Standard approaches to English grammar
usually identify two 'be verbs' – one a lexical
or copular verb and the other an auxiliary
(see Celce-Murcia & Larsen-Freeman, 1999,
p. 53; Greenbaum & Quirk, 1990, p. 36;
Berk, 1999, p. 151; Azar, 2002: A6; Teschner
& Evans, 2007, p. 51; Börjars & Burridge,
2001, pp. 166–167, to name a few). Common
textbook examples of these two uses of be
are given in (1a, b) respectively:
(1) a. LEXICAL VERB
She is a doctor.
They are hilarious.
Malcolm was the leader.
This is for you.
We're in the kitchen.
There were three children in the yard.
She is waiting.
The vase was broken by the workers.
We were devastated by the tragedy.
In this paper I would like to make two claims
– an easy claim and a more difficult one. The
easy claim is that the distinction illustrated in
(1) is spurious. All the examples of be in (1)
are prototypical auxiliaries. As described by
Huddleston and Pullum (2002, p. 92), these
are CORE AUXILIARIES. The more difficult
claim is that indeed there is a morphosyntac-tic difference between be as a lexical main
verb and be as an auxiliary, but that differ-ence is not the standard one illustrated above.
The syntactic distinction between lexical and
auxiliary be is more insightfully motivated by
the semantic distinction between ACTIVE BE
and STATIVE BE. Both of these claims have
profound consequences for theoretical ap-
proaches to the basic clause structure of
Modern English, and to English language
pedagogy. The net result of taking this ap-proach is to significantly simplify the con-ceptualization, teaching and learning of Eng-lish grammar.
The easy claim
It is very easy to show that syntactically all
the examples of be in 1 are auxiliaries. In fact
this has often been noted or assumed in the
literature, without much ado (see, e.g., Hud-dleston & Pullum, 2002, p. 114 and the refer-ences cited therein), so I am not claiming to
have discovered anything 'new' about English
grammar in this section.
Rather I would like
simply to put all the facts before the readers
of this journal in a clear fashion, and explore
how they potentially affect the teaching and
learning of English.
The lexical verb vs. auxiliary distinction
Though terminology varies widely, most dis-cussions of the lexicon of any language de-scribe a distinction between LEXICAL VOCAB-ULARY and GRAMMATICAL FUNCTORS (see,
e.g., Givon, 2001, pp. 187–237; Huddleston
& Pullum, 2002). In this characterization of
the lexicon, lexical verbs form an open class
of words that have certain syntactic features
and tend to express complex semantic con-tent. Examples of lexical verbs in English
include eat, advertise, read, dichotomize, and
a very large number of others. Auxiliaries, on
the other hand, form a relatively small, closed
set of grammatical functors. In English there
are modal auxiliaries (would, could, will, can,
etc.), aspectual auxiliaries (have and be), and
a 'dummy' auxiliary, do. Some approaches
consider auxiliaries to be a subset of verbs,
because they either take verbal inflection, or,
in the case of modals, themselves constitute
the required verbal inflection for a clause. In
any case, it is the distinction between lexical
vocabulary and grammatical functors that
underlies the traditional determination that
there are two be's in English – copular-be
(1a) is a considered to be a member of the
class of lexical vocabulary, while auxiliary-be (1b) is a member of the class of grammati-cal functors. In the following subsections I
will briefly describe a few of the morphosyn-tactic tests for distinguishing lexical verbs
Since full verbs are lexical vocabulary, they
are not amenable to contraction. Auxiliaries,
on the other hand, being grammatical func-tors, tend to be phonologically reduced, and
often do not take ordinary word stress. For
this reason some auxiliaries, including most
forms of be, may cliticize to (contract with) a
(2) WITH SUBJECT:
She would listen to you. →
She'd listen to you.
Cleo had listened to me. →
Cleo'd listened to me.
The teacher is smiling. →
The teacher's smiling.
The man in the suit is devastated by
the news. →
The man in the suit's devastated by
Around the bend will come the train.
→ Around the bend'll come the
On the wall are hanging three
portaits.→ On the wall're
hanging three portraits.
Here have fallen many fine soldiers.
→ Here've fallen many fine
This contraction does not occur with any
main lexical verbs:
(3) She owed me a dollar. →
*She'd me a dollar.
Cleo had fourteen cats. →
*Cleo'd fourteen cats.
The teacher has a Jaguar. →
*The teacher's a Jaguar.
I will her my estate. →
*I'll her my estate.
The following examples show that this con-traction also occurs with copular-be. This,
then, is the first structural feature that unites
it with auxiliaries and distinguishes it from
main lexical verbs:
(4) WITH SUBJECT:
She is a doctor. →
She's a doctor.
They are in the kitchen. →
They're in the kitchen.
The news about Australia is
The news about Australia's
Here are two children. →
Here're two children.
In the kitchen is a sink. →
In the kitchen's a sink.
These contraction facts illustrate that copular-be exhibits at least one morphophonemic
characteristic of auxiliaries – a characteristic
not shared with full lexical verbs.
Auxiliaries have the NICE properties; lexical
main verbs do not
All of the forms of be in (1) have several
properties of prototypical auxiliaries, includ-ing those identified by Huddleston and Pul-lum (2002, pp. 92–112) as the 'NICE' proper-ties (Negation, Inversion, Code and Empha-sis).
In the following sections, I will quickly
run through these properties, first showing
that they hold for auxiliaries, and not for lex-ical verbs. Then I will show that the same
properties hold for be in copular construc-tions such as those in (1a). Taken together,
this evidence clearly shows that copular-be is
a syntactic auxiliary.
