| Pages in topic: < [1 2 3 4 5] > | Market for German-English legal translations Thread poster: Robert Hess
|
|---|
| Part of it is DeepL | Nov 22, 2022 |
I know several in-house translators at legal firms in this language combination who have been saying for years that their lawyers are putting client documents (!) into the free (!) version of DeepL and then sending the results to the translators to "quickly look over". The translators feel they're fighting an uphill battle if they can't even get lawyers, who really should know better, to understand why that is a horrible idea. For these lawyers, money isn't an issue because the translation would... See more I know several in-house translators at legal firms in this language combination who have been saying for years that their lawyers are putting client documents (!) into the free (!) version of DeepL and then sending the results to the translators to "quickly look over". The translators feel they're fighting an uphill battle if they can't even get lawyers, who really should know better, to understand why that is a horrible idea. For these lawyers, money isn't an issue because the translation would either be provided by the in-house translators or, if they have to farm it out, then it's the client's money they're spending. Instead, the key element for them is time. They need those translations now, and DeepL provides a translation now. (No one needs to jump in to start convincing me that "a translation" is not sufficient in this case.)
As to the original question, one of my specializations is DE>EN legal translation. I have not noticed a drop-off in work, but that is maybe because I don't work for law firms or agencies. I work mostly for public institutions, and data protection is such a big issue in the public sector that all my clients are still highly skeptical of DeepL. ▲ Collapse
| | | | Lingua 5B 
Bosnia and Herzegovina
Local time: 20:56
Member (2009)
English to Croatian
+ ...
| Not always the time | Nov 22, 2022 |
In some companies (I don't know about legal offices), they do get the budget for translation, but someone told them it can be done for free via DeepL.
If a lawyer needs it done now/immediately, ask them if they can draft a complex contract now/immediately (in a couple of seconds, like DeepL).
[Edited at 2022-11-22 12:16 GMT]
| | | |
Lingua 5B wrote:
In some companies (I don't know about legal offices), they do get the budget for translation, but someone told them it can be done for free via DeepL.
If a lawyer needs it done now/immediately, ask them if they can draft a complex contract now/immediately (in a couple of seconds, like DeepL).
[Edited at 2022-11-22 12:16 GMT]
The lawyer would then probably respond with something like: "But drafting a complex contract is difficult and requires a lot of legal education and research! Translating isn't hard, you just have to know two languages - DeepL can just do it faster than a person!" Again - you don't need to convince me of why that statement is wrong on several counts. I'm only reporting what I've heard from numerous colleagues. These lawyers' English is right at that sweet spot of being good enough to be able to tell that DeepL has produced a very fluent-sounding translation full of great legal words while also being poor enough not to see the problems in the translation.
[Edited at 2022-11-23 12:59 GMT]
| | | | Lingua 5B
Bosnia and Herzegovina
Local time: 20:56
Member (2009)
English to Croatian
+ ...
They view DeepL as a photocopier, just run the source through it, and you get a copy in a new language. This is how much it takes for a contract in a foreign language. In my language, however, it takes months of drafting.
Until they are at a meeting blushing from embarrassment while a foreign partner holds a DeepL artwork in their hands trying to make sense of things.
| | |
|
|
|
Going by my experience of other translators' work, I would say legal transation is an area where MT probably already does a better job than most human translators.
A properly drafted contract won't contain many ambiguities to fox the machine.
Most human translators, on the other hand, seem to struggle with more than one subordinate clause at a time. They might sometimes write more elegantly, but their understanding of the source is often flawed. They also take ages to d... See more Going by my experience of other translators' work, I would say legal transation is an area where MT probably already does a better job than most human translators.
A properly drafted contract won't contain many ambiguities to fox the machine.
Most human translators, on the other hand, seem to struggle with more than one subordinate clause at a time. They might sometimes write more elegantly, but their understanding of the source is often flawed. They also take ages to do the work and make human errors, as do reviewers. So it's no wonder people are turning to machines. ▲ Collapse
| | | | Tom in London
United Kingdom
Local time: 19:56
Member (2008)
Italian to English
Ice Scream wrote:
Going by my experience of other translators' work, I would say legal transation is an area where MT probably already does a better job than most human translators.
