Pages in topic: < [1 2 3 4 5] > | Voting just for "quality of writing" is totally unacceptable Thread poster: Roland Nienerza
|
Annika Persson wrote: As far as I understand it, it's the average score that counts, not the number of votes. So if 10.000 people vote on quality of writing, and only five people vote on quality of translation, we still get an average for both. And since the determination of the quality of writing should not be mixed up with that of the quality of translation, I don't see how voting for either only one or both could have a negative effect on the average score. Am I missing something? The thread is about the idea, advanced in the very laudable intention to increase voting, that the entries might be "valued" on the merit of "quality of writing" only - "regardless of source". If this would be practised, and the discussion shows that many like the concept, entries that wish to respect the source text will in most cases have little chances against entries, that consider the source text just as very vague source of inspiration, without any distinct value of its own. Voting "regardless of source" means that a nice text about daffodils could win in a contest for a source text relating to wild water rafting.
[Edited at 2008-12-09 19:58 GMT] | | | You might try reading the thread more carefully | Dec 9, 2008 |
Roland Nienerza wrote: Voting "regardless of source" means that a nice text about daffodils could win in a contest for a source text relating to wild water rafting. That's complete nonsense. As has already been pointed out, in the later round only those working in that pair can judge the translation, and surely the source text will be taken into account. However, if by that point the badly written target texts are weeded out, that's certainly a good thing. Roland Nienerza wrote: A translation that is besides or above the source text is a bad translation. That would depend entirely on the purpose of the translation. And are you sure you are able to judge this with great certainty? I'm still wondering where you came up with "into". | | | 10.000 votes for quality of writing, none for translation? | Dec 9, 2008 |
Roland Nienerza wrote: The thread is about the idea, advanced in the very laudable intention to increase voting, that the entries might be "valued" on the merit of "quality of writing" only - "regardless of source". But that only applies to a fraction of the votes - I am pretty certain that an entry cannot win as long as it has not received votes in both quality of writing and translation. So I still don't see the problem. Does the "quality of writing" score weigh heavier than the one for translation? I don't think so (but feel free to correct me if I'm wrong). | | | It is rather - | Dec 9, 2008 |
Kevin Lossner wrote: You might try reading the thread more carefully - that you might read the thread more carefully. - Because I have started it. | |
|
|
Cagdas Karatas Türkiye Local time: 06:47 Member (2007) English to Turkish
Kevin Lossner wrote: That's complete nonsense. As has already been pointed out, in the later round only those working in that pair can judge the translation, and surely the source text will be taken into account. Well then, Kevin. How can you be *sure* that the source text will be taken into account at any stage of the contest? IMO, what one can be *sure* about is that the source text seems to be the part taken into DISCOUNT by voters or taggers. People can't help being attracted towards the ornament/ornamented (a common human behavior, should we say?), and they don't and won't care about what has to be cared about unless they are compelled to do that. One more thing; could you please clarify what you mean by complete nonsense? | | | Heinrich Pesch Finland Local time: 06:47 Member (2003) Finnish to German + ... Lets take an exemple | Dec 10, 2008 |
I probably should not again participate in this discussion, but the clock is three in the morning and the morning paper hasn't arrived yet... I had a look at the EN-DE-contest last week. There were 40+ entries displayed for rating, a quite impossible task. I started from above, read the source text and the first two translations. Then I had a look at the comments. I did not like the source text very much. It was about a man who had given up his job to be ab... See more I probably should not again participate in this discussion, but the clock is three in the morning and the morning paper hasn't arrived yet... I had a look at the EN-DE-contest last week. There were 40+ entries displayed for rating, a quite impossible task. I started from above, read the source text and the first two translations. Then I had a look at the comments. I did not like the source text very much. It was about a man who had given up his job to be able to look after the kids, while the wife was making a career. The narrator in the first person singular told, how he arrived at the airport in the hope to see some familiar faces. But there were none, so he started to cry. I did not like the attitude of the author towards the subject. But that is no important here. There were a lot of "difficult" points, and the comments dealt with these. The airport was said to be the only place where you see police with "guns". The translations differed a great deal, some had written "Gewehr", others "Pistole". I agree, that "Schusswaffe" would be the correct translation here, but is it the best choice in this context? What if someone would have translated "Kanonen"? These are guns too, but usually they shoot ammunition some 20 km towards the enemy. But