The good news about translating physics and math is that its far easier than translating newspapers, civil engineering, or shipping documents.
The bad news is that most people find it next to impossible.
If you compare physics German to German, you see right away that physicists speak a tiny subset of their language. You can read the stuff for decades without seeing a dozen words of baby talk, street slang, inventive metaphors, kinship terms or value words like splendid and sorry. Entire tenses may be missing. Whats more, a big hunk of world physics vocabulary is either 20th-century Latin or unashamed straight English.
The sublanguage of the physical sciences, let me call it Physpeak, also shows an odd correspondence from language to language: French, English, Russian and Czech physicists use equally small subsets of their tongues.
People who write these sublanguages are concerned with precise usage, which is a little different from precision. English-speaking physicists hate it when you confuse resistance and resistivity or when you say electricity is billed by the kilowatt. Physicists were the Ur-geeks who booed George Lucas after hearing a Star Wars character describe the speed of his spaceship in parsecs. Mathematicians never say sequence when they mean series or order, nor do they ever ever get order and ordering mixed. With a few exceptions, you can look up a definition in your grandmothers electricity and magnetism textbook and find the same term used today with the same meaning.
Now add together a small vocabulary, a stress on precise usage and a marked conservatism, and you probably think Who says there is no royal road to learning? A comprehensive dictionary should get you through any amount of physics. But a lot of good translators fail when they try such texts.
I can list a few reasons why Physpeak turns out to be a hard language: implausibility, jocularity, high level of abstraction, suppression of background in publications, elliptical forms of development, difficulty of finding definitions, and previous translations. (I suspect that none of these is as important as physics terror, and I will come back to this problem further on.)
Many people constitutionally do not believe that numbers like 10 to the minus 23rd or 81.577 plus or minus 0.021 really mean anything. Whats more, any reader of Science News has learned that teleportation used to be impossible but now has been demonstrated in the lab, and that there are hobbyists out there calculating the zillion-and-eighth digit of pi. What can you believe and what must you disbelieve? It is not uncommon to say the hell with it and deny credence to the whole structure. But such a decision changes the science and technology translator to a science-fiction translator, and everybody knows it is nearly impossible to translate science fiction.
Physicists (more than mathematicians, less than molecular geneticists) have a tradition of making in-jokes. Not funny jokes by and large, but they do add to the implausibility of some Physpeak texts. Consider for example the Russian word barn, which appears to be a unit of area used by nuclear physicists. You look it up and learn that it means barn in English. Its an American joke from the 1940s, derived from the idiom hit the broad side of a barn. A barn is 10 to the minus 28th square meter, absurdly smaller than a, um, barn. See, not funny but also not obvious, and thus not credible either. (I confess to thinking this one is funny, but this wonky sense of humor has gotten me in trouble more than once.)
The barn is an instance of physicists speaking in everyday words about bizarre, unexpected and unbelievable things. On a different scale, certain phenomena in the upper atmosphere during thunderstorms have received the name sprites. Names like barn and sprite give little hint about concrete objects and processes; they contribute to both the implausibility of Physpeak and its usefulness in concisely pointing to complicated and abstract ideas. Indeed, papers in some areas of physics make little use of words that are hard in themselves; readers have to prepare themselves by learning special meanings for terms like state, collision, hop, path, hue, charm, observer, space and machine.
(These very wordsin contrast to expressions like nutation, mechanics, Lagrangian and anomalygenerally take native forms in the several Physpeaking communities. While nutation becomes nutatsiya in Russian and anomaly is Anomalie in German, state is sostoyanie/Zustand and observer turns into nablyudatel/Beobachter.)
Translators would like it if every physics paper developed its concepts from the egg, but physicists would not. An article on electron levels in metal lattices might, but wont, refer the reader to a textbook chapter on valence, a standard work on spectroscopy, and the authors previous work on crystals. Every article would, but doesnt, incorporate a chain of references going back to some easy basic reading. The lack of such background means the translator has to fill in a lot, and not always in an obvious way. Do you begin your search with level, metal or what? How do you tell the difference between basic texts, advanced research, and (crucially) results that have been proved wrong?
There is a legend about the math professor who fills up three blackboards with derivations and announces, And the desired result follows trivially. He looks hard at the theorem, tells the students to wait, fetches an armful of books from his office, and spends the next 45 minutes flipping through them and making notes. Finally he slams shut the last book and proclaims, Yes, its trivial. Class dismissed. Surely translators hate to see words like trivial and obvious, but it is no less baffling when a Russian writer claims support from an English source but only after changing all the symbols, recasting the integrals as sums and quietly taking logarithms of all the terms. Arguments with the middle steps left out make rough going for the translator, who must work doubly hard to avoid turning them into nonsenseand then may not be able to check.
Related to the ellipsis problem is the very common practice of defining terms only when the writer differs with someone elses definitions. In a lot of Physpeak the best you can hope for is a parenthesis like in the Hamiltonian formalism, which points to a framework of definitions and principles that you can look up. More often the writer feels no need to say what common expressions mean, so that the translator must find out whether path in this text relates to the motion of an electron in a spark, the change of phase-space coordinates during a process, or the distance that a bombarding ion travels into a solid.
