Craig,

I was referring to your asterisked examples, not to that-clause subjects generally (which would have been strange, as elsewhere I defended the notion that it was legitimate to call them subjects).

You say that "the notion of direct object differs remarkably from verb to verb." Reformulated, I can agree. I would frame it that direct objects can bear a wide range of semantic roles, and those semantic roles are assigned by the verb. I have no quarrel with talking about the semantics of the various constructions. But objects also have a shared series of properties that don't depend on their specific semantic roles, and if we focus only on the semantics, we lose an important insight into the way language works. 

Indeed, the tendency to try to define everything in grammar by its semantic properties (a noun is a person, place, or thing; a sentence expresses a complete thought, etc.) is one of the core conceptual flaws of traditional, school-room grammar.

As to the utility of different kinds of analysis, surely that depends on your audience and purpose. I have the opportunity to teach grammar to two different groups: high school students and instructors who teach grammar to high school students. When I'm teaching high school students, I teach them the general concept of complements, and we cover the objects as types of complements. I tell them that they can assume that direct and indirect objects are pretty much always noun phrases (for this group, I try to sweep minor exceptions under the table), and when we get to content clauses, I simply teach them as complements. I make no effort to get them to try to understand the logic behind the distinction. In other words, for these students, I present the category boundaries as a given, unless I get specific questions asking why. I've found this to be a successful approach. Conceptually, it's certainly no more difficult than a traditional account of direct objects, and I find it has many pedagogical advantages in that this treatment fits seamlessly into the bigger conceptual themes that I want students to take from the course.

On the other hand, when I'm training instructors to teach grammar, I think it's important for them to understand the logic that underlies different formulations of grammar, and that means asking why we make the distinctions we do. So I do cover questions such as why it's better not treat PPs as indirect objects (e.g., "Give the book to me.") and why content clauses are normally not objects. You may consider it getting bogged down in the minutia, but I see it as a necessary training in how to handle evidence about language to construct viable arguments about the way language works.

This sort of analysis is compatible with many different theories of grammar, including construction grammar, which I gather by your mention of Fillmore that you're sympathetic to. So I'm not quite sure why you're so eager to deprecate "merely" formal analysis in a context such as discussions on this list where it seems wholly appropriate to get into the syntactic weeds. Construction grammar depends on the notion of a form-meaning pairing, and so a consideration of the form should be an essential part of the discussion.

If we think in terms of a frame, we're still left with arbitrary forms to account for. That doesn't imply, by the way, that we can reduce all grammar to a handful of elegant principles of universal grammar. I don't buy that myself, and if that's the other perspective you were hinting at, I don't think the Cambridge Grammar follows that line either. I know Geoff Pullum has written sympathetically of the approach of construction grammar and called the Cambridge Grammar an attempt at a descriptive resource that could be used for a construction-grammar research program.

You may find this paper of his, which talks about the conceptual weaknesses of both traditional grammar and (some) theoretical linguistics, interesting:

http://www.lel.ed.ac.uk/~gpullum/LTTCpaper.pdf


> On Oct 1, 2015, at 6:04 AM, Hancock, Craig G <[log in to unmask]> wrote:
> 
> Karl,
>    This is the sentence (yours)  I was reacting to:  "This is the same sort of ill-formedness that you see with your that-clauses as the subjects." My assumption was that you were saying sentences with "that clauses" as subjects (not extraposed) were ill-formed. Chalk it up to late night talk between papers. I am glad that you find the sentences unremarkable.
>    My original point was mainly that the notion of direct object differs remarkably from verb to verb, and that has implications for the grammar. Semantics and grammaticality are dynamically linked. Agreeing involves an agree-er, some sort of sentient being for the agree-er to agree with, and an area (object) of agreement. It can be a verbal process act, a cognitive act, and/or a promise of sorts. If we look at the meaning of agreement, we have to put it into a real world frame, and it is that cognitive content that forms and constrains the grammar. What we agree to is part of the structure. We can't really have agreement without it. 
>    I can't agree the cat. I can agree with the cat or agree that the cat is getting fat, with or without assertion. I can agree to give the cat less food.
>    I think one reason why traditional sentence type analysis doesn't seem useful to most people is that the notion of direct object is just a formal observation,  sometimes bogged down in minutia about what can or cannot edge over the boundaries of the category.  (Formal grammar outside the context of semantics or discourse.) 
>    If we think in terms of frame, at least grammar observations seem linked to cognition and discourse. I understand that this is not the perspective presented  in the Cambridge Grammar. It is a different frame of reference.
> 
> Craig
> -----Original Message-----
> From: Assembly for the Teaching of English Grammar [mailto:[log in to unmask]] On Behalf Of Karl Hagen
> Sent: Thursday, October 01, 2015 12:34 AM
> To: [log in to unmask]
> Subject: Re: grammar question "agreed" + clause
> 
> Craig,
> 
> I think you _are_ misunderstanding me, because I have absolutely no idea why you're citing these examples. I find them all unexceptional, and I'm having trouble seeing why you think they're relevant to our discussion.
> 
>> On Sep 30, 2015, at 6:44 PM, Hancock, Craig G <[log in to unmask]> wrote:
>> 
>> Karl,
>>   I found content clause subjects in The corpus of Contemporary American English without having to look too far, just by searching for That I as sentence opener. These are from the first page (of over 600 entries) that I looked at.
>>    "That I should even feel compelled to be here is an embarassment."
>>    "That i was psyching them out by sitting there was strangely emboldening."
>>    "That I was there at all dancing on the edge of a floor suspended over the waves I could neither see nor hear was too unreal."
>>   Are you saying those are ill-formed? One is from Antioch review, another from the Atlantic. Maybe I misunderstood.
>> 
>>   It seems to me that we have different frames of reference. We can certainly come to an agreement about how different structures act within the living language we see around us. Drawing the border lines for categories can be somewhat arbitrary.Not every element has to be a prototypical member of the category.
>> 
>> 
>> 
>> 
>> ________________________________________
>> From: Assembly for the Teaching of English Grammar 
>> <[log in to unmask]> on behalf of Karl Hagen 
>> <[log in to unmask]>
>> Sent: Wednesday, September 30, 2015 8:23 PM
>> To: [log in to unmask]
>> Subject: Re: grammar question "agreed" + clause
>> 
>> Craig,
>> 
>> I'm not arguing that there's no semantic content to subject and object. My point is that semantics alone is insufficient to account for their distribution. We can't fully account for the distribution of sentence constructions if we fail to distinguish semantic roles from grammatical functions.
>> 
>> I don't want to get into a side-argument about grammaticality judgments either.  If you want to call "Craziness ate the pie" grammatical, the same way that "colorless green ideas sleep furiously" is, I'm OK with that. But it remains semantically ill-formed without a lot of extra context and special pleading. This is the same sort of ill-formedness that you see with your that-clauses as the subjects.
>> 
>> I don't really see the point of your semantic speculations about seeking and looking as being of much help. It sounds like post-hoc justification rather than anything that could let us make predictive statements about grammaticality.
>> 
>> And no, the wh-headed clauses you cite are not content clauses; they're full NPs. That's one reason they do work as objects of prepositions.
>> 
>> Perhaps this formulation will help clarify why I push for the distinction:
>> 
>> If subordinate clauses behave like NP complements of a verb, they can legitimately be called objects. Content clauses, in most contexts, have significantly different distributions from NP complements of a verb, and so there's nothing to be gained from calling them objects in general.
