Bilingualism and penmanship can be attained and trained, respectively, but not intelligence. It may sound cruel to the translation aspirants who are perfectly bilingual and love to write, but have moderate IQ's, but truly and honestly, they will have more satisfaction in other professions. The current IQ measurement system focuses on understating and problem-solving capacity, which are two main ingredients to prepare a delicious dish of words and thoughts. Memory and creativity, two key condiments in the kitchen of a translator, are additional constituent elements of the armamentarium of a translator. The capacity to learn in the sense of problem solving (translation is actually a series of problem-solving tasks) is indeed the most important translation tool in this Internet era. A translator well-equipped with this wonderful human capacity becomes a temporary expert in any topic from any academic discipline while translating.
Every day and night, I translate dissertations and articles for journal submission. I tackle each article in the order of introduction, methods, results, revision of these three parts, discussion, abstract, and conclusion. I do as much research as I think necessary while I translate the introduction part. It takes reading of some 50 search result pages (this is utterly effective if search words are skillfully arranged) and some 50 pages from the linked or cited articles. Then I do the second round of research, not so intensely as the first round, while I reread the three translated parts. For the rest of the paper, I do the minimum necessary research. The translation time ratio of the introduction part to the rest of the paper is usually 1:2 when the length ratio is 1:6, which implies that I understand the topic in insider-level details.
With the Internet-mediated unlimited potential for producing a target text at the native-speaker/expert level, the role of a translator is no longer bridging two languages. In academic translation, in particular, the only valid translation theory is producing a manuscript acceptable by the journal. Manuscripts (representing their authors) do not mind receiving even most radical plastic surgery to enhance their chances to be accepted or even featured by the journal. Authors prefer unfaithful beauty to faithful ordinariness if faithful beauty is not among the options.
As an example of the contribution that a translator can make beyond the pure linguistic realm, let me tell a rather banal episode of one correction and one recommendation I made in a 12-page manuscript I translated yesterday. One of the words the author copied from a cited paper was "discharge" when describing the water entering into a fishway installed beside a dam. The upstream water is discharged downstream, e.g. through spillway gates, but not into a fishway. Upstream water flows into the fishway entrance (exit in this case because the fishway targeted ascending anadromous fishes). I checked that there was no mention of fishway in the cited paper and added inflow in brackets because quoted wording should not be directly corrected. The other term of concern was "visual predators." Reporting that some of the fish species were using the fishway from dusk till dawn, the author stated that this pattern was not explicable because there were no "visual predators" among the fish species using the fishway. I felt something wrong in his reasoning because birds are usually visual predators of fish. After confirming my suspicion in the cited paper, I offered the author an alternative explanation in a comment balloon.
A question may arise if such translator intervention is necessary or even allowed. Personally, I think it belongs to the basic duties of translators living in the Internet era of knowledge sharing and profiting from open access resources. Authors may not have enough knowledge of the target language and quote insufficiently or inadequately understood contents from target language literature. Translators, insofar as they have access to the related papers, should ensure that the contents quoted from those papers are not distorted in the manuscripts they translate. In this sense, translators play an active role in the knowledge-sharing global community of academia and industry. There is no room for debates over translation theories, such as faithful or communicative translation, because it is the accuracy and impact of the resulting work that matters beyond the linguistic aspect of how faithfully or fluently the source language is rendered in the target language.