• ᴛʟᴅʀ: Is it ever possible for a sentence to have a word in it that is simultaneously more than one single part of speech in that sentence under the same parse and meaning?

  (For example, a few possible pairings from lexical categories commonly ascribed to English include noun+verb, verb+adjective, adjective+preposition, preposition+conjunction, conjunction+noun, and so on and so forth.)

My hunch is that the answer to my question is no, but I have heard the contrary proposition argued. So I would like to know definitively whether it can or that it cannot, preferably backed up with references and citations supporting whichever direction the answerer chooses to take on this one. If authorities differ, please explain the conflict.

BONUS: I’m especially looking into whether an “‑ing word” can ever be more than one of a noun, a verb, an adjective, or an adverb at the same time in the same sentence under the same parse and meaning. I don’t know, but I suspect that in this case

“There can be only one!”     ⸺Highlander

Background

English is notorious for having words that can act as more than one part of speech depending on how you use them. This famous pair relies on flies and like each being a different part of speech in each sentence:

• Time flies^{[verb, singular]}like^{[preposition]} an arrow.
• Fruit flies^{[noun, plural]}like^{[verb, plural]} a banana.

That shows how flies can be either a verb or a noun and how like can be either a preposition or a verb.

Similarly, the word still can be noun, a verb, an adjective, or an adverb.

• You can still^[adverb]still^[verb] a still^[adjective]still^[noun].

Or here, using inflections characteristic of each class:

• Quickly stilling^{[verb, ‑ing]} bubbling stills^{[noun, plural]}
  still^[adverb] leaves them stiller^{[adjective, comparative]} than you’d like.

In all those examples, one can assign only a single part of speech to each word as it is used in a given sentence. Outside of a sentence, or at least of a broader syntactic context, it is often impossible to make any such assignment, since the same word has the potential to be two or more different parts of speech.

The ‑ing inflections of verbs are notorious for this property of being able to be several different parts of speech. In my previous example sentence, I used one as a verb (stilling) but another as an adjective (bubbling). These ‑ing words can also easily serve as nouns, as in savings accounts and in swimming competitions.

And I’m perfectly fine with all that.

The problem is that I’ve also been told, quite vociferously in fact, that whenever an ‑ing word in a noun or an adjective, it is also AT THE SAME TIME a verb as well!

I can find no evidence to support this proposition, and I have looked. Hard.

All the putative examples of these I’ve been able to locate seem to err by misparsing syntactic constituents. This is the same class of error as we so often see in sentences like “Give it to whoever is coming” where they mistakenly write whomever thinking that that word is the object of a preposition whereas in fact the object of the preposition is all of whoever is coming, not just whoever alone.

Looking for Evidence

Here is my thinking:

• If it’s a noun, that means it must do noun things.
• If it’s a verb, that means it must do verb things.

So the easiest way I can think of to test whether something is one or the other or both is to apply various “does it do this-or-that noun thing?” and “does it this-or-that verb thing” tests, then tally the results to see whether there’s enough evidence for a clear answer either way.

In other words, to gather evidence I have taken a sentence alleged to have these ‑ing words that “are both nouns and verbs at the same time” and applied to them various syntactic tests. These are all simple syntactic tests that should either be true for verbs and false for nouns, or else the other way around. Then I have looked at the results of this evidence. I don’t mean to limit the tests applied, but I myself have used these sorts:

Noun Tests

• Nouns can be inflected for number.

• Nouns fit into a particular slot in the larger noun phrase, which includes such things as determiners and adjectives modifying that noun.

• Nouns phrases can be connected to other noun phrases with prepositions.

• Noun phrases accept the ’s clitic used to indicate possession. (This can look like another inflection when applied to just a noun alone.)

Verb tests

• Verbs fit in a larger verb phrase, which includes such things as adverbs.

• Verbs can accept complements, like direct objects if transitive.

• Verbs can be inflected in various ways, including for tense, aspect, number, and person.

The sentences we’re going to apply these tests to are these, which are reduced from this answer:

1. Running bulls is easy.

2. Running bulls are dangerous.

The conjecture to prove or disprove is that the word running is two or more parts of speech in either one of those two sentences alone. I already know that it is a different part of speech in sentence 1 as it is in sentence 2: it’s a verb in the first one and an adjective in the second one.

