The point I was trying to make about "arrangement" is that an "arrangement", like any literary dramatic artistic or musical work has to be pass a "substantiality" (and "originality") threshold in order to amount to a copyright work.

In the case of a literary work, the alteration of a few words would be unlikely to create a new work.

It seems to me that most unaccompanied singers do not vary the actual notes of the principal melody that they sing. They may change key but a mere change of key (that is to say for this purpose a mere transcription up or down by so many semitones, it might be differnet if it was a change from major to minor or vice versa) on its own is unlikely to amount to an arrangement.   They may add some decoration but in most cases decoration is largely commonplace, following established cadences. A single singer cannot add a harmony line to himself. Likewise there is no accompaniment. That leaves timing. I suppose it leaves loudness too but I cannot really see variations in loudness being held by a judge to amount to a fresh musical work (which is the threshold that an "arrangement" has to surmount).

So what is there in timing? Can you think of a single traditional song that is likely to be recognisable if the notes were deprived of pitch so that it became a sequence of clicks? The only possibility that springs to my mind would be "Famous Flower of Serving Men", but that in turn is because I play it with rhythm rather than (as Martin Carthy said) in the time signature of "one". People do tend to say that what I do to traditional songs is distinctive (not necessarily the same as "good") but I think that's pretty much just because I grew up wanting to play "chunk-a-chunk" rock and roll so I just thump a rhythm and syncopate a bit. I wouldn't really call that an arrangement, although I might claim that for some of the guitar parts.

If that doesn't meet the threshold what will an unaccompanied singer be doing that will amount to the making of a new copyright work?   Mostly they (and, if I sing unaccompanied, I) just convey the work.

More importantly, if you were trying to prove to a judge that there was an arrangment, you would surely have to show the "before" and "after" versions of the work. Ex hypothesi that would be a bit tricky.

It's not wholly cut and dried, but that's my view. If you really want to get a feel for it you should steep yourself in copyright textbooks - Laddie Prescott and Vitoria is well regarded, likewise Copinger and Skone-James. Sterling and Carpenter is less well known but I think it shows a good grasp of principle, and there is an Australian book, now very out of date, the author of which I forget, that again showed a good grasp of the principles.

Some of the modern cases are showing what I think to be a heretical tendency to reconise an arrangement or fresh work when in truth there was no such thing. Sawkins -v- Hyperion falls, for me, in that category.