Somebody happy

Q: Why do we put the adjective after words like somebody, anyone, and nothing?

A: In a modern analysis, we can consider these compound determinatives, where the first part is the determinative (i.e., some, any, and no) and the second part is the noun. Normally, any modifying adjective would come between these, but because they have become fused, there's nowhere for the modifier to go, so it ends up after the noun (CGEL, p. 423), a location where English occasionally allows adjective phrases (e.g., the poet laureate or the person responsible).

English has a number of compound determinatives that follow this pattern. The most common today are:

	some	every	no	any
-body	somebody	everybody	nobody	anybody
-one	someone	everyone	no one	anyone
-place	someplace			anyplace
-thing	something	everything	nothing	anything
-time	sometime			anytime
-where	somewhere	everywhere	nowhere	anywhere

There are a few other words that appear to be similar, but are not:

	some	every	no	any
-day	someday	everyday
-how	somehow			anyhow
-what	somewhat

Of these, somehow, anyhow, someday, and somewhat, are adverbs, and everyday is an adjective.

Traditional grammars, too, have taken note of this phenomenon, but they don't really have an explanation or develop the various related constructions. Consider, Sweet's treatment (p. 11):

"1803. Adjectives modifying indefinite noun-pronouns follow them, as in something bad, anything good [compared with any good thing], nothing remarkable; similarly in anything else. If the adjective precedes, they become pure nouns: in her manner there was an indefinite something."

Bonus points
The inclusion of the -where words is problematic since where is not a noun. Moreover, the other compound determinatives can be split up (e.g., every little thing), but the -where words won't suffer this kind of treatment. And yet the -where words clearly function where only NPs will go (e.g., Do you have anywhere that we could stay.) (CGEL, p. 423)

As Sweet notes, else is also able to follow the compound determinatives (more does the same). In this respect, the compound determinatives are similar to simple determinatives such as little and much, which can also be followed by else and more.

Modifiers that usually follow nouns, such as prepositional phrases and relative clauses (e.g., everything that I saw, something from Canada), are similarly possible, but they have to follow any adjective. (e.g., something nice from Canada not *something from Canada nice). (CGEL, p. 423)

Other modifiers though, are less predictable in this position. Here are some examples of nouns and verbs as modifiers:

There was some school event.	*There was something school.	N
I'll have any fruit juice.	*I'll have anything fruit.	N
I've got some eating disorder.	*I've got something eating.	V
I'll take any working vehicle.	I'll take anything working.	V

Present perfect vs past simple

Q: Ian G. wrote:

"Last night a student asked me what the difference was between the following two questions, and which one was best:

1. Have you brushed your teeth? [present perfect]
2. Did you brush your teeth? [past simple] 

The context is, for example, a mother asking their child before bedtime. Both versions look acceptable to me and I was at a bit of a loss as to explain the difference."

A: There are multiple differences, but Ian's right: both questions are perfectly acceptable.

In most situations present perfect and simple past are interchangeable, but once one has been chosen, certain constraints apply.

The choice of present perfect or simple past expresses the way the speaker wants to portray the situation. It's something like the difference between the ball hit him and  he was hit by the ball or between it's under the book and the book's sitting on it. In basic cases like 1 and 2, both options are available; the speaker can choose freely between them depending on how they view the situation.

In 1, the asker is choosing to locate the question in a range of time that starts in the past and comes up to the present as in figure 1.

Figure 1. The present perfect showing the situation located at an indefinite time before the present.

In 2, the asker is choosing to conceive of the brushing (or not) as an act located at a point in past time isolated from the present as in figure 2.

Figure 2. The simple past showing the situation located in the past.

The grammatical consequences for this are not obvious in the above questions, but consider which adverbs might work in conjunction with the sentences. For example, yet and already would work well with 1 but would be unacceptable to many native speakers with 2. Conversely, at 8:00 works with 2 but is unacceptable for 1 in the situation described. In other words, although the speaker may choose to view the situation in one way or another, once the decision is made, then a number of other options close: you cannot, for example, choose both at 8:00 and the present perfect.