I remember when I was a youngster, police "die Bullen" were carrying "Kanonen". So what is a good translation here? The next stumble stone was Tie Rack. The police was said to prevent terrorists from blowing up Tie Rack. One German translator had written "Krawattenladen", others "Tie-Rack". So raters had commented on that saying that it must be "Tie Rack" without the hyphen. Relevant? Towards the end it was mentioned, that the narrator used "spreadsheets" to keep book about baby food. This was a problem again. I liked the translation which read in German "Buchführen", but there were lots of other solutions: Excel-Tabellen, Spreadsheet etc. Probably the best translation would be "Tabellenkalkulation", but does it fit here? So we cannot discuss on a general level but have to stick to the actual case. A good translation should never look like a translation. Do we agree? If the translation should no look like a translation, the best translation is one that seems to be written by a native of the target language. So I believe a native of the target language is quite able to rate an entry solely according to the result without being able to understand the source or without looking at the source. Rating the quality of the translation is much more difficult. I don't expect professional translators to come to a common conclusion about this point. The one likes the daughter, the other the mother. ▲ Collapse | | | Roland, please consider "small" languages | Dec 10, 2008 |
I think it is a good idea to encourage people that are native in language B to help evaluating entries in pairs from X to B, especially in "small" pairs. For example this time, there were pairs with only a few entries, due to the limited number of translators that work in that direction. Probably there are even less "native B" among them, so I think in the qualification round it is helpful if "native B" people can take a look at the target texts, and evaluate/comment on the grammar, style,... See more I think it is a good idea to encourage people that are native in language B to help evaluating entries in pairs from X to B, especially in "small" pairs. For example this time, there were pairs with only a few entries, due to the limited number of translators that work in that direction. Probably there are even less "native B" among them, so I think in the qualification round it is helpful if "native B" people can take a look at the target texts, and evaluate/comment on the grammar, style, etc., strictly based on the target text only. Comments posted by these evaluators are taken as an input (but not the only one) by peers and used by the vote-counting committee for calculating the average points, and I think it can help forming a better picture about the translation. In addition, the authors of the submitted translations can learn from the comments. I am not sure how the calculations work, but I am pretty sure the two categories ("accuracy of translation" and "quality of writing") are averaged separately, so even if the translation is a beautiful essay about the fragrance of roses, and it attracts high marks from people without a clue about the source text (which discusses organic fertilizers, by the way), that may increase the average for the quality, but it will not push this entry into the winning position because obviously the low points in the other category will have the necessary counter effect. Consider the other extreme: if entries in language pairs with a very limited number of people native in the target language are evaluated mostly by non-natives... (There are a good few pairs in this current contest that fit this category.) What do you think may happen? We may end up with winning translations that are absolutely dreadful to read... I think there were a lot less restrictions in the past on who could vote, and such things happened - hence the new methods. In language pair with a large population, I assume there are plenty of votes, so if someone rates only the "quality of writing", that rating has probably even less effect (positive or negative) proportionally. Just my 2 cents Katalin
[Edited at 2008-12-10 03:52 GMT] ▲ Collapse | | |
Çağdaş Karataş wrote: One more thing; could you please clarify what you mean by complete nonsense? The extreme example cited by Roland is unlikely to the point of absurdity, that's what I mean. His academic, absolutist approach to translation is very much in order for certain kinds of translation - the patent work I'm involved with is an example that comes readily to mind. However, in the context of literature, I find such an attitude absurd, and in many other types of translation, it is simply counterproductive. The need to improve on a text has already been discussed adequately in the context of textbooks or technical literature for use, but with literature there are many examples where cultural adaptation that involves rewriting the "precise" sense of the original is necessary to make sense to the audience. Tell me, Çağdaş, has a louse run over your liver this morning? If that question doesn't make sense to you, it's because I have translated it "precisely" out of Roland's native language. In my native country it would only get you a strange look. The book I cited above deals with these issues in much greater depth in a rather engaging way; I would recommend it to anyone who wants a thoughtful exploration of the subject rather than rigid dicta. Roland's call for precision lost some credibility with me too when he tried to inject more "precision" into an example in English, only to suggest an error.
