Previous: learning languageNext: coherent discourseContents

Society of Mind

26.11 grammar

How do we choose the words we speak, and how do we understand what others say? Earlier, I suggested that in the course of learning language we accumulate various processes and tactics that enable us to partially reproduce our own mental operations in other speakers. These processes affect our choices of words, the forms we select for phrases and sentences, and the styles in which we frame our narratives. There have been many attempts to study how children learn language, but psychologists do not yet have coherent theories about the underlying processes. For example, we do not yet even know whether we learn each bit of grammar only once — or whether we have to learn it twice, for speaking and for understanding what other people say.

We know so little about such matters that we can scarcely even speculate about the nature of those early language-learning steps. Perhaps the process starts with some agents that can enable a child to make various vocal sounds in response to specific internal states. These agents then become involved with built-in predestined learning processes that lead to limited abilities to imitate other sounds the child hears by using feedback from its ears. Later stages might then engage new layers of agents that connect word-sound agents to whichever polynemes are most frequently attached to certain pronomes in the language-agency. Once a suitable variety of such processes are established, more layers of frame- and memory- controlling agents could learn to support more complex language skills.

Let's try to imagine what kind of process could produce a language phrase that expresses a description of an object. Suppose, for example, that you want to draw attention to a certain very big box. To imagine such a thing in the first place, you might first have to activate your polyneme for box and then arouse some other isonomes and polynemes that modify the state of your Size agency. To express very big box might thus require grammar-tactics that express three mental operations:

--- box expresses the arousal of the box polyneme; --- big expresses a process that selects the Size agency;

--- very expresses an isonome that adjusts the sensitivities of agents in whichever agency was selected.

I do not mean to suggest that a child's earliest three-word noun phrases must be based upon such complicated processes; more likely they begin with simpler sequence scripts. Eventually, though, more complex systems intervene to replace the simple scripts by intricate kinds of frame-arrays that enable the child to make more complex rearrangements of what becomes attached to its expression-frames. Then, as the language-agency acquires more isonome-controlling skills, the child can learn to use pronouns like it or she to express other structures that are already attached to suitable pronomes. Also, as we develop skills for building chains and trees from other frames, the language-agency can learn to use corresponding grammar-tactics to express those chains — stringing together phrases and sentences with conjunction words like and

and but. Similarly, as we improve our methods for controlling memories and managing interruptions, we can learn to combine those skills with clause-interrupting forms like who and which. There seems scarcely any limit to the complexity of our social inventions for expressing mental processes, and it takes most children many years to master all the language arts their ancestors evolved.