Bonus frequency data
There are also interesting dialectal and genre differences. It's sometimes thought that the present perfect is more common in the UK and less common in North America, but things aren't all that clear. It's very difficult to find comparable data about the two varieties of English, especially when it comes to spoken English. Perhaps the best comparison would be between the British National Corpus and the Corpus of Current American English, keeping in mind that the BNC is from the 1980s to 1993 and the COCA is from 1990 to 2011. I've searched for have + past participle¹ in the two subsections of these corpora that are most similar in composition: fiction and newspapers. The results are presented in table 1.

Table 1. Comparison of the frequency of have + past participle across dialects and genres per million words.

	Fiction	Newspaper
COCA	922	1,344
BNC	1,381	1,814

There is, however, some evidence that use of the present perfect is declining over time, as table 2 shows. Thus, even these comparisons may not be valid since the BNC ends just when the COCA is getting started. The first row shows the years, the second shows the number of hits, but doesn't take into account the size of each section. The third row is the important one as it shows the hits per million words.

Table 2. Frequency of have + past participle in the COCA from 1990 to 2011.

The discussion provides a quick overview of the choice between the simple past and the present perfect as used in examples 1 and 2. It suggests that teachers need to be very careful about using tests and exercises designed to practice this distinction. In many cases either option would be acceptable, despite what the answer key may say.

---

[1] This is only a rough measure since it will also catch instances of the infinitive perfect (e.g., they seem to have made no progress), and it ignores any cases of the present perfect where something intervenes between have and the past participle (e.g., I have never seen it), when have is contracted (e.g., I've seen it), and a variety of other cases. It should, however, produce roughly the same rate of error across genres and frequencies, so should give a good idea of the relative differences between UK and US dialects and fiction and newspaper genres.

Even you surrender

Q: Why do my students constantly make this mistake:

1. Even you surrender, you have to work on the modalities and the timetables and what have you, so negotiations will take place.

A: This is common among my students as well, and my impression is that it seems to occur across a range of linguistic groups, suggesting it's some kind of misanalysis rather than a transfer error from the first language. It's even noted in the Longman Dictionary of Contemporary English (see the grammar box in this entry.)

The word even, as used here, is a focusing adverb, similar to only, also, and too. In speech, focusing adverbs are commonly marked by stress (CGEL p. 589), and even commonly comes directly before words such as though, if, when, and as. These prepositions¹ are rarely stressed. This talk by Barry Schwartz provides a good example. The waveform graph shows the string even if it, with each peak corresponding to a vowel.

Figure 1. Wave form of even if with audio.

You can see how the first syllable of even is drawn out and if gets short shrift, such that students might fail to perceive it. Also note that this kind of preposition follows even about one fifth of the time (according to a COCA search) and that the two most common words following even are though and if, and you can begin to see how students might misanalyze even as the head of the expression rather then merely the modifier. In other words, most students have unconsciously come to understand even as a preposition taking clausal complements (or, in traditional terms, a subordinating conjunction.)

If your students are adults who have studied English for some years, this error will likely be deeply ingrained, difficult for them to notice, and even more difficult to overcome. The one bright spot is that it should be easy to demonstrate to them. I'd begin with a dictation, perhaps:

I study hard / and I'm a good learner / so even if I make a mistake / I don't give up. 

Read it once all the way through, and then repeat it, pausing where indicated by the slash. You should read it in a natural speaking style or even a little quickly, not slowly and carefully. Allow students to consult with each other before asking how many, initially at least, omitted if. Point out that even is an adverb functioning as a modifier of if I make a mistake; it can't join a clause or take a complement of any kind.

A trip to a good learner dictionary's entry for even (or here for an open source option) would be a good follow up. It's also worth considering the differences in grammar and stress between, even if I make a mistake, I don't give up, where even is modifying the phrase if I make a mistake, and even I /the teachers make mistakes, where even is modifying the NPs² I and the teacher.