Negation: In negative clauses, the negative
particle not follows an auxiliary (the first, if
there are more than one):
She should eat more chelow kebab.
She is eating chelow kebab.
The vase was broken by the workers.
We have lived in Isfahan.
She should not eat more chelow
She is not eating chelow kebab.
The vase was not broken by the
We have not lived in Isfahan.
In Modern English, lexical main verbs do not
allow the negative particle to follow them:
(6) *She eats not chelow kebab.
*The workers broke not the vase.
*We live not in Isfahan.
Rather, if there is no auxiliary in the corre-sponding affirmative clause, the 'dummy'
auxiliary do is inserted, and the negative fol-lows it:
(7) She eats chelow kebab. →
She does not eat chelow kebab.
The workers broke the vase. →
The workers did not break the vase.
We live in Isfahan. →
We do not live in Isfahan.
Inversion: In certain questions the first auxil-iary and the subject must invert (exchange
She should eat more chelow kebab.
She is eating chelow kebab.
The vase was broken by the workers.
We have lived in Isfahan.
Should she eat more chelow kebab?
Is she eating chelow kebab?
Was the vase broken by the workers?
Have we lived in Isfahan?
What should she eat more of?
What is she eating?
Who was the vase broken by?
Where have we lived?
If there is no auxiliary in the declarative, the
dummy auxiliary do is inserted before the
(9) Y/N INTEROGATIVE:
Does she eat chelow kebab?
Did the workers break the vase?
Do we live in Isfahan?
What does she eat?
What did the workers break?
Where do we live?
Again, lexical main verbs do not exhibit this
(10) *Eats she chelow kebab?
*Broke the workers the vase?
*What broke the workers? (trying to
mean 'What did the workers break?')
*Where live we?
Code: In constructions that 'stand for' or
'code' a previously mentioned verb phrase,
the first auxiliary is repeated (and inverted
with the subject). The ungrammatical exam-ples illustrate the fact that lexical main verbs
do not have this property:
(11) TAG QUESTIONS:
She should not eat kebabs, should
*She should not eat kebabs, eat she?
*She eats kebabs, eats not she?
The vase was broken by the workers,
*The vase was broken by the workers,
(was) not broken it?
I should see the doctor, and so should
*I should see the doctor, and so
(should) see she.
Who should eat chelow kebab? She
should. *She (should) eat.
We were eating kebabs, and so was
*We were eating kebabs, and so (was)
If there is no auxiliary in the original verb
phrase, the dummy auxiliary do occurs in the
(12) TAG QUESTIONS:
She likes kebabs, doesn't she?
*She likes kebabs, likes not she?
The workers didn't break the vase, did
*The workers didn't break the vase,
I saw the doctor, and so did she.
*I saw the doctor, and so saw she.
Who ate the kebabs? She did. She
Emphasis: In constructions in which the truth
of the proposition is emphasized, the first
auxiliary receives emphatic stress (indicated
by all caps in these examples). Again, the
infelicitous examples show that lexical main
verbs do not possess this property:
(13) She should eat more chelow. Yes she
SHOULD. Yes she should EAT.
The vase was broken by the workers.
Yes it WAS. Yes it was BROKEN.
We have lived in Isfahan. Yes we
HAVE. Yes we have LIVED.
If there is no auxiliary in the original clause,
the dummy auxiliary do occurs and receives
the emphatic stress:
(14) She eats a lot of chelow. Yes she
DOES. Yes she EATS.
The workers broke the vase. Yes they
DID. Yes they BROKE.
We live in Isfahan. Yes we DO.
Yes we LIVE.
Copular-be has the NICE properties
The above examples all illustrate the well
known and established syntactic tests for dis-tinguishing auxiliaries from main verbs –
auxiliaries have the NICE properties, while
lexical verbs do not. It is also well known
and easily demonstrated that the so-called
'lexical' use of be in predicate nominals, ad-jectives, locatives and other copular construc-tions illustrated in (1a) has the NICE proper-ties.
Negation: Like auxiliaries, copular-be pre-cedes the negative particle, and does not re-quire or allow (in prototypical cases, but see
below for the 'exceptions') do-support:
(15) She is not a doctor.
*She does not be a doctor.
They are not hilarious.
*They do not be hilarious.
Malcolm was not the leader.
*Malcolm did not be the leader.
They are not in the kitchen.
*They do not be in the kitchen.
This is not for you.
*This does not be for you.
Inversion: Like auxiliaries, copular-be inverts
with the subject in certain questions. Unlike
lexical verbs, it does not require do-support:
(16) Is she a doctor?
*Does she be a doctor?
Are they hilarious?
*Do they be hilarious?
Is this for me?
*Does this be for me?
What is she?
*What does she be?
Who is this for?
*Who does this be for?
Where are we?
*Where do we be?
Code: Like auxiliaries, copular-be codes pre-viously mentioned verb phrases. Unlike lexi-cal verbs, it does not rely on a dummy DO:
(17) She's a doctor, isn't she?
*She's a doctor, doesn't she?
I am a doctor, and so is she.
*I am a doctor, and so does she.
22 The two be's of English
Emphasis: Like auxiliaries, copular-be re-ceives emphatic stress when the truth of the
proposition is emphasized. Unlike lexical
verbs, it does not need DO:
(18) She's a doctor. Yes, she IS.
Yes, she DOES.
We were in Isfahan. Yes we WERE.
Yes, we DID.
