Some English relative clause constuctions
Jon
constuctions, as exemplified below:
(1) John saw the man who loved Susan. -- acceptable
(2) John saw the man loved Susan. (Intended to mean the same as (1)) --
unacceptable
(3) John saw the man who Mary believes loved Susan. -- acceptable
(4) John saw the man Mary believes loved Susan. (Intended to mean the same
as (3)) -- acceptable
Linguists wondered why, in the same environment where the relative pronoun
is missing, (4) is acceptable while (2) is not, to most English speakers. Do
you guys have the same judgements?
In this very regard, I found a very interesting sentence from one of J. K.
Rowling's books, which may be either writer's (or proofreader's) mistake or
the reflection of her (British) English intuition. Here you go:
"Krum, whom Harry would have thought would have been used to this sort of
thing, skulked, half-hidden, at the back of the group." (Harry Potter,
Volume #4, p. 311 (paperback, American edition, published by Scholastic))
Do you also feel the use of "whom" (instead of "who" for a subject position)
here is correct? If it is corrrect, it can be a very important pointer to
linguists towards the solution of the problem raised by the sort of
sentences exemplified in (1)-(4).
Your input will be very much appreciated.
Jon Cha (pos...@heyum.net)
Always dazzled
Alan Crozier
"Jon p]>" <Qos...@heyum.net[Q-> skrev i meddelandet
news:2ebdf6695f98af3f...@news.teranews.com...
> Linguists have long noted the contrasts in some of English relative clause
> constuctions, as exemplified below:
>
> (1) John saw the man who loved Susan. -- acceptable
> (2) John saw the man loved Susan. (Intended to mean the same as (1)) --
> unacceptable
Unacceptable in standard English, perhaps, but OK in some dialects (mine,
for example.
Alan (Tyrone)
Brian.Farrelly
Doesn't (2) mean
John saw that the man loved Susan
rather than (1) ?
In other words, he recognised the situation rather than seeing the
person.
Brian.
--
mailto:Brian.F...@nho.hydro.com Norsk Hydro Research Centre
phone +47 55 99 68 74 ((( Postboks 7190
fax +47 55 99 69 70 2oooS N-5020 Bergen
home +47 55 13 78 49 HYDRO Norway
CyberCypher
> Linguists have long noted the contrasts in some of English
> relative clause constuctions, as exemplified below:
>
> (1) John saw the man who loved Susan. -- acceptable
> (2) John saw the man loved Susan. (Intended to mean the same as
> (1)) -- unacceptable
>
> (3) John saw the man who Mary believes loved Susan. -- acceptable
> (4) John saw the man Mary believes loved Susan. (Intended to mean
> the same as (3)) -- acceptable
>
> Linguists wondered why, in the same environment where the relative
> pronoun is missing, (4) is acceptable while (2) is not, to most
> English speakers. Do you guys have the same judgements?
I agree with the 4 posted judgments. Sentences 3 and 4 could also take
"that" in informal English.
> In this very regard, I found a very interesting sentence from one
> of J. K. Rowling's books, which may be either writer's (or
> proofreader's) mistake or the reflection of her (British) English
> intuition. Here you go:
>
> "Krum, whom Harry would have thought would have been used to this
> sort of thing, skulked, half-hidden, at the back of the group."
> (Harry Potter, Volume #4, p. 311 (paperback, American edition,
> published by Scholastic))
>
> Do you also feel the use of "whom" (instead of "who" for a subject
> position) here is correct?
No, it's not correct. The WH-pro is the subject of the embedded clause
beginning with "would have been used to this sort of thing" and should
be "who".
--
Franke: EFL teacher & medical editor.
Alan Crozier
"Brian.Farrelly" <Brian.F...@nho.hydro.com> skrev i meddelandet
news:406BEA18...@nho.hydro.com...
> Alan Crozier wrote:
> >
> > "Jon p]>" <Qos...@heyum.net[Q-> skrev i meddelandet
> > news:2ebdf6695f98af3f...@news.teranews.com...
> > > Linguists have long noted the contrasts in some of English relative
clause
> > > constuctions, as exemplified below:
> > >
> > > (1) John saw the man who loved Susan. -- acceptable
> > > (2) John saw the man loved Susan. (Intended to mean the same as
(1)) --
> > > unacceptable
> >
> > Unacceptable in standard English, perhaps, but OK in some dialects
(mine,
> > for example.
> >
> > Alan (Tyrone)
>
>
>
> Doesn't (2) mean
>
> John saw that the man loved Susan
>
> rather than (1) ?
>
> In other words, he recognised the situation rather than seeing the
> person.
Yes, in standard English it would mean that. In my dialect it could mean
both. I didn't notice the ambiguity. Flying planes can be dangerous!
Alan
Jon
"Brian.Farrelly" <Brian.F...@nho.hydro.com> wrote in message
news:406BEA18...@nho.hydro.com...
> Alan Crozier wrote:
> >
> > "Jon p]>" <Qos...@heyum.net[Q-> skrev i meddelandet
> > news:2ebdf6695f98af3f...@news.teranews.com...
> > > Linguists have long noted the contrasts in some of English relative
clause
> > > constuctions, as exemplified below:
> > >
> > > (1) John saw the man who loved Susan. -- acceptable
> > > (2) John saw the man loved Susan. (Intended to mean the same as
(1)) --
> > > unacceptable
> >
> > Unacceptable in standard English, perhaps, but OK in some dialects
(mine,
> > for example.
> >
> > Alan (Tyrone)
>
>
>
> Doesn't (2) mean
>
> John saw that the man loved Susan
>
> rather than (1) ?
>
> In other words, he recognised the situation rather than seeing the
> person.
What I meant with the contrast between (1) and (2) in my original query was,
"who" (nominative relative pronoun) cannot be left out. In contrast,
"who(m)" (accusative relative proun) can be left out without probelm, as
shown:
(1)' John saw the man who(m) Susan loved. -- acceptable
(2)' John saw the man Susan loved. -- acceptable
But, the same nominative relative pronoun "who" can also be left out when a
parenthetical clause like "Mary believes" is present in the relative clause,
as in (4), repeated below:
(4) John saw the man Mary believes loved Susan.
Any good explanation for this phenomenon?
Jon.
Lars Eighner
<2ebdf6695f98af3f...@news.teranews.com>,
the lovely and talented Jon
broadcast on alt.usage.english:
> "Krum, whom Harry would have thought would have been used to this sort of
> thing, skulked, half-hidden, at the back of the group." (Harry Potter,
> Volume #4, p. 311 (paperback, American edition, published by Scholastic))
> Do you also feel the use of "whom" (instead of "who" for a subject position)
> here is correct? If it is corrrect, it can be a very important pointer to
> linguists towards the solution of the problem raised by the sort of
> sentences exemplified in (1)-(4).
> Your input will be very much appreciated.
I've been wrong before, but as I parse it, "who(m)" is the subject
of the clause "? would have been used to this sort of thing," and
as such should be "who."
"Krum, who Harry would have thought would have been used to this sort of
thing, skulked, half-hidden, at the back of the group."
If you remove "Harry would have thought," "whom" is very clearly wrong.
*"Krum, whom would have been used to this sort of
thing, skulked, half-hidden, at the back of the group."
Restoring the "Harry would have thought" doesn't change anything.
*Harry would have thought him would have been used to this sort of
thing.
Pretty clearly it is the whole clause "? would have been used to ..."
that is the object of "Harry would have thought" and not "who(m)"
alone.
Perhaps I am missing something.
If the question is "Does this sort of thing get past popular writers
and their editors?" the answer is "Yes, it does, with some frequency."
I'm sure everyone who writes much has done something of the kind.
--
Lars Eighner -finger for geek code- eig...@io.com http://www.io.com/~eighner/
"The very essence of the creative is its novelty, and hence we have no
standard by which to judge it." --Carl R. Rogers, On Becoming a Person
Lars Eighner
<nNRac.88202$dP1.2...@newsc.telia.net>,
the lovely and talented Alan Crozier
broadcast on alt.usage.english:
It is grammatically correct, but in my dialect it does not mean the
same thing.
(2) means the same as "John saw (that) the man loved Susan" which
is not the same thing as "John saw the man who loved Susan."
--
Jon
"Lars Eighner" <eig...@io.com> wrote in message
news:slrnc6ob3e....@goodwill.io.com...
Then, I am wondering WHY this sort of thing gets past popular writers "with
some frequency". Is it because of some confusion or mix of cognitive
processes? So, the psychology is the criminal behind this?
Jon.
Lars Eighner
<c4hc46$2cqjvl$1...@ID-228859.news.uni-berlin.de>,
broadcast on alt.usage.english:
> "Lars Eighner" <eig...@io.com> wrote in message
> news:slrnc6ob3e....@goodwill.io.com...
>> I've been wrong before, but as I parse it, "who(m)" is the subject
>> of the clause "? would have been used to this sort of thing," and
>> as such should be "who."
>>
>> If the question is "Does this sort of thing get past popular writers
>> and their editors?" the answer is "Yes, it does, with some frequency."
>> I'm sure everyone who writes much has done something of the kind.
>>
> Then, I am wondering WHY this sort of thing gets past popular writers "with
> some frequency". Is it because of some confusion or mix of cognitive
> processes? So, the psychology is the criminal behind this?
The reason is that "whom" is moribund. Its proper use is no longer
native to most speakers. We learn it academically, but not deeply.
Jon
> In our last episode,
> <c4hc46$2cqjvl$1...@ID-228859.news.uni-berlin.de>,
> the lovely and talented Jon
> broadcast on alt.usage.english:
>
>
> > "Lars Eighner" <eig...@io.com> wrote in message
> > news:slrnc6ob3e....@goodwill.io.com...
>
> >> I've been wrong before, but as I parse it, "who(m)" is the subject
> >> of the clause "? would have been used to this sort of thing," and
> >> as such should be "who."
> >>
> >> If the question is "Does this sort of thing get past popular writers
> >> and their editors?" the answer is "Yes, it does, with some frequency."
> >> I'm sure everyone who writes much has done something of the kind.
> >>
>
> > Then, I am wondering WHY this sort of thing gets past popular writers
"with
> > some frequency". Is it because of some confusion or mix of cognitive
> > processes? So, the psychology is the criminal behind this?
>
> The reason is that "whom" is moribund. Its proper use is no longer
> native to most speakers. We learn it academically, but not deeply.
That's strange. If "whom" is a disappearing word, why the incorrect "whom"
in Rowling's sentence, instead of the correct "who"?
Jon.
Adrian Bailey
news:2ebdf6695f98af3f...@news.teranews.com...
> Linguists have long noted the contrasts in some of English relative clause
> constuctions, as exemplified below:
>
> (1) John saw the man who loved Susan. -- acceptable
> (2) John saw the man loved Susan. (Intended to mean the same as (1)) --
> unacceptable
>
> (3) John saw the man who Mary believes loved Susan. -- acceptable
> (4) John saw the man Mary believes loved Susan. (Intended to mean the same
> as (3)) -- acceptable
>
> Linguists wondered why, in the same environment where the relative pronoun
> is missing, (4) is acceptable while (2) is not, to most English speakers.
Do
> you guys have the same judgements?
Which linguists wonder why? Doesn't seem very difficult to understand that
(2) is unacceptable because it means something else. (There is an implied
"that" after "saw".) (4) is acceptable because it is not possible to read an
implied "that" after "saw". Can there be more than one implied "that" in a
sentence?
> In this very regard, I found a very interesting sentence from one of J. K.
> Rowling's books, which may be either writer's (or proofreader's) mistake
or
> the reflection of her (British) English intuition. Here you go:
>
> "Krum, whom Harry would have thought would have been used to this sort of
> thing, skulked, half-hidden, at the back of the group." (Harry Potter,
> Volume #4, p. 311 (paperback, American edition, published by Scholastic))
>
> Do you also feel the use of "whom" (instead of "who" for a subject
position)
> here is correct? If it is corrrect, it can be a very important pointer to
> linguists towards the solution of the problem raised by the sort of
> sentences exemplified in (1)-(4).
Looks like a mistake to me.
Adrian
CyberCypher
[...]
>If "whom" is a disappearing word, why the incorrect "whom" in
> Rowling's sentence, instead of the correct "who"?
She was a good storyteller in the first Harry Potter book, but after I
read the whole thing, I realized that she isn't that concerned with the
minutiae of grammar. She's a writer, not a grammarian nor a pedant.
Lars Eighner
<c4hf1d$2hs7np$1...@ID-228859.news.uni-berlin.de>,
broadcast on alt.usage.english:
> "Lars Eighner" <eig...@io.com> wrote in message
> news:slrnc6odcg....@goodwill.io.com...
>> The reason is that "whom" is moribund. Its proper use is no longer
>> native to most speakers. We learn it academically, but not deeply.
> That's strange. If "whom" is a disappearing word, why the incorrect "whom"
> in Rowling's sentence, instead of the correct "who"?
Exactly as I said, people are taught to use "whom," but because it is
no longer part of the language they acquire as native speakers, they
are likely to use it incorrectly - and of course by the time it is in
print it has got by not only the author but also many readers at the
publishing house. Would they be better off if they were taught not
to attempt "whom" at all? I think so. Not five of a hundred native
speakers would get "whom/who" right in speaking complex sentences,
such as in the example, and there is no point in perpetuating an
anachronism in print (especially when the wrong anachronism is so often
the result). I don't have any argument with those who use "whom" and
always get it right - although perhaps I would do better arguing with
them as they are fewer by far than those who use "whom" and get it
wrong.
Wood Avens
with p.)> wrote:
>That's strange. If "whom" is a disappearing word, why the incorrect "whom"
>in Rowling's sentence, instead of the correct "who"?
I wouldn't say it was a disappearing word in the UK, even though most
people probably don't use it in everyday speech (unless they're the
sort of person whose everyday speech is being broadcast or recorded).
I think this sentence is simply a cock-up. I read it twice before I
was sure it was wrong, and if I'd been J K Rowling I'd have re-written
it to avoid that construction altogether. But then I'm not a
best-selling novelist keen to get the latest volume to the publisher.
--
Katy Jennison
spamtrap: remove the first two letters after the @
Jon
"Adrian Bailey" <da...@hotmail.com> wrote in message
news:uHXac.7707$Id....@news-binary.blueyonder.co.uk...
> "Jon p]>" <Qos...@heyum.net[Q-> wrote in message
> news:2ebdf6695f98af3f...@news.teranews.com...
> > Linguists have long noted the contrasts in some of English relative
clause
> > constuctions, as exemplified below:
> >
> > (1) John saw the man who loved Susan. -- acceptable
> > (2) John saw the man loved Susan. (Intended to mean the same as (1)) --
> > unacceptable
> >
> > (3) John saw the man who Mary believes loved Susan. -- acceptable
> > (4) John saw the man Mary believes loved Susan. (Intended to mean the
same
> > as (3)) -- acceptable
> >
> > Linguists wondered why, in the same environment where the relative
pronoun
> > is missing, (4) is acceptable while (2) is not, to most English
speakers.
> Do
> > you guys have the same judgements?
>
> Which linguists wonder why? Doesn't seem very difficult to understand that
> (2) is unacceptable because it means something else. (There is an implied
> "that" after "saw".) (4) is acceptable because it is not possible to read
an
> implied "that" after "saw". Can there be more than one implied "that" in a
> sentence?
(Copied from one of my own follow-ups ....)
What I meant with the contrast between (1) and (2) in my original query was,
"who" (nominative relative pronoun) cannot be left out. In contrast,
"who(m)" (accusative relative proun) can be left out without probelm, as
shown:
(1)' John saw the man who(m) Susan loved. -- acceptable
(2)' John saw the man Susan loved. -- acceptable
But, the same nominative relative pronoun "who" can also be left out when a
parenthetical clause like "Mary believes" is present in the relative clause,
as in (4), repeated below:
(4) John saw the man Mary believes loved Susan.
Many linguists like me (the word "linguist" is lexically ambiguous but here
I mean the scholor who studies the structure of language as a job) are
wondering why this contrast is present.
Jon.
Robert Bannister
> In this very regard, I found a very interesting sentence from one of J. K.
> Rowling's books, which may be either writer's (or proofreader's) mistake or
> the reflection of her (British) English intuition. Here you go:
>
> "Krum, whom Harry would have thought would have been used to this sort of
> thing, skulked, half-hidden, at the back of the group." (Harry Potter,
> Volume #4, p. 311 (paperback, American edition, published by Scholastic))
It is incorrect, but a common mistake made by native English speakers
who no longer really understand how to use "whom". The mistake usually
occurs when there is a parenthetic 'he thought', 'he said' or some
similar phrase.
The speaker has deduced a 'rule' that when 'who' is followed by another
subject, it ought to be 'whom'. Mostly, this works, but not in these cases.
--
Rob Bannister
Eric Walker
> Linguists have long noted the contrasts in some of English relative clause
> constuctions, as exemplified below:
>
> (1) John saw the man who loved Susan. -- acceptable
> (2) John saw the man loved Susan. (Intended to mean the same as (1)) --
> unacceptable
>
> (3) John saw the man who Mary believes loved Susan. -- acceptable
> (4) John saw the man Mary believes loved Susan. (Intended to mean the same
> as (3)) -- acceptable
>
> Linguists wondered why, in the same environment where the relative pronoun
> is missing, (4) is acceptable while (2) is not, to most English speakers. Do
> you guys have the same judgements?
There is no deep mystery here. English conventionally--ponder
that word--allows words to be omitted from sentences for
brevity, *provided always* that the ellipsis is such that
virtually every reader or listener can and will infallibly
supply (even if silently) the omitted word or words.
In (4), there is no plausible chance that anyone will adduce a
sense--meaning an understanding of the full sentence--other than
that given in (3); in (2), almost everyone will supply a "that"
after "saw", which is a false reading. Thus, (4) is an
acceptable instance of elision and (2) is not.
> In this very regard, I found a very interesting sentence from one of J. K.
> Rowling's books, which may be either writer's (or proofreader's) mistake or
> the reflection of her (British) English intuition. Here you go:
>
> "Krum, whom Harry would have thought would have been used to this sort of
> thing, skulked, half-hidden, at the back of the group." (Harry Potter,
> Volume #4, p. 311 (paperback, American edition, published by Scholastic))
>
> Do you also feel the use of "whom" (instead of "who" for a subject position)
> here is correct? If it is corrrect, it can be a very important pointer to
> linguists towards the solution of the problem raised by the sort of
> sentences exemplified in (1)-(4).
It is, as I think everyone commenting has noted, *not* correct.
But, again, there is no "problem" in the pairs of sentences.
You can drop words only when their being dropped cannot
reasonably affect the obvious sense. The end. (Mind, what
"linguists" find as "problems" is their business, not that of
those who merely read and write sound English.)
Jon
"Eric Walker" <em...@owlcroft.com> wrote in message
news:c4j3me$2iamj8$1...@ID-90086.news.uni-berlin.de...
Hmm. Then, what about these:
(5) John killed the man loved Susan. -- unacceptable
(6) John killed the man Mary believes loved Susan. -- acceptable
(Sorry for the horrible word. I picked this one to avoid any unintended
ambiguity arising again.)
You cannot supply a "that" after the verb "killed" here. Now, how can you
explain the contrast of grammaticality between (5) and (6), with your theory
of elision?
Jon.
Robert Lieblich
[ ... ]
> Hmm. Then, what about these:
>
> (5) John killed the man loved Susan. -- unacceptable
> (6) John killed the man Mary believes loved Susan. -- acceptable
> (Sorry for the horrible word. I picked this one to avoid any unintended
> ambiguity arising again.)
>
> You cannot supply a "that" after the verb "killed" here. Now, how can you
> explain the contrast of grammaticality between (5) and (6), with your theory
> of elision?
In (5) there's still the risk of "John killed. The man loved
Susan." But that's not the fundamental point.
Here's the fundamental point: It simply isn't worth the effort to
put every possible word in every possible blank and test each
sentence individually for its various alternative readings. Over
time English speakers have developed a sense of which structures
will frequently (not "inevitably," "frequently") cause confusion and
which are unlikely to confuse. The former sort are avoided and
become unidiomatic, and vice versa for the latter.
It's easy enough to construct unidiomatic sentences that are easily
understood. But when a particular sentence structure is unidiomatic
because susceptible to ambiguity, it's best to avoid that structure
even if you can construct some unambiguous sentences using it.
After a few decades of avoiding, the structure becomes unidiomatic.
I suspect that's what happened here.
--
Bob Lieblich
Or something like that
Carmen L. Abruzzi
> Jon wrote:
>
> [ ... ]
>
>
>>Hmm. Then, what about these:
>>
>>(5) John killed the man loved Susan. -- unacceptable
>>(6) John killed the man Mary believes loved Susan. -- acceptable
>>(Sorry for the horrible word. I picked this one to avoid any unintended
>>ambiguity arising again.)
"Helped" would substitute equally well for "killed"; just a pointer for
future reference.
>>
>>You cannot supply a "that" after the verb "killed" here. Now, how can you
>>explain the contrast of grammaticality between (5) and (6), with your theory
>>of elision?
>
>
> In (5) there's still the risk of "John killed. The man loved
> Susan." But that's not the fundamental point.
>
> Here's the fundamental point: It simply isn't worth the effort to
> put every possible word in every possible blank and test each
> sentence individually for its various alternative readings.
No, nor is it possible. That's why I don't think there is any "missing
word" or elision going on here. There are simply two different ways of
constructing a relative clause whose antecedent is an object within the
relative clause. When the antecedent is the subject of the relative
clause, the relative pronoun cannot be dropped, _except_ when the entire
relative clause is also the object of an embedded clause; as it is in "John
> Over
> time English speakers have developed a sense of which structures
> will frequently (not "inevitably," "frequently") cause confusion and
> which are unlikely to confuse. The former sort are avoided and
> become unidiomatic, and vice versa for the latter.
>
> It's easy enough to construct unidiomatic sentences that are easily
> understood. But when a particular sentence structure is unidiomatic
> because susceptible to ambiguity, it's best to avoid that structure
> even if you can construct some unambiguous sentences using it.
> After a few decades of avoiding, the structure becomes unidiomatic.
> I suspect that's what happened here.
>
This cannot be the explanation. First, sentences equally susceptible to
ambiguity are not similarly avoided:
"John saw the man loved Susan for a short time" is taken to mean:
"John saw that the man loved Susan for a short time",
but "John saw the man Susan loved for a short time" is ambiguous as to
meaning. It could mean either:
"John saw the man that Susan loved for a short time"
or "John saw for a short time the man that Susan loved.
Yet, constructions of the type "John saw the man Susan loved" are not
avoided.
Second, it doesn't take "a few
decades"; such sentences are avoided within a few years of learning
standard English as a native speaker (that is, six-year-olds don't say
"John helped the man saw Susan", if they've been learning/speaking the
standard). In non-standard dialects sentences like "John helped the man
loved Susan" are indeed produced, without ambiguity or confusion, and
this has nothing to do with age; it takes "schooling" or "education" in
the standard to make such non-standard speakers believe that there is
something "wrong" with such constructions.