Some added detail
If you like syntax trees or sentence diagrams, you might find the following interesting, but don't just drop them on your students. Notice that in figures 2 and 4, each node except the top clause is labeled with both a function and a category.

Figure 2. Modern syntax tree of even if I make mistakes, I don't give up.

Figure 3. Traditional Reed-Kellogg diagram of even if I make a mistake, I don't give up.

Figure 4. Modern syntax tree of even I make mistakes.

Figure 5. Traditional RK diagram of even I make mistakes.

Notice that traditional grammar doesn't allow for adverbs to modify NPs like I, so in figure 5, even in even I make mistakes must be modifying the verb as adverbs can't modify pronouns and traditional grammar has no concept of noun phrase (NP).

---

[1] Traditionally, these are considered variously as conjunctive adverbs:


or simply as conjunctions as in the Oxford English Dictionary. But a modern approach takes them to be prepositions, albeit prepositions taking clausal complements. One reason for this change is that typical preposition phrases are commonly modified by even (e.g., even in summer she wore wool), but the central adverbs ending in -ly usually reject modification by even (e.g., *he ran even quickly).

[2] As we saw last time, adverbs rarely modify individual nouns, but it's not so unusual for focus adverbs like even to modify entire noun phrases (e.g., only the teachers). Modern grammar takes pronouns such as I to be both the head and the single member of a one-word NP.

Almost vs most

Q: What's the difference between these sentences?

1. Most people like to live in the city.
2. Most of the people like to live in the city.
3. *Almost people like to live in the city. (The star indicates an ungrammatical example).

A: First off, [1] and [2] are grammatical, while [3] isn't. The reason for this is the parts of speech (or lexical categories in CGEL terminology) that most and almost belong to.

Traditionally speaking, the simple subject in [1] is people modified by the adjective most. And the complete subject is most people. In a modern analysis, the subject is the noun phrase (NP) with the noun people as the head. Most is a determinative (many learner dictionaries use the term determiner, see here for a list of English determinatives) in the determiner function (see the endnote here). Whichever way you look at it, this is just what you would expect, and no doubt what most learners are aiming at when they say things like 3. The meaning here is that we have the set of all people—all people in the world or all that we can imagine—and most—a majority of them—like to live in cities. This is depicted in the following pie chart. (It appears that for some people, most means a large majority or nearly all).

Figure 1. Pie chart illustrating the meaning of most people.

Now consider [2]. The traditional analysis would have most as a pronoun here modified by of the people¹ which is different from the traditional analysis of 1 where most was an adjective. In CGEL, most is a determinative in both sentences, and in 2, most of the people is called a partitive construction.

Traditional or modern, though, we still have most, meaning 'a majority.' The difference is that in 2 our set has changed; it is no longer a majority of everyone. Now we have the people; in other words, we're talking about a specific set of people. It may be the people in Ontario, or the people you know, or any other set that is understood by the speaker and the listener. Thus, in Figure 2, the entire circle is all people. The light-orange and yellow sections are the people I know.

Figure 2. Pie chart illustrating the meaning of most of the people.

So, while most people means 'the majority of the people in the world' (or all those we can imagine), most of the people means the majority of a specific group.

Finally, we come to the mistake: almost people. What the learners mean here is 'most people' or perhaps 'almost all people,' and indeed, had they been speaking a few hundred years ago, almost people would have been understood in exactly that way. Ascham (1570), for example, wrote, thies giuers were almost Northenmen (i.e., 'These givers were almost all Northerners.') Today, though, almost is a degree adverb (like more, just, & only) meaning 'nearly, but not quite.' As a result, if you say that the people who like living in the city are "almost people," you mean that they are nearly but not quite people; perhaps they are humanoid robots, Neanderthals, or avatars. This kind of vivid, humorous explanation is likely to stick in students' minds, but you shouldn't expect it to cure them of the almost affliction.