Notice that other copular (or 'linking') verbs
that take subject complements, such as seem,
become or resemble, do not have the NICE
properties, and do require the presence of do
in NICE constructions. Therefore they are
lexical main verbs, and as such are syntacti-cally distinct from copular-be:
(19) N: They don't seem hilarious.
*They seem not hilarious.
She didn't become a doctor.
*She became not a doctor.
I: Does she resemble her mother?
*Resembles she her mother?
What did she become?
*What became she?
C: She became a doctor, and so did he.
*She became a doctor, and so
The situation turned ugly, didn’t it?
*The situation turned ugly, turned
E: They seem happy. Yes they DO.
*Yes they SEEM.
Among copular verbs, then, only be has the
NICE properties otherwise only attributed to
auxiliaries. We can conclude, then, that copu-lar be belongs to the same syntactic class as
There is one syntactic property of be in pro-gressive aspect constructions that seems to
contradict the claim that auxiliary-be and
copular-be are one and the same syntactic
entity. This is the fact that wh-extraction of
the complement of be in progressive aspect
constructions requires a pro-form, doing,
while extraction of the complement of be in
copular constructions is very possible and
common with no recapitulating pro-form.
Though this appears to counterexemplify the
easy claim made in this paper, I believe there
is a good explanation for this phenomenon
that does not require that auxiliary-be and
copular-be be treated as distinct syntactic en-tities. This explanation is presented briefly
below, after the apparent counterexamples.
The complement of be in a copular construc-tion can be the target of wh extraction, leav-ing no visible remnant in situ:
(20) a. What is that? It is a birthday cake.
It is baking.
b. What are you? I am a linguist.
I am singing.
These are wh questions in which the com-plement of be is the target of the wh word.
The wh word may not target the complement
of progressive aspect be. This is shown by
the incoherence of progressive aspect an-swers to these questions. Instead, if the com-plement of be in a progressive construction is
extracted, the pro-form doing must remain in
(21) a. What is that doing?
It is rolling down the hill.
b. What are you doing?
I am singing.
This non-extractability of the complement of
be in a progressive construction holds for
other auxiliaries as well, though of course the
form of the in situ pro-form varies according
to the particular complement type:
(22) He should eat chelow. →
What should he do?
*What should he?
They have eaten chelow. →
What have they done?
*What have they?
They eat chelow. →
What do they do?
*What do they?
Non-extractability of the complements of
auxiliaries has consequences in all the 'clas-sic' wh-extraction constructions, such as
headless relative clauses, illustrated here in
(23) a. *What she should is sing the national
b. What she should do is sing the
(24) a. *What she has is sung the national
b. What she has done is sung the
(25) a. *What I'm not is singing the national
b. What I'm not doing is singing the
(26) a. *What she is is singing the national
b. What she is doing is singing the
These kinds of clefts are perfectly acceptable
without a pro-form recapitulating the extract-ed complement of copular be:
(27) a. What I'm not is organized.
b. What she is is a doctor.
c. Where I am is at home.
d Why I'm here is a mystery.
This one syntactic property seems to differ-entiate the auxiliary use of be from the copu-lar use, and hence to constitute counter evi-dence to the 'easy claim' made in this paper.
However, there are at least three reasons not
to consider this property as definitive counter
First, this feature does not hold for be in pas-sive constructions. In most cases, the com-plement of 'passive be' can be extracted with
no pro-form left in situ:
(28) a. What she was is devastated by the
b. What we are is frightened by the
severity of your reaction.
In fact, a pro-form is not possible in these
contexts, just as it is not possible with the
copular use of be (29c, d):
(29) a. *What she was done is devastated by
b. *What we are done is frightened by
the severity of your reaction.
c. *What she is one is a doctor.
d. *Where I am there is at home.
The fact that Wh-extraction does not distin-guish passive auxiliary-be from copular-be
indicates that, if anything, progressive-be is
the odd one out in this typology. So-called
copular-be still has all the properties of proto-typical auxiliaries, including passive-be.
Since the class of auxiliaries is itself a mixed
bag, e.g., the modals exhibit a slightly differ-ent cluster of syntactic properties than the
other auxiliaries and even from one to the
other, it is not particularly telling that pro-gressive-be has one apparently unique feature
that distinguishes it from all the rest.
Second, there are unique features of some of
the copular uses of be as well, yet these do
not compel English grammar books to call
each one a syntactically distinct copula. For
example, only predicate adjectives can occur
in various kinds of comparative construc-tions, as illustrated in (20) (examples from
Davies 2004 – the BYU-BNC: The British
(30) a. however complicated the key
sequence is, a self-indexing function
can be found . . .
b. The more fleeting the moment, the
more poignant the emotion.
c. Depressed as he was, he managed to
ruin the mood.
These constructions are not semantically
compatible with predicate nominals and some
(31) a. *However a teacher she was, she
couldn't get a job.
b. *The more the father he was, the less
he could get done.
c. *A teacher as he was, he had to look
d. *However in the house he was, she
couldn't get him to cook a meal.
e. *On top of the mountain as we were,
we couldn't breathe.
While they do seem to be acceptable with
(32) a. However frustrated by events we
were, we never gave up.
b. The more frightened by the severity
of your reaction she is, the less likely
she is to open up.
c. Shocked by the tragedy as we were,
we couldn't bear to attend the party.
