Kingsley Amis (1922-1995), "Disappearance of Latin," in The King's English: A Guide to Modern Usage
(New York: St. Martin's Press, 1998), pp. 51-56 (at 51-52):
My own career is a good example of many things, but none more than in my experience of the language and the literature of Ancient Rome. Like millions of my fellows, I was brought up in the 1930s to study Latin. When I was seventeen I switched to English, which nevertheless meant that I continued to study the classics, though less inflexibly than before. When I secured an award in English and went up to Oxford in 1941 I had the advantage of a classical training, for all that it seldom felt like any sort of advantage at the time.
The foregoing is a mere exordium in that I have no intention of going on to say that to have studied Latin is in itself somehow good for you or for your English style. It is not that a knowledge of Latin protects anybody from making mistakes about the meaning of English words, because the meanings of words are not fixed, they change in and after their move from one language to another. It is true that defendo means 'I defend,' but a muscle is not a little mouse, which etymologically it is, nor is a pencil what its origins declare it to be, a doubly small penis. Neither is it the case that, as schoolmasters are supposed to have thought or said at one time, one was helped to think by mastering that language, as if it were a course of mental gymnastics. Nevertheless the student of Latin, as of any considerable dead language, must constantly be trying to choose the right word to give the meaning of a Latin expression in English or an English expression in Latin. And if the writing of English generally is in decline, as many would say it is, we may be tempted to say that people no longer try to choose the right word as they once did. They often got it wrong, but they tried. Do they now?
Something like the foregoing sketch might be developed to accompany an analysis of English poetry as written over the last fifty years or so. If this is seen as having become not only less formally organised but less exact in its expression, then the loss of Latin has surely had a hand in the matter somewhere. Again, I do not simply mean that an acquaintance with Propertius or Catullus in the original is beneficial to any sort of poet, though I think I do think so, but just as simply that translation into and out of Latin verse calls for exactness, and that that quality is demanded in the writing of poetry as nowhere else. Exactness, by the way, is to be understood as applying to more than denotation: a word or phrase must be suitable to its context, so that a dialectal or slang term, for instance, is on the whole unlikely to fit well into a passage of high seriousness — except for special effect, as teachers used to add.
The chances are that no particular virtue attaches to Latin as a language, although its role in our culture is unique and uniquely important. Any dead language will do as the kind of trainer I mean, such as Ancient Greek or, were it copious enough and intelligible, Etruscan. But deadness is necessary. A living language is by definition unfixed, in a state of continuous development and change, adapting and often dropping dialecticisms, provincialisms, technical terms, slang of all sorts, foreign expressions and more. It has no choice but to be useless as any sort of example.
The preceding paragraphs are no doubt speculative. What follows is all too manifest. Not just Latin itself has disappeared but in many cases any certain knowledge of what it was. A phrase like mutatis mutandis, apart from being offensively unintelligible to almost every British person, will be taken as a bit of Italian or French or (it's tempting to add) Choctaw rather than Latin. You come across it on old gravestones and monks used to sing it, or in it. The rest is silence. Latin is not only dead but cancelled.
Hat tip: Ian Jackson.