non-semantically-predicative adjectives

This is the place to post questions and discussions on usage and style. The members of the Wordwizard Clubhouse will also often be able to help you to formulate that difficult letter.

non-semantically-predicative adjectives

Post by Edwin F Ashworth » Tue Feb 28, 2012 1:25 am

Non-semantically-predicative adjectives

Like noun modifiers, classifying adjectives cannot be used predicatively:

He is a football manager. *That manager is football.
Burning processes always involve chemical reactions. *Those reactions are chemical.

However, the modifier is semantically closely linked to the head noun in both cases.

Some authorities use the terms inherent adjectives and non-inherent adjectives to distinguish those (usages) which characterise the referent of the noun directly from those which don’t .

Thus:

a heavy man (inherent usage) (ie heavy refers to the man)
a heavy drinker (non-inherent usage) (with the standard meaning) (ie heavy refers to the scale of drinking, not the drinker)

However, this classification is very tricky for two reasons.

(1) Polysemy, especially figurative, and there is a lot of it about, makes it unclear when one can claim unequivocally that an adjective is or is not characterizing the referent of the noun directly.
For instance, at elook.org are included definitions for heavy as below:

9. [adjective] marked by great psychological weight; weighted down especially with sadness or troubles or weariness; "a heavy heart"; "a heavy schedule"; "heavy news"; "a heavy silence"; "heavy eyelids"
11.[adjective] (used of soil) compact and fine-grained; "the clayey soil was heavy and easily saturated" Synonyms: clayey, cloggy
12. [adjective] darkened by clouds; "a heavy sky"
Synonyms: lowering, sullen, threatening
13. [adjective] of great intensity or power or force; "a heavy blow"; "the fighting was heavy"; "heavy seas"
15. [adjective] (of an actor or role) being or playing the villain; "Iago is the heavy role in `Othello'"
16. [adjective] permitting little if any light to pass through because of denseness of matter; "dense smoke"; "heavy fog"; "impenetrable gloom" Synonyms: dense, impenetrable

So, do we label heavy in heavy fog an inherent or non-inherent usage? Thick would be inherent (for this fog); but weighty (if it were allowed) would not describe an innate property.

(2) The sense of an adjective may shift grossly or subtly when it moves from attributive to another position:

The concerned parents v the parents concerned
My old friend v my friend is old
The wrong book v "The book is wrong"

At http://books.google.co.uk/books?id=me11 ... &q&f=false Elizabeth Coppock concentrates her research on whether or not the adjective is semantically predicative – is the adjective actually modifying the noun it seems syntactically connected to, or is there another explanation.
She lists various semanto-syntactic classes of adjectives:

Modal and veridical adjectives (A) _ a true king; an alleged gangster; a suspected arsonist;
(The adjective modifies the proposition / theory involved, not the referent {the king; the suspect})

Modal adjectives (B) _ a former president; an utter fool;
(The adjective sets parameters – time, extent – on the office tenure / degree of idiocy, etc)

Predicate-evaluating adjectives _ a mere child; common knowledge;
(The adjective characterizes the state exemplified by the referent:
He is merely a child – ie a member of an estate regarded as having little wisdom, status or power.
It is common knowledge – widely known amongst the whole population.)

Adjectives of selection _ the very man; this exact spot; the wrong trousers;
(The adjective highlights or defines more accurately, or endorses / rejects the identification or selection involved)

Event-manner adjectives _ a beautiful dancer; a heavy smoker; a hard worker
(The adjective refers to the way the person performs the event or practice involved, not to the referent, the person per se) (who may in the first example, for instance, be badly scarred)

(Many of the above may be restated using sentence (or, for the event-manner category, matrix) adverbials, eg

truly a king; allegedly a gangster; formerly President; merely a child; exactly here; she dances beautifully

Coppock risks calling such adjectives adverbial-type adjectives)

Psychological experience adjectives _ a proud day
(With this type, the adjective refers semantically not to the adjacent noun, but to the person having that proud day etc.)
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by dante » Tue Feb 28, 2012 10:45 am

