But how does the brain represent words and associate them with semantic meanings? How does it categorize words and represent syntactic rules? According to our model, the brain accomplishes this with cortical images, representations of the activity within the brain itself.
Within the brain, all representation consists of electrochemical activity. The abstract idea “neurolinguistics,” the picture of a tree, and speaking are all represented by different patterns of neural activity.
The representation of the visual image of a tree might consist of a pattern of cortical activations relating depth, spatial geometry, color, and brightness. The image of the tree does not exist in the brain; only a pattern of neural firings representing that image. Fundamentally, everything that comes into our minds reduces to patterns of neural activities. Just as the brain can produce cortical images of sensory input or motor activity, the brain also produces images of its own cortical activity. These cortical images, representations of cortical activity itself, are important for binding semantic content.
In our model, language acquisition is an adaptive process. Words are bound to semantic meaning by constructing call trees around cortical images. For instance, the word “neurolinguistics” triggers an array of connections – the phonetic sound of the word, the motor movements necessary for pronunciation, the faces of people in the neurolinguistics class, the grammatical category for nouns, etc. The cortical columns responsible for the word activate all of these relations. Words do not contain any content of their own; they are hubs that link many other parts of the cortex. The word “neurolinguistics” is a cortical image of all the activity relating to it.
The ability of an infant to produce a wide range of phonemes (babbling) is suggestive of the permutative power of the cortical columns to produce random-like results.
Semantic categories are images of the cortex recognizing patterns in itself. For example, the semantic category “colors” are directly related to the activation of a certain part of the visual cortex.
Gradual differentiation and refinement of grammatical rules by adding increasingly precise rules on top of more general ones.