What Is Deep Structure?
In the theory of transformational grammar, sentences have two ways they can be represented: deep and surface structure. Deep structure refers to the underlying meaning of a sentence as it is represented and comprehended in the brain. It serves as a counterpoint to surface structure, which is the actual written or spoken form of the sentence. This concept was created by Noam Chomsky in his 1957 book, Syntactic Structures, which formulated the theory of transformational grammar. According to this theory, humans use transformations, a type of cognitive process, to map structural relationships between sentence referents understood in the linguistic regions of the brain and the actual content of a sentence that is seen or heard.
The concept of deep structure contends that information related to each component of a sentence, such as its subjects and predicates, is codified into abstract pieces inside of the brain. Sentences that are different in terms of their surface structure, such as "The boy kicked the ball," and "The ball was kicked by the boy," can have the same deep structure. The reason for this is because the component pieces for each sentence are related in the same way in the brain, so that humans can understand the sentences as semantically equivalent, even though they are syntactically different. With ambiguous sentences, such as "I have seen driving man," with only one surface structure, multiple structural interpretations can be created by rearranging the component pieces, such as "I have seen a man driving," or "I have seen a man who normally drives."
Deep structure, as described by Chomsky, was subject to certain rules that are innate in the human brain. These include transformational rules for deriving the meaning of the surface structure of a sentence, such as adding an implied object to a sentence: the command "Just drive!" becomes an instruction for "Drive the car," through the rule of addition, for example. Through other transformations, the deep structure of a thought is converted into grammatically correct sentences that can be understood by the listener or reader. These rules, as well as the ability to maintain abstract ideas in the brain, are innate, according to the theory, so people do not have to be taught to encode language in terms of deep structure; it is a process that occurs automatically. Although the concept of structures remains important in linguistics, most linguists no longer believe that deep structure is the only way that humans derive meaning from language.
@nony - Computer software doesn’t deal with abstractions the way the human brain does. So I would venture to say that your options are limited there.
As for random writing prompts, I guess it would be okay, but I wouldn’t expect too much from it, and certainly nothing in the way of meaningful story generation. I’ve seen so called “article writing” software and the output it produced was horrendous in my opinion.
@everetra - How easily do you think that computer software could generate the structure of English sentences in a meaningful manner?
I ask because I’ve played with so called story creation software online, and it produces random sentences that are meant to be prompts for fictional writing.
Most of the stuff seems to be quite meaningless (surface structure I suppose) but now and then it produces a useful nugget.
@allenJo - This is just one theory, however. But it does address an important issue – how do we understand the meaning of what is being said?
Surface structure alone would not be sufficient. If you can’t connect the words to meaning then the sentence is useless. However, surface might be useful in some contexts, like stream of consciousness writing where the ambiguity is deliberate and the author intends to open up the sentence to different interpretations.
In that case I suppose it would be useful. But this would fall under creative license.
Well I am not a language expert but deep structure linguistics makes sense to me, at least from what I’ve read here. I can’t imagine what would happen if we couldn’t use transformations to understand the underlying meaning of a phrase. Life would be very literal.
You would have to compose commands in just the right way so that people could understand what you were saying. That means you would always have to include the subject of the sentence (like “You”) even in cases where the subject of the sentences was meant to be clearly understood or implied.
That type of approach to language would create a lot of problems and make communication very slow, to say the least.
Post your comments