A final source of difficulty in Physpeak translation lies in translations already done in the same field. But published or approved translations should help, not hurt, you say. I reply that, unless done by an expert colleague, they may contain errors, inconsistencies and off choices. Probably the most durable of these near-translations came into English late in the 19th century when the German heading Experimentelles was first rendered as Experimental. This is still one of the canonical section titles for lab chemistry papers (along with the grammatically more regular Introduction, Theory, Results and Discussion).
* * * *
No one can deny that Physpeak is a hard language. Can non-scientists translate it all the same? A delicate question. When the American Institute of Physics created a vast program to make Soviet journals available in English, it limited the corps of translators to working scientists. The quality of the results proved the wisdom of the decision. Other publishers did less well, and library shelves are (or used to be) packed with dubious cover-to-cover journals that may have served an awareness function but could not have reliably conveyed much about physics.
It is not my business to tell you that you cant do physics and math translations. Most readers canat some leveland most readers will occasionally have to in the course of other work. It will be more useful here to suggest approaches. I have three.
(1) Formal study. As the holder of an undergraduate degree in physics, I think there is no substitute for a lot of university course work. Typically (in a physics curriculum) you will take lecture and lab courses in mechanics, electricity and magnetism, modern physics (relativity, atomic and nuclear physics, quantum theory), mathematical physics, thermodynamics, plenty of calculus including differential equations, and more math and sciences. You may have the opportunity to assist in an experimental lab. You will learn to do a lot of fancy stuff with computers and lasers.
Caveat: You will never see many translation jobs on these topics. What does the course work do for you then? It helps you become familiar with scientific patterns of thinking, identify sources of information for later use, and learn Physpeak while exposing yourself to an array of specific facts and concepts. If you do well, earn an advanced degree and get a job in science, it will also provide you with the beginning translators most vital resource, a regular income.
Kindly Include Me Out
Is it better to prepare for a translation career by studying language and literature or by studying science and engineering?
I respectfully decline to get into that wrangle. Any collection of data on translators will show that some successful colleagues have taken path A, others path B, still others neither one. We are all happy and productive and we all love and respect one another, and please dont ever ask me this question again.
But some readers of this article dont have the option of going (back) to school to study physics or chemistry or math. There is a fair second choice:
(2) Informal study. A motivated student can learn enough of nearly any subject to qualify as a skilled translator. Physical sciences, as not just a body of knowledge but a discipline of thought, may be harder than some other subjects but not beyond a dedicated student. I suggest, however, that a careful and thorough plan is essential; simply reading magazines and journals can never be enough.
A good first step might be to obtain the textbooks used in 100-level math, chemistry and physics courses at a nearby universityfirst courses for science students, not survey courses for those who merely have to satisfy a distribution requirement. These books will present a range of concepts in (initially) plain language, along with a body of problems to work. This is not the time (there IS no right time) to scant mathematics; the sciences dont work without it. You may never be good at it, but you truly need to understand the interplay of matter and energy and the framework that math supplies for describing it.
The next phase is not advanced study but rather repeating the first stepusing German or Russian or Chinese textbooks. There is only one scientific community, and the electricity and magnetism course you just finished will be remarkably similar to the one you start now. You will begin noticing the international features of Physpeak as well as country-to-country differences. The repetition will affirm your new knowledge, while the parallel texts will help you build vocabulary.
Now you can go to the next level, still using textbooks and adding special materials like handbooks and encyclopedias. Continue working in two languages at once. You know enough Physpeak to ask intelligent questions of teachers and researchers and to use unbuffered resources like the Web. Maybe you can find a graduate student to give you conversation lessons. Keep in mind that the math has to go along with the physics and other studies.
Neck Out or Head Down?
A customer of mine used to give advice like Our engineers really dont use all these active verbs. Should I have tried to fit in, or was I doing the right thing by writing the best way I knew?
Not many engineers have chosen their career because they could express themselves well (in words, that is). Adequate is good enough, most would say. And yet, if you read enough Physpeak, you will see that one writer in perhaps fifty is making prose that moves, thats easy to understand, that you actually can read out loud. True, it is an incidental skill. But the presence of even a little good writing in the sci-tech literature proves two important points:
My response to this knowledge is to say, You want bad writing, go hire a bad writer.
- If not easy, clear and pleasing style is not impossible.
- It is not forbidden either.
Mathematics: Threat or Menace?
Do you experience math anxiety or physics terror? Many smart people do, and it has a crippling effect. I cant offer therapeutic help, but only some thoughts to consider:
Whatever mantra you select, you must get over your physics terror, and you must find the technique that will work for you. Talk with someone who teaches science and math to teenagers; take an evening course; pretend it is computer programming; relate chemistry to cooking or crystallography to eschatology (find the connection and win a nice award). Read the biographies of scientists and think about how many of them disappointed their fathers by not becoming lawyers. Find something that helps you stop jibbing at formulas and definitions.
