Craig,

Here are the rules governing the occurrence of "that" in content clauses, quoting most of the time from Huddleston&Pullum (652-654):

"That" must appear if
a} when the content clause is subject or otherwise precedes the matrix predicator
b) the content clause is adjunct, as in

He appealed to us to bring his case to the attention of the authorities that justice might be done.

c) when the content clause is complelemtn to comparative "that"/"as", as in

I'd rather (that) he hired a taxi than that he drove my car.

"That" must be omitted when the content clause is embedded within an unbounded dependency in such a way that its subject is reallized by a gap:

optional in She thinks (that) Max is the ringleader.
excluded in Who does she thing ___ is the ringleader?

In other cases there are a variety of conditions governing deletability that I won't enumerate here.

H&P also lay out the conditions for the deletion of "that" in relative clauses (pp. 1054-6).

"That" can't be deleted if the relativized element is the subject of the relative clause, as in "The car that 0 hit us was Ed's".

"That" can't be deleted if it is not adjacent to the subject, as in "I found I needed a file that only the day before I had sent to be shredded."

"That" can't be deleted in supplementary (their term for non-restrictive) relatives.

"That" can't be deleted if the RC is extraposed, as in "Something came up that I hadn't predicted."

You're right that the conditions governing deletion in the two structures, content vs. relative clauses.  However, I think this is a function of the grammar of the two types of clause, not a result of there being two different thats.

By the way, H&P also give an interesting set of reasons, beyond and better than those I've given, for why "that" is not a pronoun in relative clauses but is rather a subordinator.  But the arguments are pretty detailed and you're better off going to H&P pp. 1056-7 yourself for them.

Herb

Herb,


OK.  You pretty much got me now on subject deletion.  But it does still seem to me that the deletion rules for content clauses are different than for relatives, and that acts much like which and who when these deletion decisions come into play.  I guess I'm holding out for the point that it's more than the influence of traditional grammar that keeps some of us thinking that acts enough like which and who to be more than just a routine complementizer in certain instances.  The question isn't so much whether I can follow and accept the argument as it is that I still find this way of looking at it counterintuitive.  As Richard said in an earlier post, if it feels like a pronoun, wouldn't it be a pronoun?  If we can substitute it for which in some instances, does that mean it is evolving pronoun like attributes?  Like many grammatical phenomena, it is slightly different from everything else in its class?  
    Is the fact that it sometimes feels like a pronoun irrelevant?

Craig

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/