I think the sample has to be recognisable. If it's not and you cannot associate what you hearing to something that may have contributed to that sound then it becomes your sound because you have created it. And what you play from then on will also have no resemblance to the original sound.
You could take a sharp attack with a horn blast or some very short sound. Transpose it down a few tones, put it into a granular synth and reconstruct the sound particles a bit, tune it down an octave, loop it, reverse part of it and double the sound, pan left and right and detune then go in create a new ADSR shape. What you have then is something that has no bearing on the original sound. It could be argued it is your sound now.
The way the use of samples has been discussed here is only one approach. Sure you can do interesting things with them. But cut them right down and rebuild a sound and create something completely new out of it is open to being a very viable way to adding to your sound design technique. There is no copyright infringement here as far as I can see. How is it ever going to be linked back to its original source.
You don't have to disguise things so much either. I recorded a group of world musicians playing, certain things at certain tempos. If I wanted to use that music as it is as performed and add to it I would need to get the go ahead from those players and that is fair enough. I made the recording and own the recording copyright. A lot has changed since then (1995) I could re arrange every bar now and create music that was never played at the time, percussive rhythmical grooves that had no relationship to the original performance, now at any tempo and pitched to different key to what was originally recorded. Who owns that material now and has the right to overdub other parts to it and release it. I have a pretty good feeling it is me.
What about taking an original performance and creating a sample CD full of one shot sounds only. Totally new and unrelated performances can now be created using the sampler with these one shot sounds loaded. You are creating an instrument now with its preset.
What about this concept. A flute player is playing a lovely melodic lines on a deep wooden flute, phrases and a total musical performance. But every now and then lands on a note and just sustains that note. How about I sample only the long notes, perfectly loop them in the right parts to turn that into infinite sustain. These notes end up inside my Emulator say or Kurzweil and of course I can fill in all the other notes around the ones I recorded. I now have the perfect Emulator or Kurz patch that plays perfectly and chromatically over two or three octaves. Is this my sound now? I say yes it surely is because the music that I will play with it will never resemble the music from which the long notes came.
With the world musicians I recorded they all agreed at the time that anything recorded during these sessions would be basically mine to use as I saw fit. I have enough material there to say create 7 or 8 CD's of just one shot samples alone. But I did say that if I used any of the material as it was intact in terms of how the others played it, any income I made form that track would be split 50 / 50 between myself and the other players. There were many sections where the playing was incredible and does not need to be changed at all, just overdubbed with additional material. My material makes the music twice as interesting so that is the reason for the 50/50 split.
I am into the microscopic use of sound and samples from anything at all. Designing it in such a way that the original sound ends up creating something totally different. Does it matter then what the original sound actually was. If a lot of creativity goes into creating the final sound then the creator deserves the recognition for doing it.