I've been drafting a science fiction story that places a lot of emphases on the vastness of culture. One way technique I've been using is language and the hassle of translation. Earth has abut 4,500 languages spoken commonly, and that's not including slang and accents (some parts of Scotland needs to be subtitled on British television, I am told). In this particular setting, that number is multiplied by the number of species and increased further by the vast distances between settlements and colonies.
For face-to-face communication to work between radically different species, electronic translators have become extremely common. The most common type is an AI equipped with directional microphones, a vast compendium of languages, and an equivalent of an earpiece. This AI focuses on a speaker, translates in near-real time, then feeds the translation into the earpiece (or projects subtitles onto a visor, case depending). Its not a perfect system, but its well-programmed enough to handle groups of people speaking simultaneously, provident that it what its translating to what.
Example from the text: It took one of its hands off the gun to pull a electronic slate out of its belt holster. It fiddled with it a bit, pausing every few seconds to speak a few words of various Carapaced languages. Quotze thought about trying to pantomime “Artech-Westle Reformed”, but decided he would much rather not talk to this heavily armed alien.
Anyhow, I got to thinking about the limitations of this system and how that would effect both culture and the story. Body language can't be translated, that's for sure, but what really interested me was music. Because the AI translates for maximum accuracy, rhyme and meter would disintegrate across the language barrier.
"You ain't nothing but a hound dog, crying all the time. You ain't never caught a rabbit and you ain't no friend of mine."
"You are like a beast for hunting that despairs constantly. You have never caught prey and are not my friend."
Am I right my assessment of this flaw? Even assuming that the translator AI is packed with common idioms and sayings, wouldn't songs and poetry suffer with this kind of system? I'm especially curious how someone singing by themselves would sound, as well as how interplanetary "pop"-music could work.