However, clearly these comparative construc-tions cannot occur with progressive be:
(33) a. *However smiling at her I was, I
couldn't get her attention.
b. *The more singing in the rain they
were, the more they got wet.
c. *Smiling at her as I was, I couldn't get
This particular syntactic property follows
from the semantic character of the comple-ment of be – only forms that describe grada-ble attributes can be compared. One can be
more or less happy, more or less complicated,
more or less frustrated by events, frightened
by a reaction or shocked by a tragedy. It is
harder to interpret someone as being smiling
at someone, singing in the rain, a teacher, the
father or in the house to greater or lesser de-grees. These are either/or notions. In fact, it is
only to the extent that such phrases can be
interpreted as gradable attributes that they
can occur in these comparative constructions:
(34) The more to the left of and above the
dashed straight line a curve is, the
more potential exists for improvement.
Since something can be more or less to the
left of something else, and more or less above
something else, these particular prepositional
phrases can be compared with the same kind
of comparative construction as gradable mod-ifiers can.
Finally, one could interpret the requirement
that a pro-form be left in situ when the com-plement of progressive be is extracted as
simply the result of the fact that English pos-sesses no wh- word that corresponds to a pre-sent participle verb form. The same is true
when prepositional phrases other than loca-tives are extracted. For example, there is no
wh- word in English that corresponds directly
to a benefactive element. Instead the com-plex, for who(m) must be used. In this case,
the preposition must be left in situ:
(35) Whom this is for is you.
*Whom this is is for you.
This is true of other non-locative preposition-al phrase complements of copular be:
(36) a. Whom she was with is your mother.
*Whom she was is with your mother.
b. Where this traveller is from is Vulcan.
*Where this traveller is is from
c. What this road is toward is your new
*What this road is is toward your
Modern English simply lacks wh- words that
correspond to the relations expressed by
these prepositional phrases, just as it lacks a
wh- word that corresponds to a present parti-ciple.
Notice, however, that the somewhat archaic
English words, whence, and whither corre-spond to the modern from where and to
where (toward) respectively. For speakers
who still use these wh- words, the following
(37) a. Whence is that knocking? (BNC)
b. Whence that knocking is is the front
c. The final chapter –; on whither the
wedding cake –; . . .
d. 'Whither are we bound, my lord?'
Thus we see that extractability is at least
somewhat dependent on the semantic rela-tions expressed by the available inventory of
wh- words. Imagine for a moment that the
wh- word *whating existed in English. In that
case perhaps present participles could be ex-tracted with no clarifying pro-form left in
(38) *Whating is he? (meaning 'what is he
*Whating he is is cleaning the
refrigerator. (meaning 'what he is
doing is cleaning the refrigerator.')
It just so happens that such a potentially use-ful wh- word does not exist, therefore, the
composite form doing what must suffice.
This can be considered parallel to the cases
of with what, for what and modern from
where, which also requires the preposed ele-ment to remain in situ when the complement
In summary, the fact that the complement of
be in a progressive aspect construction can-not be the target of wh-extraction without a
resumptive pro-form left in situ does not en-tail that progressive be is a different kind of
syntactic entity than copular be. Many of the
uses of be that are all considered copular also
engender distinct clusters of syntactic proper-ties. These properties can often, if not al-ways, be attributed to the semantic characters
of the complements, and not to the syntactic
category of the copula/auxiliary. The wh- ex-traction characteristics of present participles
in progressive aspect clauses may simply be
due to the fact that the inventory of wh-
words of English does not include one that
corresponds to a present participle. There-fore, wh- extraction does not constitute coun-
terevidence to the 'easy claim' made in this
The myth of 'lexical verb' be
Since copular be is so clearly a member of
the syntactic category of auxiliaries, why
have pedagogical and more linguistically ori-ented works on English grammar insisted on
calling it a lexical main verb? I believe that
this strange phenomenon can largely be ex-plained by a myth of traditional grammar that
has been perpetuated by generations of Eng-lish teachers. This myth is expressed in (39):
(39) Every clause in English must have a
Starting from this assumption, all the instanc-es of be in (1a) must be lexical verbs, since
the only other element in the predicate is
non-verbal. I would like to claim that (39) is
an unnecessary and ungrounded assumption.
The more insightful generalization, I contend,
is the following:
(40) Every clause in English must have
tense, aspect and/or mode Inflection.
There are other reasons for replacing (39)
with (40), in addition to resolving the status
of copular be. First, several theoretical ap-proaches to English grammar, including re-cent versions of Generative Grammar, affirm
the assertion in (40). For example, in the
minimalist paradigm (represented by Rad-ford, 1997), the 'Sentence' is no longer the
highest node in a syntactic tree. Rather 'In-flectional Phrase' is the highest node. This
reflects the fact that the category that is the
syntactic 'head' of a sentence is its 'I-node', or
Inflection. In other words, the properties of a
sentence are projected from its Inflection – if
there is no Inflection, there is no sentence.
The actual arguments for this determination
are quite compelling, if rather complex.
Readers are referred to Radford (1997, pp.
61- ff.) for the details.
Second, the special forms traditionally
termed present and past participles that fol-low be in progressive aspect and passive
voice constructions are deverbal in that they
have lost most of their syntactic properties of
verbs; in particular, they cannot be inflected.
Therefore, like other non-verbal categories
(nouns, adjectives and prepositional phrases),
participial forms must rely on some other el-ement (a core auxiliary) to express the im-portant inflectional information when the par-ticiple itself constitutes the main semantic
content of a predicate.
Lets look at some examples that may help
illustrate this fact. Basic passive construc-tions are very similar to copular predicate
adjective constructions in which the adjective
happens to be a past participle:
(41) a. The vase was broken when the
workers moved the piano.
b. The vase was beautiful when the
artisan finished painting it.
c. As soon as I walked into the room, I
noticed that the vase was broken.
