As a translator with some linguistics training, I'm fascinated by the technical and aesthetic issues of translating. Many of the thinkers I admire think, read and write not in English but other languages - granted, most of them are European or from Ancient Greece or Latin or Sanskrit. But their works are made known to me by way of English. This being the case, I know that some translations are better than others.
There are (have been) many different versions of the Bible in English, for eg. Before the King James' Bible other versions existed (still exist). The Bishops' Bible (1568) version of Psalm 23 goes:
God is my shepherd, therefore I can lose nothing;
he will cause me to repose myself in pastures full of grass,
and he will lead me unto calm waters
which the King James translators rendered as:
The Lord is my shepherd; I shall not want.
He maketh me to lie down in green pastures:
he leadeth me beside the still waters
The same kind of process happens for classical and modern literature. In Dorothy L Sayers' translation of Dante's Divine Comedy: 1 Hell for the Penguin Classics (1949), she writes of technical issues around the translation of symbolic language in poetry and allegory:
Dante's allegory is more complex. It differs from the standard type in two ways: (1) in its literal meaning, the story is - up to a certain point and with great many qualifications - intended to be a true story; (2) the figures of the allegory, instead of being personified abstractions, are symbolic personages.
To take the second point first: In dealing with the vexed subject of symbolism, we shall save ourselves much bewilderment of mind by realising that there are two kinds of symbols.
A conventional symbol is a sign, arbitrarily chosen to represent, or "stand for", something with which has no integral connection: thus the scrawl X may, by common agreement, stand, in mathematics, for an unknown quantity; in the alphabet, for a sound composed of a cluck and a hiss; at the end of a letter, for a fond embrace. The figure X is not, in itself, any of these things and tells us nothing about them. Any other sign would serve the same purpose if we agreed to accept it so, nor is there any reason why the same symbol should not stand, if we agreed that it should, for quite different things: infinity, or a murmuring sound, or a threat. With this kind of symbol we need not now concern ourselves except to distinguish it from the other.
A natural symbol is not an arbitrary sign, but a thing really existing which, by its very nature, stands for and images forth a greater reality of which it is itself an instance. Thus an arch, maintaining itself as it does by a balance of opposing strains, is a natural symbol of that stability in tension by which the whole universe maintains itself. Its significance is the same in all languages and in all circumstances, and may be applied indifferently to physical, psychical, or spiritual experience. (pp. 12-13)
I would suggest that mathematical/scientific knowledge are systems of natural symbolism, and so are psychologically-, spiritually- and sociologically- insightful works of literature (works known to us by way of translation though the source language may not be our own). This type of symbolism is what allows ideas to be translated without losing anything significant in the translation, if not in form.
I have advocated for Inuktitut translation of classics, science and mathematics (modernization) because I know what the world has to offer can be translated into Inuktitut without detraction from either the "original" source or Inuktitut itself. In fact, both would be enrichened by the interaction.
For mathematics and science (chemistry for eg), the conventional symbolism would remained unchanged but the terminology could easily be translated using agreed-upon and already existing pedagogical principles and systematic rules and procedures of nomenclature. In fact, I did some real world experimentations recently in translating some math concepts into Inuktitut. Since I'm naturally attracted to mathematics and science, I know these concepts can and do daunt English-only speakers but not me. I know that the first principles are usually simple enough for boot-strapping to occur once mastery of them takes hold.
Using either form or function or process or end result, the Inuktitut rendition of mathematical concepts are much easier to weld where, in plain-language English, the rendition may not come across so easily (given that math and science nomenclature in English is latin- or greek- based, and can become quite cumbersome English grammar-wise). Inuktitut is a polysynthetic language, meaning that it fuses together noun or verb roots with adjectival and adverbial modifiers naturally enough to not lose grammatical integrity while retaining its descriptive power long, long after English has lost them.
In fact, Inuktitut structure is out-right mathematic (ie, structurally predictable and conceptually- / logically- productive), and - I maintain - modern North and South Baffin dialects have an Elizabethan quality (ie, English that Shakespeare used). Inuktitut is beautiful. Even the Inuktitut bible (as rendered by Moravians using Labrador dialect) has power and beauty the best of German and King James' translations have to offer. Unilingual English-speakers got the wrong impression that Inuktitut is "primitive" from poorly translated/interpretations they are exposed to.