>> 
>> There are two follow-on points that are relevant to the foregoing. First, content clauses that are external complements of the verb share almost all of the distributional properties of NP subjects, and so it's legitimate to call them subjects too. Second, there is actually one case where content clauses do have essentially the same behavior as objects, and that is when you have a content clause as an the predicand in a complex transitive construction. You see this both as an extraposed object and as a preposed object:
>> 
>> I consider it fortunate that you arrived in time.
>> That you arrived in time I consider fortunate.
>> 
>> This is somewhat tricky case, and I'll unpack it if anyone cares for the details, but in essence, this construction is parallel to content-clause subjects, in the sense that it shares almost all of the same distributional properties. All verbs that allow content clauses in the complex transitive construction also allow NPs, unlike the more general case we examined before. So for this limited subset, it makes sense to call the content clause an object.
>> 
>>> On Sep 30, 2015, at 1:49 PM, Hancock, Craig G <[log in to unmask]> wrote:
>>> 
>>> Karl,
>>>  I admit that I'm reading Charles Fillmore and trying to see how far I can extend the notion of frame.
>>>  I think craziness can eat a pie. It doesn't feel ungrammatical to me at all. It just requires a bit of metaphoric play. ("Craziness ate her family ties." "A fit of craziness came over me, and then the pie was gone. Craziness ate the pie.") To me, "Craziness ate the pie" is as grammatical as "Colorless green sheep sleep furiously."
>>>   To me, it's not useful to say that subject and object have no semantic content. The semantic content differs from verb to verb, but that is not the same as saying it's not there.
>>>  Why is it that we can admire his persistence and admire that he 
>>> persists but not marvel his persistence? What in the world would 
>>> motivate those differences? It certainly is possible that it has 
>>> something to do with a difference in how we understand what it is to 
>>> marvel and what it is to admire. (Seeing is different from looking, 
>>> and some of that difference is expressed through the grammar. "I saw 
>>> all day" and "I looked the bird" are both ungrammatical. How can we 
>>> explain that without recourse to what we understand about the nature 
>>> of seeing and the nature of looking?)
>>> 
>>>  Would your category of content clauses extend to Wh- headed clauses as well?
>>> I admire what she did.  I understand how she did it. These seem to follow prepositions more readily. ("He wrote a report about what had gone wrong.") Do you see "that" headed content clauses as a separate category?
>>> 
>>> Craig
>>> 
>>> 
>>> ________________________________________
>>> From: Assembly for the Teaching of English Grammar 
>>> <[log in to unmask]> on behalf of Karl Hagen 
>>> <[log in to unmask]>
>>> Sent: Wednesday, September 30, 2015 2:09 PM
>>> To: [log in to unmask]
>>> Subject: Re: grammar question "agreed" + clause
>>> 
>>> Craig,
>>> 
>>> In your first examples, the restrictions seem to be entirely semantic rather than syntactic. Compare:
>>> 
>>> *Craziness ate the pie.
>>> *Solitariness danced all night.
>>> 
>>> So your examples don't really tell us anything about the syntax of content clauses in the subject position because you're varying the verb, and equivalent NP paraphrases of the content clauses are also ungrammatical. In my examples, by contrast, we have a difference between a content clause and its NP paraphrase for the same verb:
>>> 
>>> I marvel that he persists.
>>> *I marvel his persistence.
>>> 
>>> This distinction cannot be accounted for solely by an appeal to semantics. Subject and object are grammatical functions, not semantic ones.
>>> 
>>> To claim that content clauses aren't direct objects doesn't entail that they can't be subjects. The behavior of subjects as regards to content clauses is different from their behavior as internal complements of the verb. There are no verbs that allow content clauses in the subject position but not noun phrases, as there are for the internal complements. It's true that content clauses as subjects don't permit subject-auxiliary inversion, but they do have the other major syntactic properties that mark the subject, most importantly their position in front of the verb, and pronoun in a tag question has the content clause as it's antecedent:
>>> 
>>> That he refuses to acknowledge the reality of global warming is unsurprising, isn't it?
>>> 
>>> 
>>>> On Sep 30, 2015, at 10:05 AM, Hancock, Craig G <[log in to unmask]> wrote:
>>>> 
>>>> Karl,
>>>> I like the term "content clause" as well. But I don't know if we want to limit the term "direct object" to structures that fulfill all the usual conditions. We run into constraints when these structures act as subjects, but would we want to call them something other than subject?
>>>> That she was crazy was everyone's opinion.   *That she was crazy ate the pie.
>>>> That she could leave without her child surprised him deeply. *That she could leave without her child danced all night.