[Edited at 2008-12-10 10:31 GMT] | |
|
|
"Quality of writing" and "Accuracy of translation" are averaged separately | Dec 10, 2008 |
Katalin Horvath McClure wrote: I am not sure how the calculations work, but I am pretty sure the two categories ("accuracy of translation" and "quality of writing") are averaged separately, so even if the translation is a beautiful essay about the fragrance of roses, and it attracts high marks from people without a clue about the source text (which discusses organic fertilizers, by the way), that may increase the average for the quality, but it will not push this entry into the winning position because obviously the low points in the other category will have the necessary counter effect. Right, Katalin. | | |
Katalin Horvath McClure wrote: I am not sure how the calculations work, but I am pretty sure the two categories ("accuracy of translation" and "quality of writing") are averaged separately, so even if the translation is a beautiful essay about the fragrance of roses, and it attracts high marks from people without a clue about the source text (which discusses organic fertilizers, by the way), that may increase the average for the quality, but it will not push this entry into the winning position because obviously the low points in the other category will have the necessary counter effect. Just my 2 cents Katalin
[Edited at 2008-12-10 03:52 GMT] Your 2 cents are as good as many a dollar. - "but it will not push this entry into the winning position because obviously the low points in the other category will have the necessary counter effect. " - as long as low points in the other category are given. Çağdaş Karataş, apparently my only reliable ally in this, has put very precisely - "what one can be *sure* about is that the source text seems to be the part taken into DISCOUNT by voters or taggers. People can't help being attracted towards the ornament/ornamented (a common human behavior, should we say?), and they don't and won't care about what has to be cared about unless they are compelled to do that." What you say, Katalyn, about the less popluated pairs is not so much different from what I saw in the more populated ones. The forum discussions are rife, it is literally all over the place, with complaints about stunningly bad entries, that often make it into the medals. For me a good solution would be that every vote has - compulsorily - to be given in both categories -Accuracy and Quality. - If a person is not qualified to vote for accuracy of translation, because he or she does not even know the source language, he or she might be qualified to vote in a beauty contest of whatever kind, but not in a translation contest. | | |
Heinrich Pesch wrote: The one likes the daughter, the other the mother. A family man and father should like both. - Others might like both two, but for different reasons - iykwim. Heinrich Pesch wrote: I had a look at the EN-DE-contest last week. There were 40+ entries displayed for rating, a quite impossible task. This is indeed an issue. - We all know, at least we should, that this is mainly about fun. But in a very curious manner in goes to the very essence of the business. The contest being fun, one should assume that filing a perfect little piece should be nothing more than a nothingness. - But, as you know, particularly in Ge, but the discussions show that in other languages things are not so different, to write an absolutely flawless little entry is trickier than it seems. Sometimes most things are nice, but a typo or two are left. At other times, no typo, no spell hitch, but grammatical, phraseological, factual or other humps. And the result is that looking at these entries, one sees invariably light and shadow merrily intertwined - and rating is really difficult, and then one sees that this might become rather a job instead of fun, and has to capitulate in face of more than 40 entries. - Straying in some tags, perfectly at random while perusing the submission is what most can afford to do. Heinrich Pesch wrote: I did not like the source text very much. It was about a man who had given up his job to be able to look after the kids, while the wife was making a career. The narrator in the first person singular told, how he arrived at the airport in the hope to see some familiar faces. But there were none, so he started to cry. I did not like the attitude of the author towards the subject. But that is no important here. You are right - it does not matter here. I also saw the apparent discrepancies in the person's behaviour. But I see too that he is in a dismal fix, so much so that one would wish for him or any person in his situation not to do more harm to himself than to ridicule himself in an airport entry hall. Heinrich Pesch wrote: There were a lot of "difficult" points, and the comments dealt with these. The airport was said to be the only place where you see police with "guns". The translations differed a great deal, some had written "Gewehr", others "Pistole". I agree, that "Schusswaffe" would be the correct translation here, but is it the best choice in this context? What if someone would have translated "Kanonen"? These are guns too, but usually they shoot ammunition some 20 km towards the enemy. But I remember when I was a youngster, police "die Bullen" were carrying "Kanonen". So what is a good translation here?. Right again. A word like nothing. "Guns". Anyone should be able to translate