On saying cat

Sampling my own activities (see reference 1) has shown that when out walking, I spend a lot of time on inner thought which is nothing to do with my walking or where I am walking. Inner thought which is usually in the form of words, often in more or less complete sentences, sometimes supplemented by images. I don’t think there has been an outdoor case of an inner picture dominating inner words, probably because the vision systems are busy enough looking after my walk, even if I am not paying much attention to what they are up to, to be able generate inner pictures. Sound is a bit different, not participating much in either the walking or the inner thought and the sound systems do have spare capacity, with the result that I quite often misinterpret stray sound stimulae, particularly when walking on a busy road, quite often managing to extract my telephone’s ring tone from the noise of passing cars and lorries. Lorries are better from this point of view as they make more varied noises.

But today’s rumination is about what goes on when I say a word in my head, without articulating it out loud.

I think I am reasonably clear that a lot, if not most of the time, I am quite near articulating the word – for example, ‘cat’ – out loud. A lot of the preparatory work has been done and while the word might not be on the tip of my tongue, it is not that far away. The relevant motor programs have been retrieved and activated.

What is not so clear is the extent to which the systems for hearing the word ‘cat’ have been activated. I dare something of the sort goes on when one actually goes to say the word ‘cat’ because the brain likes to check its output and there is probably time, at least in a longer word than ‘cat’ to correct it on the fly, perhaps correcting the tone or emphasis, and failing that, to say the word again.

It is plausible that thinking the word ‘cat’ is coupled with both speaking and hearing systems and takes time, perhaps less time than is required to say the word out loud, perhaps of the order of a one or two tenths of a second, with a quick google suggesting that the miscellaneous spoken word might come in at around 120 to the minute, or five tenths of a second each. Less for a short word like ‘cat’.

So, at this point I assert that the brain knows that you are going to think the word ‘cat’ before you get around to doing the thinking. The time-expensive conscious thinking is a bit of an afterthought, as it were. And this will be even more true of complicated words like ‘elephant’ or ‘brontosaurus’, involving more complicated speech programs than ‘cat’. All this being the subject matter of the posts at references 2 and 3.

Another wheeze is to try and visualise the word ‘cat’ written out, either in type as in a book or handwritten. Not difficult at all if one pays attention and shuts one’s eyes, easier still if one is content to do it a letter at a time, but not so easy when out walking. In any event, not relevant here, as one has to think of the word before one visualises it. I dare say, from time to time, people do visualise words in this way, spontaneously, but I cannot recall doing such a thing myself.

With entirely different considerations coming into play when one is in conversation or reading.
Another question is how the brain codes for cat while it gets the vocal apparatus going, what does it use to get that apparatus going? Does it keep an image of the word and work from that, but without bothering to make that image conscious?

We might get a clue from the animal game introduced at reference 4. When playing the animal game, with this note in mind, before I silently said each animal word, there was a sense of a pre-echo, something going on before I actually articulated the word in silent speech. I decided, possibly quite wrongly, that the brain was replaying hearing the word very quietly and very slightly in advance of replaying saying the word. From where I associated to the notion that in the growing child, understanding words comes before saying words  and learning a new word starts with hearing it – while in evolution the two skills must have come on in parallel, each driving the other forward.

All of which leads onto the illustration above.

We imagine the brain, or at least some portion of the cortical sheet, as a two-dimensional space containing assemblies in Greenfield/Buzsáki speak (see reference 5), with each assembly being represented by a point, or perhaps a small disc. Each of these assemblies is a bit of memory and each delivers an image of something: perhaps a recording of a bit of sound, an action plan to produce a bit of sound, a bit of picture or a word.

