There were two studies that were pretty much directly responsible for this web meme: (1) K. Saberi and D R Perrott. 1999 Cognitive restoration of reversed speech. Nature 398: 760; (2) Graham Rawlinson. 1976. The significance of letter position in word recognition. (A PhD dissertation at the University of Nottingham.) Mr Rawlinson also wrote a letter to the editors of the New Scientist that had an earlier version of the jumbled letter text (quoted on my blog and linked in my earlier posting). That text made it to the internet first, followed a couple of months later by the text quoted at the beginning of this thread.

Back when I was keeping track of this, there were German, Portuguese, and other versions of the text (see Language Hat), plus many scripts (mainly written in Perl) to jumble texts. It's interesting that the Saberi and Perrott study is really about reversing the middle sections of words acoustically (tape splicing / digital editing), which reminded Mr Rawlinson of his dissertation all those years ago. (I have not seen a copy of that unpublished dissertation.) His work looked at jumbling letters in words.

Ceci n'est pas un seing.