that. But are they? - You pretty well resumed the situation. And you are right, that a very good approximation for Ge would have been "Kanone". - But. That is the term for the revolvers Ge police have dangling on their waists. Even those who, like me, never "took off" from a British airport could know the pictures - one saw them recently again with the Terminal 3 calamity - of the British cops in short sleeved white shirts, black flak jackets and carrying the typical short barrelled submachine or tommy guns - in Ge "Maschinenpistolen". - That would have been the word - although, to be sure, it was not really spelled out like that in source. Heinrich Pesch wrote: The next stumble stone was Tie Rack. The police was said to prevent terrorists from blowing up Tie Rack. One German translator had written "Krawattenladen", others "Tie-Rack". So raters had commented on that saying that it must be "Tie Rack" without the hyphen. Relevant? Yes and no. It would not be relevant in a 10k translation job. In a contest it might be. I saw even for "Tie Rack Boutique für Herrenaccessoires" - the Forum rules forbid me to say in which piece, a minus for the lack of hyphens, calling this Deppenleerzeichen, with a Wikipedia link, which explained that, and made clear that the term was derogative and not acceptable. Heinrich Pesch wrote: A good translation should never look like a translation. Do we agree? No, we don't. - A good translation should never look like a bad translation. And the good thing is that a good translation never will. Heinrich Pesch wrote: If the translation should no look like a translation, the best translation is one that seems to be written by a native of the target language. Native is not automatically A-language. This has been discussed in the third thread before this one here. - The proof of the pudding is the eating. Heinrich Pesch wrote: So I believe a native of the target language is quite able to rate an entry solely according to the result without being able to understand the source or without looking at the source. Able to rate what? - That a piece is written by a native or A-speaker of the language or a person with comparative competence and that it is a nicely written text? - Most articles in "Der Spiegel" are - but hardly any one is a translation. Heinrich Pesch wrote: Rating the quality of the translation is much more difficult. I don't expect professional translators to come to a common conclusion about this point. Oops. - If that was true, there could not be a translation contest at all. But it is the curious reality that obviously many do not know that there exists a fairly reliable yardstick for what is a good, meaning an accurate, translation and what is not. Heinrich Pesch wrote: The one likes the daughter, the other the mother. s. above
[Edited at 2008-12-10 12:01 GMT] | | | @ nonsense and absurdity | Dec 10, 2008 |
Kevin Lossner wrote: Çağdaş Karataş wrote: One more thing; could you please clarify what you mean by complete nonsense? Roland's call for precision lost some credibility with me too when he tried to inject more "precision" into an example in English, only to suggest an error. [Edited at 2008-12-10 10:31 GMT] That can well happen, if one mingles with target languages other than ones own. - But I am not doing this in a job environment. Kevin Lossner wrote: The extreme example cited by Roland is unlikely to the point of absurdity, that's what I mean. His academic, absolutist approach to translation is very much in order for certain kinds of translation - the patent work I'm involved with is an example that comes readily to mind. However, in the context of literature, I find such an attitude absurd, and in many other types of translation, it is simply counterproductive. The need to improve on a text has already been discussed adequately in the context of textbooks or technical literature for use, but with literature there are many examples where cultural adaptation that involves rewriting the "precise" sense of the original is necessary to make sense to the audience. Tell me, Çağdaş, has a louse run over your liver this morning? If that question doesn't make sense to you, it's because I have translated it "precisely" out of Roland's native language. In my native country it would only get you a strange look. The book I cited above deals with these issues in much greater depth in a rather engaging way; I would recommend it to anyone who wants a thoughtful exploration of the subject rather than rigid dicta. All these effusions of translatorial insight are nice and pretty.- So much so that I let myself entice to enter into a debate about them. But they are ***off-topic*** in this thread, which is just about whether voting merely on "quality of writing", without having looked at, let alone understood the source text, should be permitted. If a person considers editing, be it slight or heavy, as necessary for a given translation, he is free to do so. He might produce something handy and practical for the reader, but will have changed the source. And if I had to vote on such a piece I would have a look at the changes made - and decide whether they were warranted or not. - If I think they were, I might rate the translation as good or even as very good. If I think they were not and just arbritrary hacking about, for lack of discipline, lack of respect for the source text - and, in most cases, ineptitude in the target language - I will rate such a piece as a bungled translation, no matter whether it *flows well* or not. Voting just on target - without collating - might be great in any environment whatsoever. Just not in a translation contest. - Why is this so hard to understand? This can indeed make me feel as having *a louse crawling over my liver*! | |