Many grammar books would say that was in
(41a) is an auxiliary because the construction
is a passive. On the other hand, was in (41b)
and (41c) is a lexical verb because the con-structions are predicate adjectives. However
the three predicates are syntactically identi-cal. Clearly there is a difference in meaning
between the passive and attributive senses of
the complements of be in these sentences, but
that difference can be attributed to the nature
of the complements, not necessarily to any
syntactic categorial difference between the
two uses of be.
Similarly, consider the following two exam-ples:
(42) a. That person is annoying me.
*That person is very annoying me.
b. That person is annoying.
That person is very annoying.
Again, many grammar books and linguists
would say that be in (42a) is an auxiliary,
while in (42b) it is a lexical verb. Of course,
the meaning difference between the senses of
annoying in these two examples is important,
and does affect the collocational possibilities
(or selectional constraints) of the two con-structions. That (42a) is progressive and
(42b) is attributive is demonstrated, for ex-ample, by the fact that the adverb very can
only be inserted in (42b) (see Wasow, 1977
for further selectional arguments for the dif-ference between passive and attributive parti-ciples). Nevertheless, if the distinction be-tween auxiliary and main lexical verb is sup-posed to be a distinction between two syntac-tic classes of items, there should be syntactic
correlates to the semantic distinction. Other-wise, there is no reason to posit anything oth-er than garden variety polysemy. Clearly
be+complement constructions may be poly-semous in a number of ways, including pro-gressive vs. attributive. However, in every
case the polysemy stems from the syntactic
or discourse context (41a, c):
(43) POLYSEMY OF COPULAR-be:
a. That person is tall. ATTRIBUTIVE
b. That person is a teacher. EQUATIVE
c. That person is in the kitchen.
d. There is a rat in the kitchen.
e. This is for you. BENEFACTIVE
g. This is mine. POSSESSIVE
(44) POLYSEMY OF AUXILIARY-be:
a. That person is eating a banana.
b. That banana was eaten by someone.
Looking first at the examples in (43), we see
that the semantic relations expressed are sig-nificantly different from one another, yet tra-ditional and pedagogical grammars typically
find no reason to posit syntactically distinct
'copulas' for each relationship.
(44) two quite distinct meanings are ex-pressed, both of which depend on the seman-tic properties of the complements, rather than
on any syntactic category difference among
the forms of be – the present participle form
of a verb expresses an ongoing activity, while
a past participle refers to a resultant state.
The auxiliary in all these examples is func-tioning in exactly the same way – to express
the Inflectional information required of every
In summary, insisting that there is a funda-mental syntactic difference between copular
be and auxiliary be introduces a number of
unnecessary conceptual and pedagogical
complexities. Adopting the alternative asser-tion, suggested in (40), resolves these com-plexities. From this point of view, every in-dependent clause must contain an element
that is 'Inflectable' with whatever Inflectional
information is appropriate for that clause's
syntactic function (e.g., as an independent
assertion, a question, a relative clause, an ad-verbial clause, a clausal object, etc.). One job
of the first auxiliary in an Inflected verb
phrase is to carry the necessary Inflectional
information. Most auxiliaries participate in
expressing various aspectual and modal cate-gories as well, but be basically just serves as
a 'platform' for Inflection when the lexically
rich element – the one responsible for most
of the semantic content of the predicate – is
de-verbal or non-verbal, and therefore cannot
express the Inflectional information directly.
This function unites the uses of be in copular
as well as progressive aspect and passive
The harder claim
In the previous section, I have shown that the
distinction between copular-be, and auxilia-ry-be is spurious. The only arguments against
this claim are based on different interpreta-tions and selectional properties that arise be-cause of different semantics of the comple-ments that follow be. But, as I have shown,
many similar semantic differences may arise
between be and its complement that tradi-tional grammars do not attribute to a syntac-tic category difference between types of be.
Given the overwhelming fact that copular-be
and auxiliary-be have exactly the same syn-tactic properties, there is no reason to suggest
that the different uses of be are due to a cate-gorial distinction between two lexemes.
The more difficult assertion I would like to
make is that in fact there are two syntactical-ly distinct be verbs in English, and that one is
a lexical main verb and the other is an auxil-iary. Furthermore, I will claim that the syn-tactic distinctiveness of these two be's (evi-denced by syntactic properties) is motivated
by the semantic difference between stative be
and active be. The reasons that this assertion
is more difficult are 1) the argument may
give the impression that the harder claim ac-tually contradicts the easier claim. In fact it
does not. 2) Corroborative evidence for the
harder claim is based on data from 'non-standard' forms of English. Some examples
given below would definitely be 'ungrammat-ical' to most English teachers, and should not
be used as examples in ESL classrooms.
However, such examples are attested in natu-ral discourse, and are logically coherent. This
fact lends additional support for the hard
claim, though it does not constitute the major
Semantic Stativity vs. Activity
The semantic distinction between STATES
and ACTIVITIES is mostly determined by
volitionality and change. Situations that are
presented as involving change, and normally
initiated and controlled by some entity acting
with volition (on purpose) are ACTIVITIES.
Situations that do not involve change, and
have no controlling entity are STATES. This
is a very general characterization. As with
any semantic distinction, there is in fact a
continuum between prototypical states and
prototypical activities – there are very good
examples of states and very good examples
of activities, but a large number of situations
fall somewhere in between (see Vendler,
1967; Chafe, 1970; Comrie, 1989 for fuller
characterizations). However, the grammar of
English tends to discretize (make into dis-crete categories) the semantic difference be-tween stativity and activity in a number of
ways. In this section I will describe two of
the 'tests' for whether a situation is being pre-sented as a 'state' or an 'activity.' These I will
refer to as the habitual test and the progres-sive test.
