I’ve been using computer-aided technology for translation since 1998. I use Transit, currently TransitNXT, from STAR. It’s a very good program and has many excellent features I am too lazy to learn about. It is also much hated by many translators. Anyway, it has helped my legal translation in many ways, but it has never saved time. Over the years I have often heard that such tools are useless for legal translation, but that was just something said by people who had no idea what they were talking about. My CAT tool lets me see source and target text at the same time and links to all the terminology I have stored (the STAR terminology database, Termstar, is the best I know).
Percy Balemans has a post on her blog in which she has gone to the effort of describing all the reasons she likes CAT tools (post dated February 2013 but still relevant today), so I am linking it here and can save myself the effort: The usefulness of CAT tools.https://pbtranslations.wordpress.com/2013/02/11/the-usefulness-of-cat-tools/
Maybe I’m admitting too much here, but here it goes.
While I use CAT tools whenever necessary, I don’t like them for legal translation, because, so it has been my experience, translations prepared in them read like they’ve been prepared in them. It is one of the most unpleasant experiences to subject oneself to. Whenever I prepare a legal translation in a CAT tool, I end up spending hours correcting transitions, intertextualities, and countless other things outside the CAT tool. It’s even worse when the translation has been prepared in a CAT tool by others.* This has not been my experience with translations not prepared in CAT tools, though there are other issues, but they are preferable to the issues caused by CAT tools.
The one thing that puts me off CAT tools in general is the segmentation carried out in CAT tools; the segmentation leads to the fallacious belief that the true and complete translation of each sentence engenders a true and complete translation of the entire document. Translators even talk this way (just listen to them); it drives me crazy. When asked about this or that translation, they say things like, “It’s not wrong per se” (the correct answer is: it’s not correct per se either) or “But the sentence was translated correctly.” No it wasn’t; the segment was translated correctly; the sentence means something else entirely. We’ve known for over 2,500 years that such reasoning is bad. It has a name: it’s called the fallacy of composition.
What is more: I hear over and over again (especially from CAT tool marketers/salespeople) that legal translations are repetitive, but no one has ever been able to tell me how; I’ve been translating legal documents since 2006. Repetitiveness might be true if one is working with one particular law firm for one of its particular clients on one of their particular matters, but not otherwise. Mostly, my work requires subject-matter specific knowledge, not terminology databases, although client glossaries are welcome and useful. To overstate the issue more than a little: clients generally expect that I know how to translate terms of art, they do not expect that I have to look them up in a database I’ve complied over my tenure as a translator. Of course I need to do research, and very often too, but once I’ve done it, there is a reasonable expectation that I have learned something and committed that something to more than just short-term memory.
And last, a large number of projects in legal translation on are from PDFs. I cannot rip the PDF in the CAT tool and deliver OCR-like formatting, as it would only lead to legacy formatting issues later on, for which clients are never thankful — especially (and this case comes up at times) if the translation is based on a prior agreement that exists only in PDF form where bits and pieces are intended to constitute part of a large cross-border transaction or litigation at the drafting stage. Additionally, most projects do not have turn around times that enable me to format the source text at the outset. That is the reality on the ground or at least that is the reality on the ground in my experience.
CAT tools are at best tools for process-oriented quality control measures able to be used for a small number of projects. At worst, they cause more work in the legal arena than they’re worth.
* Please don’t say: “Well, that’s because of your subcontractors.” It’s not. I see work product from lots of translators, freelance, in-house, etc. And all the CAT tool work product is to varying degrees unpleasant to read.
Well, we can disagree on this, can’t we?
I don’t discuss any issues with agencies or CAT rool developers so I haven*t had your experience there. I use a CAT tool only for myself.
I really don’t understand what you are saying about PDFs. If I get a text in a Word file, I preserve the formatting in Word. If I get a PDF, I use OCR but *not* to copy the formatting. I just OCR the PDF file without formatting and clean it up as a simple text file. I may change the font or basic paragraph formatting. I suppose you mean by ‘OCR-like formatting’ that you try to OCR the formatting, which results in many different font sizes and paragraphs. Well, don’t! In the days when I did more certificates, which have complex formatting, if I used OCR at all, it was only to import the text. There’s a lot of formatting work to be done there, but not for my usual legal translations.
I agree that repetition is not a big help in legal translation. For example, many German contracts and English contracts have a clause about supervening illegality: if one clause turns out to be invalid, this is not to make the whole contract invalid. One would expect that clause to be phrased similarly in different contracts, but it never is – the wording is always different, so I just translate freely, keeping close to the German original. As Percy Balemans writes, repetition is helpful when one client sends an amended version, or in my case, if I am working for one firm, its clauses may be similar.
If you prefer to work in a word processor, of course you can use search-and-replace, for example, to replace a German abbreviation by the name of the statute in English. You don’t need to see source and target texts for that. But if on reading through the exported document I want to change or revise, then returning to the bilingual display in TRANSIT is the ideal context.
Of course, we can disagree.
Re: OCR/PDF: I mean just what you describe. Cleaning up the formatting that results from an OCR is a nightmare, which also is a cause of legacy formatting issues in Word documents. The problem is the formatting code, which of course we don’t see, can get all jumbled, and this is not fixable. One has to copy the text from the OCR, paste it into another document stripped of all formatting, and reformat from scratch. This can be a time-consuming task, depending on the length of the document and the complexity of the formatting (lists inside lists with words and numbers, four or five levels of headings, etc.). Generally, turnaround times don’t account for this. Translating outside a CAT tool reduces the time and effort considerably; I simply format (using style sheets) as I translate.
Maybe you have to supply documents with more complex formatting than I do. I use an old version of Abbyy FineReader (11) and set it to ‘Text only’. Very rarely is there a problem with the kind of texts I do, The only big problem I have is if a client wants a contract to be in two columns, with the English on the right, and I mess up the automatic numbering. But that’s just me. I have taken to asking clients to do the two-column format themselves, and I just supply one column. After all, it’s the translation they want.
I do have some projects with complicated formatting.
As it happens, I know how to use lists in so-called bilingual translations: the trick is to use style sheets in Word. You need to duplicate each style, one for the SL, one for the TT — create one list template and duplicate it for the other language, and you’re good to go. Just make sure you have content (i.e., words) where you need them before you apply the list for the first time.
Normally I am given a grid to enter things into, but I sometimes have problems with the automatic numbering, which has to be set to start again in the right-hand column. I don’t know why – it may be because I use an older Word verison. Anyway, I hate Word almost as much as you hate CAT tools.
As for the complex formatting, since you don’t want to use CAT tools, you obviously don’t have to. If I have very complex formatting *and* a significant amount of text in between, I do have to apply the formatting myself before using the CAT tool. My decision if I want the text to go into the memory.
Well, I believe CAT tools are useful in nearly every field. If we create good termbases and translation memories, they will certainly save us time in the long run. Not only search time, but even typing time, for those who don’t use Dragon speech or similar solutions.
I do encourage translators to learn how to use CAT tools and learn how to create and organize termbases and translation memories. They make a big difference in the long run.