3.2.3 Rules as Mental Representations



Children represent the rules they detect in their ‘mental grammar’

Evidence from studies of children like Barbara show that learning English (or any other language) outside of classrooms involves the active (but unconscious) construction of it in the learner’s mind. Infant learners are not explicitly taught, and neither do they just passively absorb and copy what they hear, like parrots. Novel forms like Barbara’s drinked and ringed demonstrate that they work out patterns from the data that’s available in the speech used to and around them.
This means that they are storing, analysing and creating knowledge in their minds, in what cognitive psychologists call mental representations of English. A mental representation is the temporary or permanent form that information takes when it’s stored in human memory. When we recognize something that we have experienced before, we’re matching its features with those that are stored in a mental representation.

  In Depth

We have mental representations of words (in a mental lexicon) like we have mental representations of concrete entities (e.g. clocks) and abstract entities (e.g. time). And we have mental representations of the meaningful patterning of words (in a mental grammar), just as we have mental representations of the meaningful patterning of concrete entities (e.g. the orientation of the hands on a clock face) and of abstract entities (e.g. ‘six minutes before the hour’). See Jackendoff (2012) for an entertaining introduction to the mental reality of language and its relationship to thought and meaning.


activity  Activity

Look at the photograph in Fig. 3.8 and think about the process by which you recognize what the objects are, even though it’s highly unlikely you’ve ever seen these particular ones before. Be as explicit as possible about what your mind is doing to recognize the objects portrayed.

  • Figure 3.8: Helsinki city bikes

    [Source: Tomisti]






Children ‘construct’ their language gradually through experience

In the same way that people learn about bicycles by seeing them (and maybe using them to get around on), they learn their first language(s) as children by hearing them and using them in interactions. This is a very gradual process, because languages are so much more complex than bicycles!

For one thing, it’s relatively easy to distinguish the different component parts of a bike, and to see how they function: the saddle is what you sit on, the pedals are what your feet operate, the wheels are the two round objects which spin over the ground, etc. Language is a very different phenomenon. For children like Barbara, language is the exchange of countless streams of speech which:

    • ● unfold moment by moment in constantly shifting acoustic shapes;
    • ● sound different each time they are produced (depending on who’s talking and under what circumstances);
    • ● accompany all sorts of communicative acts and communicated events, some of which may not have immediate visible manifestations (e.g. when someone is talking about something that happened in the past).

Children learning their first language(s) are in some respects faced with the same challenge as an adult learning a new second language (L2). Recall the last time you heard a new language which is unrelated to the one(s) you already know. It’s likely that you didn’t even know where one word finished and the next one started. (Learning about new physical objects, like motorbikes once you know what bicycles are, is much easier!)

But children learn languages remarkably fast and they do this by storing and analysing input automatically, at an unconscious level. As they distinguish different units within the stream of speech they’re exposed to, they store them in memory, and each time they hear a unit again, its memory trace gets stronger. As they detect units co-occurring with other units, they also store mental representations of these groupings, and gradually these get more and more abstract.

In effect, they construct their language bit by bit, developing prototypical units out of millions of different tokens experienced in many thousands of interactions (Tomasello, 2003). Although massively more complex, it’s essentially the same process as the one by which you have developed a prototypical bicycle in your mind on the basis of seeing wheels, saddles, frames, handlebars, pedals, etc., observing how they function separately and together, and using them yourself.


activity  Activity

Let’s make this a bit more concrete by returning to Barbara again. Read through the extracts below and see if you can identify a multi-word unit that occurs in all of them. How does this unit pattern?

1. INV: Oh look at your lovely teeth have you got all your teeth oh aren't they gorgeous!
2. INV: Have you got one of those [a new toy]?
CHI: yeah
INV: have you got one?
CHI: yes
3. INV: I'm going to talk to your Mummy and then I'm going to read you some of those xxx have you got good toys to play with as well?
4. INV: That's right have you got your own bedroom?
CHI: yeah
INV: have you got your toys in your bedroom and all too?


She also notices that another verb, are, occurs before the pronoun you, just like have. In the transcript we’ve been looking at she hears dozens of examples of the combination. Here are the first five:

  • Figure 3.12: The [Are you …] pattern in speech used to Barbara

Just as she found that have you is regularly followed by got + entity in questions, she finds that are you is regularly followed by going to + action in questions. She stores all these similar bits of language together in her memory. It’s a massive task, yet somehow Barbara extracts a set of regularities from the speech streams she’s exposed to and involved in, which together begin to form a mental lexicon and mental grammar.

We can call these bits of stored language constructions (Hilpert, 2014), some of which are strings of particular words, like have you got, I have, and are you going, and some of which are more abstract. For example, from sequences like have you got all your teeth and are you going to count, she builds the prototypical question depicted in the top row of Fig. 3.13:

  • Figure 3.13: Example of an abstract construction and two instances of it

Like bicycles, not all examples of a construction will follow its prototypical form:

    • ● Barbara will also encounter additional units in the sequence (e.g. adverbs between the subject and main verb, as in Have you really got your own bedroom?). Similarly, the prototypical bike doesn’t have mudguards, but the Helsinki bikes do.
      ● Sometimes Barbara will hear the first three units of the construction but no complement (e.g. Are you coming ___?). Similarly, the prototypical bike has visible spokes, but the Helsinki bikes don’t.

And the prototypes are abstractions from actual tokens in both cases: the specification of auxiliary in 3.13 doesn’t tell you which particular one it might be (have or are) , just as the frame on the prototypical bicycle doesn’t tell you what particular colour it might be (bright green in the Helsinki tokens).

These abstract constructions are what many linguists recognize as ‘rules’ of English: not external statements about what is proper or correct or part of ‘good English’ or ‘the language’, but mental representations in individual users’ minds which constitute the language itself. These rules are constructed on the basis of what learners detect in the input they’ve been exposed to; they therefore govern the way users speak and listen to those they interact with.