2

I am learning about English grammar, but as a programmer, I have natually gravitated towards learning about syntactic structure. I am learning from university lecture notes which I found through Google.

The production rule for a sentence is:

S → NP (Aux) VP

However, the author does not address sentences which are of the general form:

NP Aux Adjective  (where the Aux is a copula, like is, are etc)

For example:

John is tall.
He is gentle.
They are unwell.

Edit: another unaccounted case would be sentences of the form:

NP Aux NP  (where the Aux is a copula)

For example:

Donald is a programmer.
John is an artist.
He is a kung-fu master.

Can the production rule above S → NP (Aux) VP account for such cases, or do I need another production rule?

3
  • 1
    You're right, the original rule can't account for your examples. And you have (almost) stated an additional rule that can account for them.
    – robert
    Dec 4 '13 at 16:17
  • However, you ought to find a syntactic theory that's machine washable. PS rules are not, and they're also arbitrary. Assuming, for instance, that Aux is in the base structure instead of inserted when necessary by rule is an assumption that leads to lots of imaginary phenomena.
    – jlawler
    Dec 4 '13 at 16:32
  • JLawler: I had a quick look at your "coursepack part 2", and although it does not mention PS rules, it shows that a sentence can be broken down into constituents (like in a syntax tree), just like PS rules. What is the difference between your approach and PS rules? Dec 4 '13 at 23:02
3

The phrase structure rules you are learning stem from the 1960s. Those rules are still taught by some in introductory linguistics courses, and they seem to still be accepted by many in computational linguistics. Theoretical syntax, however, progressed beyond those rules decades ago.

The confusion with the examples you are considering stems from the status of AUX. My guess is that the person who wrote the notes does not view the copula be as an AUX. What this means is that for the author of those notes, a sentence such as John is tall does not contain an auxiliary. The copula is construed as a main verb, which means the VP is is tall. So there is no reason to introduce a new phrase structure rule to cover such cases.

I want to emphasize that those old phrase structure rules will not get you very far when it comes to building an understanding of the syntactic structure of sentences. What happens to the VP, for instance, when subject-auxiliary inversion occurs, e.g. Is John tall?

13
  • 1
    Exactly. You'd have to have a special rule to cover it, and maybe some invisible trace to feed that rule. That's what I meant in the comment above by "imaginary phenomena"; angels dancing on pinheads. Syntax is not that difficult.
    – jlawler
    Dec 4 '13 at 19:50
  • Tim, JLawler: PS rules are quite easy to learn and understand and can be extended to include the special cases we are discussing. What is the latest, greatest approach to theoretical syntax, then? Dec 4 '13 at 22:12
  • 1
    @Tahir, my suggestion is to take it look at how dependency grammar (DG) parses sentences. DG has become quite prominent in computational circles, so since you are a computer guy, it would be the natural place to look. If you want me to back that statement up, I'll be happy to do so (with arguments, insights, and links to various sources). Concerning the simplicity of phrase structure rules, that is the only redeeming value they have. It may not seem like it at first, but the DG understanding of syntactic structure is much simpler than what you get with those phrase structure rules. Dec 4 '13 at 23:28
  • Tim likes dependency grammars, and indeed they are useful for some things. But you'll still need constituents; it's just that PS rules are not the best way to get there. Read McCawley 1998 to see how to find them.
    – jlawler
    Dec 5 '13 at 3:44
  • @Jlawler, DGs acknowledge constituents! Given the standard definition of the constituent (a complete subtree), DGs acknowledge constituents just like PSGs do. Take 2 minutes of your time and look at the last part of the following article: en.wikipedia.org/wiki/Constituent_%28linguistics%29. If you want an entertaining debate about the matter, see the talk page for the article. The notion that DGs do not acknowledge constituents is outadated. It's time to renew one's exposure and understanding of old notions. I'd love to go over this with you! Dec 5 '13 at 4:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.