On saying cat again

Rather more than a year ago, I was ruminating on the business of saying the word ‘cat’ to oneself, without actually articulating it out loud. Ruminations which I find on inspection of the notes I made at the time triggered many happy digressions into the related literature, with reference 2 being a fair sample.

This morning I played with the phrase ‘toy boat’, a play offered here as a footnote to the original post.

The phrase entered my life as a task posed by my mother when I was young: say ‘toy boat’ three times quickly and see what happens. Her point being that it nearly always went wrong. She did not have, or at least did not offer, any explanation of why this might be, it was enough for her for us to share her observation that it did go wrong. I now believe the problem to lie in the relationship between the two vowels concerned, with it being difficult for the vocal apparatus to quickly switch from one to the other. A difficulty which carries over, for example, to the variation ‘bang boat’. Perhaps any pair of vowels will present the difficulty, perhaps to varying degrees.

However, this was not my concern this morning.

Rather, on waking up, I was interested in the extent to which the difficulty with saying ‘toy boat’ manifested itself when one said the word silently, in inner thought, rather than out loud. With the short answer being that it did manifest itself, even with the lips carefully shut. Then, a bit later, on the morning Ewell Village anti-clockwise, I tried doing silent ‘toy boat’s while walking along.

The story there seemed to be that if I synchronised my sayings to my strides, which allowing four miles an hour or two yards to the second,  I made about one full stride to the second, I could get my sayings right. But if I stopped synchronising and then tried to say ‘toy boat’ twice quickly, the vowel for ‘toy’ usually slipped into the ‘boat’ following where it did not belong. Perhaps the difference is that when synchronising, there is a short gap between each saying, a short gap more or less consistent with the average of 120 spoken words a minute suggested at reference 1. Without any gap, the brain is short of time,  tries to cut corners and deploys the long motor sequence needed to do the ‘oy’ in ‘toy’, a sequence which has already been computed and remains available, for the ‘oa’ in boat. Which might have save time but which also fails to produce the right sound. The motor sequence is not near enough.

In any event, in the course of silently articulating the words, I was clearly hearing something in some way or other, which brings us to the figure at the beginning of this post.

Red for the motor commands to the mouth and other parts of the vocal apparatus. We assume a cascade, perhaps comparable to the famous OSI seven layer model of communication from ISO (noticed by Wikipedia at reference 6), with Shakespeare in at one end and zeroes and ones out of the other, perhaps thousands of miles away. In this case, words left and neural firing right. Blue for the signals coming back from the ears, whereby one hears what one says. We assume another cascade.

We also give steps to our ladder, with copies of the signals from the red strand cutting across to the blue strand where they are turned into guesses at what we should be hearing and with copies of the signals from the blue strand cutting across to the red strand where they are turned into corrections to what we are trying to say; a dynamical system with both feed-forward and feed-back. Guesses which result in our seeming, after a fashion, to hear what we are thinking. And which may also result, on a bad day, in what we hear not matching what we are saying very well at all, perhaps correcting the northern accents with which I actually speak to the southern accents to which I aspire. I suppose we are a little way off knowing whether such a thing makes sense in the case of inner thought.

With the curved orange arrow on the right standing for sound waves moving, through the outside world, from mouth to ear, in the event that things get that far.

Lastly, we suppose that we have some control over how far down the ladder our inner thought should go. Do the guesses get better the further down we go?

At which point, a little experiment suggested that one could vary the inner thought in the direction of speech in a reasonably controlled way, with one key step being when one started to move one’s lips and another when one started to voice the sounds. There was an impression that this slowed one down, but further experiment involving beating the time with my hand, suggested that there was not very much slowing down. Inner speech was certainly not a lot faster than speech. So generalising, one cannot think words that much faster than saying them, certainly not ten times faster, perhaps not as much as five times faster. In so far as humans are concerned, business does not go at the speed of light, whatever Mr. Gates might say.

At which point, I thought of M. Dehaene and his event related potentials (see references 3, 4 and 5).

Maybe if one connected IBM’s Deep Blue (about which one does not hear so much these days) or Google’s Deep Mind to an EEG machine which was connected to my head, one could train the EEG machine to recognise the onset of my toy boat inner thoughts, thus enabling one to do some proper experiments on times and timing.

As is it, I think one has to do hundreds of tests and then take their cunningly synchronised EEG average in order to find a bit of brain signal which is reliably linked to my saying toy boat. One then has the quite possibly insoluble problem of mapping that back to, finding that marker in a single test. Quite possibly, there is just too much noise in the brain for this to work.

Maybe the answer would be to not bother with Deep Blue or Deep Mind, and to just plug me into an EEG machine organised so that I could see something on the screen. Maybe my Mark 1 brain could then spot the feature which marked the toy boats? Indulge in a bit of interactive work with brain waves, a game which I think finds favour in California, along with cranky diets and life styles.

Lots of scope here for all kinds of interesting experiments. Perhaps they have already been done.

Conclusions

Further evidence, should any be needed, that the business of inner speech, speech which is not actually spoken out loud, is still intimately mixed up with the vocal apparatus.

References

Reference 1: http://psmv3.blogspot.co.uk/2016/11/on-saying-cat.html.

Reference 2: A comparative analysis of speech rate and perception in radio bulletins - Emma Rodero – 2012.

Reference 3: https://www.college-de-france.fr/site/stanislas-dehaene/index.htm.

Reference 4: http://psmv3.blogspot.co.uk/2018/01/what-is-consciousness-and-could.html. Included for the papal connection noticed at the beginning.

Reference 5: https://en.wikipedia.org/wiki/Event-related_potential.

Reference 6: https://en.wikipedia.org/wiki/OSI_model. A bit more accessible than the ISO web site.