Are words classified (PoS) according to their use in a sentence, or does classification precede usage?

Question

This is a rather broad question, so I'd like to limit this to verbs, at least in this explication of the question.

Verbs take many forms and roles in sentences. Present participles can take the role of subject, object, adjective, and non-finite verbs in reduced clauses. Past participles can function as adjectives, passive verbs, and are used in the perfect tenses. The role of subject and object can be played by several types of words and word phrases (and even clauses); is it really necessary to reclassify them based on the roles they take?

Isn't classification independent of the role a word plays in a sentence; aren't all of the uses still properly classified as VERB in terms of their PoS?

I get so many questions from students about PoS, and I just want to tell them that these are all verbs, that role/usage/ and function do not necessarily correlate to parts of speech.

But in some cases they could. For example the words swim and look can be a noun or a verb.

I had a nice swim this morning.

She gave me a dirty look.

As a teacher, I would classify swim and look as nouns in these sentences.

But in other uses where the words take a distinctly verbal form in a subject or object I would still classify them as verbs.

Swimming is good exercise.

I like swimming.

I like to swim.

To swim to Kathmandu is a fool's errand.

I need to look for my dog.

I enjoy looking at old pictures.

Swimming, to swim, to look and looking are verb forms (PoS) that take the role of subject in or object in these sentences. At least for instructional purposes, they should not be classified as nouns just because they take the role of a thing in a sentence - they are still participle or infinitive forms of a VERB.

I understand that in linguistics this reclassification is called conversion or zero-derivation, but for students, requiring them to understand this concept seems onerous (as with many concepts in linguistics that are used in instructional or prescriptive grammars - too often in my opinion). I also get that even the noun swim is derivative (but one would need to understand etymology to know that, and this is well beyond the requirements of language learners.)

The human brain is a very capable language processor - we learn our first language without any comprehension of grammar concepts, and we can rely on this capability to learn second languages too, albeit with a bit of grammatical help but in a significantly reduced way than that employed for the purposes of linguistics. For learners, the primary thing they need to grasp is the nature of relations, which are primarily descriptive and adverbial, and the many ways they are produced and the different forms these relations take.

So, would I be committing some kind of linguistic or grammatical sin if I instruct my students to see these forms as their root PoS, and that these PoS can take multiple roles in sentences?

Answer 1

Short, snappy answer: parts of speech are a lie perpetuated by Big Syntax.

Longer, actually useful answer: parts of speech are an abstraction created by linguists to explain how syntax works. There's no universal definition of what it means to be a "noun": some ancient Roman grammarians didn't distinguish between nouns and adjectives, for example, but did distinguish between nouns and pronouns.

In English, for example, is it more useful to call jump and kick both "verbs" (putting them in the same category), or is it more useful to call jump an "intransitive verb" and kick a "transitive verb" (making two different categories)? Both approaches have their merits!

So, parts of speech are labelled in whatever way makes the theory more complete, concise, and elegant. You could argue convincingly that swim is underlyingly a verb which can be zero-derived to a noun in I had a swim this morning, because the meanings are connected in a clear way and this same type of zero-derivation can be used on other words. Or you could argue convincingly that the verb and the noun are separate, because you can't say *I had an eat this morning. Both are valid positions to take! So it comes down to what you think works best for your students.

But personally, I'd emphasize that parts of speech come down to how words are used in sentences. Swim can act as a noun, an intransitive verb, or a transitive verb, for example, and it's useful to be able to say which one is being used in a particular sentence: I had a swim versus I want to swim versus I swam fifty laps.

Answer 2

First, linguists generally prefer that language teachers tell the most linguistically credible lies that deal with the demands of their students, but also tell the student to sign up for a linguistics class if they want more details. Mis-stating "what linguists believe" is probably a bigger sin than saying "swim is a verb" (if we presuppose that there are sins).

There are two general trends in linguistics regarding PoS, the universal trend and the language-specific trend. In the language-specific trend, it is assumed that categorization is based on behavioral patterning. If butcher, swim and cow act one way and car, sandwich and lake act another, in terms of syntactic distribution, then they have some categorization difference. Then you just need a theory of how classes of words are categorized. In the universal trend, there is a fixed set of categories and some kind of metarules about why "lake" isn't a verb (I can't actually explain it because I don't practice that approach). This trend may rely on a set of semantic properties.

The lg-specific trend allows for a richer set of intersecting properties, so that a morpheme may be a "verb" at one level of analysis but an "adjective" at another and a "noun" on top of that. That could be an analysis of "living" as in "This will benefit the living". I think it is useful to get students to understand this "flexibility", but it's not obvious what is a good way to do that in a language class rather than a linguistics class. I think it depends very much on accepting the notion of "hierarchical structure": if you don't have that and you just have a string of words, the syntactic distribution of "living" is really baffling.