Nice analysis except for this part:
(Many of the above may be restated using sentence (or, for the event-manner category, matrix) adverbials, eg

truly a king; allegedly a gangster; formerly President; merely a child; exactly here; she dances beautifully

Coppock risks calling such adjectives adverbial-type adjectives)

For the same reason one can't say that "football" in He is a football manager is an adjective, it is not possible to say that allegedly in "allegedly a gangster" is an adjective either. "Allegedly" is an adverb functioning as a pre-modifier of the noun "gangster". It is "..an external modifier occurring at the periphery of of the NP" (CGEL p436). I think that adverbs cannot function as internal dependents - complement of the noun i.e their use as noun modifiers is always "peripheral" to the noun's meaning. As was said on the forum many times before, adjectives are by far the most common pre-modifier of nouns but obviously not the only one.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by Edwin F Ashworth » Tue Feb 28, 2012 1:59 pm

The antecedent for (such) adjectives is Many of the above not sentence adverbials; if you want a precisionist statement:

Many of the above examples may be restated using sentence (or, for the event-manner category, matrix) adverbials ...

... Coppock risks calling adjectives used in constructions where such is the case adverbial-type adjectives.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by dante » Tue Feb 28, 2012 4:45 pm

This book you've linked to Edwin is an interesting reading. The author's thesis with similar (or the same) content is available here http://eecoppock.info/CoppockThesis.pdf I'm not sure though that it helps if certain semantic classes of adjectives are named after syntactic elements or other word classes because of their semantic properties. If I take this citation from the book, for example:
The adverbial class of non-predicative adjectives contains those that can be paraphrased by an adverb. For example, a true poet is truly a poet.
Does this mean that "truly" belongs to an adjectival class of adverbs because "truly a poet" can be rephrased as "true poet", or it works only in one direction?

Here are two examples from literature, which I've adapted to illustrate the point about pre-modifying adjectives and adverbs in a noun phrase, similar to the explanations on page 436 of CGEL :

1. Allegedly, the most conservative administration since that of Calvin Coolidge generated those deficits.

2. Allegedly the most conservative administration since that of Calvin Coolidge generated those deficits.

2a) Those deficits had been generated by allegedly the most conservative administration since that of Calvin Coolidge.

3. The alleged poor conservative administration had generated those deficits[/i][/b].

3a) Those deficits had been generated by the alleged poor conservative administration.


In 1, the adverb "allegedly" is a (modal) adjunct in the sentence structure. If comma is left out after "Allegedly" it becomes part of the noun following noun phrase - an external modifier in its structure, as shown by sentence 2. The passive form of this sentence 2a clearly shows that "allegedly" is an integral part of the noun phrase "Allegedly the most conservative administration since that of Calvin Coolidge", as the subject is now part of the by-agent phrase. Finally, sentences 3 and 3a illustrate the syntactic function of the adjective "alleged" similar to that of the adverb "allegedly" in sentences 2 and 2a. Instead of categorizing the adjective "alleged" as "adverbial-type adjective" (or possibly categorizing "allegedly" as adjective-type of an adverb), maybe it could be analyzed as transferred epithet ( purely semantic term and more concise and to the point description of these adjectives) as suggested in CGEL:

"Traditional rhetorical analysis uses the term transferred epithet or the word hypallage (from the Greek for "exchange) for such cases.

Since the term refers to the situations where "..the adjective does not apply literally to the head nominal.", it is probably applicable to most situations where the adjective is of "adverbial-type", that is, when it refers more to event-manner, as the author noted, than it has to do with the properties of the noun it modifies.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by Edwin F Ashworth » Tue Feb 28, 2012 9:39 pm

As often seems the case, Dante, you refuse to ditch / amend an explanation that was quite a good attempt in its day, when presented with a better one.

(A) If it makes you happier, forget the term adverbial class of non-predicative adjectives and use the descriptors 'that subset X of adjectives where the adjective does not reference directly the head nominal' and then 'that subset Y of X where an easy recasting to an equivalent statement using an adverbial, whether sentence or matrix, is possible'. It's just nicer to have more appropriate names than X and Y. And I'm allowing mere, very (in the senses below) etc to retain their traditional name adjectives. For now.

(B) Hypallage in its broad sense describes constructions alien to modern ears:

Virgil was given to hypallage beyond the transferred epithet, as "give the winds to the fleets" (dare classibus Austros, Aeneid 3.61), meaning "give the fleets to the winds." (Wikipedia)

I had not intended to cover classical rhetoric, but, rather, common modern English usages (albeit grammatically odd ones).

The subset of hypallages, transferred epithets, seems to largely overlap with what Coppock calls Psychological experience adjectives*:

One kind of hypallage, also known as a transferred epithet, is the trope or rhetorical device in which a modifier, usually an adjective, is applied to the "wrong" word in the sentence. The word whose modifier is thus displaced can either be actually present in the sentence, or it can be implied logically. (Wikipedia) [emphasis mine]

A proud day. Who is proud? The person/s mentioned (a proud day for the Welsh) in the rest of the dialogue.

* There are other examples of transferred epithets, such as 'Disabled Toilet'; Coppock seems to have missed this category, unless she regards disabled as a noun here (Toilet for the Disabled rather than Toilet for Disabled Persons). I can't really include it with her Nominal adjectives category, though it is virtually a compound noun. The missing referent is obvious:
Who are disabled? Intended users of the toilet.

However, there does not seem to be an easy-to-deduce 'word whose modifier has been displaced to appear next to a noun that is related but hardly the referent' in any of the other categories:

A true king (Who is true? Nobody (in this sense). True refers to the underlying assertion, not to a hidden person.) True is not a transferred epithet; neither is alleged, by the above definition (and I'm not aware of an alternative).

An utter fool. Who is utter??!!

A mere child. Who / what is mere?!

This very man. Who is very?

A beautiful dancer. Who are we claiming to be inherently beautiful? Nobody - the person's named activity is beautifully performed.

Coppock covers the whole topic of 'adjectives-or-are-they which do not modify the adjacent noun' much more satisfyingly, in terms of thoroughness and providing sensible explanations for the idiosyncratic meanings the structures convey, than the CGEL.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by dante » Wed Feb 29, 2012 11:58 am

..it is probably applicable to most situations where the adjective is of "adverbial-type", that is, when it refers more to event-manner, as the author noted, than it has to do with the properties of the noun it modifies.
Or, on the second thought, probably not applicable, especially not as a general rule. Something can be "alleged", can be "possible" or "apparent" or anything along that line, and neither of these look like a transferred epithet to me now. I may change my mind again on this, it isn't so straightforward thing obviously. Now I've cleared things up with myself Edwin, I can comment on your writing :)
If it makes you happier, forget the term adverbial class of non-predicative adjectives and use the descriptors 'that subset X of adjectives where the adjective does not reference directly the head nominal' and then 'that subset Y of X where an easy recasting to an equivalent statement using an adverbial, whether sentence or matrix, is possible'
It does make me happier when it's put that way because it helps keeping the concepts of word categories apart as well as keeping form and function apart. When we say adverbial adjective, or adjectival noun or so, I feel that it's not clear where one word category stops and the other begins.

As I said, the book you linked to seems a very good reading , only it would take me half a year at least to go through it cover to cover :) Pity I don't have the time at the moment to read it carefully. How I feel about it is that taking semantic aspects into the account makes the analysis much more complicated and demanding. To take the sentences from my previous post as an example, it doesn't take much thinking to realize that "allegedly the most conservative administration" and "alleged poor conservative administration" in 2 and 3 are both noun phrases - they are passivized as a whole, to name the most apparent reason for such claim. Only a little more attention should be given to the analysis of the elements of these phrases to find out that "allegedly" is a bit more peripheral to the meaning of the head noun than "alleged" is in the other phrase. These broad considerations are a framework for any further, more detailed analysis of the elements of these phrases. And from this point in analysis on, you suddenly realize what "the devil is in the detail" means actually :) One would need to read a whole book like this one to learn the nuances of "allegedly the most conservative administration" and "alleged poor conservative administration", or another book, on the semantics of adjectives, to learn why "alleged poor conservative administration" and not "poor alleged conservative administration" or "conservative poor alleged administration" and so on.
So, as an ESL learner and a language layman I'm obviously not in a position to assess the value of the Coppok's book. I mentioned this point with using the term "adverbial adjective" because it seems inconsistent with the terminology used in CGEL, for the reasons the authors stated in their book.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by Edwin F Ashworth » Wed Feb 29, 2012 6:46 pm

Care, Dante.

'Verbal noun' is considered by many to be the best way of labelling -ing forms somewhere along the continuum from present participle to deverbal noun. As Phil White pointed out from Quirk, gerund only covers a smallish (but important) portion of this cline.

YOU (not Coppock) have introduced the similar-looking term "adverbial adjective"; she, undoubtedly anticipating the possible confusion, risks merely (!) "adverbial-type adjectives". I'd have probably coined "adverbial-substituting quasi-adjectives", since I consider them to be corruptions of adverbial expressions (ie I'm guessing the adverbials came before the asqa's) and don't consider them near enough in properties to belong in the adjective class. They may even be further from ('true') adjectives (ie adjectives) than determiners are, and they (determiners) are never regarded as adjectives nowadays.
But then I didn't do the excellent research, so I'll use Ms Coppock's term - to general relief.

I'm glad you see that by no means all of these strange moieties can be classed as adjectives whose referent is a noun, whether visible or to be understood, other than the one which follows them.

It is interesting to note that alleged, reputed, purported, possible, probable and certain can all be readily replaced by equivalent adverbials, but almost certainly suspected and certainly known (I'd never use suspectedly) cannot.

I started thinking about (and researching) the topic of non-semantically-predicative adjectives(-or are they) when trying to pin the meaning and grammatical behaviour of mere - surely the most outlandish (some would only go as far as peripheral) of the lot. I don't think you can give a reasonable definition in the usual sense. You're stuck with a re-write such as:

usage exemplar:

He is a mere child – He is a member of that lowly estate - regarded as having little wisdom, status or power - we know as children.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by dante » Wed Feb 29, 2012 8:58 pm

'Verbal noun' is considered by many to be the best way of labelling -ing forms somewhere along the continuum from present participle to deverbal noun. As Phil White pointed out from Quirk, gerund only covers a smallish (but important) portion of this cline.
I agree that the term is widely used, and it is noted in CGEL:
Dictionaries tend to define the gerund as a verbal noun, but there are strong grounds for analysing destroying in "Destroying the files was a serious mistake" as a verb, and for drawing distinction between such words and others ending in -ing which genuinely are nouns and which we refer to therefore as gerundial nouns:

He was expelled for killing the birds form of verb
She had witnessed the killing of birds. gerundial noun

The grammatical differences between these are explained in the following text. The point is that the authors of CGEL avoid the terminology "verbal noun", "adverbial" etc to keep the definition of word categories as clear as possible and to divorce form and function terminologically. In other words, you won't find verbal nouns in H&P grammar, which I was referrring to.
YOU (not Coppock) have introduced the similar-looking term "adverbial adjective"; she, undoubtedly anticipating the possible confusion, risks merely (!) "adverbial-type adjectives".
COPPOCK not me used the term Edwin, take a better look :) I've discovered that there are even "nominal adjectives" :)
Many of the adjectives in the non-predicative category can be classified as “adverbial,” because they can be paraphrased with an adverb. These include adjectives of veridicality (as in true king, real friend), the so-called “modal” class of adjectives (as in alleged criminal), along with other adjectives of modality (old school, erstwhile friend), adjectives expressing a degree of veridicality (perfect ass, pure nitwit), adjectives evaluating the predicate they modify (mere kid, common soldier), adjectives of selection (the very man, the same reason), and event-manner adjectives (hard worker, beautiful dancer). Alongside adverbial adjectives, there are adjectives of psychological experience (as in sorry sight) and so-called “nominal adjectives” (as in criminal lawyer). I will address each of these classes in turn.
Anyway, even if I'd invented "adverbial adjective", "adverbial-type adjectives" still sounds mouthful and unnecessary hyper-production of terminology. I mean, I've read the surrounding text more carefully this time and the point of "adverbial adjective" makes more sense to me now, but was a separate term really necessary here? The class of adjectives which belong to different semantic categories but have in common that can be replaced by an adverb and they are all non-predicatives? Maybe it makes sense within the wider context of this chapter - transformational analysis of prenominal adjectives, I don't know. I should read the whole book I guess to come to a definite conclusion.

Another problem is the term"adverbial" itself, because of its similarity with formal category of adverbs. Quirk and the other authors say in their Comprehensive grammar that the category of adverbials is a functional category that can be realized by different formal categories - adverb phrase, prepositional phrase, noun phrase. But why name it "adverbial" then? Only to make a confusion I guess :) Now, when you say "adverbial-type adjective" one isn't sure if the kind of adjectives referred to are similar to adverbs or to adverbials (functional element). As the quote above shows, the author had in mind "adverbs", using the term adverbial as an adjective, derived from "adverb", not referring to the functional category of adverbials. Whatever the case is, the things are cleared up in CGEL by using the term "adjunct" for "adverbials", much to the point in my opinion.

Finally, the similar cross-category synonymous realizations of syntactic elements are most probably abundant, and I could even think of an example:

It is a possibility/ possible that I don't understand the idea well enough.

A possibility belongs to the class of adjectival-type nouns and possible belongs to the class of noun-type adjectives.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by Edwin F Ashworth » Wed Feb 29, 2012 10:22 pm

Apologies; you're right. The descriptors adverbial adjectives and adverbial-type adjectives are both used in the piece by Coppock. Of course, even the first term is not inaccurate, being licensed by the adjectival usage of adverbial (!):

The word adverbial is also used as an adjective, meaning 'having the same function as an adverb' (Wikipedia)

So, in for instance He is an alleged criminal, the modifier alleged is equivalent to the (sentence) adverbials Allegedly or It is alleged that (though I'd rather call these 'pragmatic markers subset modality' myself). I'm still working on how to describe what we're talking about - pragmatic-marker-subset-modality-like pseudo-adjectives?

One problem you address, the confusion over terms like adverbial, where different grammarians / dictionary compilers use the term in conflicting ways, is sadly commonly encountered in studies of English. It urgently needs addressing.

The other major problem, the lumper / simplifier v splitter / fine-grinder debate, is, in my opinion, less tractable. There will always be differences of opinion. However, when a word is not functioning anything like a typical member of 'its class', I think it needs re-classifying.

With regard to the admitted (!) complexity of the whole issue, in the piece P White refers to by Quirk et al, there are no less than 13 different intermediate grades claimed by the authors between
(1) I am going to burn this old painting. and
(15) He is painting his daughter. (See 'fishing' post)

Where do the 'verb /-al /-ish /-y /-ious ones' start along this scale?
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by dante » Wed Feb 29, 2012 11:20 pm

Where do the 'verb /-al /-ish /-y /-ious ones' start along this scale?
I remember Phil's explanation of gerund here on WW Edwin, I'll look for it and add more thoughts on this next time, in this thread or in some of the previous threads about gerund/participle distinction.
The main point where H&P think the traditional analysis of gerunds and participle is flawed is in taking the functional role the ing form has in the sentence to decide if it is a noun or a verb. They illustrate this approach on this example:

Destroying the files was a serious mistake.
I regret destroying the files.

In the first sentence "destroying the files" is in the slot typically filled by a noun - therefore "destroying" is a noun in traditional analysis. It also can be rephrased as "destruction of the files", meaning that the ing form derives from a noun, another reason to analyze it as a noun in this phrase. The same idea is applied in the second example, where "destroying the files" is in the position of an object - again, a noun, because the object is typically realized by a noun. "The primary difference between a gerund and a participle, therefore, is that while a participle is functionally comparable to an adjective, a gerund is functionally comparable to a noun" CGEL p81

The authors of CGEL give four compelling reasons against this analysis. The first one is that nouns are post-modified with of-PP phrases, and don't take other nouns as objects. Really, is there a noun, apart from this ing form analyzed as a noun in traditional account, that can take another noun as an object? Can you table the dinner or student a subject or anything similar where you would have a noun perform some action on another noun? Not that I can think of.
Another reason for taking the ing form to be a verb in similar situations is that they are in such cases, as other verbs, modified with adverbs, not adjectives. Adjectives are noun-modifiers.