The habitual test: When an independent
clause occurs in the so-called 'present' tense
form, the temporal reference may be inter-preted as habitual aspect or as a 'true present,'
i.e. a situation that is in effect at the time of
utterance. Activities are normally understood
as habitual (45), while states are normally
understood as true present (46):
They sometimes build their eyries on
inland lava pinnacles.
You exercise to look good.
In the senses intended in the naturally occur-ring examples given in (44), build and exer-cise describe ACTIVITIES in that they refer
to situations that involve intentionality, voli-tion and change. When occurring in the
'present tense,' as in these examples, these
verbs do not assert that the activities are tak-ing place 'now,' i.e., at the time of speaking
(though they incidentally may be), but rather
that they occur from time-to-time over a long
period that includes the time of speaking. No
particular finite event of building or exercis-ing is referenced.
On the other hand, the verbs see, love and be
in (46), below, express STATES, in that no
movement or change is asserted. In the pre-sent tense, these examples assert that the state
holds 'now,' i.e., at the time of utterance. The
specific current instance of the state is being
referenced, rather than the possibility that the
state holds true from time-to-time (all exam-ples from the BNC):
(46) TRUE PRESENT:
I see you are troubled at something.
I love you.
The room is red now.
Thus 'present tense' for activities expresses
'habitual aspect,' while 'present tense' for
states expresses a true present. In order to
express the idea that an activity is taking
place 'now', progressive aspect must be em-ployed. This constitutes the next test for
whether a situation is being presented as a
state or an activity.
The progressive test: There is an apparent
semantic anomaly between stative situations
and progressive aspect. This is because the
progressive aspect construction evokes an
image that involves 'progression,' i.e., pro-gressive change and/or movement. A state,
by definition, does not involve movement or
change, therefore prototypically stative situa-tions are not semantically amenable to ex-pression in the progressive aspect:
(47) I see the airplane.
?I'm seeing the airplane.
She likes ice-cream.
?She is liking ice cream.
We know the answer.
?We are knowing the answer.
The room is red.
??The room is being red.
However, the question-marked examples of
(47) are not 'ungrammatical.' Rather, they
constitute less-than prototypical expressions
of the stative concepts of seeing, liking,
knowing and being red. In fact, stative con-cepts can occur in the progressive, but when
they do, a different, non-stative, sense is ex-pressed. Because of the cognitive schema
evoked by the progressive aspect construc-tion, the construction itself evokes the notion
of activity. The examples in (48) through
(50), from the BNC, illustrate the same verbs
as in (46) above, but this time in the progres-sive aspect. The active interpretations of
these situations, as made clear by the context,
are given in caps following each example:
(48) We're seeing already, that Health
Authorities haven't got the money . . .
Yes they were seeing how much
more they could eat and take home.
I was seeing them one after the other.
(49) Football is a game of chance and I am
loving every minute of it.
(50) They 're being rude up that end.
they're being silly. ACTING SILLY
I'm being honest. SPEAKING
I thought I'd persuaded him that he
was being foolish. ACTING
Another piece of evidence that be in the pro-gressive aspect is active is the fact that it
doesn't seem to work with subjects that are
incapable of acting with volition:
(51) a. The children are being quiet.
b. *The river is being quiet.
Thus we see that, like other stative verbs,
when be appears in the progressive aspect it
takes on an active, volitional meaning. But
wait – isn't this a syntactic property of lexical
verbs that distinguishes be from auxiliaries? I
don't think any other auxiliaries can occur in
the progressive aspect:
(52) *They are shoulding eat more chelow.
*They are having eaten more chelow.
*They are doing eat more chelow.
Furthermore, active be can occur with do
support, as in the following:
(53) Careful! No don't be silly Amy.
DON'T ACT SILLY
Don't be stupid Stuart! DON'T DO
My dear, do be quiet –; he may be
listening now! CEASE MAKING
Do be careful, love . . .ACT
This is another property that be does not
share with other auxiliaries.
(54) *They do should eat more chelow.
*They do have eaten more chelow.
*They do do eat more chelow.
Finally, compare the examples in (53) above
to the following stative situations expressed
with the same lexical items, but without do-support or progressive aspect (examples from
(55) They're silly buggers though aren't
They are stupid that lot!
Toads are quiet and harmless and
Usually she is careful,
While it may be a stretch to think of states as
being 'habitual,' it should be clear that these
clauses in the present tense make assertions
about the general character of their subjects,
rather than to any particular instance of their
being silly, stupid, quiet, etc. that is asserted
to be true at the moment of speaking.
These examples show that indeed there is
something strange about be. It has all the
properties of auxiliaries, but it can occur in
the progressive aspect and it can occur with
'do-support.' It just so happens that whenever
be occurs in the progressive aspect, or takes
do-support, it expresses an activity rather
than a state.
Stative be vs. Active be
In the above section we have seen that be in
copular constructions that express STATES
has all the properties of auxiliaries. However,
be may have properties of lexical verbs ex-actly in those situations that express ACTIV-ITIES – acting quiet, acting silly or acting
stupid, etc. It passes the syntactic tests for
lexical verbs exactly and only when the se-mantics involves an ACTIVITY, usually ini-tiated and controlled by an agent acting with
volition. This is the basis of the hard claim
made in this paper: that in fact there are two
syntactically distinct be's in English, one sta-tive/auxiliary be and another active/lexical
In addition to the evidence presented so far,
is there any independent evidence of the dis-tinction between the two be's? Consider the
following naturally-occurring example from
one of my daughters when she was 12 years
old. The context was the behavior of one of
her friends who attended a birthday party:
(56) He's not silly; he just be's silly when
he's around girls.
