WHAT IN FACT A GRAMMAR IS

© 2002 by Orchid Land Publications

C.-J. N. Bailey

[20020614]    

i

     I believe that hordes of authors of grammar books and even more hordes of teachers of grammar do not know what grammar is.  It certainly is not a list of rules.   Grammar consists rather of STRUCTURES consisting of items related in certain ways discussed below.  If this is not understood, one cannot "do grammar."   The kinds of items and structures that exist are very few; when explained properly, they are quite easy to grasp.  But even the very young can intuit the difference between "Cats hate mice" and "Mice hate cats" and between "good people" and "bad people" or "move quickly" and "move slowly."  One needs simply to lead a youngster into understanding how the parts related to one another--and what they are called.  One needs to discover what the RELATIONS in questions are and which ELEMENTS--PARTS OF SPEECH--can function in each given relationship.    
     Grammar may include word-formation and, in some languages, inflectional morphology.  The latter is not part of English grammar except with respect to forming regular participles, and notably the irregular anterior participles.  See the Excursus on morphology at the end. 

ADMONITION TO THE GRAMMAR TEACHER

      At the very beginning, one has to ask oneself two questions:
--Does one view the students as sponges (or parrots); or does one regard them as on a voyage of discovery, deducing or inferring the system one is teaching when properly guided and prompted?
--Is one going to approach grammar the way a literature-teacher approaches literature--feeding information to be remembered for an examination, as well as, with some kind of luck and only for a few, some concept of taste and style?  Or is one taking two or three building blocks and showing the students how to use them to build up a system of relations among those elements by simple steps?

     Next, a word on terminology--where it is important and where it is not.  Terminology is not an end in itself; it is a means of varying utility.  The only reason for the existence of terminology is the technical is the avoidance of confusion--a purpose violated by the inept terminology that abounds much more in English grammars than in, say, Romance grammars.  How can the term "continuous tense" be adequate for was standing if a continuous thing like what we read  in "Troy was standing 600 years" is ungrammatical?  What follows uses the traditional terminology for the parts of speech (some of which are grouped as modifiers, predicators, or functors; subject, object, indirect object, predicate complement).  Where the writer prefers as more transparent term, it is often used along with (and separated by a slash) the traditional but less useful term; e.g.  nominal/substantive, operator/functor (though functor has a slightly wider sense that includes connectors/conjunctions . . . along with a very few quasilogical symbols like &  and = that should pose no problem.  The term force is the only one used in language studies. I distinguish reality force from ontic force (see deontic and epistemic force below). 
     What is presented here is simpler than traditional presentations; terms inappropriate for English like case are avoided, and modality is the cover-all term for verb forms (tense, mode, voice).  The building blocks of grammar are shown to be hardly be small and simple to grasp.  What is presented here is not a linguistic analysis in the sense of what is taught in linguistic courses, where the object is quite different from that of teach children grammar.
     If an instructor does not know what a term means, one should either look it up in the dictionary or see what is said here; the object is to learn, not to memorize.  If one cannot build up a system from its elements, one should teach oneself before trying to teach children.  What follows here shows how this is done.  If one cannot do this, one would be better off teaching some other aspect of English.
    Two extremes are to be avoided.  Stressing authority or intuition.  A grammatical construct is right or wrong not because some academy decrees it (English is fortunate in having no such academy, but some textbook-authors try to fill the void); nor is it correct because someone does it.  Arbitrarily inverting "Cats hate mice" to "Mice hate cats" won't mean anything.   The marginal value of rules is discussed below.

    Past efforts to teach grammar have failed for some obvious reasons, aside from presenting grammar as a set of unexplained rules that parrots or sponges are supposed to absorb and be able to apply, rather than as an intellectual adventure.  Past efforts have failed because of the inevitably failed enterprise of squeezing English grammar (and its terminology) into the mould of Latin or Anglo-Saxon or any other language with a very different system from that of English.  The French (and therefore Romance grammarians) have done much better--not balking at double negatives, distinguishing definite from indefinite pasts (rather than miscalling either a "perfect"--that is "completive"), understanding the way pronouns change forms without changing functions (e.g. subject, or object) according to context, etc.  English should be approached similarly.  Incidentally, grammars don't include epicene they, them, their (meaning "she or he," etc.) among the personal pronouns, or mention impersonal one.

