Do you mind the gap?

[fluffyhamster]

Do you think that the notion of gaps in relation to relative clauses is really, really useful?

For those who aren't sure what I'm on about, here's a bit from the Longman Student Grammar of Spoken and Written English:

9.8 Postmodification by relative clauses

When discussing relative clauses, we will focus on three key components: the head noun, the relativizer, and the gap.
> The head noun is the noun modified by the relative cluase.
> The relativizer is the word, such as who or that, which introduces the relative clause. It refers to the same person or thing as the head noun.
> The gap is the location of the missing constituent in the relative clause. All relative clauses have a missing constituent, which again corresponds in meaning to the head noun.

Thus, consider the relative clause construction:

the diamond earrings that Mama wore ^. (FICT)

> The head noun is earrings.
> The relativizer is that, referring to the 'earrings'.
> The gap here occurs in the direct object position, after the verb wore. The underlying meaning is that 'Mama wore [the earrings]'.

I think I am developing an allergy to "underlying meanings" and the like (is the notion of gap derived from Chomskyan grammar, or does it have a longer or more mainstream pedigree?). Give me basic statements ('the earrings killed him'), relative clauses where the lack of any obviously competing (i.e. different to the preceding subject; salient) word between the subject and verb makes it obvious what the subject still is ('the earrings that killed him...'), and relative clauses where a present and competing word now clearly signals a new subject ('the earrings (that) he killed for...' (versus 'He killed for the earrings' (?!'The earrings, he killed for')), not to mention the remnants of case in English ('him' versus 'he', and their usual places in word order) to ably assist us, plus let's remember that RCs aren't ever followed by a repetition of a supposedly missing element but something else far more interesting entirely (such as a description: ...were poisoned/very beautiful etc'). Is it only because of the supposed possibility that subject and object might be unclear or get mixed up in bits such as 'the earrings (that) he killed for...' that there is talk of subject versus object gaps (if so, all that hardly seems the clearest way to explore and explain this area of grammar! Kind of self-fulfilling prophecy thingy?).

But I haven't really thought about RCs for years, so I could be talking a load of rubbish!

[Metamorfose]

It's difficult to pin down a simple answer. I'd say 'no, that will only make things difficult for students' at basic level or if they are not interested in grammar at all, but maybe there are some people around who really like to see things moving like the mechanical universe and it might be useful for them...but thinking again, at the end of the day it sounds really unnecessary unless you have someone who really knows what they are talking about and on the other side people willing to learn this.

Under Generativist tradition the idea of 'move alpha' over the idea of erasing and filling the gaps has been used since Goverment and Biding theory (in the 80s) I guess. Alpha can be any element and it can take any place whithin the string of a sentence according to the language, type of word and the like)

[woodcutter]

Presumably the Longman Student Grammar went onwards and did something or other with these gaps. What was this thing? I can hardly wait.

If that isn't the case then of course they shouldn't mention it.

(As to the London Underground, the gap is generally tiny anyway. What a fuss!)

[fluffyhamster]

I forgot to add something about case - went back and edited my OP above.

There doesn't seem to be much more to the notion of 'gap' or use made of it in the LSGSWE than the above quote, and the following (which is more or less more or less of the repetitive same old same old as above LOL):

gap: the location of the missing constituent in a relative clause: ...a great athlete, which I believe I am ^. (9.8 )
(From the 'Glossary of terms')

9.13 Noun complement clauses

9.13.1 Noun complement clauses v. relative clauses

On the surface, noun complement clauses (NCCs), such as the following, can appear to be identical to relative clauses (RCs) with that:

The fact that it can be done is important. (ACAD)

However, NCCs and RCs are actually very different structures. Their differences are summed up as follows:

Function of clause: RCs identify reference of head noun; NCCs present the content of the head noun or add descriptive information.
Structure: RCs are incomplete and contain a 'gap'; NCCs are complete and therefore have no 'gap'.
Function of that: in RCs, a relative pronoun; in NCCs, a complementizer.
Omission of that: Is possible in RCs with object gaps; impossible in NCCs.
Types of noun modified: In RCs, almost any noun; in NCCs, only a few nouns.

(NCCs are similar to verb and adjective complements clauses, discussed in Chapter 10).

Compare the following two sentences, both with the noun report as head:

Postmodifying RC:
1) Peter reached out for the well-thumbed report that lay behind him on the cupboard top. (FICT)

NCC:
2) Other semiconductor stocks eased following an industry trade group's report that its leading indicator fell in September. (NEWS)

The RC in example 1 identifies which 'report' Peter is reaching for. It has a gap in subject position, which corresponds to the head noun report. The underlying meaning of the RC is that 'the report lay behind him', but that takes the place of the report in the RC. (Since the gap is in subject position, omission of that is impossible in 1, but it can be omitted in other RCs; see 9.8.2).

In contrast, in example 2 the NCC gives the actual content of the 'report': that 'the trade group's leading indicator fell'. The NCC is complete structurally. It does not include a reference to the head noun in any way (i.e. it does not have a gap corresponding in meaning to the head noun). In addition, the complementizer that cannot be omitted in this or any other noun complement clause. Finally, report is one of the few nouns that can be modified by complement clauses (see 9.14 below for more on head nouns).

I can't believe I just typed all that out.

Anyway, the last paragraph there is all well and true, but does it actually justify the comparatively more involved/less easy treatment given to RCs?

I suppose I should read up on Generative grammar, just so I can understand what "move anything anywhere" actually means (that's what 'move alpha' seems to be in my mind, from what little I've read about it in dictionaries like Crystal's, IIRC). I must say though that I prefer the idea of "start from somewhere" (i.e. a word, then follow it with another word appropriate after the first etc, in for want of a better word, Markovian fashion (I still need to really suss/read why finite-state grammars supposedly can't work - should "snuggle down" with Chomsky's original "arguments" one day, eh! )).

[woodcutter]

I don't recall any previous explanations bringing these noun complements into the equation. Are these kind of nouns all related with speech? Is that kind of "that" a kind of quotation mark?

If we have the sentence "The bread which I ate was good" then by this reasoning I get a lot of bread.

You have "The bread (2the bread) I ate (3the bread) was good",

where the pronoun "which" represents the second bread and the gap is the third bread. Assuming that we have to create number 2 by moving something from 3 means that unless an utterance is SVO we can't process it. And one of the problems is that sometimes syntacticians assume that this need for SVO applies to all speakers of all languages, and they reduce all sentences of all kinds to SVO, which is pretty strange. (Some people reduce every single thing to other orders as well)

I don't see that it explains too much. I explain relative clauses like this. You sometimes need to put in a relative pronoun in English because if it is missing it is confusing, because the relative clause is not flagged.

The dog ate the bone was big
The dog which ate the bone was big

If the subject of the sentence is the same as the subject of the verb in the relative clause then the absence of which throws you off in the first sentence.

But "The dog I ate was big"
"The dog which I ate was big" (guess where Woodcutter works!)

are both OK because word order flags the relative clause when the subject of the sentence and the clause are different.

[fluffyhamster]

It's not quite a the (good ol' crusty white bloomer RC postmodified) type, but I spotted a space to add a further batch of 'bread' to our overheating oven, Woody:

1The bread (2the bread) I ate (3the bread) was good (bread).

[fluffyhamster]

I don't recall any previous explanations bringing these noun complements into the equation.

There's probably something similar in each of the other more egghead tomes that I have, but I guess I've always skimmed through such discussion pretty swiftly, not given if too much thought, because I also couldn't recall (reading) much about it. For some reason though the LSGSWE treatment of gaps caught my attention, and thus the NCCs have accordingly crept more into view this time. Anyway, if I come across something that makes this stuff appear of really earth-shattering importance, I'll try to remember to post about it here at some point (but don't hold your breath waiting!). Or maybe someone will beat me to it!

[fluffyhamster]

Rather than start a completely new thread, I thought I'd bung this here.

I got a PM recently asking for grammar help (from me - as if! LOL ). Here it is:

I am analysing two newspaper texts. The first step is to break the piece into clauses. And BINGO ... I don't know how to do that. This is where I need simple advice and clear instruction.

My idea of a clause is this -
'Is a fundamental unit of meaning. It is structured around a verb phrase which describes some kind of process. The verb phrase is usually with one or more elements indicating the participants and the attendant circumstances'

I have this definition in my self-made grammar glossary... I kind of understand that.

Now ... the first sentence in my text is this:

'A collision between US and Chinese military aircraft has led to Pentagon accusations that China is intercepting US military aircraft in an "unsafe manner"'.

And I don't have a clue on how I can identify clauses within this sentence...and when I do attempt it ... I don't know if I am correct or totally wrong. I don't get the rules or manner in which I apply the definition to completing an action. When I see examples with clauses already inserted, I can see why ... but I am totally unable to complete the task myself...or lack confidence to do so.

Any pointers or rules I need to follow?

At the moment ... I am thinking it is just one clause anyway.

I'm going to try and look it up and find an answer myself...but thought I would run it past you anyway.

Most people seem to find it best to think of a clause as something akin to a simple sentence i.e. as a grammatical unit with at least a subject and a finite/tensed verb (verb phrase beginning with a finite/tensed verb) e.g. I've written this reply for you; You are reading it now.*

When there are two or more clauses in a sentence, then it is a complex sentence (as opposed to simple sentence) containing at least one dependent/subordinate clause besides its main clause. (I'm ignoring co-ordination, as that doesn't seem difficult to recognize).

According to Graeme Kennedy, 'Most subordinate clauses begin with a word or words which can mark the fact that the clause has subordinate status. These include...' (Click on the following GBS link for the listing):
http://books.google.co.uk/books?id=wE5X ... #PPA269,M1
(NB: Make sure that you compare Kennedy pp 269-270 with the LSGSWE pp 31-36 inclusive, esp pg 31).

The order of the two types of clause is not fixed - subordinate can precede main (If you like, I can give you an example), or the main precede the subordinate (I can give you another example (,) if you'd prefer), but let's assume that main will usually precede subordinate.

Having made this potentially helpful theoretical assumption, and in light of all the stuff above (especially the definition of clause as minimally subject followed by tensed verb), let's now look at your sentence.

'A collision between US and Chinese military aircraft has led to Pentagon accusations that China is intercepting US military aircraft in an "unsafe manner"'.

I'm going to go through it word by word. Tell me when we reach something verby, especially if it looks tensed.

1) A... (nope, no verb yet, just a determiner/the indefinite article (according to mainly the OALD7: http://www.oup.com/elt/catalogue/teache ... ?cc=global ))
2) A collision... (no verb, just det + noun)
3) A collision between... (no verb, just det + noun + preposition)
4) A collision between US... (no verb, just det + noun + prep + "adjective")
5) A collision between US and... (no verb, just det + noun + prep + "adj" + conjunction)
6) A collision between US and Chinese... (no verb, just det + noun + prep + "adj" + conj + adj)
7) A collision between US and Chinese military... (no verb, just det + n + prep + "adj" + conj + adj + adj)
8 ) A collision between US and Chinese military aircraft..(no verb, just det + n + p + "adj" + conj + adj + adj + n)
9) A collision between US and Chinese military aircraft has...STOP! WE HAVE A VERB!

Indeed, "our" verb, the finite/tensed one we were looking for. It actually starts a verb phrase that is a 'compound tense' ("present perfect") i.e. 'has led' ('led' is a past participle/non-finite form, by the way, and the main/lexical verb, with 'has' the finite auxiliary that adds perfective aspect. Skip on from that last bracketed sentence or two and don't come back to it if it's causing you to start to freak out! LOL;).

So we now know that we have a subject and a finite verb, the two essential elements of a clause according to our definition. The subject is the long noun phrase 'A collision between US and Chinese military aircraft', the verb phrase is 'has led'. Obviously that sounds a bit incomplete, so we ought to see if there is an object or whatever completing the clause...

The preposition 'to' bridges/links the verb phrase to...

Pentagon ("noun", but kinda adjectival, seeing as there is no determiner)
Pentagon accusations ("compound noun/NP", and plural, should suffice for now!)
Pentagon accusations that - NP + a conjunction.

The OALD states: that: as conj. Used after some verbs, adjectives and nouns to introduce a new part of the sentence: She said (that) the story was true. It’s possible (that) he has not received the letter. The fact (that) he’s older than me is not relevant.

So, what is this 'that' introducing in the rest of our sentence? (T)hat...

China
China is
China is intercepting
China is intercepting something - all of which is, yup, basically another clause.

Now, you might think this embedded clause is a subordinate one, but its function is actually more to just postmodify/complete the noun phrase that began with 'Pentagon accusations' (and ends with 'manner').

So I'd say that you "basically" have a sentence there with two noun phrases, the first relatively simple, the second relatively complex (due to the embedded clause), which are hung each side of a pivotal central verb phrase and which function respectively as subject and object of a single clause overall.

I don't know if that answer is satisfactory (you might fear that some will shout, 'Wrong, there are TWO clauses ultimately in that sentence!', even though the latter is embedded within and postmodifying a noun phrase/object rather than clearly/immediately subordinate to any main clause), but I think that saying 'One clause' and providing a hopefully sophisticated enough (and obviously hopefully correct!) explanation as to why that should be is better than "wimping out" with 'Two, Sir!' simply for fear that here is a trick question absolutely requiring the latter answer. Which is all a way of saying that if you do answer 'Two clauses', you should probably make sure that you explain how one is part of something functioning as the object of a bigger clause.

Hope this helps, and please tell me if I am wrong! (I'm never 100% sure about things myself - this could all be a load of twaddle!).

By the way, the word-by-word linear build-up isn't really necessary once you are more confident in your parsing abilities; quite a few people would go straight for a clean binary subject-predicate/NP-Chosmkyan "VP" parse/division ('A collision between US and Chinese military aircraft / has led to Pentagon accusations that China is intercepting US military aircraft in an "unsafe manner"'), then divide the VP into its further constituents etc.

*I won't bore you with non-finiteness except to say that in contrast to finiteness, non-finiteness is traditionally reserved for phrases that "lack" a/"THE" subject within them themselves and therefore don't have their verbs inflected for tense (compare a finite verb phrase without a subject - either an imperative (Stop! vs ?You stop!) or apparent nonsense (*Stopped!)). For example, in My father travelled by two buses each day / to get there on time, / leaving home at 5am / and usually returning after 10pm, there is a clause (finite ~) followed by three non-finite phrases (=verb phrases at least: 'to get', 'leaving', and 'returning'). Note that non-finite phrases logically are a type of subordination/dependent. Other joys to get your head around eventually are 'verbless "clauses"', 'absolute "clauses"' etc (see the Chalker & Weiner Oxford Dictionary of English Grammar that I'm always recommending - a bargain at around a tenner).

Is my explanation sound, or dodgy? (Or how about clear versus unclear, too complex (or not complex enough?!) etc). Feedback (for both our sake's!) will be appreciated!

[woodcutter]

I would have thought that there were obviously two clauses, (even if defining clauses is problematic) but that giving a title to the "China is intercepting" clause is tricky without clearly identifying the role of this kind of "that", which we haven't been able to do here, but which here again seems to function as a kind of quotation mark for the accusation.

You are saying that there is only one clause in the whole thing?

[fluffyhamster]

Well, yes, there are identifiably two finite clauses, but at the level of functioning as 'elements of "the clause"', are there still two?

(I)f you do answer 'Two clauses', you should probably make sure that you explain how one is part of something functioning as the object of a bigger clause.

So I'd say that you "basically" have a sentence there with two noun phrases, the first relatively simple, the second relatively complex (due to the embedded clause), which are hung each side of a pivotal central verb phrase and which function respectively as subject and object of a single clause overall.

[fluffyhamster]

This should be of interest, Woody (and whoever else might be reading):
http://img29.imageshack.us/img29/8373/pearce.jpg
(from Michael Pearce's The Routledge Dictionary of English Language Studies).

(I'm referring of course to the That-clause entry, specifically the part in point 3) that begins 'There is another group of nouns...' and ends with the example ending '...nothing can be done.' I like Pearce's phrasing/the idea that the noun controlling the clause is the very source of the information in the clause. Simple stuff really, but it can and does help when reference books actually remember to state the obvious sometimes!).

I also noticed and now quite like the notion of clausal sentence as expressed in Huddleston's 'A Short Overview of English Syntax' (see section 1, right at the top of the document): http://ling.ed.ac.uk/~gpullum/grammar/otherstuff.html .

[woodcutter]

I can't read the meat in either of those I'm afraid.

This is what anonymous person wrote in wikipedia about sentence elements and it is in line I would have said - things that follow a complementizer can be a clause too, the definition that we generally use in reality is the simple one given below. (I don't see that whether the "that" as a complementizer follows a noun or verb makes much difference)

(in these cases the that seems a kind of quotation mark for internal ideas)

clausal sentence elements
A clause consists of a subject and main verb. Not all clauses function as sentence elements. Look at the two contrasting examples below.

We know that he is a fool. (SUBJECT + VERB + CLAUSAL OBJECT)
The man who is a fool knows nothing. (The clause is qualifying the man and is not a sentence element in its own right)
The clause can function as a subject, object and adverbial sentence element.

Whether he is guilty is the issue. (CLAUSAL SUBJECT + VERB + COMPLEMENT)
We know that Mr Jenner eats cabbage. (SUBJECT + VERB +CLAUSAL OBJECT)
We arrived before they left. (SUBJECT + VERB + CLAUSAL ADVERBIAL)

[fluffyhamster]

Are the following any meatier?

Regarding the "number of clauses" issue:

main clause A CLAUSE which is capable of making a complete sentence by itself; a clause which is not a SUBORDINATE CLAUSE. A sentence always contains at least one main clause, and a SIMPLE SENTENCE consists only of a single main clause. In the simple sentence Susie finished her drink, the whole sentence is the main clause. In the COMPOUND SENTENCE Susie cooked dinner, and Natalie did the washing-up, there are two main clauses connected (conjoined) by and. See also MATRIX CLAUSE.

matrix clause A clause which contains a SUBORDINATE CLAUSE within it. In the COMPLEX SENTENCE The employees who were dismissed are suing the company, the matrix clause is The employees...are suing the company, while the remainder is the subordinate clause (a RELATIVE CLAUSE) contained within it.
Does the subordinate clause itself form part of its matrix clause? Traditional grammarians usually answered 'no' to this question, but contemporary linguists more often prefer to answer 'yes'.
A matrix clause is often a MAIN CLAUSE, as in the example above, but it need not be: it can itself be a subordinate clause. In the sentence The victim told police that the man who attacked her had had a beard, the subordinate clause who attacked her is contained within the subordinate clause that the man...had had a beard.

Nice and chatty though Trask is, some may however in this case prefer the succinctness (though in other cases it can border on terseness) of Chalker & Weiner's 'matrix clause' entry in their Oxford Dictionary of English Grammar:

matrix clause
A superordinate clause minus its subordinate clause.

This is what in popular grammar is called the main clause of a complex sentence.

In grammatical descriptions that use the term main for the entire superordinate clause, matrix is used to distinguish the part from the whole:

I was ten (matrix) when I got my scholarship (subordinate).
Nobody had expected (matrix) that I would get one (subordinate).

Regarding the "NCCs":

4.22 Noun clauses, especially 'that'-clauses, can also be used in apposition to a noun group, as an explanation of that noun. This often happens with nouns related to reporting and thinking (e.g. 'fact', 'belief') and with some other nouns (e.g. 'danger', 'possibility').

The fact that many of us eat too much junk food can hardly have escaped anyone's notice.
He had condoned it at first in the belief that he was genuinely helping further the cause of science.
The producer may then use the logo on the wine label as evidence that the wine is 'verified' as organic.
There was a danger that many of the country's important foreign workers would decide to leave.
They called their child Indiana, prompting the question, what's wrong with a good old English name?

4.23 Note: 'that'-clauses sometimes look like relative clauses, but their function is different.

He had heard rumours that old Fritz's heart wasn't as good as it used to be. ('that'-clause in apposition)
I am aware of the rumours that have recently been circling about me. ('that' introducing relative clause)

(Then there is the term 'content clause', as apparently used in Huddleston & Pullum's CGEL in preference to 'that'-clause).

[Heath]

Sorry, but as soon as I saw the 'gap' it screamed "Chomsky" at me, and I stopped reading.

So, um, was the answer to Fluffy's question about it stemming from Chomsky answered? I guess it did - but to save reading everything, would anyone mind confirming it?

[fluffyhamster]

Hi Heath, how's it goin'?

The purpose of this thread was (and still is!) actually more to question than champion the notion of 'gap', so you might like to go back and give it a chance (can't absolutely guarantee though that it will still be that enjoyable or even interesting! ).

Anyway, to answer my/your/our/the question, the term indeed appears to be (as probably everybody suspected) of 'Generative' origin - for example, a quick search on Google for 'gap linguistics term introduced by' produced the following link (near the bottom of the first page of results); the quote of specific interest is about halfway down the page it is taken from.

http://www.umich.edu/~cfc/rosslatin.htm
>

During the 1960s Sweet’s (Waldo ~ - FH) and Seligson’s interest in transformational grammar gave rise to an emphasis on sentence-level phenomena, most notably gapping. Gapping, traditionally called ellipsis, is the absence of common elements in one of two clauses, or, in linguistic terms, 'unfulfilled syntactical expectations.' This notion was included in the 1966 LASA, but the term gap came into use only following an 1970 article by John Robert Ross (the husband presumably of the Deborah Ross who wrote the webpage's content. Ross himself though has since the mid to late 60's and early 70's always been a pretty "big"/oft-quoted name in Generative linguistics - FH) on gapping. Like the kernel and metaphrasing, the concept of the gap proved highly significant, as it allowed a systematic treatment of an area of Latin which follows rules distinctly different from those of English.