It's a bit too long for my liking but it is very coherent.

'Neither the semantic content of an item, nor its syntactic role are defining characteristics of its part of speech.' is very contentious. Dictionary.com has: 'part of speech noun, Grammar 1. any of the classes into which words in some languages, as Latin and English, have traditionally been divided on the basis of their meaning, form, or syntactic function [this latter tying in with distribution].' Has there been a pronunciation that we must all accept one model?

@EdwinAshworth Well, when I want to know about grammar, I always check with dictionary.com! Erm, not :) I'm not saying that parts of speech are not associated with different functions, they are. I'm not saying that distributional data isn't important, it is. I include in my description some of the prototypical syntactic functions of nouns. But there's no syntactic function apart from Predicator which is entirely carried out by one particular category of word. So carrying out any single syntactic function is not going to make an item part of a specific category by definition.

But unless you have an authority deciding on which particular criteria a word must satisfy (whether they be semantic, distributional, formal or size of font ...) to qualify as POS No 17, no final classifications can ever be agreed. And some linguists prioritise tests differently from how others do.

@EdwinAshworth Sure, people disagree about individual items and the nature of the categories, for example whether some categories are super-categories or sub-categories. There are some troublesome and difficult items. But this does not mean either that this enterprise is subjective or that we should do away with parts of speech, or that we should not take a principled approach. Modern linguistics is still in its infancy, so we've got a long way to go. There's the bathwater and there's the baby.

But 'Neither the semantic content of an item, nor its syntactic role are defining characteristics of its part of speech.' strongly suggests that there is/are other defining characteristics. The enterprise is shown to be subjective by the fact that there are at least three different schools re gradience / hybridity / you-wouldn't-like-it-if-I-said -toss-a-coin, as stated by Aarts.

@EdwinAshworth No, that's not what that means. The post starts of with the sentence There is no single syntactic or distributional property of nouns which is sufficient to guarantee the inclusion of an item in the word class. Just because Bas mentions those four, doesn't mean that he subscribes to them. They aren't equal. At the end of syntactic gradience, he basically says that there's intercategorial sharing of features and there are more and less prototypical members of categories, but that items move towards the boundaries between categories and then sometimes hop over or split ...

@EdwinAshworth But there is no doubt about it that Bas Aaarts believes in word categories. I know this for a fact. Quite apart from this though, if you read any of his other papers or for example his book, English syntax and argumentation, or his published work on "determiners", you'll see many examples of Bas, basically running through the types of test explained above. His preposition analysis of there in ESAA particularly interesting.

At the end of your monumental effort we seem to have reached the conclusion that there is neither a sufficient nor a necessary characteristic for an item to be classified as a noun - which rather supports @Edwin's contention: There is no consensus among linguists on how the parts of speech issue is best handled for the trickier cases. Hence, I'm waiting expectantly for your promised tests! PS. Did you mean 2005 as the H & P reference? The CGEL was published in 2002.

@Shoe I said there wasn't one single sufficient characteristic. I certainly did not say there were no necessary characteristics. [Who gives a stuff about the trickier cases in particular? It's not really what the question's actually about] The point is there are tests we can do, there are properties that nouns share. For a start they can all head NPs! :) SITEG, 2005; CaGEL, 2002. SITEG's the one you don't need a mortgage for! But, actually, I need to go through my post and add relevant refs for each one ...

@Shoe SITEG's quite useful, because it shows what H&P consider to be of central importance. You can also carry around with you without your arms falling off:)

I think it's the edge cases where the proving happens (in the sense of the exception proving the rule). No-one is going to dispute that "wug" is a noun if I point to a strange object on the table and say That's a "wug". Or that "plow" is a verb if I manipulate the object and say: I'm plowing the wug. But what about "plowing" in John's plowing the wug annoys his wife? (I presume SITEG is H & P's A Student's Introduction to English Grammar?)

@Shoe Sure, but it's not the pointing or the manipulating that's making a wug a noun. The point with the difficult cases is to be principled about it. H&P's take on this would be: a) plowing takes an object there, b) it cannot be modified by an adjective c) it must be modified by an adverb d) the subject, although it's genitive cannot be replaced by another definite determiner such as the e) the clause can be passivised f) if passivised it requires a passive auxiliary BE and a past participle like any other verb (and a two verb noun would really be stretching things!). VERB!!