These images will vary in the extent to which they can be made conscious. A picture of a cat would be a good candidate, while a list of instructions to the kidneys to tell them to do something or other would be a bad candidate. While, somewhere in between, it might be possible to make a reduced form of some of the images conscious by means of the emotions or feelings that they come with. In mathematics-speak a many to one map: many images might deliver much the same emotion. So maybe there is a funny smell in the jungle and I am afraid, with the under-the-hood, unconscious link being a fear of funny smells in general. I have been taught to play safe and all smells are bad unless known to be good. But I don’t make anything, any ideas about what the smell might be, conscious at this stage, because the field is too wide, I need more information to narrow it down a bit, to something which can sensibly be made conscious.

We note that these assemblies are apt to be changed by use, by the context of that use. They are not read-only discs in the way of DVDs. However, I don’t think that this aptitude is relevant here; for present purposes they are, they can be thought of as read-only.

Such assemblies are clustered, so, for example, we might have one cluster of assemblies to do with domestic cats in general. We might have another to do with Joey, the newly arrived Siamese cat next door. Such assemblies are also pork-pied, with segments for the various modalities – with the collection illustrated not being exhaustive. Note that the segments are labelled with verbs because the images they contain are active. They have to be played to work; just sitting there is not enough at all.

It seems likely that there will be a great deal of overlap between these clusters in our two-dimensional space, with the separation not taking place in the space occupied, rather by their connections. The upside being that everything is potentially connected to everything else.

Each cluster has a centroid (the allusion to the centre of mass of an object is deliberate), perhaps just a single neuron which fires when any member of the cluster is activated. And it works the other way: fire it up enough and the cluster as a whole is activated. With this activation probably qualified by context dependant synaptic weights. The synapses are everything, with the centroid neuron by itself not being of much interest. A sort of neural identifier, in much the same way as the string YL010160M might possibly identify me to the tax people – but without it, in itself, telling them very much about me.

Note also that the activity of the centroid does not necessarily mean that any of the conscious-capable imagery associated with that cluster is actually conscious. Such activity is a necessary but not sufficient condition – at least in the sense that an image could not be active for long enough to be conscious without activating its centroid.

We also allow degrees of consciousness; with one thing being more conscious than another, with some other thing being only just conscious. And it may well be that as consciousness drops, the quality of the image may degrade, the power needed for a good display is not there. Rather as a poor video signal can break down (or perhaps break up) on a television. And it all maps quite neatly onto the threads and frames advanced at reference 6, that is to say with the evolution of an important assembly in time mapping neatly onto a thread.

There is also a higher level network whereby one centroid activates another. Activating the centroid for the Joey cluster is likely to activate the centroid for the domestic cat cluster, and vice-versa.

So, finally, getting back to saying cat, the centroid for that word would need to be activated for the word to be said, but what had caused the centroid to be activated, what else had been activated at around the same time and what would be made conscious and when would depend on circumstances.

But I think one could say one was doing UT – unsymbolised thought – in Hurlburt-speak: and as soon as one is poked, perhaps by one of his bleepers, one knows that one had been thinking about cats, it was just that one did not know it at the time.

PS: I was reminded in the course of this post that most of the English words for domestic and farmyard animals are short and monosyllabic, presumably to reduce the brain and mouth labour involved in using these frequently used words: cow, bull, pig, boar, bull, hen, duck, sheep, goat and so on. Concerning which a quick google suggests plenty of technical complications, and under that a law discovered by one Jean-Baptiste Estoup and named for one George Kingsley Zipf, according to which the frequency of use of a word is inversely proportional to its length.

Reference 1: http://psmv3.blogspot.co.uk/2016/08/descriptive-experience-sampled.html.

Reference 2: http://psmv2.blogspot.co.uk/2015/03/but-he-that-hath-steerage-of-my-course.html.

Reference 3: http://psmv2.blogspot.co.uk/2015/11/free-will-1.html.

Reference 4: http://psmv3.blogspot.co.uk/2016/05/new-game.html.

Reference 5: http://psmv3.blogspot.co.uk/2016/11/a-day-in-life-of-brain.html.

Reference 6: http://psmv3.blogspot.co.uk/2016/08/describing-consciousness.html.