How the quest to learn what is a classiomatic led me to discover the speech-to-song illusion

Listening, today, to one of my favourite Duran Duran songs, "The Chauffeur", I found myself unable to remember the lyrics from start to end (hardly surprising, as I was doing something else at the time), and, heading off to a lyrics website, was surprised to find a few extra lines on the end compared to what I was expecting, concluding with "Who's going now, in to a classiomatic?"

Going back to the music, I discovered these lines were spoken, not sung, and not very loudly, over the music. But what on earth was a "classiomatic"? Heading off to Google, I discovered this blog post, which investigated trying to make out these lyrics in an entertaining manner:

Google |what is a classiomatic| and you’ll get this page first: a foray into that dungeon of the benighted, Yahoo Answers, where the blind lead the blind and other blind people vote on which road looks best.

In the course of this, the author goes into a most interesting digression:

But this is a real exercise in top-down effects on speech perception. If you think you hear something, it sounds more like what you think you hear. It’s similar to how things in the dark can look exactly like human forms and faces, and then when you turn the light on they look nothing like them. Categorical perception is a fascinating issue in linguistics; it’s why anglophones often aren’t even aware that there’s an aspiration on the p in pan that’s not there on the p in span, even while some other languages would treat those two sounds as different, and why speakers of Spanish merge some English vowels and speakers of Japanese have a hard time with l and so on. It’s also responsible for the speech-to-song illusion, whereby a snippet of speech, taken out of context and repeated, can come to sound enough like music that many people will repeat it as music.

I'll second the author when they add:

(Check out the link. It’s really interesting. For the record, I heard the speech the same all the way through, as did some others I know, but others – no less linguistically trained, to be sure! – hear the illusion.)

Be sure to listen to the audio samples; they're most amusing.

FWIW, I was halfway in between the sampled people and the blog post author: listening to the snippet over and over again made me much more aware of its prosody, and its the absolute pitches, but it didn't "burst into flower" as song for me, and if I'd had to reproduce it, I'd have reproduced the prosody but as speech, not song.