The form be's (pronounced bees), though ut-terly non-standard, is logically coherent in
this context. It shows that this native speaker
has two be's in her lexicon. The stative be is
the irregular one that is really an auxiliary
whenever it occurs (as demonstrated in the
first part of this paper). The active be, on the
other hand, is morphologically regular, taking
the regular third person singular present tense
–s ending. Thus active be and stative be are
formally, as well as semantically, quite dis-tinct. This example is particularly telling in
that it explicitly contrasts stative be – He's
not silly – with active be – he just be's silly,
thus showing that the speaker had internal-ized both be's in her lexicon, and considered
them to describe distinct states of affairs, one
of which she presented as true and the other
Example (56) is so sensible in this context
that I was curious to determine how wide-spread this usage was. Unfortunately, the
BNC provides no clear examples of the 'regu-lar' active be illustrated in (56). So, I turned
to an even larger corpus – the internet. There
I found much more fertile ground. Below are
a few of the several hundred examples of the
morphologically regular, active be. Examples
(57) through (59) are a few of the results of a
Google search for 'he just be's' (845 total
(57) Sometimes he just be's like that.
(58) he dosent really dress up he just be's
himself and wears bermuda shorts,
headband, sandles and plain shirts
(59) He doesn't hold one side or the other,
he just be's himself and I admire that.
The following is from a Google search for
'she just be's' (428 total hits):
(60) If she just be's herself...people will
stay add her!
Clearly X just be's Xself is a relatively com-mon collocation. Other examples of morpho-logically regular be used in an active sense
are also attested on the internet. However,
they are eclipsed by many instances of the
regularization of auxiliary be in AAVE (Af-rican American Vernacular English) – though
many varieties of AAVE generally exclude
the -s ending in 3rd person present tense.
While the regularization of active be may or
may not have originated with AAVE, it is a
totally reasonable formation based solely on
the internal syntactic character of so-called
Standard English. Consider the following ex-ample:
(61) If she just be's herself, she'll do fine
in the debate.
The 'standard' way of expressing this would
(62) If she just is herself, she'll do fine in
According to my native speaker intuition, this
doesn't capture the sense of volitionality and
activity that is nicely expressed in (61). This
distinction is reminiscent of the distinction
between other pairs contrasting stative and
active be (constructed examples):
(63) a. Why aren't you the leader?
b. Why don't you be the leader?
In example (63a) the speaker just questions a
state of affairs, while (63b) is a suggestion
that the addressee act in some volitional way
to take a leadership position. Again, this il-lustrates that auxiliary be (63a) is stative,
while lexical be (63b) is active.
Example (64) is one last example of regular
active be, this time occurring in the major-class past tense with -ed:
(64) I gave the monitor to her while she
‘beed the doctor’ using the monitor to
poke around my feet.
This is an example of an adult quoting a
child, and so may be dismissed as simple
overgeneralization. Nevertheless, it is inter-esting to me that this usage clearly implies
the child was actively acting like a doctor.
The regular form, she was the doctor, simply
would not have expressed the same sense.
Consequences for pedagogy
The consequences for English grammar ped-agogy of spuriously uniting copular be with
the lexical copular verbs, and distinguishing
it from auxiliaries are manifold. In particular,
every discussion of the NICE constructions
must be qualified in a disjoint way: Auxilia-ries and be work one way; lexical verbs ex-cept be work the other way. If ESL/EFL
teachers and grammar books would consider
stative be to be an auxiliary, the number of
special cases that students would have to
learn and assimilate would be reduced by al-most half. After all, a significant number of
rather complex constructions are sensitive to
the auxiliary/lexical verb distinction as mani-fested by the NICE properties, namely:
Do-so (recapitulated verb-phrase
And perhaps others.
Another consequence of calling copular be a
lexical verb is that it renders the basic clause
structure of English mystifying to many
SLLs. My contention and my experience as a
TESOL and EFL teacher is that the assertion
given in (40) (repeated and slightly modified
here for convenience) goes a long way in
helping students conceptualize and internal-ize basic English clause structure:
(65) Every clause in English must have
one expression of tense, aspect
and/or mode Inflection.
There are several reasons for this fact. First,
many languages do not employ auxiliaries to
the extent that English does. Such languages
(Russian, Burmese, Tagalog, Indonesian, to
name a few) require no lexical verbal element
in copular constructions (predicate nominal,
predicate adjectival, locational and existential
constructions, etc.). Typologically, the clause
structure of English (and many other Indo-European languages) is rendered quite 'exot-ic' by the supposition that a lexical verb is
used in such constructions. This is a major
and unnecessary conceptual hurdle for many
Second, be is so common in English that
many students become confused as to when
to include be and when not to, as well as
when to inflect it and when not to. For ex-ample, the use of spurious be is common, as
well as double inflection constructions such
as the following (actual examples from ad-vanced Korean SLLs of English):
(66) Did you brought the forms?
She is went to the store.
They already were came.