But I really think that that’s all it is. It isn’t also a noun in either of them. Just because running bulls is the subject of sentence 1 doesn’t mean that running is a noun; it can’t be a noun or it wouldn’t be able to take a direct object like is happening there. Only transitive verbs take direct objects; nouns never do.

Possible Origin of Confusion?

I think this confusion may stem from people being told that a complete sentence “must have a noun and a verb”, which isn’t a “real rule” in English. Rather, a sentence must have a subject and its predicate, and lots of things can be subjects other than just plain nouns alone, including clauses like “Running^{[verb, ‑ing]} bulls is hard” or “What they told^{[verb, past]} you is wrong”, where is hard and is wrong are the respective predicates and everything preceding those is each sentence’s respective subject as a clause. Neither running bulls nor what they told you is somehow a noun, since those are clauses. But they’re still subjects nonetheless, and we don’t need to pretend they’re nouns to make them be a subject. That’s the main argument for saying that running is somehow a noun there, and I can’t see it.

See my reasoning here for the application of noun tests and verb tests to these two sentences. I cannot come up with any way to make the running verb from sentence 1 also be a noun or an adjective in sentence 1, nor to make the running adjective from sentence 2 also be a noun or a verb in sentence 2.

It has been argued that you cannot use syntactic tests to determine the part of speech of a word, and that broader historic traditions of assigning these things allows them to be simultaneously multi-parted even when no syntactic test can find any such evidence. I do not pretend to understand those arguments, and I am not making them. I simply know no other way to do this than to apply syntactic tests.

Motivation for the Question

Many answers on this site allege that ‑ing words used in non-finite verb clauses are not only verbs alone as their clause usage proves already but also “actually” nouns when said clauses are used substantively and also adjectives or adverbs when those clauses are used as adjuncts modifying something else.

I believe they mistake the verb clause as a syntactic constituent replaceable by nouns or adjectives/adverbs for those respective lexical categories. I think only the clauses can function as substantives or modifiers, but each word’s lexical class is still that of a verb alone.

Here are some examples that seems to suggest otherwise:

• What's the difference between a gerund and a participle?
• Is “programming” not a noun?
• “What led to you doing this thing” grammar?
• Participles? Present participles? Are they nouns too?
• Does Gerund act as noun always?
• Referring to a gerund with a pronoun
• Ving (gerund) as a Noun
• Is it a gerund or a verb?
• Define Gerund, Noun and Verb

As well as:

• Cambridge Dictionary
• Merriam-Webster
• Collins Dictionary
• Oxford Dictionaries Online

So we end up telling people that things belong to two distinct and oppositional lexical categories at the same time. This is extremely confusing, so I’d like to find evidence that it does or does not ever occur, or even can.

The Question, Again

So I again ask: is it ever possible for a sentence to have a word that’s simultaneously more than one single part of speech in that sentence (under the same parse and meaning)?

I’m particularly looking for whether an ‑ing word can ever be more than one of a noun, a verb, an adjective, or an adverb at the same time in the same sentence under the same parse and meaning.

ᴛʟ;ᴅʀ
Is it ever possible for a sentence to have a word in it
that is simultaneously more than one single part of speech
in that sentence, under the same parse and meaning?

So, if a grammatical English sentence contains a word A, can A be more than one POS?

Parts of speech are grammatical terms and have varying meanings for different grammarians.
Let's rule out quantum superposition of POS, so no Schrödinger's Gerund that's noun and verb.

There certainly are sentences where it's impossible to tell which of several possible categories a word falls into, like the first sentence below, where exhausted can be either a predicate adjective, as in the second sentence, or part of a passive construction, like the third one.

• I was exhausted.
• I was exhausted and the bed was soft; we suited each other well.
• I was exhausted by the irritable conversation and left early.

But that's not "in the same sentence". In the first sentence, there's just no way to know what the speaker intends about POS; it could be either one. And there's no way to know if one speaker might feel it was an adjective, but another speaker might think it was a participle. Or the same speaker might do both, to the same sentence, on different occasions.

So, the key word in the question is Simultaneously. And the answer to the question is No.

If anything in a parse changes from one POS to another, that makes it a different parse.
Thus, if A has two different POSs, they will occur in two different parses of the sentence.
And therefore the sentence can't simultaneously have two POSs in the same parse.