Bonus material
One interesting side note to this is the question of what may be modified by adverbs. Traditional grammars tend to explicitly limit adverbs to modfiying verbs, adjectives, or other adverbs:
Implicitly, though, they allow that certain adverbs can modify noun phrases (NPs). For example, in Teach Yourself English Grammar, Phythian discusses the correct placement of adverbs. He writes that in the sentence only the bridegroom hired a suit for the wedding, "the sense is that only the bridegroom (and no-one else) hired a suit" (p, 77). Clearly, here, only is modifying the NP the bridegroom. The adverb almost is unusual in that it does occasionally modify nouns as in 4 and 5.

4. It was as he began drawing a third and realizing the strange feeling of the almost stillness about him that he heard a new kind of sound at the side of the garden nearest his hiding-place.
5. ^?I am tired from the almost week of walking under a pack.

Since this isn't that common, it's unlikely that learners have noticed it, but it is one more possible source of confusion surrounding this construction.

It is, however, quite common for almost to modify entire NPs as in 6:

6. He made a choked sound, almost a sob.

---

[1] I couldn't actually locate a traditional grammar that discussed this particular construction. If you know of one, let me know.

Who and `do support'

Q: Masoumeh A. asks,

"I have some grammar questions which look very simple but somehow they make me confused… someone told me that we can't use do, does, did after who, for example:

1.
a) Mary helped me to come to Canada.
b) Who helped you to come to Canada?
c) Who did help you to come to Canada?"

A: Unfortunately, it seems that Masoumeh's informant has only explained part of the rule. A more complete version is:

An auxiliary verb is needed in open questions, unless the question is about the subject. When the question is not about the subject and there is no auxiliary verb, then we add do.

The details and the missing piece follow:

What we're looking at here is something called do support or do insertion. The fact is that in most open questions (questions using a question word: who, what, how, etc.) we need some kind of auxiliary verb, and when there isn't one, we recruit do to do the job for us. Consider these examples:

2.
a) Where would you get a forty-foot sword sharpened, anyway?
b) *Where ___ you get a forty-foot sword sharpened, anyway?
c) Where do you get a forty-foot sword sharpened, anyway?

In 2a, the auxiliary verb is would. In 2b, there is no auxiliary verb, and the sentence ends up being ungrammatical, so we insert do, resulting in the grammatical 2c version. This is always true with open questions formed with when, where, why, and how, which are not pronouns and therefore cannot function as subjects: these sentences always need an auxiliary verb. But it is only sometimes true with who, what, and which, and it depends on whether or not the question word functions as the subject of the sentence (or part of the subject, see example 5 below).

If the question word replaces an NP (or part thereof) that is not a subject, then usually the NP becomes who, what, or which. Next, if there's no auxiliary verb, do is added in front of the main verb. Third, the subject and the auxiliary verb switch places, and fourth, the question word is fronted as in examples 3 and 4.

3.
a) They like pie.
b) They like what? (change the noun phrase (NP) to a question word)
c) They do like what? (there's no auxiliary verb, so add do before the main verb)
d) do they like what? (switch the subject and the auxiliary)
e) What do they like? (front the question word)

4.

a) They like her.

b) They like who? (change the noun phrase (NP) to a question word)

c) They do like who? (there's no auxiliary verb, so add do before the main verb)

d) do they like who? (switch the subject and the auxiliary)

e) Who do they like? (front the question word)

Sometimes, though we not just asking what in general (e.g., what was it?). We know the noun, but we don't know its specific referent: it's pie, but is it this pie, or that? Here, we can use whose or the determinatives what and which by themselves (e.g., which?) or together with the noun (e.g., which pie?). Either way, this NP with the question word moves to the front as in 5c and do is added.

