I don't know anything about Philip's path to get where he is now, but there are a number of videos of Philip using his current method of communication, which appears to be typing on a bluetooth keyboard to an ipad app that repeats his letters and words back to him.
The mom holds the keyboard, but Philip moves his own hand. Mom seems to use a lot of cues to help him keep on track--shaking the keyboard to bring him back to it, repeating his words back to him--but I really don't see her writing it for him, moving the keyboard under his finger to spell things, or "interpreting" vague gestures.
https://www.youtube.com/watch?v=MK_Q60LhgzA
This one has several minutes of typing where you can see the keyboard and his hand. It's slow, slow going, but it really looks like he's the one doing the writing.
A clear experiment is to replace the mother with a automatic keyboard shaker. Some device with a button that moves the keyboard randomly to grab his attention again, but most of the time the keyboard is completely still. Also, during the experiment he should not see other persons, because it can send another subconscious cues.
A well known case with less emotional weight is "Clever Hans" http://en.wikipedia.org/wiki/Clever_Hans . It was not a hoax or a scam, it was the unsuspected ability to read some subtle clues. And Hash was a horse and he is a child. He is much more intelligent than a horse!!! He probably is better reading expressions and sublte clues. With some optimism, perhaps he even realized "Etaoin shrdlu" http://en.wikipedia.org/wiki/Etaoin_shrdlu are the more frequent letters and that makes the typing easier.
I used to work for a startup that (among other things) made very innovative teaching tools for autistic children. We worked with actual autism experts who run autism treatment centers, and we designed lessons and a curriculum for our product to teach basic social skills to autistic kids (by basic, I'm talking things like "when you greet someone: smile, nod, say hello").
Yes, I'm being vague about what the product actually is because I don't want to 100% tie this account to my real-life identity, but I can tell you that it's unique and got a lot of buzz.
I had a lot of issues with the company's management (for both technical and non-technical reasons), which is why I quit, but one of the reasons I stayed as long as I did was because I believed in the idea of the product, and I felt that it was going to truly change people's lives for the better. It really made me feel the warm fuzzies when a friend of mine (well, more of a friend of a friend, but we knew each other) who has two autistic children one day sent me a message on Facebook saying "I just noticed that you work for [company]!", and she was super-excited to realize that this person she knows works for a company that she follows in the news (and we were a tiny startup, so it's not like I worked for a big megacorp or anything).
I know many parents of autistic kids, and so many of them are just so terribly desperate for any sign of improvement in their child that they're easily led by charlatans. They'll desperately believe and see patterns of improvement where there are none because they want to believe that they can lead their children to be independent and strong.
I'll start with why: this may not be Philip writing to us. I would like to believe it is, and those of us who may be on the autistic spectrum will not find it difficult to believe that this writing is of the ability they had when they were his age. That confirmation bias is but one of a number of cognitive biases that impel us to find Philip in these words. We are, of course, social creatures, and generally when we read heartfelt things like this we will set aside our critical nature. When presented with new and plausible information in the absence of anything that might indicate contrarily, we err to assume that it is true.
But it likely is not. This blog highlights the "Rapid Prompting Method". This method is not strongly supported in blind tests. There is little empirical evidence that it allows persons with nonverbal autism to communicate, and in fact, decades of failed trials that suggest that this and other methods of facilitated communication do not work. It tugs at the heart strings, and it pains me to say this, but it is sadly almost certainly not Philip expressing these thoughts.
What is rapid prompting? It is a process of rapidly presenting someone with options until they have produced sensible text. In practice, it looks like someone guiding a person who is stimming as a human Ouija board[1]. The practitioner silently omits incorrect options, just as superstitious teens neglect digrams that don't make sense when they are found with a Ouija board.
This may be more than that, I would very much hope so. But the American Psychological Association's position on all forms of facilitated communication is firm: they are fraudulent[2]. Comparing rapid prompting to facilitated communication has already been done, as has been written about in NIH-supported journals[3]. These and other methods are woo, as the skeptical community describes it: not science. This woo is perpetrated by unwitting charlatans[4].
That harsh language is necessary, because we are doing something, I think, much worse than merely mistakenly giving a voice to persons with nonverbal autism. Instead, there are loving, adoring parents and "rapid prompting" experts who are unwittingly casting their voice as another's. Instead of undoing the silence of autism, they are turning people into dummies with a textual form of ventriloquism.
What the hacker community should take from this is this: the desire to help those persons who have nonverbal autism is going to be very strong, but you can do great harm by making something that doesn't work. But if you're hacking something together, if you think you have a way to help, be cautious. Don't give in to the desire to "disrupt" this space. Instead, work with the psychological community, work with those who study neurological development. Don't just make something parents and loved ones want, make something that works.
What society should take from this is that when people have their ability to communicate taken from them, it is very easy to put another's voice in their place and have everyone be none the wiser. It is very sad, but that is what happens to many of those "assisted" with facilitated communication. And as always, be skeptical of cures and success stories. It's easy to be swept away by what we want to be true.
And finally, there are people who want to help those with communication disadvantages of all kinds. For those people working on this right now: Take that desire to make something that works, and rigorously apply the skepticism the public should have - but doesn't - to what you're doing. That is vital, it is key, it is the most important thing you can do. Don't allow wishful thinking to turn a child with autism into a puppet.
[1] https://www.youtube.com/watch?v=xp3G8yM2vTg - One of many, many similar videos you can find.
[2] http://www.apa.org/research/action/facilitated.aspx
[3] http://www.ncbi.nlm.nih.gov/pubmed/24102487
[4] http://www.science20.com/countering_tackling_woo/blog/why_ra...