|
|
Editing is part of the translation process (especially in literature) | Dec 10, 2008 |
Roland, Correct me if I am wrong, but I think editing is part of the translation process, especially in literature. I don't think any serious publisher would skip that step. What I mean is that there is a translation, which (by default) should be true to the original, but then it goes through editing, where the focus is on the target text, often times without the input from the source. I don't do literary translations for a living, so I am not familiar with the details, but I heard that ed... See more Roland, Correct me if I am wrong, but I think editing is part of the translation process, especially in literature. I don't think any serious publisher would skip that step. What I mean is that there is a translation, which (by default) should be true to the original, but then it goes through editing, where the focus is on the target text, often times without the input from the source. I don't do literary translations for a living, so I am not familiar with the details, but I heard that editors are often monolingual. They evaluate and edit the text for the target audience, and that sometimes results in better text, than the original. (I have heard anecdotes about this.) Is that bad? I don't think so. The publisher's interest is to sell the book, otherwise they would not invest into the translation. If the translation does not sell in the target market, because it is hard to read, no matter how accurate it is, it will be a loser. In my opinion, nobody wants that. How is this related to the contests? Well, so far the texts were literary pieces (not books, but mostly articles, essays, but still, literary work). In a contest, you want to see completed works, don't you? Complete in a sense as you would see complete work in a published book. In other words, not just a raw translation. So, translators are expected to present polished work, in a sense that they would edit their own text (or participate as a pair, and work together on editing it) to a level where it would require as little editing in the target language as possible. I think the "quality of writing" category is supposed to evaluate that. If you are on the opinion that this step is not needed, or it is impossible to do without reading the source, then we just have to agree to disagree. ![](https://cfcdn.proz.com/images/bb/smiles/icon_wink.gif) In the qualification round (and remember, this is the only time where natives that don't work in the pair can rate the quality of writing) the tags that one can place do not focus only on "flow". You can tag grammar errors (those are not "accuracy" type problems), inconsistency in style, etc., and also, you can tag positive attributes (flows well, good term selection). By the way, good term selection does not necessarily refer to accuracy - in many cases there are multiple choices in the target, that are equally accurate, however, some choices are better due to subtle nuances, and sometimes those can only be felt by natives. (At least in the languages I work with, this is my experience.) What I am saying is that native input on evaluating these issues is valuable. If you are worried, that a translation that was "too creative" would grab the points just because it is nice to read, think about the numbers. If you are reading 5-6 or more translations in your native language where you don't know the source, wouldn't you notice if one of them was significantly different from the others in terms of the meaning at certain places? I think you would. You would get suspicious, that either the translator took too much liberty, or all the others misunderstood the source. I have seen both cases happening, even in this current contest. What can you do in this case? Well, you could stop the evaluation altogether and not rate anything. Or, you could mentally ignore those parts from all translations and compare the remaining text. You, of course, wish that those questionable parts will be evaluated by others that work in the pair - and my experience shows it does happen. (I would not assume that the majority of people that evaluate contest entries would be ignoring mistranslations, but we are in different language sub-communities, so our assumptions are based on different experiences, I guess.) The only risky case would be if you are evaluating entries in a pair that only have 2 or 3 translations. In that case, if one that is different from the other would not stand out as much, or you could be less sure who is right without reading the source (you can't use the majority method). However, pairs where such low number of entries were submitted, would NOT enter the qualification phase anyway, so there is no way that people who do not understand the source would vote. In the final round, only people that work in the pair and native in the target can vote. All that said, the contests are not perfect. There are technical issues and/or conceptual issues that generate debates, and of course, there are lots of cultural differences and different attitudes towards the whole thing, so it is hard to make everybody happy. Katalin (with an i, not a y) ![](https://cfcdn.proz.com/images/bb/smiles/icon_wink.gif)