A properly drafted contract won't contain many ambiguities to fox the machine.
Most human translators, on the other hand, seem to struggle with more than one subordinate clause at a time. They might sometimes write more elegantly, but their understanding of the source is often flawed. They also take ages to do the work and make human errors, as do reviewers. So it's no wonder people are turning to machines.
It's time we got rid of people completely. They're just getting in the way.
| | | | Adieu
Ukrainian to English
+ ...
| Yes, but I don't see why you guys hyperfocus on DeepL | Nov 23, 2022 |
It's not the only one or even the best one.
There are a bunch of them.
And yes, it is slowly becoming a problem. Even despite the usual problems companies get when they try to integrate or upgrade these things.
For example, I have repeatedly watched one major client attempt to sneak away to some new semi-automated fad 1-2 times per year, then slink back with ruined (corrupted/missing/"poisoned") translation memories a week or two later and accept rate incre... See more It's not the only one or even the best one.
There are a bunch of them.
And yes, it is slowly becoming a problem. Even despite the usual problems companies get when they try to integrate or upgrade these things.
For example, I have repeatedly watched one major client attempt to sneak away to some new semi-automated fad 1-2 times per year, then slink back with ruined (corrupted/missing/"poisoned") translation memories a week or two later and accept rate increases.
On the other hand, they'll sadly probably eventually get it tolerably close to right. I would be surprised to get more than a couple more years out of them.
Unless of course they manage to screw it up so badly that they drop the idea for a while.
[Edited at 2022-11-23 12:54 GMT] ▲ Collapse
| | | | | Because that's what they use | Nov 23, 2022 |
Adieu wrote:
It's not the only one or even the best one.
There are a bunch of them.
And yes, it is slowly becoming a problem. Even despite the usual problems companies get when they try to integrate or upgrade these things.
For example, I have repeatedly watched one major client attempt to sneak away to some new semi-automated fad 1-2 times per year, then slink back with ruined (corrupted/missing/"poisoned) translation memories a week or two later and accept rate increases.
On the other hand, they'll sadly probably eventually get it tolerably close to right. I would be surprised to get more than a couple more years out of them.
Unless of course they manage to screw it up so badly that they drop the idea for a while.
This thread was specifically asking about the DE>EN English pair, and DeepL is a German-based company that delivers its best output in DE>EN according to everything I've seen (not just DeepL's own assessment). I've also looked at a few other systems' NMT outputs, and DeepL is by far the best I've seen in this language combination. In all of my contact with clients and colleagues, so far no one is mentioning integrating Google's or Microsoft's NMT into their workflows. Again, though, I'm sure this is specific to the DE>EN language pair, but that's why it's being mentioned here to the exclusion of other systems.
| | |
|
|
|
| Three month ago... | Nov 23, 2022 |
...the head of IT of the company I work for conducted a survey on who uses what cloud-based translation system (DeepL, GT or any other). I couldn't help trying to start a discussion with him about the dangers to use these systems for professional purposes when you are not able to understand the source or the target language at least passively to control the output the AI system delivers. He answered that it would be the responsibility of the user whether the output is right or wrong...aha.
... See more ...the head of IT of the company I work for conducted a survey on who uses what cloud-based translation system (DeepL, GT or any other). I couldn't help trying to start a discussion with him about the dangers to use these systems for professional purposes when you are not able to understand the source or the target language at least passively to control the output the AI system delivers. He answered that it would be the responsibility of the user whether the output is right or wrong...aha.
[Bearbeitet am 2022-11-23 17:21 GMT] ▲ Collapse
| | | | Adieu
Ukrainian to English
+ ...
I hadn't realized that DeepL had a main "native language".
No wonder many people sing its praises and many others are baffled.
Rita Translator wrote: Adieu wrote:
It's not the only one or even the best one.
There are a bunch of them.
And yes, it is slowly becoming a problem. Even despite the usual problems companies get when they try to integrate or upgrade these things.
For example, I have repeatedly watched one major client attempt to sneak away to some new semi-automated fad 1-2 times per year, then slink back with ruined (corrupted/missing/"poisoned) translation memories a week or two later and accept rate increases.
On the other hand, they'll sadly probably eventually get it tolerably close to right. I would be surprised to get more than a couple more years out of them.
