Yes, I’ve found those ‘backers’ so tricky too! A little break won’t hurt if you all need it to break the cycle – but I ‘d have thought it’s unlikely to make him more stimulable.
If you decide to keep going here are a few ideas I’ve found helpful from an Anne Harding -Bell course:
•Forget the lateral for the moment – easier to get the central airstream going once you’ve the placement of the alveolar correct.
•Rather than calling it ‘t’ use a novel ‘non – speech sound ‘ label like ‘ tiny tap dripping noise ‘. So at this stage the ‘new sound’ he’s going to be hearing and making has no phonological meaning and goes into his phonetic motor programming. No past failure with this sound either because it’s new!
•Articulation wise, like Nikki said, I’d model the sound much further forward and give him as many cues as possible to help him to see and feel what’s going on with his articulators. Model it as a soft ‘t’ with audible aspiration, and give a tactile tap on the back of his hand as you model it . The visual focus is ‘tip through the lips’ tap noise – a relaxed tongue between lips.
•To start with I did all the modelling – “You watch my tongue as I make the tiny tap dripping noise “ . The ‘discrimination’ task was just “Did you see my tongue peeping through when I made my sound ? (every time) – great ! You win a flag to pop the dragon “ etc So, all input work at first, to strengthen the neural pathways and drive a new motor programme from just watching (and listening) rather than focussing on motor execution which may lead to lots of false ‘wrong’ outputs and makes him feel he can’t do it .
•I got Mum to video bits of me doing the sound modelling on her phone and watching it back was his home practice during the week. Something goes in subconsciously; he isn’t a risk of failing and he gets loads of input work watching you do it.
•Then I added in “You put your tongue just like mine as I do the tiny tap dripping noise” So he makes the articulatory gesture (and gets some proprioceptive feedback but no pressure to say the sound ) I make the sound , he watches and gets his token or whatever. He’s not failing. Output is voluntary – what I found was he spontaneously joined in with the output when he was ready “Great, I think your tongue just managed that new sound!” It’s a case of “See it, Store it, Say it ! “
•Then I moved to modelling CV nonsense and then CV real words STILL with a ‘tip through the lips’ ‘t’. The emphasis was still on him watching me, whilst making the articulatory gesture himself – by which time he was saying it without really realising. At this point he could watch himself in the mirror/phone and check what his mouth was doing. At any point a velar crept in (which it may well do once you hit real words already stored with a wrong motor programme) I went back to a model with a visible tongue tip.
•We then moved on to short words and gradually I started to move my articulation back as I modelled words ,until it was alveolar, and his own articulation followed quite naturally and bingo we had a good ‘t’ . You could add in any contrastive minimal pair work with velars at this stage, but we didn’t need to.
•Once all that is sorted and you have a good ‘t’ he will be stimulable for ‘s’ and you can use the ‘t’ as the start of a new novel sound with a central airstream . I use Anne’s ‘quiet train slowing down noise ‘ ‘ttttt tsssssss’ into ‘ ts ts ts sssssss’, similar I guess to your ’flat tyre sound. ‘