5.

a) They like this pie.

b) They like what/which/whose pie? (change the determiner¹ to a question word)

c) They do like what/which/whose pie? (there's no auxiliary verb, so add do before the main verb)

d) do they like what/which/whose pie? (switch the subject and the auxiliary)

e) What/which/whose pie do they like? (front the question word)

This all applies when the question word does not function as the subject or its determiner. Asking about subjects is much easier, since the subject is already at the front of the clause. In this case, there is only one step: change the subject (or its specifier) to the question word:

6.
a) She likes pie.
b) Who likes pie? (change the NP to a question word)

7.
a) This works.
b) What works? (change the NP to a question word)

8.
a) This machine works.
b) What/which machine works? (change the specifier to a question word)

Since there is no subject-auxiliary inversion, there's no requirement for do insertion. So, the rule isn't that you can't use do with who, but rather that:

when the question word is the subject or its determiner, do insertion isn't needed.

That's not the whole story though. Notice that, unlike 2b, 1c isn't starred. In other words, 1c is grammatical. Even though we don't need do insertion, do can be used when there is no auxiliary verb to add emphasis, as in the following conversation:

A: Somebody helped me come to Canada.
B: Was it a lawyer?
A: No, I didn't have a lawyer.
B: Was it your parents?
A: No, they tried to stop me.
B: Then who did help you come to Canada?

In sum, then, an auxiliary verb is needed in open questions, unless the question word functions as the subject or its determiner. When it is not, and when there is no auxiliary verb, then we add do. We can also use do insertion for emphasis when the sentence has no auxiliary verb.

---

[1] Determiner is a function in noun phrases. It is usually performed by words belonging to the category  of determinatives such as this, many, a, and which. Possessive NPs such as Juan's, the dog's, my, or whose also commonly function as specifiers. Traditional grammar would simply call this an adjective functioning as a modifier.

What are linking verbs?

Q: What the heck is a linking verb, and why would I teach it?

A: When learning any new verb, it's important for learners to know its complementation pattern, or what can follow it. Some verbs are intransitive, some transitive. Some verbs take infinitival complements (e.g., I want to go,) and some don't (e.g., *I enjoy to go; that star indicates something ungrammatical). Some allow content clauses or "noun clauses" (e.g., I hope that it works), and on it goes. Linking verbs present a number of other combinations that students need to be aware of.

A linking verb is a verb that takes a predicative complement (remember: complements complete & compliments praise) but not an object. Typically the complement is in the form of an adjective phrases (e.g., it looks good), noun phrases (e.g., it became a problem), or preposition phrases (e.g., it is against the law). In other words, it is a verb that can be followed by a phrase and which attributes the properties represented by that phase (identity, size, location, etc.) to the subject.

Interestingly, the term linking verb doesn't appear in any of the grammars mentioned in the introduction. It's a 20^th century gloss of copulative, which means basically the same thing¹. The CGEL uses the term complex intransitive both for clauses and the verbs that head them. This has some advantages, as we'll see at the end, but is so unfamiliar to teachers and learners that it is unlikely to be adopted.

The most common linking verb is be. Others include sensing verbs: feel, look, smell, sound, taste, seem, and appear; verbs indicating a continuation of state: remain, keep, continue, and stay; and verbs of change of state: become, get, go, grow, and turn. This covers the most central members, but it is not a complete list.

Here are some more examples with the linking verb underlined and an AdjP complement in bold:

1. He seems very nice.
2. It was good.
3. She became pregnant.
4. The soup smelled more delicious than ever.

Most linking verbs can take all three types of complement (i.e., AdjP, NP, or PP), but that is not always the case as illustrated in 9 and 11 below:

5. It seems a crying shame. (NP)
6. He doesn't seem up to it. (PP)
7. It is pie. (NP)
8. It smelled like pie. (PP)
9. It smelled pie. (here the NP pie is an object, not a predicative complement.)
10. Did she become president? (NP)
11. *Did she become on the desk? (PP)

Linking verb smell does not take NP complements. In 9, we have the transitive verb smelled, not the linking verb of 8. You can see that in 7, it and pie are the same thing, but in 8, it is something else, perhaps an animal, but not pie. Linking verb become doesn't take PP complements, and so 11 is simply ungrammatical.

