I did some months ago intend to write something about my experience of using corpora for translation purposes, especially legal translation. (See earlier entry and footnotes by John Kuti there)

At that time, it appeared the free programs I might have recommended had lost their value for me because they had access to fewer corpora.

Then again, one could get fairly similar results with a Google CSE (custom search engine).

I followed a webinar on the topic last year, and it ended with a contribution by Juliette Scott, who is a legal translator who is doing a Ph.D. on the subject. She now has a weblog, called Translation & the Law: from words to deeds, which is certainly a good place to find out more.

There was also a blog post by Kevin Lossner in Translation Tribulations, entitled A NIFTY method for legal terminology (I thought NIFTY was a play on NIMBY, but I found out it is the name Juliette Scott gives her method) – Here you will find the links to use if you want to help in Juliette’s research/find out more.

From a real-life seminar in London a couple of years ago I also have a most wonderful and useful book on the subject of corpora; Working with Specialized Language: A practical guide to using corpora” by Lynne Bowker and Jennifer Pearson (dated 2002 but still useful in 2012: you can look inside at amazon)

The basic approach to making a corpus of legal texts is to collect them on the internet or from other sources and convert them all into a format readable by the corpus program. This takes a bit of time. It also raises copyright problems unless you just use it for your own purposes. This was the problem with the free software BootCat, which had lost the right to use certain sources from the Web. The free software AntConc is for a later stage of the process.

Here’s an article by Michael Wilkinson: Compiling Corpora for Use as Translation Resources

I did have some rapid success in one field of legal English late last year. I sometimes translate lawyers’ websites and also extracts from directories in which law firms are described in glowing terms. Here’s an example from a firm I have nothing to do with:

CMS Hasche Sigle
Aufbruch in eine neue Zeit – und zwar mit Schwung. Unter dieses Motto könnte man das vergangene Jahr bei CMS stellen. Schon lange gehört die Kanzlei in Hamburg zu den führenden Adressen, jedoch monierten Wettbewerber, CMS sei zu breit aufgestellt, um im Markt wirklich hervorzustechen.
Diese Zeiten gehen zu Ende: V.a. die M&A-Praxis hat zuletzt einen deutlichen Schub erhalten und sorgte für Schlagzeilen, als ein Hamburger Team zusammen mit dem internationalen CMS-Verbund Takeda bei dem €10 Mrd schweren Erwerb von Nycomed beriet. Dies spiegelte sich auch im Markt wider, die Gruppe erntete in diesem Jahr spürbar mehr Lob. Gemeinsam mit Dr. Marc Riede betreute er zudem die HSH bei der Restrukturierung von Hapag-Lloyd.

It’s quite easy to collect this kind of thing in English from UK, USA and other sites and to search it for useful expressions. I might find more ideas for words like betreuen.

But I still have the feeling that a corpus would not help me with most legal translations, because I am not trying to create a text that looks like it was written in English about English law, but one that is clearly about a foreign legal system. If I created a collection of contracts, for example, every potential match of phrase would need to be checked legally to see if it meant the same thing. I have the feeling that I’d love to computerize my vocabulary work, but it would then bypass my own brain and experience.