@Shoe ...Conversely, plowing in John's plowing of the field annoys his wife a) there is no direct object b) there is just the kind of PP we expect to see with nouns c) it must be modified by an adjective d) it can't be modified by an adverb e) we can use the instead of John's f) the version with the thematic roles swapped over doesn't require a passive auxiliary or a past participle: the plowing of the wug by John annoyed. g) It falls in exactly with sentences like the massacre of the vogons by the daleks annoyed J's wife and The dalek's massacre of the vogons ... NOUN!

@Shoe But there's still room for disagreement on that : ) Yes, SITEG is that book! Don't know what the normal acronym for it is though.

Your answers are convincing! To summarize: In Mary dislikes John plowing the wug, plowing is a verb (with certain noun-like properties) and in Mary dislikes John's plowing of the wug, plowing is a noun (with certain verb-like properties).

PS. I just checked back on the famous wug-test. The pseudoverb should be "spow" not "plow"!

...And it should be 'Mary dislikes John's frumily spowing of the wug'.

@EdwinAshworth not frumy?

It's very unfortunate that you haven't yet learned to distinguish nouns from noun phrases. Much of what you say is about noun phrases and not about nouns at all.

@GregLee As I said, still being polished. Can you elucidate?

@Araucaria, begin with your sentence "Huddleston & Pullum, 2005, describe four syntactic functions that may be carried out by nouns. Within the clause they function freely as Subjects, Objects and Predicative Complements." But those are all functions of noun phrases, not nouns. (That's why we can infinitives as subjects, for instance.)

@GregLee A large number of linguists do not recognise clauses in a particular function as a kind of NP. Except in very special circumstances, they regard XP as a phrase where there is an X as the head of the phrase. As there is is no Noun in "to go", we wouldn't recognise that as an NP whether it's a subject or not. Now, you may be working from a different paradigm, but there is no principled reason to enforce a you can use it as a subject therefore it's an NP on other linguists.

@GregLee However, I agree that it's probably more helpful to talk about NP's there in any case - and when I get some time, I will! so thanks. It just is gonna take a bit of time. I'd just point out though that if your particular grammar regards an XP as a phrase with X as head, and it doesn't posit any null or empty categories, then an noun without any dependents is an NP. If you say which NP is the subject of people are generally friendly, the answer is quite reasonably: "the noun people" :-)

I would never, ever say such a thing, @Araucaria. The subject is the noun phrase "people". Nouns cannot be subjects.

@Araucaria, I have no opinion about null determiners.

@Araucaria Should have dropped the of.

@EdwinAshworth Ah, that makes sense then! :) Was a bit confused.

@Araucaria [comment 3 above] I said I'd listen to Aarts, not necessarily subscribe to all his views:-) The fact that there are at least three schools of interpretation for difficult cases is what I was majoring on. I'm not saying that word classes aren't useful and don't work well in explaining structures most of the time (though that can easily become circular reasoning). I'm saying that classifications are not standardised for known elements (contrast the periodic table). I'm saying that a hierarchy of diagnostic tests is not universally accepted. I'm saying people who claim they are ...

@Araucaria, this is still not right: "Huddleston & Pullum, 2005, describe four syntactic functions that may be carried out by phrases headed by nouns (including phrases consisting of just one word)." NPs are not "phrases headed by nouns". There are lots of NPs that do not have nouns as heads. And for fulfilling these four functions, it doesn't matter if a NP has no noun as head. You just have to give up the idea that there is some intrinsic connection between the category of the head and the category of the phrase with that head. Try to forget X-bar "theory".

@GregLee Ok, I thought I'd explained that in my comment a bit further above, but maybe I'm wrong. What are the "lots of" different types of non-noun headed NPs that are recognised by H&P? Because as far as I know there are only fused-function type constructions like "the poor" ... What are the others?

No, @Araucaria, I'm not saying that, and I see that what you've said is not false. It is, however, totally misleading, since it implies that there is a connection between the category of the head and whether a phrase can fulfill the four functions you mention. And this is false. Did you read my answer where I give examples of subject NPs with verbs as heads? Am I just beating my gums here?

@GregLee And more importantly, that wording was put in there so it deliberately couldn't cause affront to someone of your persuasion. It doesn't say that those functions can only be carried out by phrases headed by nouns. It says phrases fulfilling those functions can be headed by nouns. Are you saying that those functions can't be headed by phrases headed by nouns? Because your current criticism, even from your own (apparently narrow) paradigm seems unfair.