It is of course very easy to find sentences that have two interpretations; this is one way to make jokes, and certainly such ambiguity is common. But each interpretation represents a different parse. That's one of the purposes of parsing, in fact -- to distinguish ambiguous sentences and make their differences distinct. But that doesn't mean they're simultaneous, in any sense.

Yes

There are constructions called zeugmas (after Greek ζεῦγμα, 'a yoking') where a word or phrase is intentionally made to apply to two or more others in a sentence despite functioning differently for each. The horde of words in English that can function as various parts of speech mean that zeugmas can be created where a single word simultaneously acts as two separate parts of speech.

@DavePhD had a particularly nice one in his reply

She is and runs fast.^{[adj.; adv.]}

that could also be reworked into the still stranger

She is but stands fast.^{[adj.; adv.]}

The doubled, incompatible, but grammatically functional senses of the word—usually requiring a double-take to figure out—are the appeal of the rhetorical device. It's simple enough to create such expressions using -ing words:

He is and likes running.^{[v.; n.]}

They are and enjoy knitting.^{[v.; n.]}

although anything creative is likely to be rather awkward:

He's being and is a hit.^{[v.; n.]}

He was, has, and eats standing.^{[v.; n.; adv.]}

and No

If you consider the doubled parts of speech in those zeugmas irrelevant because they're different 'meanings'... well, as Mr Yeats pointed out in his post, nouns and verbs are definitionally different "meanings". A noun used as a verb may intend act as (a) ~ ("man/crew/staff"), act like (a) ~ ("cat/horse/slut/fool around"), make (a) ~ ("sound/peep/bridge"), use (a) ~ ("hammer/canvass/google"), provide (a) ~ ("water/house/board"), &c. A verb used as a noun may intend an instance of ~ ("a run/try/fuck") or several instances of ~ ("a beat/pulse/record"), something that ~ ("a shoot/hit/fuck"), &c. They'll all be slightly different as a matter of course.

If you exclude any such form of doubled meaning, the only form of word that could work as a verb and noun simultaneously would be verbal forms.

It [sic] isn’t also a noun in either of them. Just because running bulls is the subject of sentence 1 doesn’t mean that running is a noun; it can’t be a noun or it wouldn’t be able to take a direct object like is happening there. Only transitive verbs take direct objects; nouns never do.

If you peremptorily define your opponents' contention out of existence, of course, it is impossible for them, me, or anyone else to continue to argue their case.

and Maybe

You can rest there and be content that everyone who disagrees with you is making some form of a category error. Ambiguity between adj. and adv. senses of participles could be dealt with by similar definitional adjustments, like Aml's "adjectives-being-used-adverbially" in his reply to DavidPhD's post.

I'd think it should be unsettling that you're just shifting definitions out from under your opponents. It certainly seems odd you'd put up a 500 point bounty just to hear how right you are. The actual definition of noun

A word used as the name or designation of a person, place, or thing; the class or category of such words...

and verb

That part of speech by which an assertion is made, or which serves to connect a subject with a predicate.

don't remotely support your contention that there is a magical bright line between the "noun" in "I like eating" and the "verb-in-a-noun-phrase" in "I like eating bacon". In fact, they don't support the idea that either "eating" is a verb at all, let alone universally.

The essential nature of the gerund in Latin—and, by extension for many grammarians, its -ing equivalents in English—is that it is a noun which retains the verb-like regimen (taking adverbs, objects, &c.) of its origin. RaceYouAnytime says as much both with his examples and his research describing the way generic English learners intuitively approach this structure. Aml claims to agree with you completely, but still understood your "running" as a noun with an object. Both historically (Latin and Old English's grammatical suffixes make verbals' noun status much more emphatic) and popularly, that's exactly how gerunds are approached.

You can absolutely define away your opponents' points. It doesn't really make them 'wrong' in any meaningful sense, though. You've just redefined 'noun' and 'verb' into something they and most other English speakers don't intend. You shouldn't expect such arguments to be convincing to them or anyone who doesn't share your specialized definition of these terms.

It's not that I think you're wrong on the basis of modern linguistic processing. It's just that no one is 'right' here. You're using a separate definition of noun and verb from theirs and you're just talking past each other. Buy 'em a beer and change the subject. They can suffer in the Hell of Poor Grammatical Understanding in the afterlife of your choosing.

