I think most people agree that we do not actually think in words. If we did: we would not search for the right word to express a thought; we could not think as pre-verbal infants; we would not have integrated thought if we were bilingual - and on and on. We do not think in actual words but perhaps in something like concepts. Those concepts could be represented by words or pictures or sounds or whatever. But do we actually think in concepts. Some would say that we think in symbols and with them we can ‘compute’ or think. Perhaps we have symbols standing for concepts and we manipulate these with some sort of rule system.
Here is a quote from Schlenker. “This does not mean that thought is not a system that manipulates symbols; in fact a widespread contemporary model, the ‘computational model of the mind’, suggests that the mind should be analyzed by analogy with a computer, which manipulates abstract symbols. On this view, thought is just symbol manipulation. But the symbols in question need not be part of verbal language; they may be part of what Pinker calls ‘Mentalese’, which is just another term for ‘language of thought’.”
I do not find this very convincing. Are we not just guessing here? What do we actually know about thought? We know what we are conscious of, but not necessarily its original form, and we know what we can observe in some tricky experiments. We also can make some assumptions. That does not add up to much understanding.
Let us make some assumptions from the bottom up for a change. If we climb phylogenetic history, we have no brains, then increasingly complex brains. Somewhere in the simplest brains are the simplest thoughts and in more complex brains are more complex thoughts. Thought is after all what brains do – they are matched, inseparable and just two aspects of the same thing. The only processes of thought that are possible are those that the brain can do; the nature of thought and the structure of the brain must evolve together. As the brain does not look or act like a computer and cannot (so far) be one-to-one mapped to a computer, the whole computer-brain metaphor has to be taken as approximate.
There are many things that the brain does that we can be relatively sure of. Take ‘objects’ as an example. We know that in reality objects are very hard to put a finger on. They do not have crisp boundaries. But our brains create very separate, clearly bounded objects; it places them in very definite places in space and tidies them up with more consistent colour, size, texture etc. It attaches the input from various senses and memories together so that by seeing an object we may know how it will feel, smell, handle etc. The object can acquire a meaning. Is an object a concept or a symbol? Do we care about this label? The difference is hard to pin point – has it to do with completeness or implied meaning or what?
If we are thinking about an object in the category of ‘tree’, does it matter whether we are thinking the word ‘tree’, or the image of an archetypal tree, or the memory of a particular tree? The brain seems to use some fixed categories (such as ‘face’) and seems to have the ability to create categories when it is useful to group similar objects. The objects have attributes (or they bind to characteristics such as colour) and it appears that similar attributes allow categories to be formed. We can call objects and categories of objects by names like concept or symbol if we like but it does not change their nature as facilities of the brain.
Besides objects we have places and the brain’s way of creating and dealing with space. We have events, those units of memory that are closely causally and temporally connected. Events seems also to be created out of an undivided reality. They can be strung together to make larger events. We create a method of motor understanding that goes something like: need/opportunity – goal – plan – intent – decision – action – outcome. And when we find that we can understand a particular event in this way, we assume the actor is animate and thinks; we do ‘theory-of-mind’. This idea of thought is starting to resemble language with its nouns, adjectives, verbs, subjects and objects, sentences and so on. But the arrow is in the direction of language resembling thought and not thought resembling language. By language I am including forms of “mentalese”. If there is mentalese then it would resemble thought.
But why should the brain do it’s thought this way. Why create objects, places, events out of an undifferentiated reality? I find an answer, maybe a correct one, in the need for consciousness and memory in order to integrate brain activity and learn from experience. It is an awareness and storage problem that leads to creating these concepts/symbols as a way of accessing information.
So, in summary, my assumption (just a guess really) is:
language developed to be compatible with the nature of
consciousness/memory/learning-from-experience which, much earlier, developed to be compatible with the
low level operations of the brain in sensing and acting (without these operations forming a integrated model of reality).