sprachwaffe

The larger question, of which this is an instantiation (I’ve been dying to use that word for days now–here’s my opportunity), is how English, or any language, integrates borrowed words into its ordinary vocabulary. The more fully integrated, the more “popular” a word is, the more likely that it will (come to) be treated more or less as a native word; conversely the more restricted and more technical a word is, the more likely it will be that it retains its native morphology. Thus the Latin plural of vacuum is vacua and such a plural is listed in the dictionaries beside vacuums, the only one I’ve ever heard or seen used in English. On the other hand, crisis is a pretty common word and yet it retains its Latin (< Greek) plural crises. Compare also quantum and its plural quanta. There also seems to be a phonological constraint operating here too: a word is more likely to preserve its borrowed plural if the word stress is on the syllable immediately before the ending and to replace it with a regular English plural if the stress is further away. So a one-size-fits-all or one-size-fits-all-to-the-same-degree answer won’t work.

So with metropolis we have a reasonably common word with the word stress on the third-from-the-end syllable, so its not surprising that its plural is metropolises (and likewise acropolis and necropolis). Polis, however, is not so common and is used only in “non-English” contexts, i.e., when talking about the Greeks or about Greek history or social customs. It is also stressed on the second-to-last syllable. So its retention of the Greek plural poleis, in those rare instances where more than one is talked about, is not too surprising (note that it is a real Greek plural, too, not one that has been Latinized).

So, it’s not a simple answer but one where the actual shape of any particular item is dependent on the exact weight of a number of multiple causes. I’ll let you physicistae write the exact equation.

We might be able to give a prize to the world’s most efficient language if we only knew what efficiency meant in this context. Is it the language which expresses things with the fewest words? With the fewest syllables? In the fewest milliseconds? Is it more efficient to have, say, a definite and indefinite article to specify, inter alia, old and new information (e.g. ‘a guy in my class….’ [i.e., 'guy' is new information in conversation] and later in the conversation ‘the guy told me…’ [i.e., now 'guy' is old information, common knowledge to those in the conversation]) and reduce potential ambiguity at the cost of introducing a lot of words, albeit short ones, or is it more efficient to not have articles and let context sort out (most of) the ambiguities and have fewer words? Is a language more efficient if it has no redundancy, at the cost of potential mishearing in noisy environments, or if it is redundant and is thus subject to less mishearing?

These are questions which generally lack intuitively satisfying answers. But even in the absence of a good metric for judging efficiency, linguists are pretty sure that languages are probably about equal in overall efficiency, though different languages may differ in where they concentrate their efficiencies, hence the (correct) perception that it takes longer, or shorter, to explain a particular concept in a particular language.