[Edited at 2008-12-10 14:53 GMT] ▲ Collapse | | | relativity and quantity - | Dec 10, 2008 |
Katalin Horvath McClure wrote: Roland, Correct me if I am wrong, but I think editing is part of the translation process, especially in literature. I don't think any serious publisher would skip that step. What I mean is that there is a translation, which (by default) should be true to the original, but then it goes through editing, where the focus is on the target text, often times without the input from the source. I don't do literary translations for a living, so I am not familiar with the details, but I heard that editors are often monolingual. They evaluate and edit the text for the target audience, and that sometimes results in better text, than the original. (I have heard anecdotes about this.) Is that bad? I don't think so. The publisher's interest is to sell the book, otherwise they would not invest into the translation. If the translation does not sell in the target market, because it is hard to read, no matter how accurate it is, it will be a loser. In my opinion, nobody wants that. I had also not much experience with these editors, but the little I had confirms that that is a lottery of its own. - But we should not confound here the translation process with post-translation, largely commercially motivated intervention. At least I think we should agree that, on average, an editor will have less work to do if dealing with a well done, competent translation. Katalin Horvath McClure wrote: How is this related to the contests? Well, so far the texts were literary pieces (not books, but mostly articles, essays, but still, literary work). In a contest, you want to see completed works, don't you? Complete in a sense as you would see complete work in a published book. In other words, not just a raw translation. So, translators are expected to present polished work, in a sense that they would edit their own text (or participate as a pair, and work together on editing it) to a level where it would require as little editing in the target language as possible. I think the "quality of writing" category is supposed to evaluate that. If you are on the opinion that this step is not needed, or it is impossible to do without reading the source, then we just have to agree to disagree. ![](https://cfcdn.proz.com/images/bb/smiles/icon_wink.gif) . I agree that we have to. And I do my best to it.![](https://cfcdn.proz.com/images/bb/smiles/icon_wink.gif) Katalin Horvath McClure wrote: In the qualification round (and remember, this is the only time where natives that don't work in the pair can rate the quality of writing) the tags that one can place do not focus only on "flow". You can tag grammar errors (those are not "accuracy" type problems), inconsistency in style, etc., and also, you can tag positive attributes (flows well, good term selection). By the way, good term selection does not necessarily refer to accuracy - in many cases there are multiple choices in the target, that are equally accurate, however, some choices are better due to subtle nuances, and sometimes those can only be felt by natives. (At least in the languages I work with, this is my experience.) What I am saying is that native input on evaluating these issues is valuable. I have no issue with the tagging process. I consider that it is now very greatly set up. - Spelling, grammar and similiar aspects could of course be tagged without looking at the source. My only wish would be that one might have the possibility to take back erroneous tags. For the rest - allow me to repeat here. In my opinion, nobody should be permitted to judge on quality of writing independantly from the source. Everyone who wishes to vote on overall "quality of writing" should be stopped by the system automatically, if he is not able or willing to give a vote on "accuracy of translation" too. On clicking "Overall Vote", two fields should open at once - and close only if both have a value in. - This will be the only way of stopping the abritrary free for all, as it is now. Sure, deliberate abuse could still happen. But arbitrary voting would overall be reduced considerably. As the system stands, the humble, straight worker will always be outshone by the jugglers and dazzlers. With just a little exacerbation one could imagine cases in which "nice flowing and nice reading" pieces are "bad translations". | | | wonita (X) China Local time: 23:47 Purpose for participating in a contest | Dec 10, 2008 |
Roland Nienerza wrote: As the system stands, the humble, straight worker will always be outshone by the jugglers and dazzlers. I agree with you at this point. If your purpose for participating in the contest is to win, why not try to adjust yourself to win? As I always believe, the best one adjusts himself to the majority; the second best one tries to change the majority in favor of him; the third best one just withdraws. There are quite some colleagues who stay away from the translation contest for this reason. | | | 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 » Voting just for "quality of writing" is totally unacceptable CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
| TM-Town | Manage your TMs and Terms ... and boost your translation business
Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |