Stop looking at the subject matter and look at the process of learning behind it.
Textbooks and materials can be gotten through libraries or even bookstores. (Washington DC had a great place for such stuff called Reitner’s, even though it was pricey as all get out). The resources are out there.
Professional standards and ethics in translating push translators to ensure that imperfections are minimized, and any translator worth his salt knows when the job is beyond them and will have no compunction about turning it down.
At my university we were expected (no official requirements) to understand English and German - either because the texts were in those languages or the teachers only spoke English (and sometimes only barely). This was in the Netherlands, by the way.
Pretty much all of the German requirements were for formal logic and philosophy courses. The English was mostly for programming and some math.
I would imagine translating an in-depth text book is not cheap. Certainly, I buy all my current technical books in their original language if that language is English or German, because the Dutch translations suck (as in, are rife with errors), don’t exist, are more expensive and/or are pressed years after the original.
I think the lack of a market is the largest issue here. My training is in algebraic topology, and most of the foundational material was originally done in French. (And generally by the same authors as the algebraic geometry work mentioned above).
As an example that I’m intimately familiar with, Serre’s thesis underlies all post-war work in the field, and was printed in French. As far as I know, it has never been officially translated into English (granted, most of his ideas have been, though). My advisor sent me off to read Serre’s thesis, with two goals in mind: 1) to develop my skill with spectral sequences, and 2) in order that I learn to read French. This might be the most important work in the field, and there isn’t much of a demand for translation, in part because many (or most) topologists can already read the French, why bother with English?
And yet both can be learned. If law can be learned and understood enough by someone otherwise untrained in law to provide a quality translation, it is therefore possible that the same can be done for math. Honestly, it’s not that hard to see the forest here.
ETA: President Johnny, what’s this thesis by Serre you spoke of? I work in French to English, and this has got my curiosity roused enough to experiment.
Olentz, yeah they are interpreters. But there’s some wiggle room on the definition. No, they don’t do text to text, but lawyer to M (interpreter), M to client, client back to M to lawyer and then to reporter. In front of a jury. Seems like translating to me.
Flipshod, it’s translation insamuch as it’s going from a language to another, but interpreting is its own specialty and very, very different from written-text translation:
You cannot go back and review your work,
if you ask for clarification, you’re a lot more likely to run into “just translate!” than someone working on written text,
dictionary, what’s that?
on a written text, neither your accent nor that of the author are a problem; when speaking, it can be very much a problem,
etc.
Also, something else that can be a problem is that some judiciary systems consider the translation to be the “official” deposition. If the witness answers rapidly and unhesitatingly and the translator has problems finding the right word, the transcript shows the hesitation, but not that it wasn’t the witness who hesitated (this came up in one of my classes, I’ve been taking a MSc in Translation… should be working on my dissertation, actually…)
Olentzero, do you know anybody who translates advanced math papers? I’m not saying your answer would prove anything, I’m just curious.
This debate seems to have two parts. First, how well does a translator need to understand advanced math to translate it? and second- how much study would be required to gain that level of understanding?
I suspect that in order to accurately translate an advanced mathematical article, the translator would need to have a pretty good grasp of the material. Isn’t it true that translation isn’t an exact science? Don’t you need to have a sense of the meaning and style of the text in order to come up with a comparable text in another language? And the more precise the wording of a text is, the better you need to understand it in order to give an accurate translation. Given that math often requires a great level of precision, this seems to indicate that a translator need a very high level of familiarity with the topic.
As for the second question- yes, it is theoretically possible for someone to educate themselves enough to understand the most advanced math papers. However, even if a person did this without going through an academic institution, they would have to put in at least as much work as someone who did go the normal route. And there would have to be more than one such person, so that they could check each other’s work. Also, there is the undeniable fact that some people are better at some things than others, and people tend to avoid the things they’re not good at. Nava’s anecdote is a good illustration of how it rare it probably is to find a professional translator who is also inclined towards mathematical/technical fields. It’s not impossible, no. It does, however, seem that the rarity of these people, and the difficulty they would have learning advanced math without the benefit of outside guidance, make it not surprising that a significant number of math papers are untranslated.
and, for what it’s worth, the PhD in math at Hunter College does require the ability to translate a section of mathematical text from a foreign language into English.
I fail to see why the nuances of either expression make the slightest difference to the mathematical meaning. Besides which, only a native speaker or someone with nearly native fluency is likely to appreciate what difference there is. I cannot appreciate much the difference between “werden” and “become”, but I would not have noticed. The second example is worse. The definition of prime number is that it has no factors and there are no possible connotations that would illuminate it in any way.
Getting back to the OP, does he have any idea how how expensive it would be to translate all of mathematics? Translation of mathematics really requires two people, one who is an experienced translator and one who is a mathematician. As a case in point, my wife is a professional translator, French to English, and together we are translating, as a retirement project one of Grothendieck’s most famous papers, over 100 pages long, that appeared in a Japanese journal. Another example. A colleague was interested in a paper that appeared in a French journal in 1840 that beat Hamilton by a few years in a complete analysis of rigid motions of a sphere. I found it online and printed it for my colleague, but it was in French and no one will translate a paper that is 170 years old.
A few years ago, I was interested in a paper in Russian and sure wished I could read it. My wife can read Russian and made a stab on translating it, but it didn’t go very well because she knows no mathematics.
But it is true that more and more schools are abandoning language exams. At McGill, we have kept a French requirement, but that is more politics than science.
Incidentally, it is not always the case that you are given a passage in math. When I took a French exam at Penn maybe 50 years ago, some sadist in the French department gave us something essentially untranslatable. It was a popular science article making analogy between (electromagnetic) “ondes” and (water) “vagues”. Since both words translate into English as “waves”, this was an essentially impossible. I don’t recall now what I did, but I passed.
The problem is not whether translators could learn enough math to translate papers. But, who is going to pay these highly qualified folks to translate every paper from every language into every other language? (And they’d need more highly qualified folks to check the work.)
Much of this stuff probably has a limited market. Unless these super-trained translators want to do the work free, it probably makes sense for mathematicians to learn enough of another language that they can understand the papers themselves. Especially when one language might have the most important work in a particular specialized sub-field.
As a translator, and as someone who knows both translators and interpreters, there is no wiggle room. It’s agreed on within the professions - written word is translation, spoken word is interpretation.
I don’t know anyone who translates advanced math papers, although I had considered looking in the communities I’m part of. And I was fully prepared for the possibility that translators who do translate mathematical papers came to translation from mathematics and not the other way around. Still doesn’t eliminate the possibility that the reverse is true.
But the thesis by Serre mentioned upthread gave me an idea for a ‘proof of concept’ exercise, and I guess I’ll spell it out here. If PJG or Itself can give me either a link to the text or a way of finding it, I’d like to try translating a max 200-word chunk of it and having them look at the English translation. If I - a translator who only got as far as high-school calculus and a long-forgotten course in mathematical modeling at Georgetown - can put out work that they, as algebraic topologists, agree doesn’t completely suck, then I would regard that as strong evidence of my contention. In other words: if I don’t know a whole lot about algebraic topology but can provide a decent translation, then it is clearly possible that a translator who knows the subject better - but still not as thoroughly as a trained mathematician - can provide a decent translation as well.
That’s been part of my point all along. The other part of my point is that the level of familiarity with the topic a translator would need doesn’t require a degree in mathematics. A translator is merely facilitating communication between trained, skilled professionals separated by a language barrier. The metaphor that comes to mind immediately is that of a local guide leading a foreign scientific expedition through an unexplored area. The botanists and biologists are highly trained in the scientific techniques of their respective fields, but the guide is the one who knows the way around. We don’t need to know taxonomic classification inside out, but we’re the ones who can find the animal or flower you’re looking for.
Bridget reinforces a good point I tried to make earlier. Any translators who do mathematical papers don’t rely on those jobs to make a complete living. The market lies elsewhere. But that doesn’t prove translators can’t adequately translate mathematics papers without having an MS or PhD in mathematics.
ETA: Hari, PM me. Your projects sound interesting.
My bachelor’s is in math, and I can’t make heads or tails of algebraic topology, so I can’t imagine this being successful. But if you’re up for the challenge and you can translate from German, I’d be interested to see what you can do with some text from Landau’s Grundlagen der Analysis, which is a fairly elementary text.