    Grammar-teaching is doomed to failure when a teacher cannot understand the failure of past efforts, the failure of past terminology, the failure of teaching grammar as rules rather than as intelligible structures.  If a teacher cannot make the paradigm-shift, then the outlook is parlous.  Einstein said that insanity is doing the same thing over and over and expecting a different result.
     CLICK HERE for a sample teacher's grammar (of about than forty pages) laying out how to help pupils build up the grammar of English
on their own in eighteen lessons--one per week.  The system is based on Bailey, Essays on time-based linguistic analysis (Oxford University Press, 1996), being a non-theoretical and pedagogical translation of that theory based on the writer's long experience in teaching the subject--and taking care to try to head off the violations of the system that users of English are sometimes or often prone to commit.  This approach is not for teachers unable to think out of the box or those unable to connect the nodes.  Fortunately, normal children lack these impediments.   Many authors of grammar books go astray because they labor under the demonstrably false assumption that English is a Germanic language--a lineal descendent of Anglo-Saxon.  They  need to learn how the various daughter languages of English have been born.  
     The complexity of many languages lies in the inflectional morphology; in English, it lies in system of syntax.  With the four building blocks on two dimensions, it is easy to  build up the very complex grammar of English. 


     Grammatical structures of more than one word are called CONSTRUCTS."  An independent or main clause is formed with (i) a subject/object, (ii) a predicator (finite verb), and optionally  (iii) a complement--which may be an object or a nominal/substantive or adjectival predicate complement.   Each of these three items may be modified by a modifier.  (See below on appositives.)  That is what a grammar is.  Subjects and objects have to be nominals/substantives--names for things, including abstract things and acts--but may be single words or entire subordinate/embedded clauses.   Predicate complements are heard in:  “It was Sue” “It was Sue’s” “It seemed good”  “I feel good”  “It smells bad” (“It smells badly” would refer to, say, an animal that, say, had lost its ability to smell any but the strongest odors.)”  The verbs that link a subjects with their predicate complements are copulas or copular verbs.  An example of an infinitive is “She wanted them to be happy”; with a participle, “He heard of their being happy.”  Examples of copular verbs are be, become, seem, feel, taste, smell.  Finite verbs are distinguished from verbids or non-finite verbs:  Infinitives can be  nominals (substantives) or modifiers--adjectives or adverbs); gerunds (which are nominals--in effect, nouns); and participles (which are adjectives).  (See boxes below on their reality and deontic/epistemic force.)  Like the predicator of a main clause, these  may be preceded by have (in which case the verb is changed to an anterior participle like seen, gone, won, slept).  A verbal item preceded by have indicates an act or state that has existed in a block of time not ending before the time of the statement.  

    "She wrote that article."  (past:  in a block of time ending before now)
    "He has written a new article."  (anterior:  in a block of time not   
         ending before now.

    With not yet and already, one uses the anterior form. Just takes either, since "We've just done it" is now often (and more logically) "We just did it."  But "We already did it" is as un-good as is "We didn't do it yet."  Among the priot-time forms of English finite verbs are these:  has seen, had seen (see below on had've seen in hypothesis clauses), saw, was seeing, useta see, would see (which is characterizing).  See more below on would see.

     Here are the past forms of see (no passives, only actives):

saw 
was/were seeing (progressive--not "continuous")
have seen, have been seeing 
had seen, had been seeing
useta see (habitual-iterative)
would see (characterizing, as in "She would quickly see the point of a   
      logical argument even when she was young.")

     Posterior finite verb forms are preceded by a form of be plus going to; verbids are preceded by [be] going to; e.g. 
     "They were going to eat early last night."  [a past posterior]
     "Having been going to eat early, they stopped playing before sundown."
    "Their having been caught doing that caused an uproar."  [Here, doing is like a predicate complement, i.e. "having been caught" is one of the various passive verbs that can be treated as a copular verb like be, become, seem, smell, sound," etc.]

    To see how an infinitive is used as a predicator, consider these examples of a nominal and a modifier (adverbial) infinitive construct:

     FOR THEM TO DO THAT would create chaos.  (nominal)
     The order was FOR THEM TO STAY. (nominal)
     The commander ordered THEM TO STAY.  (nominal with virtual for)
     It was easy FOR THEM TO DO THAT.  (modifying adverbial)

In "I saw them doing it," doing does not have to be regarded as a predicator, though in some instances it seems that it could be.  In other places, it can be taken simply as a modifying participle with its own object.

      Note that any verb form--not least, any compound verb form--is referred to with the neutral term MODALITY.  The Latinate terms mode, tense, voice (Greek diathesis is also used and would be preferable) and so on do not suit English grammar, which has modal verbs (discussed below) requiring, as already observed, deontic/epistemic force.

     Grammars err is calling the exochronous modality a “present tense,” the anterior modalities “perfect tenses,”  and progressive modalities “continuous past/present.”  As some of the proofs are tech­nical, I will simply pose quandaries that call this nomenclature into question—taking them up in reverse order.

