The museum worker looked at me in bewilderment. My three-year-old, Louise, had just explained that her favourite animals were crocodiles and dinosaurs, but hardly a word had made sense to him: in Louise's own private language, animals were "nyownies", crocodiles became "fottyows" and dinosaurs were "dowdies". For Louise, the word "car" began with a "t" and "gate" began with a "d". And she could not pronounce "s" - the most common sound in English.
I asked the GP to refer us to speech and language therapy: Louise's excited flurry of made-up words and inexact sounds made no sense to outsiders. The GP looked sternly at me, then at Louise, who was quietly scratching behind one knee. "That child's worst problem is her eczema," she said, "and she doesn't need speech therapy." Under pressure, she checked Louise's hearing and referred her anyway, warning me of two-year waiting lists.
The eczema got better: the speech didn't. The problem became more painfully obvious. Figures from the education charity I-CAN show that one in 10 children in Britain has some speech impairment. Louise - who is now seven - was clearly one of them.
Six months after the referral, Louise had an initial assessment session, and the therapist gave us lists of sounds to work on while we waited. We worked through pictures, asking her to name them: "gate" not "date", "girl" not "dirl", "cook" not "tutt". Just 10 minutes a day, the therapist advised, or Louise might rebel. Slowly, painfully, her speech changed, and then we were on to the next level: if Louise could say "g", could she then learn to say "gr" and "gl"? And could she put the sound at the end of a word, and say "pig" and "bag"? Then we were on to the next stage: working on "ch" and "j".
We also started working on the baby words. When we told Louise quietly that "crocodile" begins with a "c", not an "f", she was amazed: in her own intensely intellectual world, she was simply not hearing the way other people spoke.
We were still on the waiting list - the city had just a handful of speech therapists, they told us. In the meantime, Louise started school. Other children were starting to make critical comments, though the teacher reassured us that three children in the class needed speech therapy. We comforted Louise, but we quietly despaired. For although Louise's hard work had transformed her speech over the years, one sound eluded her. She just could not say "s" or the related sounds "z" and "sh". Therapists sent us more practice sheets, but there was no point until she could produce the sound.
At a further assessment visit, the therapist explained what was happening. Instead of putting her tongue behind her teeth and producing a clear air stream down the middle of her mouth to produce an "s", Louise was raising the middle of her tongue so that the air hissed out of the sides of her mouth. Unfortunately this harsh, slushy sound - therapists call it a lateralised "s" and say it is very hard to change - was louder than the rest of her speech.
Desperate for help, we paid £50 for a session with a private speech therapist. In the quiet of our kitchen, Louise's face distorted as she strained to make the sound. The therapist asked her to say "t", then to extend it - "ttttt . . ." And for the first time ever, Louise produced a tiny "ttttt . . . s".
We were left with exercises to practise, but Louise could not make the "s" sound again. I badgered the therapist, who was reassuringly frank. "I could have six good sessions with her, but after the session she might still not be able to make the sound," she said. "You've got a lively and intelligent child there, so why spoil the rapport you have with her by correcting her all the time? And she might correct the sound herself when she reaches her teens and feels the peer pressure." Why had Louise's speech gone so badly wrong, I asked. The bottle of milk Louise loved when she was two may have done the damage, she suggested: a child with a bottle or dummy stuck in its mouth may learn to talk round it and produce a lateralised "s".
Then - finally - we were at the top of the NHS waiting list. Suddenly, miraculously, we were offered four two-hour group sessions in the holidays. These were an eye-opener. All the children there had the slushy "s" and some were painfully ashamed to attempt it, hanging their heads. One sat frozen while her mother told us of the bullying she endured daily at school. Over those sessions and another four some six months later, the children played games that focused on "s" sounds. We went home with exercises - first "s", then in a short word ("sit", "say"). Louise drew sandals and sun cream in a suitcase, described seaside scenes, gave herself 10 ticks for each time she repeated the word "seven", and our lives revolved around speech therapy. We played games in the bath - "seven objects starting with 's', now close your eyes and tell me which one I removed" - and over meals.
Speech therapy was ruling our home life: I asked for more sessions. But the therapist said research shows that endless repetition is crucial in changing children's speech - and anyway, they had those huge waiting lists to think of. After the eighth group session they discharged Louise. Set aside 10 minutes a day to remind her, they counselled, because only then will she generalise her clear "s" in her everyday speech. And so we finished the job as we had started it - at home, with the minimal help that the hard-pressed NHS therapists could offer us. Louise has slogged at her exercises on beautiful spring days, forgone museum visits, put her language exercises before her school reading and maths. Her own determination has changed her speech. Now when she chats to museum staff about school, fossilised sharks' teeth and the snakeskin she found in Cyprus, they never look at me to translate.
• To find out more about I-CAN, call 0870 0104066 or e-mail info@ican.org.uk