Modern English Grammar

My post today is uncharacteristically devoid of baseball content. It is about grammar, one of my many unremunerative interests. Specifically it is about modern English grammar. I don’t mean by this (except incidentally) the grammar of modern English. Rather, I mean modern grammar of English. Also, modern grammars of English.

What distinction am I making between grammar and grammars? Grammar is a field of study. The term is often applied broadly to the study of language, but I mean it in its narrow sense: the study of inflections and syntax. (If you just read “the study of of blah blah and blah blah,” don’t worry about it for now. Just go with it.) A grammar, on the other hand, is a model of a language’s inflections and syntax (or, if you prefer, of its “blah blah and blah blah.”) The whole point of a grammar is to model a language so well that, using just this model, you can tell if a word string is grammatical or not. There is more than one model out there, hence “grammars.”

Who gets to decide whether a word string is grammatical or not? That is a big question that leads to shouting matches. I’ll save the shouting for another day and stick here to uncontroversially grammatical or ungrammatical word strings. The important point here is that a grammar–a model–that can correctly identify the grammaticality of more word strings is an objectively better model than a grammar than only do this with fewer word strings.

Furthermore, (here is where I get annoying, rather than merely pedantic) unless you have either received formal training in linguistics or you are an eccentric hobbyist, any grammar, in the ‘model’ sense, that you have learned almost certainly sucks.

A closely related sense of the word “grammar” is of a book laying out a grammar in the ‘model’ sense. These tend to come in two varieties: pedagogic grammars and comprehensive grammars. The former typically are school textbooks, plus the occasional book aimed at the self-improvement crowd. The latter are large, academic tomes with massive tables of contents and indices, and with no expectation that any sane person would read the book cover to cover.

My eccentric habits include collecting old grammars, in the ‘book’ sense. These run from the late 18^th to the mid 20^th centuries. Most are pedagogic, but the prize of my collection is the mid-19th century Grammar of English Grammars by Goold Brown: a doorstop of a book with tiny print, with Brown delighting in correcting (or, often, “correcting”) other grammarians’ grammar.

An interesting feature of the earlier books is that they don’t merely repeat one another. They constitute a century-long discussion about how to model the English language. They often disagree even on pretty basic points. How many parts of speech does English have? The answers range from seven to ten. How many verb tenses? Somewhere from two to six.

This discussion was a healthy one. Then it stopped in the 1890s, at least for pedagogic grammars. Look at a grammar from this era and it will feel very familiar to anyone who has studied grammar ever since. This is what in the 20^th century came to be known as “traditional grammar.”

Modern comprehensive grammars are a different matter. The discussion continued there. We read of Noam Chomsky and the like, but the discussion was not restricted to such esoterica. A book like Randolph Quirk, et al., A Comprehensive Grammar of the English Language has much that is familiar, but then goes off with stuff like determiners and adjuncts and so forth.

How do we explain this divergence of pedagogic and comprehensive grammars? I don’t have a fully satisfactory answer, but I do know a bad one: that grammarians of the 1890s finally, after years of fiddling, got it right; and that later grammarians are a bunch of pointy-headed ivory-tower types who can’t leave well enough alone. Many self-proclaimed language mavens hold to this position. Any amendment to traditional grammar is viewed with suspicion as accommodations to the inexorable decline of the English language, and indeed of Western Civilization. I will write about the supposed decline another time. My goal here is show that this explanation is inadequate. I will use a non-controversial construction that traditional grammars are at a loss to explain, yet modern grammars handle easily.

Consider the following four sentences:

(1) John put his shirt on.(2) John put it on.

(3) John put on his shirt.

(4) *John put on it.

The asterisk at (4) is the conventional notation among linguists to denote an ungrammatical construction. And ungrammatical it is. No adult native English speaker would use it. I noticed my daughters (now five and seven) using this construction when they were younger, but this has dropped out about the same time they started getting the hang of irregular verbs.

The task at hand is to, using traditional grammar, explain why (4) is ungrammatical. For all that it sounds just awful, why this is so is not at all obvious.

