Gerald,

I agree with you that we need a terminology, and that "direct object" is meaningful, even if we disagree about its boundaries. I take it that the import of your example is "what she said" is a clause, ergo clauses _can_ be direct objects. But the what-construction behaves like an NP, not a clause. Notice the following points:

There's an important difference between ordinary relative clauses and the what-construction:
 
(1) I believe the story which she told us.
(1a) *I believe which she told us.
(2) I believe what she told us.

In (2), "what she told us" is not equivalent to just the relative clause. It's equivalent to both the relative clause _and_ its antecedent. "Believe" cannot take an ordinary relative clause as a complement, as shown by (1a). The direct object of (1) is clearly an NP headed by "story." The relative clause is the complement of the noun.

Ordinary integrated relatives also permit pied piping, whereas the what-construction does not:

(3) What she referred to is important.
(3a) *To what she referred is important.
(4) The story she referred to is important.
(4a) The story to which she referred is important. 
(4b) *To the story she referred is important.

This inability to front the preposition is identical to the preposition's inability to appear before an NP (cf 4b).

Like NPs and unlike complement clauses, the what-construction is sensitive to subject-verb agreement:

(5) What she told us is important.
(5a) What stories she told us are important.
(5b) That she told us stories is important.

The what-construction (again unlike complement clauses) can be the object of a preposition (and indeed can go _anywhere_ an NP can):

(6) She alluded to what she told us.
(6a) *She alluded to that she told us the story. 

It also readily admits subject-verb inversion:

(7) Is what she told us reliable?
(7a) *Is that she told us the story reliable?

And the what-construction can only appear as the complement of an adjective or noun in those same exceptional cases where the word also takes an NP complement:

(8) I'm happy that she told us the story.
(8a) *I'm happy the story.
(8b) *I'm happy what she said.
(9) It is worth what she said.
(9a) It is worth her money.

In short, the distribution of "what she said" is identical to that of a noun phrase, and so it has a good claim to be a direct object, but that doesn't tell us anything about content clauses, because this what-construction, which the CGEL calls a "fused relative," is an NP that contains a clause as a subordinate element, not a bare clause.


> On Oct 4, 2015, at 8:18 PM, GERALD W WALTON <[log in to unmask]> wrote:
> 
>    I believe what she said. I believe that she said it.
>    To me, the fact that "that she said it" can fit in the same slot as "what she said" is strong motivation for declaring it a direct object. I know "that" is a complementizer (and an optional element), but I'm not sure why the boundary for what gets accepted as direct object can't be wide enough to include it. 
>      ********
> Regardless of the approaches we take, we have to have some kind of terminology:
> 
> "I" serves as a subject, a pronoun, an NP
> 
> "Believe" is a verb
> 
> "She said what" and "she said it" are clauses that are complements of the verb "believe" (VP)
> 
> In the second case "that" serves only as a word that introduces the noun clause
> 
> Yes, of course Direct Object is a meaningful term.
> Gerald
> 
> To join or leave this LISTSERV list, please visit the list's web interface at:
>     http://listserv.muohio.edu/archives/ateg.html
> and select "Join or leave the list"
> 
> Visit ATEG's web site at http://ateg.org/

To join or leave this LISTSERV list, please visit the list's web interface at:
     http://listserv.muohio.edu/archives/ateg.html
and select "Join or leave the list"

Visit ATEG's web site at http://ateg.org/