He was expelled for wantonly killing the birds. ( ing form is a verb, it is modified by the adverb wantonly, takes "the" birds as its object)
She had witnessed the wanton killing of the birds (ing form is a noun, it is modified by the adjective "wanton" and followed by an of-phrase)

The and similar determiners combine with nouns not verbs, that's why "*the killing the birds" doesn't work, but "the killing of the birds" does.

Finally, nouns typically have plural: These killings must stop. You can't do this with verbs : *Killings the birds must stop.

However, when a word is not functioning anything like a typical member of 'its class', I think it needs re-classifying.
Then the question is: if it walks like a duck, swims like a duck and quacks like a duck is it a duck? The -ing forms in these examples function as nouns typically do but don't have other syntactic properties similar to nouns, which means that they may quack like a duck but then they walk and swim like some other animal :)
It seems essential and the starting point in the analysis of the sentence to accept that anything can function as anything. We've shown in the previous posts that adverbs can modify nouns, nobody disputed that. It is not anything like their typical function, but we find them in that position, don't we? I don't think it would help if we named adverbs differently to explain their peripheral use as noun modifiers.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by Edwin F Ashworth » Thu Mar 01, 2012 12:12 am

But, again, why this analysis? Why the examples that illustrate only this explanation?

In the 15 examples Phil quotes from Quirk & Greenbaum, A University Grammar of English, is the grammatical example (their No [6]):

Brown's deftly painting his daughter is a delight to watch.

Here, painting (a) occurs in a genitive construction and so 'must be' a noun.
and (b) is modified by an adverb and therefore 'can't be'.
(I don't accept that adverbs can modify nouns in such constructions. In the Conversion - adverbs functioning as adjectives (!?) post, the claim was for the conversion (hence the title) of certain words from one class to another:
The underlined words [which are better known as members of the adverb class] do not function as adverbs
; as adverbs, they would have nothing to modify. Clearly they are used as adjectives. [not as if they were adjectives]
[(cf he drives too fast; fast = adverb
his car is not fast enough; fast = adjective).
And perhaps at one time, fast was only used as an adverb, but has later undergone conversion to a useful adjective. While the adverb, also useful, still exists.]

And how are you going to define say adverbs if not by agreed properties? Have Quirk et al, or Pullum et al, or Crystal an authorised and unchangeable list? Who gave it to them? Does the list say on which days 'fast' is an adjective and on which an adverb?
Or should we look at the way it is being used in a sentence?
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by dante » Thu Mar 01, 2012 11:32 am

"Allegedly" it becomes part of the following noun phrase - an external modifier in its structure, as shown by sentence 2.
I'll need to correct myself once again. When we remove the comma, "allegedly" does become part of the following noun phrase, but not as an external modifier of the head noun - it modifies the adjective phrase "most conservative". So, the example I chose is not an appropriate illustration of the point that adverbs can function as noun modifiers.
Brown's deftly painting his daughter is a delight to watch.

Here, painting (a) occurs in a genitive construction and so 'must be' a noun.
and (b) is modified by an adverb and therefore 'can't be'.
Adverbs can modify nouns, so that criteria wouldn't prevent analyzing "painting" as a noun. But let's apply other criteria for nouns, why do you stop with these two in the analysis Edwin? If "painting" is a noun, why can't you make it plural, as it is supposed to be a singular noun in this sentence? Why can't you use an adjective to pre-modify the noun "painting": Brown's nice painting his daughter...should work if it is a noun. Why can't you say "The deftly painting his daughter..", when the fact is that singular countable nouns take definite or indefinite article?
Brown's is analyzed as the subject in the absolute construction/ non-finite clause, which itself is the subject in the main sentence. I guess that the nominative would be acceptable too : Brown deftly painting his daughter .., as an alternative, where we don't have this problem with genitive being a subject in the clause structure. I don't know if analyzing genitive in this construction as the subject is theoretically defensible, but the fact that nominative is almost a perfect replacement for it makes me happy enough with this analysis.

