Symbolic AI – Deep and Surface Structure Trees

Finally, you get to see a tree 😀

This post will cover the basics of what Deep and Surface Structure trees are, as well as a little bit on moved words – only little though!


Before we mention anything about Deep and Surface Structure trees, we need to see what a typical tree in Symbolic AI looks like.

Let’s use that old favourite, “every girl”.

The trees are constructed by using the syntax rules for grammar; such as having a “Noun –> Verb –> Noun”.

In order for “every girl” to work, we need to start off by using a “Noun Phrase”, that is, all the words together produce a Noun. Now a Noun Phrase can be broken up into a Determiner and an actual Noun, where the Determiner is “every” and the actual Noun is “girl”.

Properly, these syntax rules look like the following:

  • Noun Phrase (NP)  –>  Determine (Det)  /  Noun-bar (N’)
  • N’  –>  Noun (N)
  • Det  –> “every”
  • N  –> “girl”

The tree for this would look like the following:


Now, from this, you can probably see better why in Basic Lambda Calculus 2 we applied the meaning of “every” to the meaning of “girl”. This is because the NP is computed by applying the Det to the N’, that is, the meaning of “every” to the meaning of “girl”.


Deep Structure

A Deep Structure tree, is a tree that has been drawn before any movement of words occurs. Let us, for the moment, assume we have the sentence “who every girl loves”. Now don’t worry, we are not going to be performing lambda calculus on this sentence in this post 😉

Here, “who” is a Noun, and so is “every girl”. But remember, for a sentence to be syntactically correct, we need “Noun Verb Noun”, where “love” is the Verb. But what we have is “Noun Noun Verb”!

Let’s say, for now, that the word ‘who’ is ‘Matthew’. So that now we have “Matthew every girl loves”, now this sentence looks completely wrong! So what we need to do it move the for Noun occurrence to after the Verb, so that we obtain “every girl loves Matthew”. Now i really wish that this were true :D, however, we now have a sentence that makes perfect sense. So replacing ‘Matthew’ back to ’who’, we have “every girl loves who”, which is exactly what “who every girl loves” is trying to say.

This sentence is called a Complementizer Phrase (CP). And the Deep Structure for the sentence looks like:


In this Deep Structure tree, we show where the movement of words would go when we draw the Surface Structure (more later). So, we have the word “loves”, however, “loves” is not a verb, where as “love” is. So we need to show that the verb “love” will move to the Inflection (I) position, to join up with ‘’s’ and become ‘loves’.

Who, will have to be moved from the NP to the CP-Spec.


Surface Structure

The Surface Structure shows the tree after words have been moved. So for the Deep Structure tree above, we would get:


And that’s it 🙂 The ‘t’ is a trace, and not very important for the time being. But that should give you the general idea. Also, the tree at the very top for “every girl”, that would be a surface structure tree, and there are no movements at all.


About Badgerati
Computer Scientist, Games Developer, and DevOps Engineer. Fantasy and Sci-fi book lover, also founder of Cadaeic Studios.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: