2.2 Verb Phrase (VP) [-Fin]

There is much we can initially talk about regarding the Non-Finite [-Fin] Verb well before we start to muddle our way through its functional nature (MVP). First of all, let's clear up the issue of linear order--where do we find this non-finite [-Fin] Verb in a sentence? The short answer is: the [-Fin] VP always follows the functional Main Verb Phrase--keeping to its functional-to-lexical structural relationship. So, using a certain terminology, we could say the [-Fin] verbs are always verb-second within the predicate--that is, they always occupy the second verb slot of a sentence [functional MVP, lexical VP]. (Recall, this same functional-to-lexical relationship holds for D to N as presented above). Relegating the [-F] Verb to second fiddle (since it holds no functional feature values), however, doesn't mean that the verb can't provide an array of interesting distributional displays. The prosaic VP can be quite dynamic, wielding an embarrassing wealth of different projections and distribution. Of course, the VP's main task (void of its functional features) is to establish the Predicate of the sentence--providing the essential information about the topic or Subject of the sentence. This core information is seen as real referential meaning, tethered to aspects of the real outside world. So, in a sense, the lexical Verb along with the lexical Noun bear a certain flavor of truth--whereas the Noun introduces the "who/what" of a sentence, the Verb returns and asks "So, what about it?". Clearly, when all is said and done, we could "all just get along" (to use a certain phrase) with little else, and this was what the main topics were concerned with in our introductions of Lexical Grammar.
Moving onward from the sole verb, let's consider this diverse class of Verb Phrase projections below.

(73) Three Infinitive Verb Forms

The Infinitive"To" Form
The proto-type Infin(itive) Verb Form in English is the so called "to" infinitive which bears an infinitive {to} inflectional quasi-prefix marker (coming off a preceding MVP functional phrase): e.g., to walk/to study/to write, etc. Such infinitives are to be found in the second verb slot following the MVP. Consider some token infinitives below:

Token Infinitive {to} Verb: Structure: [MVPV1 + Infin VerbV2]
(a) I liked to walk.
        (b) John needs to study.
(c) I prefer to write.
  (a') like{d}     + to walk
   (b') need{s}   + to study
  (c') prefer{ø} + to write

One crucial underlying syntactic structure to note here is that all verb functional material is to be found in the V1 MVP slot, the V2 Infin Verb serves a sole lexical role. Any notion of Tense and Agreement of Person or Number (see §2.3 below) has to be attributed to the first Main Verb of each structure--in this case, to like, need, prefer (respectively). Consider below what the tree diagram would look like focusing in, for the time being, on the Infinitive Verb only. (See (97b) below for: (i) a separate treatment of subjectless infinitive clauses as derived sentences as well as for (ii) couplet [+Fin] & [-Fin] projections):

(75) Tree diagram of Infinitive {to} Verb:

  write ...

Since the [-Fin] marker {to} is a (functional) Inflection marker of sorts, and, as argued throughout, must be housed in a Functional Phrase, we include here a Functional Auxiliary node adjacent to the Lexical Verb. This Aux(iliary) node allows us to establish a place from which we can host and project functional material--via Inflection onto the verb. Similar to what was suggested above in (55) regarding the DP "Born & Deliver" (service) of Inflection, the MVP likewise goes through its own Inflectional process--albeit, a process that doesn't directly bear the moved inflected element or morpheme onto the stem--(like what we find with verb inflection of Tense: V+{ed} => walked). Nonetheless, the [-Fin] {to}marker is inflectional (as a quasi-prefix marker) and so does need to be housed in the verb's counterpart functional node. In one sense, unlike (55) above regarding the DP inflection, there is no reason to suspect any movement here--at least of the overt type, since the infinitive {to} marker remains detached from the stem (and never acts as a clitic e.g., * I need t'go/t'study/t'write...If anything, an argument could be made that it seems at times to act as clitic ("wanna" contraction) to the preceding first verb--as with the negative 'not' in Aux+Negative+Verb constructions: e.g., I don't work (Neg. do not) and I wanna work (Infin: 'want-to' work). (Also, see movement section below on the "wanna contraction").

The Infinitive" ing' Form
A second form of Infinitive Verb is the so called {ing} Infinitive since the inflectional {ing}marker is posed via movement onto the main verb stem. In this case, there is overt morphological movement inflection. Consider the tree diagram below noting that an "exchange" (substitution test) of inflections {ing} with {to}is possible:

(76) Tree diagram of Infinitive {ing} Verb:

walk -ing
  study -ing
  write -ing


Token Infinitive      {ing}Verb   /   Infin     {to} Substitution:
(a) I like walking in the park / I like to walk..
(b) Mary likes studying / Mary likes to study...
(c) John prefers writing at night / He prefers to write





The above {ing} Infinitive verbs are quite special and some generative grammarians opt to give them an extraordinary place in our grammar outside of its typical dual role: (viz., on one had, {ing}-verbs can serve as a progressive/imperfective suffix e.g., He is smoking a cigar, and on the other a DP-gerund e.g., [DP His smoking] is going to kill him.). In any event, the {ing}-verb form might not be viewed here as a "gerund" since a gerund is traditionally defined as a Verb that takes on the role of a Noun. There is no reason to believe that the above {ing} verbs cannot be interpreted as true verbs--as witnessed in our substitution test. (The typical driving force behind treating them as gerunds has to do with the fact that {ing}-verbs and their phrases adhere to another substitution test that permits an exchanged of a gerund by a pronoun it--e.g., I like 'reading this book' => I like 'it' (where it = reading this book). However, this test might be too strong in light of the fact that the same substitution seems to hold (for some linguists anyway) for the Infinitive {to} verb form as well (a non gerund)-- e.g., I love 'to dance' in the morning = ?!I love 'it' in the morning, where 'it' = 'to dance' .)

Moreover, there appear to be times when both inflections {to} and {ing} may surface simultaneously on/with the same verb (as in (77b)--e.g.,

(77) I am accustomed: (a). to sleep with the window down
  (b). to sleeping with the window down.






=> ...to sleep-ing... : [ {to}=> affix, {ing} => suffix]


Similarly, main verbs such as intend/suppose/believe, etc. can either be followed by an Infinitive {to} verb or an Infinitive {ing} verb with no change in meaning--e.g.,

(79) I intend: (a). to go to the meeting.
  (b). going to the meeting.

Moreover, much of what is behind the selective nature of the distribution seems to be affected by grammatical constraints: for instance, (i) Active vs. Passive voice, or (ii) Auxiliary vs. Main Verb. Consider the selective nature of the Main Verb advise and Aux/Main Verb need--e.g.,

(80) (a).  (i) John advised buying a house.
(Active voice)
         (ii) John advised to buy a house.  
  (b). John advised me *buying a house / to buy a house.
('me' object)
  (c). I was advised (by John) *buying a house / to buy a house.
(Passive voice)


(81) (a) Need we follow the course?
(need = Aux)
  (b) John advised to follow a house.
(need = Main Verb)

It would also seem that inherent properties of the two infinitive verb types can also affect the semantics of the preceding main verb--consider the following sentences:

The verb: <stop> (a) I stopped to talk to him => (stopped in order to talk)
  (b) I stopped talking to him => (quite talking)
The verb: <start> (c) I stopped to talk to him => (began conversation)
  (d) I stopped talking to him => (same as (c))

In (80b), it would not be too far off the mark to suppose that the pronoun Object-DP me in selects (specifies) an Infinitive {to}-verb as its complement (as opposed to the {ing}-verb) in the same way that the Aux verb in (81a) selects for a Bare Verb.

Notwithstanding our ungainly meandering through a myriad of gerund formations and definitions, in any event, let's consider such {ing} verbs for now as having the flavor of true verbs with their own unique distributional qualities and features and leave the debate on gerunds for the time being. Some aspects of this discussion will reappear in the section of Infinitive Verb Types mentioned below.

The Bare Infinitive Form
The third and final infinitive verb form to consider is the so called Bare Infinitive. It is referred to as 'Bare' simply due to the fact that it hosts no such overt inflection--these bare shape verb forms are what we find as verb entries in our dictionary: e.g., go, visit, listen, speak, eat, drink, sleep, return... Consider the tree diagram of such Bare [-Fin] Verbs:

(83) Tree diagram of Bare Infinitive Verb:


Token Bare Infinitive Sentences:
(a) I might walk in the park.
(b) Mary should study today.
(c) John can write at night.




As you might now be aware, there are some very specific distributions that accompany these three Infinitive Verb Forms. Let's flesh the distributions out in the paradigm below and then discuss them one at a time.

(84) Table: Infinitive Verb Distributions & Phrases

Infinitive Verb Form: Token sentence [Main Verb1 + [-fin] Verb2]:
(i) {to} Verbs
(ii) {ing} Verbs
(iii) Bare {ø}Verbs
I like/want/need/ *can to read this book.
I like/ *want/*need/*can reading this book.
I *like/*want/*need/ can read this book.
(iv) Infinitive with subject
(v) Infinitive without subject
(vi) Infinitive preceded by for
(vii) Infinitive without to
We believe him to be innocent.
We believe [X], [X] => He is innocent.
She wants to be alone.
They are eager for him to win.
We heard him open the door

First, the Infinitive {to} form along with the {ing} form seem to be able to cross over and serve as complements to the main verb like rendering (I like to read/reading). (NB. The distribution seems to hold with other verbs that share a common semantico-grammatical relation of belief, desire or volition: e.g., feel, believe, love, hate, prefer, etc.).

Second, whereas a verb like like can have as its complement either {ing} or {to} verb type, a verb like want cannot and must select a {to} infinitive verb.

Finally, modals (can, could, will, would, etc.), like all other structure-class words, can bear no inflection whatsoever (either as a prefix or suffix) and so must select a 'Bare infinitive verb'. Occasionally, Bare or {ing} forms (which are reduced from a Finite Main Verb carrying Tense) surface after an accusative pronoun in what could be analyzed as an embedded or small clause:

(85) (a) Did you see Mary/(her) leave/leaving the book on the desk?
        --> reduced from
        (b) Did you see Mary/(her) <(while)> She left the book on the desk?

All of these distributions as cited above--and they by no means make-up an exhausted list--are very interesting from a syntactic perspectives and reflect inherent semantico-grammatical relations internal to the verb's make-up (pertaining to sub-categorical features). One in a handful of such internal sub-categorical features helps to determiner whether or not a verb can select a certain Infinitive Form as its complement.

(86) Three Infinitive Verbs Types

In conjunction with the three elaborate infinitive verb forms presented above, let's briefly talk about the three Infinitive Verb Types.

Adjectival Infinitives
Adjectival Infinitives have the distinct quality of an adjective embedded in the form of an infinitive verb. One way to see through this adjectival quality is to note that they can be restated as relative clauses. Consider the examples given below:

(87) (a) John is building a table to fit into the corner  
  (b) John is building a table that will fit into the corner (relative clause).
  (c) John is building a "corner table" (Adj + N]

In (87a) above, the infinitive verb to fit is considered to have adjectival qualities in that it can modify the Noun that precedes it--i.e., it describes (as an adjective) the kind of table being built (a 'corner-table').

Adverbial Infinitives
Adverbial infinitives can either be turned into a Wh-question (by using why) or be paraphrased with in order to...

(88) (a) Mary first needs to have completed Grammar 302 to take this class.
  (b) In order to take this class, Mary needs to complete Grammar 302.
  (c) Why does Mary need to take this class?

Clearly, the above examples are not adjectival--one couldn't say a "completed Grammar 302 class" like one would be able to say a "corner table". These adverbial Infinitives speak more closely to the action or state of the verb than to the substantive qualities of the noun.

Nominal Infinitives
Nominal Infinitives are interesting because once again, our substitution test can make claims about the sort of modification taking place. Consider the token sentences below:

(89) (a) She wants to ski in Italy during her winter break.
  (b) She wants <x>, (<x> to ski, equates to the pronoun 'it' or something)
  (c) Mary likes to clean glasses at her work at the café
  (d) Mary doesn't mind doing 'it' (it = to clean glasses).

So, on top of the three Infinitive Forms talked about earlier, there are three Infinitive Types as well--all entering into very sophisticated distributional schemes that could only be accounted for by means of a very elaborate matrix of lexical sub-categorical features.

Having examined many of the roles Non-finite [-Fin] Infinitive verbs play in a sentence, let's now turn our attention to the Verb's functional counterpart---the Main Verb Phrase.

2.3 Main Verb Phrase MVP [+Fin]

The (functional) MVP plays a critical role in shaping the sentence. Above all, what we must realize is that the first verb in a SVO ordered sentence automatically takes on the responsibility of projecting functional features. Recall that with the Noun, the functional counterpart is the Determiner and it is the Determiner that spells out and ultimately projects all the functional features onto the Noun (via Inflectional) (see 55 above). And so the functional-to-lexical relationship holds between D & N. Well, regarding the Verb Phrase, the verb's functional counterpart is labeled as a MVP and, embedded in this MVP, the verb forces an Aux to surface in order to fulfill its own unique functional-to-lexical counterpart (Aux & V). So, using a certain syntactic terminology...

(a) What the functional D is to the lexical N    => [D + N],
(b) the functional Aux is to the lexical V         => [Aux + V].

But before we talk about the Aux, and flesh out all of its functional features in connection to the VP, let's spot exactly where the MVP occurs in a typical SVO sentence. As we have noted earlier in our introduction, English is an SVO language--meaning that the Main Verb comes after the DP-Subject, hence, SV (or more accurately, S+MVP). So, one of the easier things to remember here is that the first verb in an SVO sentence is going to be responsible for whatever functional material is going to be found for that sentence. Having said this however, there may be times when what appears to be some form of an Infinitive Verb [-Fin] first surfaces in the sequence of a sentence and you might ask yourself: what is going on here? Well, first consider some token sentences below showing such first verbs as Non-finite [-Fin] verbs (and not [+Fin] MVPs)--e.g.,

(91) Non-Finite Verbs
Moved element
(a) To be or not to be, that is the question.
( Infin Topic)
Vi, SiVO
(b) Studying all day makes Jack a very unhappy boy.
(Gerund Subj)
(c) To get a student loan, you must keep your grades up.
(Infin VP-move)
(d) Running for the bus, Mary fell and bruised her leg.
(elliptical clause)
*d.c., SVO
(e) Drink all day and night, I would never.
(Bare VP-move)
V, S*m
(* d.c. = depended clause.) (* m = modal)    



(a') The question is to be or not to be. SVO
(b') The studying makes Jack unhappy. SVO
(c') You must keep your grades up to get a student loan. SVO
(d') Mary fell while running for the bus =>
      (Mary fell while she was running for the bus)
(e') I would never drink all day and night. SVO

(NB. We have omitted here important distinctions between certain aspects of the predicate to facilitate illustration: e.g., Predicate Infinitive Phrase, Direct/Indirect Objects, complex sentences.)

What one immediately sees from the above examples in (91) is that they are not core un-derived sentences with basic SVO orders, but rather are instances where certain phrase constituencies have been uprooted and moved (or fronted) from out of the back of the sentence and into the front position--making it appear that we have lost altogether our core SVO structure. This up-rooting of phrases from the basic SVO order is called Movement (see movement in §4 below). There will be more to say about this later. For now, however, suffice it to say that whenever a core SVO order is maintained, the First Verb/Aux-Modal (shown in italics above) of that sequence serves its proper role in fulfilling the functional requirements of the sentence--it is these requirements as in terms of features that will be the topic of our next section.

Main Verb Features

Recall, as a recap, that the main functional features regarding the DP were Definiteness, Case, Number, Person, (Gender). Well, regarding the MVP the most important features that will be on the table here include the following: Agreement features (having to do with Person & Number), and Tense. There may be arguments for more elaborate MVP features such as Mood--e.g., imperative, indicative, subjunctive--as well as Aspect, etc., but for the time being, let's keep to the most obvious functional features. That's right--you've guessed it: there indeed is some overlap of D-features onto MVP features having to do with Agr(eement). By the way, that's what the term 'agreement' is all about--an agreeing relation of features between the DP and the MVP. The Agr-features between the DP-Subject and the MVP are Person and Number. Let's restate these features:

(93) Table: DP Person & Number: A Recap

1st [1P]
2nd [2P]
3rd [3P]



Recall, that (any and all) DPs can be restated as a reduced pronoun capturing all inherent features of that pronoun--actually, this is just another version of our beloved "substitution test"--e.g.,

(94) Reduced Pronoun Substitution

My body/self (=I), Your body/self (=You), The boy/girl/dog(= He/She/it),
John and I (=We), You guys/two (=You), The children (=They) , etc...

Let's see how a strict agreement relation holds features between a DP and a MVP:


(a) [DP We] [MVP *is/are linguistic students]
(b) [DP The children] [MVP *plays/play in the park]
(c) [DP John] [MVP *take/takes the exam]

In the above sentences, note that there exists a Person & Number Agr(eement) between DP-subjects and MVPs. For example, since We in (95a) has the DP features of [1P, +Pl] (first person, plural), the agreeing main verb must also cast the same features of [1P, +Pl] in order to save the projection (in this sense, "save" refers to keeping the projection grammatical). So how is it that we must project are and not is for sentence (95a)? In order to explain the correct verb choice here, we must have a parallel Person & Number paradigm for the Main Verb as well:

(96) Table: MV/Copular Person & Number

Number: Singular:   Plural:
Verb "write"
1st [1P]
2nd [2P]
3rd [3P]
(I) write-ø
(you) write-ø
(he/she) write-s
  (we) write-ø
(you) write-ø
(they) write-ø
Copular "be"
1st [1P]
2nd [2P]
3rd [3P]
(I) am
(you) are
(he/she/it) is
  (we) are
(you) are
(they) are

To make matters more concrete, the reason sentence (95a) doesn't grammatically project with the copular verb is (and must project the copular are) has to do with the DP features of We. As shown above, the DP We contains a [3P] feature along with a [+Pl] feature. The copular verb with the same corresponding (agreeing) features must be are--being that are also contains [3P] and [+Pl] MV features. Regarding Main Verbs, the only overt morphological inflection that surfaces due to Agr-features is the third person/singular/present tense {s}--e.g., she talk-s/ swim-s/ sleep-s/ work-s, etc. etc. (See Tense below for our brief discussion of {s} as a tense inflection marker). Note that all other verb forms take a verbal zero allomorph {ø} similar to what we find at times under the D. This zero allomorph allows us to maintain our inflectional 'delivery process' of functional to lexical despite the fact that no overt marker is realized in the phonology (cf., 83).

For a recap, the born & delivery service of Inflection found in (55 & 83) is restated here. (More will be said about the inflection of Pres(ent) Tense and Agreement later on). 97b below shows a Recursive VP Structure [+Fin] & [-Fin].

(97) Inflection: Born & Delivery



Let's follow this up more closely by diagramming this very interesting relationship between DP and MVP below:

(Note that we are only diagramming the DP and MVP Agr-features [Person & Number] of the sentence, excluding all other relevant Predicate material such as DP-Object, Preposition, or secondary Infinitive Verb).


In the above examples, the DP and MVP Agr-features of [3P, +Pl] must correspond--example *(e) contains a feature mismatch (termed Feature Crash) between the [3P,+Pl] DP (friends) and the [3P, [-Pl] MV (play-s) and thus cannot project. Even though the Person feature is preserved [3P], the number feature is misspelled causing a feature crash. Note that in (*e), the morpheme {s} shows up both on the D for plural (friends), and on the MV for [3P,-Pl Present] (play-s). (Don't forget that the morpheme {s} serves actually three distinct grammatical functions in English: (i) the possessive {'s}, (ii) the plural {s}, and (iii) the 3Person/singular/present Tense {s}. Although the S's may look (and sometimes sound alike), it is important to understand that they play very different grammatical roles).

Below, consider a proto-type [3P, -Pl] construction:

For further exercise, try to construct your own trees using the complete paradigm of Person & Number features (as presented above).


Along with Agreement of Person & Number--making-up the classic Subject Verb Agreement analogy--Tense is the next Main Verb feature in need of discussion. Unlike Agreement features that are shared by both Verb and Noun, the Tense feature is an exclusive Verb feature. However, there still is some debate over whether or not the verb {s} inflection marks only Present Tense or both T(ense) and Agr(eement) of [3P -Pl]. This debate is strictly theoretical in nature and need not concern us here--for ease of exposition, we shall take it that verbal {s} marks for both present tense and agreement. Consider the Tense/Agreement paradigm below--noting that the zero allomorph non-inflection{ø} surfaces in all but the 3P singular Present (where an overt {s} inflection surfaces):

Table: MV/Copular Tense & Agreement
E.g., Verbs "Walk" and Copular "Be"

Number     Singular     Plural
Person   present past   present past
(he / she)
(he / she / it)



(Note that we are using regular verbs here marked with past tense {-ed}. See section on irregular tense below).

In the table above, the two well known inflectional suffixes for Tense are established: the 3P Present {s} and the Past Tense {ed}. These two inflectional markers are motivated by a grammatical rule (a rule that is an abstraction of possible variables):

(102) The Grammatical Rules for Tense:

(a) Present Tense Rule = [V + {s}] for [3P -Pl]          --> e.g. She walk-s.
(b) Past Tense Rule = [V + {ed}] for all regular verbs --> e.g. She walk-ed.

The rules should be understood as an algebraic equation in the sense that a variable makes-up the classification of Verb--in theory, any word can insert into a verb slot, hence, rendering any word a verb. This was what was behind our early discussion on the fabricated novel word "Sib" found in (4). Also, the over-generalization of rules is witnessed by children acquiring their first language--e.g., He writed, falled, goed etc.

How many Tenses does English Have? This is an interesting question on a number of fronts. Firstly, it digresses our discussion back to the notion of Prescriptive vs. Descriptive linguistics--remember, we opted for the latter. For example, linguists shouldn't be too concerned about authoritative prescriptive argumentation that seeks to uncover language relics, say, of ancient Roman civilization and try to attribute them to modern English. In other words, we must be particularly careful not to take formal categories of Latin and attempt to apply them to modern English (as was practiced in 19th c. Grammar). English is not Latin (and it never was). (Recall our discussion of prescriptive Case (e.g., "It is I" vs. "It is me") in (51) above). Rather, what we must surely do is keep vigil in attending to empirical observations about the here and now of linguistic matters. Secondly, gut intuition about grammatically correct English sentences can't be nicely packaged and memorized from out of a leather-bound manual--language simply doesn't work that way (it is much too messy). In a very real sense, we are our own authors of our dictionaries, and besides, language is in constant flux. It is rather disheartening to find that there remains--in traditional (prescriptive) English Grammar classrooms--a seemingly endless regurgitation about how English has maintained their three proper grammatical Tenses. It has not! The English language, unlike its close relative "Romance" has only two Tenses--Present and Past (or more specifically, using binary notation [+/-Past] ).

Let's examine more closely this common misconception about a possible future tense for English. First of all, it is crucial to understand that grammatical Tense is a functional feature that can only be borne from an Aux and inflected onto a V. In other words, only main Verbs in English can carry Tense as an inflection. I'll even take it one step further: Only Main Verbs can establish Tense. The old story goes that English has Future tense via the modal "will"--as in the example, I will see you tomorrow. She will take the class next term. The president will talk on the economy this evening. etc. These are indeed all good English sentences, and there is a flavor of truth in that they constitute some notion of a Future time. However, notion of time and grammatical time are two very different entities. For instance, consider how readily accessible it is to say a mere lexical word such as Yesterday or Tomorrow (lexical/form class in that it has semantic meaning, as opposed to functional/structure class) without any discourse to Grammatical Tense on the main verb and still make out the notion that the event took place in the past or future (respectively)? (This is an aspect of speech common both in Child Language and Pidgin: e.g., Yesterday, I go there. Tomorrow, I talk to him). The argument that English has Future Tense relies on a confusion about (i) notion of referential time/tense as compared to (ii) grammatical time/tense--the former belonging to semantics (and perhaps pragmatics) while the latter exclusively belongs to the functional category of tense morphology [+/-Past]. The idea that modal verbs can project tense would mean that all modals would be able to project tense--as a category of 'part-of-speech'. Language is not structure independent, nor is it piece meal with isolating words taking on structure independent tasks. Rather, language is rule driven and if we attribute a class of words with a grammatical role, then we must look very carefully at assigning that role to the entire class of words. Morphologically, English has two tenses only--e.g.,

(103) He likes/ He liked, He takes/He took. [+/-Past]

It may be that other verbal categories assign other forms of time--aspect, progressive, perfect--notwithstanding these complex constructions formulated with Auxiliary verbs, tense is still reduced to being either present or past. (See Aux and their Grammar in the section below). As stated, if we establish the modal "will" as a future marker, then we ought to establish other modals as well--e.g., can, should, may, etc. There is no sense to be made in taking such modals as tensed: token elliptical sentences e.g., I should... today, I should...tomorrow, I should have... yesterday seem to skirt the full tense paradigm with little problem. Moreover, the proposed future modal "will" is mostly used for functions other than time reference. Consider the examples in (104) below:

(104) (a) I will come, if you want me to => (willingness)
  (b) She will (would) typically study all night => (habit)
  (c) (door bell) Right, that would be John => (probability / expectance)
  (d) We will surely all die one day => (truth)

Most important of all, there seems to be no shortage of ways to express the notion of 'future' even in full absence of a so-called proper tense inflection. Consider some creative means below:

(105) Example of future time reference without future tense

(a) "going to": I'm going to talk (tomorrow).
(b) Progressive: I am driving to Cambridge tomorrow morning.
(c) Simple Present: John starts work Monday. Mary speaks at noon.
We leave tonight. I am cooking for lunch. etc.

All of the above examples make it very hard to justify any real rule formation of a future tense in English. Quite the contrary, the examples show that grammatical present tense verbs can equally pertain to future references of time. On the other hand, just imagine trying to find a way out of saying past tense rule [add {ed}] for a regular past tense verb reference--e.g., *Yesterday, I walk/talk/visit... As you quickly discover, one can't simply break the add {ed} rule, and if one can (as in the examples above of future without "will"), then there was no rule to begin with.

Note that the example in (105a) can phonologically reduce to / / e.g., I'm gonna talk (tomorrow) => (I will talk). But also note that a seemingly similar progressive aux construction e.g., I'm going to class cannot be phonologically reduced in the same manner (* I'm gonna class). This might demonstrate that this "gonna" expression has developed into a quasi-mode marker of future time reference (but not grammatical tense). (Also, see "wanna" contraction in (236)). In addition, it seems that some Romance languages like Spanish, Italian, French can opt for a similar future marker using the verb "go". This is note worthy since the three aforementioned languages all have available proper future tense morphologies that inflect on main verbs.

A word needs to be offered here regarding general morphological inflection on Modals (tense or otherwise). In short, modals can't take inflection in any way, shape or form. Part of their unique intrinsic feature value (sub-categorization) is that they, as a class, (i) select to have no inflection onto their stems, and that they (ii) allow no inflected morphology on an adjacent stem positioned as their complement. (Recall, this was shown in table (84) above regarding infinitive verb distributions): e.g., (i) * She can-s, can-ed/ may-s, may-ed/, (ii) *She can "to" play/play-"ing" ... etc. There are a number of reasons for this--the main one being that modals are functional/structure class words, so if you attribute some reference of tense to them, that still doesn't buy you any meaning. And it goes without saying that Tense without a meaningful stem carries very little proposition worth. Having cleared up some common misconceptions about grammatical tense vs. referential tense, let's proceed in examining how one should go about drawing tree diagrams that incorporate the three features of T(ense)/Agr(eement) as presented above (see 108)--first pausing to take note of the inherent problems associated with Irregular Tense.

Irregular Tense

As an alternative to a "Rule-Based" grammatical tense, English affords a variety of means for establishing past tense. This second order course to grammatical tense is termed Irregular Past Tense since the regular rules are not being applied. In English, these irregular verb constructions for past tense come in a variety of forms. Consider the irregular constructions below (ordered from 'most common' in frequency to 'most rare'):


Irregular Grammar: Token sentence:
(a) Vowel Change sing > sang >, see >saw, take>took, write>wrote
(b) Word change be>was>were, go>went, bring>brought, buy>bought
(c) No change put>put, set>set, cut>cut, bid>bid, hit>hit, split>split

What is important to realize here is that even irregular construction must enter into an Inflectional process. The vowel change can be viewed as a form of inflectional morphology in itself--a kind of infix (where infix here means a morphological item internally put inside the stem as opposed to in front of (affix) or at the end of (suffix) e.g., sp-ea-k > sp-o-ke. (A similar vowel change process shows up in Spanish verb morphology for present vs. past--e.g., habl-o > habl-e). I one sense, we could postulate that the stem of "speak" (without a morphological infix) is the radical root V````sp-k.

At any rate, a similar process is at work in delivering irregular inflection. On the other hand, for the examples which include word change and/or no change, arguments for inflectional processes are harder to defend: it seems that such irregular formations derive neither from any patterning (as in sound patterning of e.g., sing>sang, ring>rang, drink>drank, etc.) nor from semi-rule construction, but are rather instances of lexical memorization--such constructions simply must be memorized. And as a note, children clearly prefer rules over memorization (both processes occupy different parts of the brain (see (2)). This is the reason we find such over-generalizations of regular rules for Tense and Number: e.g., I goed, I wented, I singed, I seed, I eated, I nevered, I have two tooths/ teeths, two fishes, two furnitures (=mass noun), etc.

(107) Inflection of Irregular Vowel Change

Features: [3P,-Pl +Past]
Irreg vowel changes (a) : { o }    wr-o-te, sp-o-ke
Irreg vowel changes (b) : { a }    s-a-ng, r-a-ng
Irreg vowel changes (c) : { u }    h-u-ng, r-u-n
Inflectional Proces           



Verbs: write speak sing ring
(a) John wrote / spoke...
(b) Mary sang / rang...
(c) They hung / run...







(108) MVP Tree diagram showing Present Tense Inflection

/      \
/ \
speak-s / believe-s
Inflectional Proces :                
Token Sentences: [-Past]
(a) She speak-s to him
(b) She believe-s in him








Again, it is ever so important to stress that it is the functional Aux(iliary) node of the MVP ( a [+Fin] Phrase) that generates Tense (as well as Agreement features) and then proceeds to carry the morphological element onto the lexical verb stem via inflection. Recall, that both D and Aux play similar roles in that they generate functional features for their respective lexical counterparts: [D to N, Aux to V]. As a recap, the abstract morphological Aux features thus far presented include (i) Person & Number features (Subject-Verb Agreement) as well as (ii) Tense. Later on in subsequent sections, we'll come to discuss an array of other Aux elements and features that can overtly project under Aux

(109) MVP Tree diagrams showing Past Tense Inflection

Although most of the time the Aux is "empty" of an overt phonological item--during such times as when it simply houses the abstract functional features of T/Agr (zero allomorph) as shown above--there are times, however, when the Aux slot will be occupied by either an actual Auxiliary Verb (=Do, Be, Have), Modal (=can, will, may, etc.) or a combination of both (e.g., I will/should/must be going).

The following section examines the Auxiliary and Modal Grammars.


2.3.1 just Auxiliaries, Modals and their Grammars

(110) Auxiliary Verb
The Auxiliary Verbs "Do-Be-Have" each play a particular role and have individual tasks in English Grammar. In addition to their specific grammatical tasks of Do--simple, Be--progressive/passive, Have--perfect, all three auxiliaries (as well as modals here) may serve out two general tasks of providing operations that include (i) supporting Question formation, and (ii) supporting Negation formation. Taking Question formations first, there are two types of question operations that need to be discussed:

(111) Questions

(a) (Yes-No Question): Does Mary like pizza?

(requires a Yes/No response)

(b) (Wh-Question): 'What' does Mary like to eat?

(requires a stated response to 'what'--e.g., pizza.)

Yes-No Questions
A "Yes-No" question is obviously defined by the way it elicits a response of 'Yes' or 'No' to the question.. If we first examine the kernel original word order of the 'Yes-No' question in (111a)--maintaining our basic (kernel) SVO order--we would find (112a') to read as follows:

Token Sentence Word order Grammar
(a') Mary does like pizza?
=> Kernel Interrogative.
(b) Does Mary like pizza?
=> Derived Interrogative
(c) Mary likes pizza.
=> (no "Dummy-Do" aux insertion)
* (v = aux 'do')   => Declarative (Non-interrogative)

Note that in order to form a proper 'Yes-No' Question in English, some movement has to ensue. (Also see the discussion of Movement in §4 below). Specifically, the moved element we are on about here is the Auxiliary "do". By restating the interrogative sentence in its original core SVO order, we quickly come to realize that in fact the Aux "do" has been up-rooted from its original sentence internal position (positioned after the Subject) and has subsequently been moved across the subject to a pre-subject frontal position--the precise movement is indicated and recorded with a trace-(t) subscript intended to mark (i) the moved element, and (ii) the position from which it was moved. It is especially note worthy here to mention that only functional structure-class Auxiliary verbs (and Modals) in English can undergo such abstract movement--the abstract qualities of the functional categories auxiliary and modal serve us well here in accounting for such possible movement. The fact that main verbs can't undergo movement operations creates a tidy account for the classifications of structure vs. form class words and such operations provide interesting theoretical analyses. However, not all languages abide by the same tidy parameters. It seems that it is enough for some languages to simply move the main verb from out of the second position (=SV) and into the front position rendering a Non-Auxiliary (V, S) word order--e.g.,

(a) Hablamos nosotros inglés bien? ( Speak we English well?)

(Do we speak English well?),

(b) Estudia usted las reglas? Spanish : (Study you the rules?)

(Do you study the rules?)

English, in any case, minimally requires an Auxiliary Do insertion (called "Dummy-Do" by linguists since it carries no real semantic meaning) to be positioned in front of the subject--creating the token sentences below:

(114) 'Dummy-Do' Insert for Questions and Movement "t" Trace marker

(a) Doesi Mary ti like Pizza? 't' stands for trace of a moved item.
(b) Mary "does" like pizza! (emphatic)  

Without this auxiliary movement (=Auxiliary Inversion), it is impossible to formulate a question operation. Consider the following improper outcomes when no Auxiliary is provided for a question operation: (a) *Like Pizza John? (b) *What John likes? Note that this same 'Yes-No' Question construction can be posed with the other two Auxiliary verbs--"Be" and "Have". (See (149) below for a summary table of Aux grammars):

(115) Yes-No Question with Aux. "Be" (Progressive)

Token sentence Word Order Grammar
(a) Is Mary eating pizza? (vSVO) => Interrogative /Progressive
(a') Mary is eating pizza. (SvVO) => Declarative/Progressive
(b) Are the students studying (vSV) => Interrogative/Progressive
(b') The students are studying. (SvV) => Declarative/Progressive

(116) Yes-No Question with Aux. "Have" (Perfect)

Token sentence Word Order Grammar
(a) Has Mary eaten pizza? (vSVO) => Interrogative /Perfect
(a') Mary has eaten pizza. (SvVO) => Declarative/Perfect
(b) Have the students studied? (vSVO) => Interrogative/Perfect
(b') The students have studied. (SvV) => Declarative/Perfect

Let's pause here to see just how we can accommodate an overt Auxiliary word into the MVP of our previous tree diagrams. (See appendix on tree diagrams).



Wh-Q(uestions) are exactly that--question that are formulated by a variety of Wh-words: (Who, What, Where, When, Why, Which, and How). The Wh-Question is identical to the above Yes-No formation except for the one additional element of the Wh-word. Whereas Yes-No questions entailed only one movement operation as such, with Wh-Q, we now entertain two movement operations which breach the SVO word pattern--(i) the Auxiliary Inversion (= Aux-move), and (ii) The Wh-word movement (= Wh-move). Let's consider below some token examples of Wh-Q by examining first their kernel (original word order), and then examine the moved elements making up the Wh-Q derivation.

(118) Wh-Questions using "Simple" Grammar: "Do"

Token sentence Word Order Grammar
(a) Mary does like what? (SvVO) => Kernel Interrogative
(b) What does Mary like ? (OvSV)* => Derived Interrogative
(c) Mary likes what?. (SVO) => (no "Dummy-Do" aux insertion)
* (v = aux 'do', O = wh-word) => Declarative (Non-interrogative)

(119) Wh-Question using Progressive Grammar: "Be"

Token sentence Word Order Grammar
(a) Mary is eating what? (SvVO) => Kernel Interrogative
(b) What is Mary eating? (OvSV) => Derived Interrogative
(c) Mary eats what? (SVO) => ( no Aux insert / simple grammar)

(120) Wh-Question using Perfect Grammar: "Have"

Token sentence Word Order Grammar
(a) Mary has eaten what? (SvVO) => Kernel Interrogative
(b) What has Mary eaten? (OvSV)* => Derived Interrogative
(c) Mary eats what? (SVO) => (no Aux insert / simple grammar)
(See (§4, 215) for Wh-diagramming)  

In addition to Question Operation, the three Auxiliary Verbs also enter into a second functional operation having to do with Negation. The Negation operation in English must involve the Auxiliaries--unlike other languages like e.g., Spanish which can negate the Main Verb of a sentence with a simple pre-verbal "no"/"non". Compare the two languages English vs. Spanish with regards to Negation operation:

(121) Negation
English Negation vs Spanish Negation  
(a) I do not/don't like...   (a') No me gusta... (= no me like)
(b) They do not buy...   (b') No compran nada... (= (They) no buy nothing)
(c) Don't open them!   (c') No los abran ustedes (= (You) no open them!)
(d) We don't study much   (d') No estudiamos mucho (= (We) no study much)

Consider what would go wrong if, in English, we followed the Spanish functional grammar of Negation:

(122) Examples of Erroneous Negation without Aux

(a) *I no like pizza. (c) *No open the door!
(b) *They no buy books. (d) *We no study much.