I believe that a strong emphasis on the im-portance of the INFL (Inflection) slot in Eng-lish goes a long way toward helping students
overcome such difficulties. In many ways the
initial position in the predicate phrase is the
pivot for English syntax. Part of mastering
the 'character' of English syntax, and thereby
developing fluency, is capturing a sense of
how central the INFL position is. Of course,
it is not necessary to emphasize the unity of
copular-be and auxiliary-be in order to help
students assimilate this important fact about
English. However, I believe that keeping the
two distinct actually introduces unnecessary
confusion which makes understanding of the
overall clause structure of English much
Finally, this approach underscores the pro-found importance of the distinction between
activities and states for English grammar.
While this is a semantic distinction that can
undoubtedly be expressed in every language,
not every language pays quite so much atten-tion to it grammatically as does English. In
this paper we have seen how the activity/state
distinction helps explain the different usages
of the 'present tense' and 'progressive aspect'
forms. In addition to this well-known feature
of English grammar, activity vs. stativity
helps to explain the use of perfect aspect
forms in discourse. In particular, the distinc-tion between simple past and present perfect
is one that many SLLs find perplexing. This
distinction can largely be understood in terms
of the difference between an active event and
a resultant state – the simple past tends to ex-press an active event, while the perfect ex-presses a state that results from an earlier
event. A full exposition of this manifestation
of the state/activity distinction in English is
the subject matter for a different paper. Suf-fice to say that, again, part of assimilating the
general 'character' of English grammar is in-corporating the state/activity distinction into
one's unconscious cognitive framework for
This paper underscores what I consider to be
an important and often overlooked considera-tion in second language learning – namely
that each language has its own typological
'character', or profile, that students must in-corporate into their subconscious model of
the language in order to feel comfortable
speaking it, and to develop a 'natural' or 'na-tive-like' written and conversational style.
Sometimes ways of conceptualizing and dis-cussing grammar that arise within traditions
of first language education ('grammar
schools' in the traditional sense) are not intui-tive for second language learners. In particu-lar, many second language learners of Eng-lish in the present century come from first
language backgrounds that are typologically
very different from that of English. In such
situations it is very important, in my opinion,
to focus on the features of English that are
particularly perplexing and which seem, on
the surface, to be very different from those of
the students' first language. Second language
learners often perceive English as quite exot-ic, not necessarily because it really is so dif-ferent or strange, but mostly because of the
way it is presented in second language clas-ses. I believe that a typologically informed
approach to English grammar will go a long
way toward helping students comprehend the
essential character of English syntax, and
thereby enjoy the rewards of becoming con-fident second language speakers.
In conclusion, I have shown that the assump-tion that every English clause requires a lexi-cal verb is unfounded. Like most languages
of the world, the main predicating element in
copular constructions is not a verb at all, but
a de-verbal or non-verbal complement. The
be that occurs in such constructions functions
mostly as a 'platform' for expression of the
all-important Inflectional information. As
such, it has all the syntactic properties of core
auxiliaries, and none of the properties that
distinguish lexically rich verbs. Thus, copu-lar, passive and progressive aspect construc-tions are unified in requiring an auxiliary be.
The second claim is that there is, in fact, a
lexical verb that, in its base form is identical
to auxiliary be. Semantically, it has lexical
content in that it expresses activity; in most
cases it may be paraphrased with the lexical
verb act or act like. However, for some
speakers this lexical verb belongs to the ma-jor inflectional class, taking the present tense
form be's and the past tense form beed. This
lexical verb does have the syntactic charac-teristics of lexical verbs in general, and as
such is united with copular verbs such as be-come, seem and resemble. This non-traditional, but syntactically and semantically
highly motivated approach to the basic clause
structure of English significantly simplifies
the conceptualization and teaching of English
1. Several parts of this proposal have
been previously considered in the lit-erature. Lyons (1977) notes the syn-tactic commonality between copular
and auxiliary be. Bach (1967) as-sumes that English copular be is
transformationally inserted, i.e. that it
is not a lexical verb. Williams (1984)
suggests that the copula be is inserted
in INFL; in other words it does no
more than express the Inflection of
2. Grammatical functors may take em-phatic or contrastive stress, e.g., 'I
AM going.' In which case contraction
is precluded. However, in the absense
of special pragmatic features, gram-matical functors are usually un-stressed, and therefore tend to bind
phonologically to a local host.
3. The 'semi-auxiliaries' ought to, have
to and used to do not have all of these
properties, but that is a topic for an-other paper. So called copular be is a
core auxiliary in that it has all the
properties of the best examples of
auxiliaries in the language. When I
use the term auxiliary in this paper, I
mean core auxiliary, as described in
Huddleston and Pullum (2002, p. 92).
4. The 'frowny face' symbol () indi-cates the following utterance is infe-licitous in the context provided,
though not strictly speaking
5. Expressions in which the verb phrase
alone is emphasized do allow emphat-ic stress: Yes she should EAT more
chelow kebab (rather than MAKE
more chelow kebab). The property in
question here is verum focus, when
the truth of the whole proposition is
being emphasized. Only in this case
may the complement of the empha-sized element be ellipted, as illustrat-ed in the examples in 0.
6. In most of the examples of whence
referencing an ablative (from X) rela-tion in a copular construction in the
British National Corpus, the copula is
omitted, e.g., Whence this insolence?
. . . whence the name --; Aubeterre,
etc. However, there are a few exam-ples, such as 0a, in which the copula
7. Some languages actually do have dis-tinct copulas that are used to express
portions of the range of semantics ex-pressed by English be. Mandarin, for
example, uses the form shì for attribu-tive and equative clauses, zài for loca-tional clauses, and yŏu for existential
and possessive clauses. Spanish uses
estar for temporary attribution, all lo-cational clauses and progressive as-pect, haber for existential clauses, and
ser for permanent attribution, equa-tive and passive clauses.