For myself, as a side note, looking at both sides from the outside, I don't see the analytical benefit in calling "eating" a verb-in-a-one-word-noun-phrase instead of just treating it as a noun and saying verbal nouns can take objects. That approach seems less complicated in analytic languages and simply correct in synthetic ones.

How about "splashing is forbidden?" Splashing seems to function as either a noun or a verb.

It could be modified by an adverb to fit the verb test:

Loudly splashing is forbidden.

Or it could be modified with an adjective to fit the noun test:

That loud splashing is forbidden.

A psycholinguistic perspective

Ultimately, to address the meat of the question, I don't believe that a word can truly function as multiple parts of speech with the same meaning and parse. It's important to consider how a reader perceives the parts of the sentence.

Studies of language acquisition in young children indicate that gerunds, or words ending with -ing generally, tend to be comprehended as two separate lexical units. An example lies in a study of 15-month-olds by Toben Mintz:

Taken together, the experiments in this study demonstrate that English-learning 15-month-olds represent the suffix -ing as a discrete unit. Thus, although previous experiments failed to find evidence that 15-month-olds have acquired morphosyntactic dependencies involving -ing (Santelmann and Jusczyk, 1998), infants may nevertheless be in the process of learning these dependencies at this age. Specifically, having a discrete representation of an affix allows infants to notice dependencies between that affix and other forms.

...

The experiments reported here demonstrate that English-learning 15-month-olds represent -ing as a distinct form. When processing novel words that end in -ing, they segment the suffix from the stem. This allows them to notice morphosyntactic and morphosemantic patterns that involve that form, and that will form a part of their acquired grammatical knowledge

• Mintz, Toben H. "The segmentation of sub-lexical morphemes in English-learning 15-month-olds." Frontiers in Psychology

This conclusion is supported to varying degrees by other studies of language acquisition. See also:

• Pliatsikas, Christos; Wheeldon, Linda; Lahiri, Aditi; Hansen, Peter C. "Processing of zero-derived words in English: An fMRI investigation" Neuropsychologia 53 (2014) 47–53

By this logic, one could argue that a gerund functions as two parts of speech by virtue of being processed first as a verb and then as a noun. In the sentence, "I enjoy running," a young child processes "run" first, with "-ing" acting as a modifying unit. In this sense, it's not unlike verbs in infinitive form. "To run is fun." Here, we're comfortable calling "run" a verb and "to run" a noun clause. "Run" can still be treated as a verb -- it is free to operate on an object or be modified by an adverb. "To run races quickly is fun."

However, the case of infinitives also illustrates an important distinction between nouns and noun clauses. Even though "To run" is a noun clause, it doesn't have the same flexibility as a typical noun. It cannot be modified by an adjective or be made countable.

Given that early processing of gerunds tends to treat them as two separate lexical units, there seems to be an implication that the parsing occurs twice by young children as they learn the function of the phrase. This notion is confirmed in another notable study approaching the question from a psycholinguistic angle:

The abstract can be found here:

• "Noun or verb? Adult readers’ sensitivity to spelling cues to grammatical category in word endings." Kemp, N., Nilsson, J. & Arciuli, J. Read Writ (2009) 22: 661.

Among other things, Kemp et al. discuss the level of computation required to process phrases with ambiguous function as measured by response time, a common method in psycholinguistic research. Some passages seem highly relevant to how we mentally parse these parts of speech:

Laboratory studies confirm that having to resolve a word’s grammatical category can delay reading time, in such phrases as She saw her duck and chickens vs. She saw her duck and stumble (e.g., Boland, 1997).

The study and others cited in it also conclude that orthographic suffixes, like -ing, provide clues as to the part of speech of a word, even when used as suffixes in "fake words" where research participants were asked to identify the part of speech of made up words with differing orthographic endings. The use of segmented lexical clues to determine a part of speech parallels the earlier mentioned research related to how children process words like "splashing." It is parsed in two parts, first as a verb, but its ultimate function similar to a noun clause doesn't imply that it is in itself two parts of speech at once.

Since the same words can be adjectives and adverbs, sentences could be constructed where the word modifies a noun and a verb:

She is and runs fast.

While others cheat, my children are and play fair.

In Yellowstone National Park, bison roam, and truly are, free.