Traditional grammar has a hard time analyzing any of those sentences. It wants to interpret “on” either as a preposition and find an object of the preposition, or as an adverb modifying the verb. We might analyze the preposition in (1) and (2) as having an implied object:

(1′) John put his shirt on [himself].(2′) John put it on [himself].

So far so good. Invoking implied words is a dangerous game. Go down this road and you end up with torrents of implied verbiage that explain everything and nothing. But in small, carefully administered doses, tacit bits can be useful. But this doesn’t really help us here:

(3′) ?John put on [himself] his shirt.The question mark is like the asterisk, but less decisive. (3′) is awkward at best, but perhaps not impossible. But this still doesn’t explain the last sentence:

(4′) *John put on [himself] it.If you can stomach (3′), then there is no good reason to reject (4′). If you can’t stomach either, then this leaves the problem that (3) is fine while (4) is not.

Then there is the interpretation of “on” as an adverb modifying “put.” This doesn’t help us. Indeed, it makes things worse. It does nothing to explain why (3) is grammatical while (4) is not, and it adds the problem of how do we interpret “put on” even apart from the object of the verb? Ordinarily a verb-adverb pair is transparent:

(5) John ran quickly.Go to any standard dictionary and look up “run” and “quickly” and put them together and you have all you need to understand what this sentence is saying. But go to that dictionary and look up “put” and “on” and try to work out from this what “put on” means. It simply won’t work.

The sad truth is that traditional grammar offers no interpretation for this construction. Indeed, when traditional grammar was developed, no one had noticed the peculiarity of the construction. A few decades later there are academic grammars showing an awareness that something was going on, but it wasn’t until the 1960s that linguists figured this out. And the thing is, it turns out that these constructions are all over the place. English is just lousy with them.

The construction is, in modern grammars, analyzed as a “phrasal verb.” A phrasal verb consists of two elements. The first is taken from the general pool of verbs. The second, called a “particle,” is taken from a shortish list of candidates. These are split between words you would ordinarily think of as prepositions (on, off, up, over, etc.) and words you would ordinarily think of as adverbs, specifically spatial adverbs (aside, away, apart, etc.).

We have seen the two peculiarities of phrasal verbs: If the phrasal verb is transitive (it need not be) and the object is a noun, said object can be placed either between the two parts of the phrasal verb or after them, as in(1) and (3). If, however, the object is a pronoun, it can only be placed between the two parts, as in (2). If it is placed after the two parts, as in (4), the result is ungrammatical. (Why is this? Heck if I know. Deep explanations would be lovely, but right now we are just describing the rules, not explaining why they are the way they are.)

The second property is that the meaning of the phrasal verb usually is not transparent. It can’t be derived by looking at the meanings of the two parts taken separately. Here are some examples to show this. In some cases you can work out a possible derivation, but just looking at dictionary definitions won’t do it:

Susan hung up.

Shut up!

We’re just hanging out.

Mike is burning up.

Jane is stressing out.

Fred is thinking it over.

Matilda pulled away.

You can ask around.

The batter struck out.

The pitcher struck out the batter.

You could compile an entire dictionary of these. In fact, this has been done several times. This is because it turns out that phrasal verbs are a bitch for non-native speakers to learn. If you are a native speaker with no connection to ESL/EFL, it is entirely likely that you have never heard of phrasal verbs. They aren’t something that native speakers have any problem with. We at an early age unconsciously internalize their grammar, and we learn the meanings of phrasal verbs individually, without trying to analyze the meanings of the individual parts. But to non-native speakers, phrasal verbs are a big deal. It is like there is an extra dictionary of words to learn. Strike that: It’s not “like” there is an extra dictionary. The extra dictionary is quite real. Search on “phrasal verb” on Amazon and you will find innumerable dictionaries and workbooks and implausible attempts to make learning these fun and easy. If English is your first language, these aren’t aimed at you.

And this is why it matters (Part I). Grammar can seem pretty ivory-tower: a fun puzzle game with no practical application. Sometimes it is. But sometimes it turns out to be very practical. This is one of those times, and in an area where traditional grammar just doesn’t cut it. Fortunately, linguists didn’t stop in 1890, even if people stopped noticing.

Part II will show how bad grammars can trip up native speakers as well. Stay tuned.