@Araucaria, what NPs without noun heads are there? (1) action nominalizations with Poss-ing complementizer. (2) Nominalizations with for-to complementizer ("To err is humsn"), (3) headless relative constructions ("What I said was ignored"), (4) indefinites with only a determiner ("Some like it hot"). I'll think of more, if I must.

@GregLee Well, you might be beating your gums against yourself. In H&P's grammar, as I've explicitly said in my post several times, many types of phrase can function as subject. I don't see how that could be any clearer than I've made it. Exactly how am I implying that subject/objects/PP's can't have other heads apart from nouns?

The question you're answering is about proving something is a noun. Why are you telling us about properties of NPs if you don't think this has something to do with identifying nouns? From the fact that a phrase is a NP, you cannot infer that its head is a noun. This is what I'm trying to tell you.

@GregLee In case it isn't clear, one of the major points of my post is that different types of phrase and word categories can fullfil many different types of function. Is that not clear from the post?

@GregLee (and Araucaria), in this paragraph in GregLee's answer post: To illustrate, here is a sort of minimal pair between noun and verb: "Eating lobster is forbidden." (This is the first half of the pair.) The subject noun phrase, "eating lobster", has "eating" as its head, so you might suspect that "eating" is a noun. Yet it isn't -- "eating" here is a verb, which you can tell from the fact that it takes the direct object "lobster". Nouns can't have direct objects in English. <== That seems to imply that in his grammar that a NP can have a verb as its head. Am I understanding that right?

@GregLee I went out of my way in the edited first paragraph of my post to describe some of the functions freely carried out by phrases headed by nouns, as opposed to saying by NPs - specially for people of your persuasion. Now, if I describe the properties of phrases headed by nouns - does that not say something about nouns????? And be fair when you answer

So you think Pullum disapproves of identifying subjects as NPs? I just reviewed your quote from Pullum, and I can't find that. What Pullum does do is criticize identifying syntactic function (e.g., subject) with word category (e.g., noun). And that is exactly my complaint about your essay. What is the relevance of identifying subjects to identifying nouns?

"Now, if I describe the properties of phrases headed by nouns - does that not say something about nouns????? And be fair when you answer" I'll try to be fair. No. If you mean that the head of a phrase which is headed by a noun is a noun, that tells us nothing, since it is tautological. And what else is there? Maybe from the fact that the head of a phrase is a noun one can infer that the phrase is a NP, but of course that doesn't help identify nouns (which was what you were doing, I thought).

@F.E., yes, that's right, with just two qualifications. One could define head in such a way that the head of a NP must be a noun. In that case, the subject NP in the eating lobsters example would, presumably, not have a head. The other qualification is that one could take the head of the subject in such a case as an abstract noun ACT which is is not pronounced (that has been proposed, I think by the Kiparsky's in an old article "Fact").

@GregLee Is there a grammar framework that you are basically relying on here? One which labels the expression "eating lobster" in the main clause "Eating lobster is forbidden" to be a noun phrase (NP) where the NP's head is the verb "Eating"? Perhaps the name of a linguistic textbook (and its author?) that uses that grammar framework?

@GregLee You seem unable to critially analyse your own thinking. You call anything that can function as a Subect or Object an NP, regardless of whether it has a noun in it or not. So your classifying NP according to it's function!!! I am not, I am describing what Huddleston and Pullum say. According to them, an NP is a phrase headed by a noun. An infinitive isn't headed by a noun. Neither is a clause. So these aren't NPs. At all. In any way for them. In any circumstances. Full stop.

@GregLee Actually, I should have said a phrase headed by a nominal, which is a phrase headed by a noun.

@Araucaria, I don't believe that Pullum would ever say "an NP is a phrase headed by a noun." He's too good a syntactician. Show me where he ever said that. (It is quite different to say "A phrase headed by a noun is an NP", which might be true, though I don't think it is.)

@GregLee (and Araucaria): There's this excerpt in Huddleston and Pullum's 2005 textbook, A Student's Introduction to English Grammar, page 13: "Expressions such as all things and some people are called noun phrases -- phrases with a noun as their head. The head of a phrase is, roughly, the most important element in the phrase, the one that defines what sort of phrase it is. The other elements are dependents." They used the word "defines" in that excerpt. (Aside: Of course, there's also the related stuff about fused heads, which lack the overt presence of a head noun.)