The above examples indicate that English has a functional grammar--a grammar whose parameters cannot be violated--that follows from a particular Rule:

(123) English Negation Rule => [Subject + Auxiliary + "not" + MVP]

Whereas "do" is the Aux element par excellence functioning in most cases exclusive to simple grammar, the other two Aux verbs also can negate--dependent upon whether or not their specific grammars are engaged (e.g., Progressive/Passive and Perfect). Consider below examples of negative formations formed by auxiliaries "Be" and "Have":


Aux Be: Progressive Negation Passive Negation
(a) John isn't writing a book. (c) The book wasn't written by him.
(b) Mary is not reading the paper. (d) The paper was not read by her.


Aux Have: Progressive Negation Past Perfect
(a) The students have not prepared. (c) John hadn't smoked for years.
(b) She has not slept well. (d) Mary had not been well.

In the following section, let's turn our attention to the specific grammars associated with each Auxiliary Verb--making up the Simple, Progressive, Passive and Present/Past Perfect Grammars.

(126) Auxiliary Grammar: The "Do-Be-Have's" of English Grammar

(127) Do
The Auxiliary "Do" is responsible for carrying out the Questions and Negation formations as presented above in the absence of any other complex grammar. In other words, in the absence of the complex grammars Progressive/Passive/Perfect, the Aux. "do", by default, inserts into the simple grammar construction in order to instigate the Question/Negation operation. That's is what is behind the term "Dummy-Do": "do" does nothing in the way of any real grammar--unlike its Be/Have counterparts--and basically inserts as a quasi filler for no more important reason than to separate the Subject from the negative element "not". This line of reasoning, by extension, could also account for a seemingly adjacency condition placed on Wh-word operations as well--specifically speaking, "do" could also be seen as inserting between the Wh-word and the Subject. (It may be that English subjects simply do not like to sit in adjacent juxtaposition to the lexical operator items.

(128) Adjacency Condition of operator words:

Wh-words Negetive "not"
(a) What does he like? (b) He does not like math.
(a') *What he likes? (b') *He not likes math.

So, in a real sense, the Aux Do's main role is to preserve the surface order of--


(i) [Wh-word + Aux + Subject + MVP] (for Wh-formation), and
(ii) [Subject + Aux + Not +MVP] (for Negation formation).

With regards to a possible Adjacency Condition as sited above, let's pause for a moment to reflect on some interesting clitic formations having to do with the Negative Question clitic {n't} moved along by Aux word inversion.

(130) Negative Question

Clitic {n't} [+Movement] vs. Non clitic "not" [-Movement] *Non-standard
(a) Don't you speak French? (a') Do you not speak French? ? a" *Do not you..?
(b) Can't he work alone? (b') Can he not work alone? b" *Can not he...?
(c) Won't she come? (c') Will she not come? c" *Will not she...?
(d) Isn't he working? (d') Is he not working? d" * Is not he...?
(e) Haven't they (already) eaten? (e') Have they not eaten e" *Have not they..?

As the above examples show, there also seems to be an Adjacency Condition placed on clitics. (A clitic is defined here as a word-like element (morpheme) having the property that it must attach itself (cliticize) to another word. Also, see "wanna" contraction as a clitic in (236) below). It also must be noted that clitics are deficient on at least two grammatical levels: (i) at the Phonological Level, the clitic {n't} cannot maintain its full phonological infrastructure (syllabic stress, etc.), and (ii) at the Morphological Level, the clitic has certainly lost its lexical form--e.g., {n't} cannot stand alone as a free morpheme and be uttered in isolation and still maintain its lexical meaning (as opposed to "not"). For instance, in the clitic "will + not", the form changes to "won't"--clearly, this apparent change of stem from will to won is not a productive stem and is rather viewed as a phonological adaption. Conversely, sometimes inherent grammatical considerations work in shaping the outcomes of clitics--as in the "gonna" contraction sited above in our discussions of Tense (105).

(131) Be
The Auxiliary Be, outside of its scope in formulating question and negation, has the task of providing (i) the Progressive Aspect as well as (ii) the Passive Voice. Rules and tree diagrams to the two functional grammars are stated below:

(132) Progressive Rule è [Be + Verb + {ing}]

(a) John is running on the track. [is + run + {ing} ]
(b) I was preparing for the track meet. [was + prepare + {ing} ]
(c) The freshmen were using the long jump. [were + use + {ing} ]
(d) I am looking at our times [ am + look + {ing} ]

(133) Tree diagrams showing Auxiliary: Progressive

(The tree above has been reduced of all its formal features only showing the Aux progressive grammar procedure)

Note that the Inflectional process of moving the {-ing} Bound Morpheme onto the Verb Stem invokes the identical process of Inflection presented throughout. The Aux node now reflects all the components of the functional Progressive Grammar: Be and {ing}. Also note that while Be (here, inflected in its finite form of "is/was" ) stays put under the Aux node--since it is the actual Aux word--the bound morpheme {ing} is obliged to "move" from out of its mother Aux-node and position onto the main verb stem via inflection. Recall, that since the {ing} morpheme is bound, by definition it must attach onto another word (much like what happens to clitics). As a rule, only Lexical Form-Class Words (i.e., Nouns and Main Verbs) can take Inflection. Of course, all other features apply as well for the structure regarding Tense and Agreement (T/Agr).