s my take on introducing a rule-of-thumb for charging extra for legal texts:

Run an alaysis of your source text. Each CAT tool can quickly count the number of segments and the number of words. Divide the word count by the segment count. If the result is over 10, aim for 5% higher than your regular rate. If the ratio is over 12, aim for 10% higher, if the ratio is over 14, the fair rate would 15% higher, and above 16, you may want to charge 20% above your regular rate.

Disclaimer: I haven’t tried this myself, but I will try to do so with a new client if a complex legal text comes up next time.


About bancsaba

Discovering the world, step by step, word by word
This entry was posted in Uncategorized and tagged , , . Bookmark the permalink.