@GregLee (and Araucaria): I'd think it would be safe to say that the authors of the 2002 reference grammar, The Cambridge Grammar of the English Language, who happen to be Huddleston and Pullum et al., would consider that the prototypical noun phrase has a noun as its ultimate head. That's the impression that I've gotten. :)

Okay, @F.E., I lose, since Pullum apparently did say such a thing. Shame on him. He was probably thinking of N-bar. Anyhow, see my previous comment for a list of several NP types which do not have noun heads. For a careful account of NP structure that does distinguish NP from N-bar, see McCawley's wonderful text The Syntactic Phenomena of English.

@GregLee (and Araucaria): Since I've already looked this up, I might as well put this in. :) In the 2002 CGEL, page 326: "1. Distinctive properties of nouns and NPs … Summary of defining properties of NPs … [1.ii] NP structure: Except in what we refer to as the fused-head construction, ([Two of them] were broken; [Many] would disagree; It benefits [the rich]), NPs consist of a noun as head, alone or accompanied by one or more dependents." That excerpt is from the year 2008 reprinting of the reference grammar.

@GregLee I'm currently in the middle of going through that linguistic textbook of McCawley's, The Syntactic Phenomena of English, 2nd edition, the single volume, 800 pages, paperback printed in 1998. It seems to be a generative linguistics type of grammar, and the info was used to teach his students. There are lots of interesting info in it, but there's also errors and typos. (One of the most confusing parts of his framework is his labeling system, especially his use of "noun phrase (NP)", imo.)

Thank you, @F.E., for the additional reference. We have to deal with a veritable comedy of errors. It gets worse. The fundamental assumption of the popular theory dependency grammar is that syntactic phrases inherit their properties from their heads. That can't be right. That assumption also seems to be behind X-bar theory.

@F.E., why does McCawley's use of "NP" confuse you? (It doesn't confuse me.) Maybe you could formulate this as a question, so we admirers of McCawley's text wouldn't have to cram replies into these little comments.

@GregLee His use of "noun phrase (NP)" can be confusing, especially to a student, when they then go out and try discussing grammar with others. McCawley's NP--let me refer to it as "sNP" so as to keep it distinguishable from the "noun phrase (NP)" as used by grammars such as the one by 2002 CGEL--well, his sNP can be realized by a NP or a clause or perhaps even other categories (PP? AdjP?) when that sNP functions as subject or object of a clause. And that's real similar to traditional grammar's confusion with parts-of-speech.

@F.E., if other grammarians assume that subjects, verb objects, preposition objects must have noun heads, whatever you call the relevant phrase category, since that's not true, comparing those other grammarians' frameworks to McCawley's is comparing wrong to right. If the truth confuses you, I think you ought to just get used to it.

@GregLee if other grammarians assume that subjects, verb objects, preposition objects must have noun heads, whatever you call the relevant phrase category They don't! They just assume that clauses, VPs, AdjP, AdvP, PPs can be subjects. So in "a little more humble is how I'd like him to be" the subject is an Adjective Phrase, not a Noun Phrase. It's not that right-headed people think that everything that's a subject is an NP and must be headed by a noun - it's that SUBJECTS DON'T HAVE TO BE NP's! Neither do objects, neither do the complements of prepositions.

@Araucaria, so then what category is the subject in the example I discussed "Eating lobster is forbidden". If it's not a NP, what is it? If it's not a NP, why is it that it can be coordinated with another subject which does have a noun head? It's interesting that you can find subjects (and objects??) that are not NPs, but I think it just confuses the issue. NPs have several well-agreed on syntactic properties other than being subjects. They can be coordinated with other NPs, and they can be pronominalized using definite pronouns. (cont. ...)

... Whether the head of the NP is a noun or a verb makes not the slightest difference, so far as I know, to whether a phrase has these NP-like properties.

@GregLee If the truth confuses you, I think you ought to just get used to it. <== Say what? Where's that coming from? You realize that I can be doing other things than communicating with you, right? I'm not being paid to be nice to you or to teach you basic grammar.

@GregLee As H&P and many other grammarians note, when it comes to co-ordination ,phrasal category is less important than syntactic function for determining acceptability. Off the top of my head, without thinking to hard, I think H&P would say that in your sentence "Eating lobster is forbidden", 'eating lobster' is a clause, or VP, functioning as subject.

But, I'm not sure your co-ordination claim stands up to scrutiny. Could you give us an indisputably felicitous version of that sentence where "eating lobster" is coordinated with a noun headed NP?

Because for example "Children and eating lobster are forbidden" doesn't come over too well ...

@Araucaria, "Eating lobster and the consumption of children are forbidden." But I don't see anything wrong with your example, either. For the definite pronoun test: "Eating lobster is forbidden, but it's fun."