I never could understand science makes a poor excuse for delivering a bad translation.
- Ordinary people do this.
- Physics is just how things work.
- Math is a formal language.
- We invented this stuff to satisfy the curiosity we all find natural.
- When you find regular rules or laws, you look for a way to state, remember and work them; thats what science and math do for you.
- Language works for you, you dont work for it.
- Ordinary people do this.
- If it makes you unhappy, there are plenty of other ways to live.
(3) Spot study. Just as a technical translator may run into regulatory texts and contractsand has no option but to handle themanybody in the business can expect to see material on physics, chemistry, engineering. Catastrophe? Not every time. Given a little science background, some access to tools and just a bit of nerve, most any translator can get through these tight spots.
Again, I have a method to suggest. It is not perfect, it will not turn you into a real physicist, and occasionally its wheels may fall off, but it is better than no method and it does improve your odds of success. I describe it with four long words: identification, orientation, exploitation, evaluation.
You first identify the problem you have to solve. I once worked with an anal-retentive radio engineer who printed little trouble-report slips. If a tape deck wouldnt play, you were supposed to fill out a slip and carry the device to the shop. The biggest space was headed Nature Of Problem. About half the slips came in with the notation Does not work right. This was not helpful, and it will not be helpful for you to say I dont know what this passage is about. You need to pin down the reason you cant translate it.
This is the spot for a word about dictionaries. They are all wonderful, and they will not get you out of this fix because you have to be an expert to use them correctly. Finding that your Mystery Word means homotopy does not mean that you can now go ahead with the translation.
Now and again comes a proposal for a dictionary containing usage information enough that it could, in principle, substitute for knowledge of what the discourse is about. You could look up homotopy and find all the kinds of expressions that use the word. (Does one, for example, apply, impose, diagonalize, derive, break, convolute, lift, take or subtend homotopies, or does one pray that medical science will bring relief from them?) One or two such works exist, but they have limited usefulness; even tiny Physpeak defeats them. And with an ordinary dictionary youre lucky to see one or two phrases that the compiler happened to note.
When the problem is identified, it still takes some skillful work to get oriented, that is, to learn the basic features of the ground. In the homotopy example, you need to keep looking at definitions until you find one that has the marker [math], which is not an answer but points to an answer. You have now put yourself in position to consult basic and advanced works on math, which will quickly narrow themselves down to point-set topology.
This is not good news, is it. While the orientation step has not failed, it may have revealed that you have a great deal of hard slogging in store. It is hard, but your exploitation of resources should at least make it a paying proposition. The first thing to do when you pick up a new reference is WRITE DOWN THE NAME OF IT. Translators have wasted more effort by using sources they cant find again than they have by forgetting how they rendered Einsatz in that last job.
Your note-taking style is your own and nothing I suggest will change it. I can only urge that you consider retrieval as well as storage. If your habits allow you to make and use marginal notes in dictionaries, fine; if (like me) you quickly lose anything smaller than your fax machine, find a medium that will keep track of itself. My electronic glossaries hold quite a lot of this bibliographical information, together with terms and phrases collected from old jobs and reference works.
Invest some time in recording what you learn; words, sure, but also entire sentences (usage examples), pictures, references to job and customer, pointers to related terms and conflicting definitions, and anything else that it occurs to you to put down. I have found it useful to include in my terminology notes the date of making the note, for it does happen that 1998 sees more clearly than 1984 did. If you have authority for a note, record the authority; this is especially vital when you have customers sending you terms. (The question whether your customer is an authority is matter for another somewhat less temperate article.)
It is useful to go through your references twice, once early in the translation process and once nearer the end. You learn answers more easily when you ask the right questions, and every sentence translated will teach you new questions. Your last revision should find you smarter than your first draft, so you will get better notes.
Now we come to the hardest step, evaluation. After all, if you were able to evaluate Physpeak translations, you would not have to be doing this spot study. Give up? No, you assess everything else you do, you must not shrink from assessing this. But you have a limited range of points that you can evaluate.
- Are all the words spelled right? Do subjects and verbs agree in number?
- Does the text contain slurred portions, the written equivalent of speaking behind your hand? These may not be fatal; they may not even take away from the value of your work; but they are weaknesses where you wanted to show strength.
- Does the translation read like original publications you have consulted? Consider vocabulary range, sentence length and complexity, character of verbs, and general level of stuffiness.
The caution that goes with this kind of checklist is obvious: You can score well with a rotten translation! But unless you have a Physpeak expert close by, it probably is the best you can do. So you dont rank as a specialist; you dont rank as a fool either.
And I did warn you that this last approach would not work every time.
* * * *
I have described some features of a key sublanguage and approaches to the pitfalls it holds for nonscientists. What I have not tried to do is set you up in business as a finished Physpeak translator; you will find it a hard job becoming one, even if you have good academic training. The languages of science and technology have become part of the mainstream realm, and more than just sci-tech translators have to cope with them; with craft and a bit of luck, you can.