write_canonical() and natural language interpretation

Question

Implement the following two operators has and of in such a way that with phrases like: peter has car of john answers to questions such as: Who has What of X

I have this simple program that define two operator: has (that specify who have what) and of (that specifies who owns what)

This is the solution write in a Prolog source file:

:- op(100, xfy, of).
:- op(200, xfx, has).

peter has car of john.
peter has car of mary.

car of john of mary.

As I can read on the official SWI Prolog documentation related to write_canonical() function: http://www.swi-prolog.org/pldoc/man?predicate=write_canonical%2F1

Write Term on the current output stream using standard parenthesized prefix notation (i.e., ignoring operator declarations). Atoms that need quotes are quoted.

So if I launch a Prolog query inside write_canonical() function:

15 ?- write_canonical(peter has car of john).
has(peter,of(car,john))
true.

I obtain how Prolog interprets the natural language form, is it right?

In the previous case it is pretty simple and appear clear that first evaluate the sentence: (car of john) and then evaluate the sentence: peter has (result of previous evaluation)

having the following interpretation: peter has (car of john)

Ok, this is clear...now we come to my question.

If I do:

16 ?- write_canonical(peter has car of john of mary).
has(peter,of(car,of(john,mary)))
true.

How can I interpret it in natural language? (if that makes any sense in natural language)

At the beginning I I had initially interpreted as the car belongs booth to john and to mary...so I have try to introduce the fact 3 facts of the previous example but if in the Prolog shell I try to execute:

10 ?- peter has car of john of mary.
false.

This appear false.

What is the correct interpretation in natural language?

Unknown Unknown · Accepted Answer · 2013-03-29T18:01:15

I did a double take while reading your previous question about this and CapelliC's answer. It should be actually read,

"Peter has a car, which belongs to John, who [John] belongs to Marry."

Your interpretation should have been (using write_canonical) something like:

has(peter, of(car, [john,marry]))

or maybe

has(peter, of(car, john)), has(peter, of(car, marry))

depending how you want to implement it, but also, if it is actually the same car (which is not completely clear from your example).

"Peter has car of John of Mary" is of course not conventional English.

write_canonical() and natural language interpretation

1 Answers