@Chris W - Fair Use
As an individual you have the right, under the Fair Use provisions of copyright, to make personal copies and derivative works for your own, non-commercial use. You could purchase a book, then painstakingly copy it by hand to a blank volume, illuminating the words and illustrating to your heart's content. Likewise you can read the book aloud into a tape recorder and allow your family to listen to the tape as they like. This is protected under law.
However, the Kindle falls afoul of the same laws that are being applied (wrongly, in my opinion) to BitTorrent and other forms of unlicensed reproduction and distribution. This is illustrated by the way other eBooks that include text-to-speech disable the feature for DRM-encumbered works.
Text-to-speech implemented in a device which has the sole purpose of delivering copyright content for VISUAL reading (ie, the eBook derivative license) does NOT automatically allow audio reproduction (the audio book derivative license) - unless the copyright covering the specific work explicitly includes the audio reproduction permission as well. It is a case that text-to-speech has NO NON-INFRINGING USE in the device.
If the Kindle allowed playing audio files (music, audio books) this would be fine. Fair Use allows you to make a copy of a music track and play it for your personal use on another device, and this would include you making an mp3 copy of an audio book for later personal play-back.
The Kindle doesn't just reproduce something that has already been purchased in the licensed format: it PUBLISHES a derivative work on-the-fly. Unless that right has been purchased by the licensee that makes the work available as an eBook derivative and the audio derivative is extended to Amazon as part of the copyright included with the work, then Amazon's inclusion of a feature that can ONLY be used to infringe on the derivative work copyright is illegal. And the use of text-to-speech with non-copyright works, while completely legal, is only NOT infringing because the original copyright - and therefore all of the derivative works - is no longer in force.
This is really a huge, HUGE can of worms being opened. The audio book derivative is just the first hair from the camel's nostril poking under the tent. I can envision in the near future a home super computer with an AI that will take a book text as input and create a lavish, high-def anime-like video of the story, complete with editing and dialogue. This would completely transform the motion picture and television industries overnight. Notice I didn't say "wipe out": these industries would be just like the buggy-makers of the 19th century who transformed into auto makers.
My contention is that if Amazon pays a license fee for the text-to-speech derived works NOW, it both solves the problem AND sets precedent for the NEXT wave of innovation. The MPAA would get HAMMERED by the above movie scenario, but they couldn't argue against the precedent...and the author of the story - who normally gets only a few hundred thousand for their story AT BEST from a studio - would see automatic additional revenue for the "movie" rights built right into their copyright payment.
I hope this clarifies some of the points in question here.
(BTW, I'm "well read" in this topic, but I'm NOT a lawyer or solicitor, so check with you legal counsel for a more technical explanation.)