In the introduction, we saw linking verbs defined as "verbs that express state or feeling rather than action," and I said that this needed some reconsideration. There are some elements of truth in it: some linking verbs are indeed more common in stative verb phrases (VP). We don't typically say *the table is seeming brown or *the soup is feeling hot. But lots of verbs that are not linking verbs (e.g., have, & like) typically appear in stative VPs, and there are many linking verbs that make dynamic VPs (e.g., The kids are getting rowdy/tired.) So it's best to forget the notion that linking verbs expresses states rather than actions and focus just on the complementation pattern.

And certainly, a number of linking verbs do express feelings (e.g., I felt safe) but many don't (e.g., goldfinches turn brown in the summer) and conversely, many verbs expressing feelings aren't linking verbs (e.g., I'm enjoying the concert). Again, we should stay focused on type of complements allowed.

Some complications
It can get a little complicated though, because some verbs take multiple complements, usually with a direct object, and these are not consistently included with lists of linking verbs:

12. That makes me + happy. (Direct Object:NP + Predicative Complement:AdjP)
13. That makes us + part of something bigger. (Direct Object:NP + Predicative Complement:NP)

This is why the CGEL uses the terms complex intransitive and complex transitive. Verbs like make that take complements of the form Direct Object + Predicative Complement would, in their terminology, be complex transitive.

Also, as we saw with 8 and 9 above, some verbs have different meanings, some of which are linking and some of which aren't². One example is feel, which takes predicate AdjP complements in the sense of one's emotions and direct objects in the sense of touching something:

14. She felt miserable. (AdjP; linking)
15. She felt his hand. (Obj; transitive)

---

[1] Reed and Kellogg address this point very briefly calling these verbs copula and their complements predicate adjectives and predicate nouns.

[2] The Cambridge Grammar of the English Language treats all these as simply "verbs taking predicative complements" (p. 263).

Introduction

In this column, which will run roughly twice a month, I'll tackle grammar question from two viewpoints: first, traditional school grammar, as exemplified by the Oxford English Dictionary 2^nd edition and grammars such as Reed and Kellogg's Higher Lessons in English, Henry Sweet's A new English Grammar, and their progeny; and second a more modern analysis following The Cambridge Grammar of the English Language (CGEL) and its companion A Student's Introduction to English Grammar.

Although there are various approaches to grammar, a glance at published textbooks and school curricula, and observations of various classrooms shows that "current traditional grammar" dominates English language teaching in Ontario. This is typically a catch-as-catch-can mingling of traditional school grammar with various modern innovations and ESL-specific rules of thumb. The LINC grammar items listed in the LINC 5-7 Curriculum Guidelines exemplify this nicely.  On page 336, we find phrase defined as

"A group of related words that does not contain a subject and a predicate. A phrase cannot stand alone as a sentence. There are several different kinds of phrases: prepositional phrases, participial phrases, adjective phrases, gerund phrases, infinitive phrases, verb phrases, and so forth. A phrase is named for the word that introduces it." 

This begins with the traditional definition but continues with a thoroughly modern idea of a phrase. You don't find various types of phrases, such as verb phrase or adjective phrase, discussed in traditional grammars. There, a phrase is roughly a syntactic unit larger than a word and smaller than a clause and lacking a main verb. Of course, this conflicts with the modern idea of a verb phrase. This seemingly unconscious blending is similarly evident in the Longman Dictionary of Language Teaching and Applied Linguistics (discussed here) and TESOL's "Grammatically Speaking" column (discussed here). On the other hand, the LINC document also includes adjective clause, a term that would have been familiar to Reed and Kellogg but is not used in the CGEL.

Beyond this mix of the traditional and the modern, the LINC document includes definitions that reflect neither approach. Consider the definition of linking verbs as "verbs that express state or feeling rather than action," (p. 335) which could do with some attention and which brings us to our inaugural question: what is a linking verb anyhow?