Joe Fineman
It shows up in songs & poetry once in a while, perhaps for the sake
the meter:
There's a man comes to our house every single day.
--
--- Joe Fineman j...@TheWorld.com
||: Prose: Earth is turning us into its shadow. :||
||: Poetry: The sun is setting. :||
John Lawler
>"Eric Walker" <em...@owlcroft.com> writes
>> Jon writes:
>> > Linguists have long noted the contrasts in some of English relative
>> > clause constuctions, as exemplified below:
>> > (1) John saw the man who loved Susan. -- acceptable
>> > (2) John saw the man loved Susan. (Intended to mean the same as (1)) --
>> > unacceptable
>> > (3) John saw the man who Mary believes loved Susan. -- acceptable
>> > (4) John saw the man Mary believes loved Susan. (Intended to mean the
>> > same as (3)) -- acceptable
>> > Linguists wondered why, in the same environment where the relative
>> > pronoun is missing, (4) is acceptable while (2) is not, to most
>> > English speakers. Do you guys have the same judgements?
>> There is no deep mystery here. English conventionally--ponder
>> that word--allows words to be omitted from sentences for
>> brevity, *provided always* that the ellipsis is such that
>> virtually every reader or listener can and will infallibly
>> supply (even if silently) the omitted word or words.
>> In (4), there is no plausible chance that anyone will adduce a
>> sense--meaning an understanding of the full sentence--other than
>> that given in (3); in (2), almost everyone will supply a "that"
>> after "saw", which is a false reading. Thus, (4) is an
>> acceptable instance of elision and (2) is not.
Eric is correct (inevitably, as he might say :-).
There's another piece to it, however, which governs exactly which
kinds of constructions are parseable, and therefore reconstructible.
English is a right-branching language, which marks the beginnings of phrases
and clauses so that the parsing algorithms we use can push the current parse
on a stack and start a new one, popping back at the end to finish the
previous parse. That means that adverbial clauses, relative (i.e,
adjectival) clauses and complement (noun) clauses all need some kind of
pre-clausal marker for the push. In adverbial clauses that's the
subordinate conjunction that begins the clause ("because, when, although,
after, ...").
In complements that's the complementizer, for instance "for...to" for
infinitives and "that" for tensed complements. Note that:
a) the "for" marking the subject of infinitives is normally deleted,
*except* when the infinitive starts the sentence:
I wanted (for) him to leave early.
For him to leave early would be a mistake.
*Him to leave early would be a mistake.
b) the "that" marking tensed complements is optionally deletable
*except* when the complement starts the sentence:
I said (that) he should leave early.
That he left early was unfortunate.
*He left early was unfortunate.
In relatives, that's either a wh-word or "that". These are moved to the
beginning of the clause if they're not already there because the relative is
the subject. However, it turns out that we can regard two noun phrases in a
row as a push marker, so the relative marker may be deleted *except* where
it doesn't already come before the subject because it already *is* the
subject, i.e, precisely in the cases presented.
This explains why this case is an example of the principle that Eric quite
correctly cites, without the necessity of having to try all possible parses
to determine it.
>Hmm. Then, what about these:
>(5) John killed the man loved Susan. -- unacceptable
>(6) John killed the man Mary believes loved Susan. -- acceptable
> (Sorry for the horrible word. I picked this one to avoid any
> unintended ambiguity arising again.)
The technical term is 'grammatical'. 'Acceptable' has an implied
experiencer ('to X') that begs the question. 'Grammatical' begs a
question, too, but it's a question that eventually has an empirical
answer, and therefore can be ignored if one is careful.
>You cannot supply a "that" after the verb "killed" here. Now, how can you
>explain the contrast of grammaticality between (5) and (6), with your theory
>of elision?
This case falls out from the principle I gave above.
"Who" or "that" in (5) would be the subject of 'loved Susan',
hence not deleteable.
-John Lawler University of Michigan Linguistics Dept
---------------------------------------------------
http://www.umich.edu/~jlawler/aue/
"... and, who knows? Maybe the horse will sing."