You simply can't take a single functional role of any of the word categories to be the sole criteria for its classification. There's no point in it, no classification is possible if we don't include everything we know about the syntactic behaviour of any individual word, on all the levels of the analysis. How I understand it, defining a word category consists of analyzing its characteristic syntactic and semantic properties, comparing it with the properties of other words/word classes and examining how it combines with other words in different syntactic structures. That's how we get the whole picture, and a meaningful word categorization.
What the word does, that is, its function on the level of a clause or phrase, is another thing. Knowing word functions helps of course, as any of the word classes has their typical syntactic function, but none of these functions can be said to be their exclusive property. In other words, syntactic functions are realized by heterogeneous concepts, which can't be meaningfully categorized together as one and the same formal category.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by Edwin F Ashworth » Thu Mar 01, 2012 1:13 pm

Could you explain where, in the following series of statements, you first disagree, please, Dante?:

(A) In Jeremy drives a fast car, fast is an adjective.

(B) In Jeremy was driving too fast, fast is an adverb.

(C) In Jeremy has destroyed his painting of Captain Slow, painting is a noun (specifically, a deverbal noun as it has converted completely to noun from verb in this usage).

(D) In Jeremy is painting Captain Slow, painting is a verb (specifically, a present participle).

(E) In Brown's deftly painting his daughter is a delight to watch, behaviours typical of both noun and verb are displayed by painting.

(I could point out nounal behaviour, with supporting references, but I think you agree on the noun-like behaviour to the extent of saying painting as used in (E) actually is a noun. I'll stick to giving support for its verb-like behaviour):

(ad·verb (dvûrb)
n. Abbr. adv.
1.
The part of speech that modifies a verb, adjective, or other adverb.
2. Any of the words belonging to this part of speech, such as so, very, and rapidly.
The American Heritage® Dictionary of the English Language,

adverb [ˈædˌvɜːb]
n
(Linguistics / Grammar)
a. a word or group of words that serves to modify a whole sentence, a verb, another adverb, or an adjective ...
Collins English Dictionary – Complete and Unabridged

An adverb is a part of speech that changes the meaning of verbs or any part of speech other than nouns (Wikipedia)

(I myself would prefer a more rigorous definition, to limit adverbs solely to the modification of verbs. But I can't find the less rigorous definition you claim in any respectable - indeed, in any - dictionary. So, deftly in the example in (E) is modifying something other than a noun.)

To be fair, I think that these definitions do need a certain amount of improvement to be more comprehensive. In expressions like The concert tomorrow will feature Herb Alpert, the postposed modifier tomorrow is ellipted from 'which will take place tomorrow' , 'being held tomorrow' etc. BUT I know of no such constructions with preposed modifiers, or with modifiers not obviously related to clauses. I think this sort of construction is also limited to indicators of time and place (http://www.usingenglish.com/glossary/adverb.html):

ADVERBS MODIFYING NOUNS

Adverbs can modify nouns to indicate time or place.

EG: The concert tomorrow

EG: The room upstairs

So deftly would fail on three counts (to be classifiable as modifying a noun).

Adverbs are, of course, used with -ing forms, as with deftly here and quickly in eating quickly will give you indigestion, but I'd argue that this shows the verbishness of the -ing forms in those constructions rather than any strange behaviour of adverbs. The nounishness of painting in (E) and eating here is also plain to see.


(F) Not every grammarian would assign examples such as the sense of painting in (E) above unequivocally to either the noun class or the verb class. Indeed, some seem to argue that such an assignment would be inherently inaccurate.