Unless of course they manage to screw it up so badly that they drop the idea for a while. This thread was specifically asking about the DE>EN English pair, and DeepL is a German-based company that delivers its best output in DE>EN according to everything I've seen (not just DeepL's own assessment). I've also looked at a few other systems' NMT outputs, and DeepL is by far the best I've seen in this language combination. In all of my contact with clients and colleagues, so far no one is mentioning integrating Google's or Microsoft's NMT into their workflows. Again, though, I'm sure this is specific to the DE>EN language pair, but that's why it's being mentioned here to the exclusion of other systems.
| | | | | DeepL is best EN to GER | Nov 23, 2022 |
In my experience, DeepL - when used for legal translations - is better from English to German than the other way around.
The reason, it seems to me, is that most legal translations in DeepL's GE-EN database should be EN translations, since there is much more demand for that direction. That means that the GE documents in the database were mostly drafted by attorneys, whereas the EN translations were, to a large extent, done by non-attorney translators. Hence EN legal translations fro... See more In my experience, DeepL - when used for legal translations - is better from English to German than the other way around.
The reason, it seems to me, is that most legal translations in DeepL's GE-EN database should be EN translations, since there is much more demand for that direction. That means that the GE documents in the database were mostly drafted by attorneys, whereas the EN translations were, to a large extent, done by non-attorney translators. Hence EN legal translations from DeepL read a little funny to my eye, while GE translations look much better to me, at least as far as terminology and style is concerned. Even into German, DeepL still makes plenty of errors, of course. ▲ Collapse
| | | | Lingua 5B
Bosnia and Herzegovina
Local time: 20:56
Member (2009)
English to Croatian
+ ...
| What and how much was fed into it while the machine was being “trained” | Nov 23, 2022 |
I would say that EN>DE market is bigger (eg. annual translation volume) than the DE>EN. Therefore, the machine ate and digested more of that pair, hence more refined results. Just my observation, I don’t have official statistics.
| | |
|
|
|
| Demand depends on niche | Nov 23, 2022 |
I was referring only to GE-EN legal translations. In that market, demand is much stronger for English translations, as English is usually the contract language even if some of the contracting parties are from countries other than Germany, e.g., Belgium, Italy, or China. Plus most German corporate clients have no problem reading and understanding English, obviating the need for German translations in most cases.
But I realize, of course, that this may hold true only for my particula... See more I was referring only to GE-EN legal translations. In that market, demand is much stronger for English translations, as English is usually the contract language even if some of the contracting parties are from countries other than Germany, e.g., Belgium, Italy, or China. Plus most German corporate clients have no problem reading and understanding English, obviating the need for German translations in most cases.
But I realize, of course, that this may hold true only for my particular niche.
Robert ▲ Collapse
| | | | | I don't panic anymore because of DeepL, ... | Nov 23, 2022 |
...because here I have learned for now that, as soon as Germany finally has undergone a crisis like Brazil allegedly has experienced, as we learnt elsewhere here, the effects of DeepL will eventually disappear and business will be as usual.
[Bearbeitet am 2022-11-23 21:23 GMT]
| | | | Dan Lucas
United Kingdom
Local time: 19:56
Member (2014)
Japanese to English
| Some better news | Feb 24, 2023 |
Dan Lucas wrote:
My statement being that Germany is in recessionary territory? Okay. Well, let's wait and see. Neither the PMI nor the consumer confidence index counts as "hard" data (unlike, say GDP).
The ifo Business Climate Index was released a couple of days ago and rose slightly, which is positive. The ifo economist was quoted as saying that the German economy will not get around a recession, but it will be a mild one. If that is what happens it would be a good result for Germany and all who trade with her.
Regards,
Dan
| | | | | Pages in topic: < [1 2 3 4 5] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Market for German-English legal translations | Trados Studio 2022 Freelance | The leading translation software used by over 270,000 translators.
Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop
and cloud solution, empowering you to work in the most efficient and cost-effective way.
More info » |
| | Trados Business Manager Lite | Create customer quotes and invoices from within Trados Studio
Trados Business Manager Lite helps to simplify and speed up some of the daily tasks, such as invoicing and reporting, associated with running your freelance translation business.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news