>>>> 
>>>> It seems to me that direct objects differ radically depending on the verb in focus. We have direct objects that express the content of a speech act, direct objects that express the content of a perception or the content of a thought, direct objects that express the content of a promise--all these would be common with agree.
>>>> "The mayor agreed that the problem was serious."  ("The mayor said, 
>>>> "yes, the problem is serious."")  "The mayor wouldn't say so, but he 
>>>> agreed with what she was saying." (The mayor thought that she was 
>>>> right.)  "The mayor agreed that he would fix the problem by next 
>>>> Tuesday." (The mayor promised to solve the problem.)
>>>> 
>>>> I like to think that these can function like the bubbles in a cartoon--something we can perceive, believe, say, promise to do, and so on. They can't be replaced easily by noun phrases because they are not singular things that can be acted on in a material way.
>>>> If we look at this from the perspective of frame semantics, the constraints are not formal ones so much as they are cognitive, motivated by the cognitive structure of the frames themselves.
>>>> To agree presupposes another sentient being and an area of agreement. The grammar can't violate that understanding without tension.
>>>> 
>>>> Craig
>>>> -----Original Message-----
>>>> From: Assembly for the Teaching of English Grammar 
>>>> [mailto:[log in to unmask]] On Behalf Of Karl Hagen
>>>> Sent: Wednesday, September 30, 2015 2:04 AM
>>>> To: [log in to unmask]
>>>> Subject: Re: grammar question "agreed" + clause
>>>> 
>>>> Of course it's evidence for revising the traditional account. You may disagree with the conclusion, but it points to a difference. Let me unpack the logic a bit more, as perhaps I was too terse.
>>>> 
>>>> My basic point was that "the vacation," which is unequivocally a direct object, does not behave the same way, in terms of linear position, as the content clause. The default position of the direct object that no one argues about (i.e., a noun phrase) is immediately after the verb. There are exceptions to that rule, for example, indirect objects and particles can intervene, but they aren't relevant to the structures we're looking at here.
>>>> 
>>>> She revealed the answer.
>>>> *She revealed unexpectedly the answer.
>>>> She revealed that she had eloped.
>>>> She revealed unexpectedly that she had eloped.
>>>> *She revealed to me the answer.
>>>> She revealed to me that she had eloped.
>>>> 
>>>> Notice that neither adverbs nor preposition phrases can ordinarily intervene between verb and the NP, but there's no problem at all with them between the verb and the content clause. If both the NP and the content clause have the same grammatical function, why the difference?
>>>> 
>>>> Now I can understand that in isolation you may not be inclined to change your view of categorization from this one point, but this isn't the only difference. (You can find a full, technical account of these differences in the Cambridge Grammar of the English Language.) As far as I can tell, the primary motivation for calling this sort of content clause a direct objects comes from the label "noun clause." But that label implies that "noun clauses" are freely substitutable for noun phrases, which they are not. For example, some verbs license content clauses but not NP direct objects. And a content clause cannot be the object of a preposition:
>>>> 
>>>> I marvel that he persists.
>>>> *I marvel his persistence.
>>>> I marvel at his persistence.
>>>> *I marvel at that he persists.
>>>> 
>>>> Notice that the first pair of examples here militates against the idea that we should regard content clauses as direct objects because we can substitute noun phrases for them. That substitution is possible with some verbs, but with not others.
>>>> 
>>>> To preserve the traditional account, we would need to erect a whole host of exceptions for content clauses. Or we can avoid trying to shoehorn content clauses into a category for which they are not well suited and simply call them complements. The latter choice, I submit, is both simpler and more accurate.
>>>> 
>>>> 
>>>>> On Sep 29, 2015, at 10:34 AM, GERALD W WALTON <[log in to unmask]> wrote:
>>>>> 
>>>>> I do not think that "the position of the PP here is evidence that the content clause is not adequately analyzed as the direct object of the verb." In my opinion "the vacation" and "that they would fly to Paris" are very definitely the direct objects.
>>>>> 
>>>>> She arranged the vacation.
>>>>> *She arranged with the family the vacation.
>>>>> She arranged that they would fly to Paris.
>>>>> She arranged with the family that they would fly to Paris.
>>>>> 
>>>>> To me such labels as "noun clause" are helpful and do not cause the "what must be so" reasoning.
>>>>> 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/
>>>> 
>>>> 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/
>>> 
>>> 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/
>> 
>> 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/
> 
> 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/