I’ve sent a link to the paper by PM. I’m curious to how it works out. My advisor gave me many horror stories of mistranslations and mistranscriptions that he had seen. In one case, the person typing up his paper helpfully corrected a recurrent “typo” - every time he had written the word “functor”, she replaced it with “function”, completely changing the meaning of the paper.
Thanks for the link! As for your horror story, that’s as much a result of the typist not being careful and checking with your advisor on the matter before proceeding as it is unfamiliarity with the subject. Typists and translators need to have the rule “check with someone when uncertain” ingrained into their habits early on, to be sure.
Good morning! Back from a lovely weekend out on the archipelago with the Swedish branch of the family, it’s now time to get this little experiment on to the next stage. Herewith, the first 250 words of Serre’s thesis in the original French, followed by my translation. President Johnny Gentle and Itself, as the resident algebraic topologists in the thread, your judgment is final. How’d I do?
[QUOTE=The original French]
L’objet essentiel de ce mémoire est d’etudier l’espace Ω des lacets sur un espace donné X. L’intérêt de cette étude est double: d’une part, Marston Morse a montré que, si X est un espace de Riemann, les propriétés de Ω sont étroitement liées aux propriétés des géodésiques tracés sur X; et, d’autre part, on peut, avec Hurewicz, utiliser Ω pour donner une définition récurrente des groupes d’homotopie de X et, par suite, tour renseignement sur les groupes d’homologie de Ω entraînera une meilleure conaissance des groups d’homotopie de X.
Mais l’étude directe de l’homologie de Ω s’était avérée difficile, et n’avait guère pu être menée à bien que dans le cas où X est est une sphère. Nous utilisons ici une methode indirecte, suggerée par la relation π[sub]i[/sub] (Ω)=πsub[/sub] (X), qui consiste à considérer Ω comme la fibre d’un espace fibré E qui est contractile, la base étant l’espace X donné. En appliquant alors à E la théorie homologique des espace fibrés développée par J. Leray, on obtient des relations étroites liant l’homologie de Ω et celle de X, relations que l’on peut appliquer avec succès aux deux problèmes cités plus haut.
La théorie homologique utilisée ici étant la théorie singulière (seule adaptée aux problèmes homotopiques), il nous a fallu montrer que la théorie de Leray était valable dans ce cas, et pour cela, il nous a fallu en refaire complètement la partie topologique. Notre exposé ne nécessite donc pas la lecture préalable des mémoires de Leray sur le sujet.
[/QUOTE]
[QUOTE=Olentzero’s Translation]
The main object of this paper is to study the space Ω of braids in a given space X. The significance of this study is twofold: firstly, Marston Morse has shown that, if X is a Riemann space, the properties of Ω are closely linked to the properties of the geodesics traced on X; and, secondly, with Hurewicz, Ω can be used to provide a recurrent definition of the homotopy groups of X and, therefore, all information on the homology groups of Ω will lead to better knowledge of the homotopy groups of X.
But direct study of the homology of Ω has proven to be difficult, and has scarcely been conducted except in cases where X is a sphere. Here we will use an indirect method, suggested by the equation π[sub]i[/sub] (Ω)=πsub[/sub] (X), which consists of considering Ω as the fiber of a contractible fiber bundle E , the base of which is the given space X. In thus applying the homology theory of fiber bundles developed by J. Leray to E, the close relations linking the homology of Ω to that of X is obtained; relations which can be successfully applied to the two problems cited above.
The homology theory used here being the singular theory (adapted only for problems of homotopy), we had to show that Leray’s theory is valid in this case; for that we had to completely rewrite the part on topology. This article does not, however, require preliminary reading of Leray’s papers on the subject.
[/QUOTE]
I’ll just get this out of the way right off: the first 250 words of any scientific work (math, physics, biology, whatever) are generally much more descriptive and less technical than the meat of the work. To really demonstrate the contention that no major mathematical training is necessary to translate a mathematical work, you’d need to translate something from the middle.
I had considered that, but seeing as how I didn’t have a subscription to the JSTOR site and that the introductory paragraphs still have some clearly technical language - specialized enough that I had to do some research to try to find what I believe to be the right phrases in English - that this should still be proof enough of concept. I defer to PJG and Itself on this point, however, and if they feel the same perhaps we can work on getting something further in.