At http://books.google.co.uk/books?id=JjGb ... er&f=false, is found, in Rolf Kreyer's Introduction to English Syntax:


Gradience

While I used the term fuzziness to refer to indeterminacy within one category of description, the term gradience is used here to denote the merging of two categories in one another. These two categories are understood as the endpoints of a 'gradient', as Quirk et al make clear:

A gradient is a scale which relates two categories of description (for example two word classes) in terms of degrees of similarity and contrast. At the ends of the scale are items which belong clearly to one category or to [the] other; intermediate positions on the scale are taken by 'in-between' cases - items which fail, in different degrees, to satisfy the criteria for one or the other category. (Quirk et al 1985:90)

That is, we have a set of elements which clearly belong to category I [eg painting in (C) above where category I = nouns] and another set of elements which clearly belong to category II [eg painting in (D) above where category II = verbs]. In addition, there are also tokens which show features of both categories [I'd certainly include painting in (E) above - I'm not getting into the other 12 shades of meaning for painting Quirk mentions] so that it is impossible to assign them [ie each of these 'hybrid tokens'] to one category only. [Underlining mine]

If one cares to refer to the link, one will find that Kreyer uses seven tests to assess 'noun-ness' and / or 'verbness' for five chosen tokens of painting (including the participle extreme) and gets a differing set of results for each of the five (Table 7:4). He does not presume to draw a line defining the noun usage - verb usage divide (based on a token's passing tests 2, 3 and 6 say).

On the other hand, with 7 tests, each with a pass / fail possibility, there are 2 to the power 7 = 128 possible classifications, and Kreyer says wisely that it would be ridiculous to have 128 word classes in this area - or more, when different tests are thought of. I read his recommendation to be three designations - nouns, verbs, and in-betweens of varying flavours (for which the term verbal nouns might be acceptable to some. As Phil White points out, gerunds are a subset of these in-betweens, nearer the noun end.)

(By the way, I don't think mere is close enough to any accepted word class to warrant inclusion, joint or otherwise.)
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by dante » Thu Mar 01, 2012 3:25 pm

But I can't find the less rigorous definition you claim in any respectable - indeed, in any - dictionary.
Many reasons for that I guess. One of them is maybe that the use of adverbs as noun-modifiers is quite marginal and dictionary makers thought that simplifying things a bit might be useful. Another reason is probably that changes in the linguistic mainstream thought do not occur easily or all that fast. The Cambridge Grammar of the English Language was published in 2002. and it obviously takes time for people to adjust to the contemporary grammatical theory. A different definition of the category of adverbs can be found here for example: http://www.lel.ed.ac.uk/~gpullum/E1750124510000486.pdf

I agree with your analysis of these sentences Edwin of course. The "painting" in E) is best analyzed as verb, as I said in the previous post.
A) In Jeremy drives a fast car, fast is an adjective.

(B) In Jeremy was driving too fast, fast is an adverb.
The distinction between adjectives and adverbs is not always easy to spot as both are gradable and can be similar in form ( e.g -ly adverbs are common, but there are -ly adjectives, homonymous adjectives and adverbs , manner adverbs losing -ly in informal speech: I saw it clear, He drove slow..etc). The difference between adjectives and adverbs is more visible when we use those which are not homonymous, and have regular form:

Jeremy drives a slow/*slowly car. --> slow is an adjective, adjectives are noun modifiers

"Slowly" is not allowed, because it is an adverb and adverb do not easily modify nouns. Adverbs also do not appear in a predicative position:

* Jeremy's car is slowly.

Adjectives, on the other hand, can be used both predicatively and attributively:

Jeremy's car is slow.

Adverbs can modify adjectives and other adverbs, adjectives cannot.

Jeremy's car is terribly slow. / Jeremy drives terribly slowly.
ACCESS_POST_ACTIONS

Re: non-semantically-predicative adjectives

Post by Edwin F Ashworth » Thu Mar 01, 2012 4:08 pm

dante wrote:... changes in the linguistic mainstream thought do not occur easily or all that fast. The Cambridge Grammar of the English Language was published in 2002. and it obviously takes time for people to adjust to the contemporary grammatical theory. ...

I agree with your analysis of these sentences Edwin of course. The "painting" in E) is best analyzed as verb, as I said in the previous post....
Yes, I'm waiting for people to take on board Quirk's 1985 assertion and Kreyer's 2010 endorsement that

"it is impossible to assign them [ie each of these 'hybrid tokens'] to one category only" - ie it is wrong to label painting in (E) above as simply either noun or verb.
ACCESS_POST_ACTIONS

Post Reply