Put simply PPL holders will not be able to train candidates for licences or
ratings, but they will be able to train Licenced pilots for some of the more
esoteric endorsements, such as spinning and aerobatics, formation flying, banner
and glider towing and parachute dropping
I'm not sure what the current regs are for glider towing, but I had "Glider Towing Checking and Training" approval on my licence in 1984--on a PPL.
It would be onerous to expect prospective tug pilots to find someone who has a CPL + IR to issue a towing rating. The expertise for such a rating already exists at you local neighbourhood gliding club and the instruction would be sufficiently comprehensible and practical to train a competant tug pilot in-house.
The Gliding Federation of Australia has been running its own show without too much interference from the government regulator for decades. The GFA sets standards and develops training programs for glider flight training and maintenance/repair to a very high standard. You don't hear of many gliding accidents, so they are obviously doing something right.
My personal experience with gliding instruction was that the instructors (who did a 50 hr GFA course= same as a Grade 3), were mostly excellent trainers. I can't say the same for some of the disinterested G3 and G2s I had who simply wanted to log hours for a Qantas interview.
It was a GFA requirement in the 1970s that glider pilots be able to enter and a recover from a fully developed spin - with 1 complete turn- prior to first solo. The nature of gliding is that thermaling and wave soaring is mostly done at slow speed, in a turn in gusty conditions. Whilst GA light aircraft are not often operated in this way, I think that the confidence that comes from being able to effect recovery from a stall/spin is of benefit. A G1 instructor who is afraid of a stall probably shouldn't be in their position!
As has been said before, the UK experience with Assistant Flight Instructors has been statisfactory. In a club enviornment to PPL would probably work ok. Is there any evidence to suggest otherwise?