     If “was standing” or “useta stand” were a “continuous tense,” why can we not say “Troy was standing there 600 years”?
     If “has done” were a perfect (completive) modality, how could we say “She’s always done that and probably will keep on doing it for some time.”  Moreover, how is such a grammarian to explain why we cannot say, “He has done it at noon,” whereas “He has always done it at noon” is unexceptionable?
    If “is” is a present, what is the difference between “The child is good” and “The child is being good” or between “That person seems to be clever” and “That person is seeming to be clever”?  If it were a tense, why does it have the same reversals in marked (non-default contexts) that we find in anterior and posterior modalities?
    Till a grammarian deals with these issues, don’t believe him or her.  Till one can say why the exochronous mode is unmarked for copular and modal verbs as well as conceptual and perceptual verbs—but not for other verbs—that grammarian doesn’t have much going for him or her.

   The present-anterior is dying out (deservedly in “have just left” but not very felicitously in “already did” or “didn’t yet”).  But speakers should not replace this verb form with the past in “always true/untrue” or posterior contexts.   “We’ll know what was dis­cussed” can be used only if the discussion is a past fact; if not—i.e. if one does not know the discussion has occurred, one should say “We’ll know what has been said when it is discussed.”
     What is the difference?  A past is used for an event or state occurrent in a block of time ending before now; a present-anterior is used for an event or state occurring in a block of time last up till the time of the statement—“now”—whether it has just ended or is still going on.

ii

      If a grammar is a system of relations among words, a construct is an item longer than a single word, in short a prepositional phrase, a main or subordinate clause, or an infinitive construct (see examples above).   Aside from  appositives, exclamations (including vocatives like "Hey Jan!"), and sentence modifiers (including absolute constructs not attached to any part of a sentence with a conjunction), there are only two kinds of grammatical relations:  

(A) There is a horizontal relation of PREDICATION, i.e. of subject to predicator to (optional) complement--either a (direct) object or so-called predicate complement--the nominal in "That's me" and sufficient in "It seems sufficient").  An indirect object has a virtual to or for (e.g. "They gave [to] us a book"); an imperative predicator has a virtual or real you as its subject.    

A subject or object is a nominal/substantive--a word (noun, pronoun, gerund, infinitive, articular adjective [e.g. singular the good, plural the young].  A predicator is a verb, which may be compounded (e.g. had been going to have been getting watched--hardly a "tense."  Verbid predicators have been illustrated.

A predicator is normally a verb--often comprised of a compound consisting of one or more auxiliary verbs plus a verbid--viz. a durative participle (seeing) or an anterior participle (seen).   The verbid is an infinitive in "I am to be on hand tomorrow."  Note in connection with "We saw them do it" and "We commanded them to do it," there is a virtual for before them; there is a virtual to before do in the first of the preceding examples.  Cf. "That caused her to worry" with virtual for and overt to.  It is a major problem of grammar to know why and when examples like the preceding are not overtly parallel with "We made plans for them to do it,"  and "It was easy for them to do it" with overt for and to.  See below for more on operators/functors real and virtual.   One generalization that is possible is that modifying infinitive constructs having a subject take both overt for and overt to.  Cf. "eager for them to be on hand" and "easily enough for them to be harmed by it."

An active predication is one in which the subject does something; a passive predication is one is which the subject is acted on.  There are also contraponents in which an overtly active verb has passive import:

Dogs  -->  eat dogfood.
Worms <-- get eaten [by birds--a phrase modifying get eaten].
This book <-- translates easily.

(B) There is a vertical relation of MODIFICATION, which can be symbolized by an arrow pointing from the modifying word, phrase, or clause UP TO what is modified.  (Sentence modifiers might best be placed before a sentence, like an absolute, on which see below.)  Adjectives modify nominals/substantives; adverbs modify anything else.  (Most intensifier adverbs lack the suffix -ly.  Contrast "It was most agreeable" and "It was mostly agreeable"; "It was clear out of reach" and "It was clearly out of reach."  Other examples include hard : hardly, just : sure : surely, awful : awfully, justly, bloody : bloodily, right : rightly, pretty : prettily.)
    Besides (A) and (B), there are (C) auxiliary verbs mentioned above, there are connectors/ conjunctions and operators/functors.  There are connectors representing the logical symbols for "and" and "or,"  negatives like not and never and the compound negative unless (=  "if not)."  (In the principle of inverting subject and verb when not begins a sentence, Hardly and Only are clear negatives that trigger the rule, as in "Only twice did that happen" and "Hardly two minutes ago did they arrive.")   

Note that in Latin, Germanic, and Slavic languages the word for "but" is like and; and the word for than is a conjunction.  In Romance languages and English, but is a subordinating conjunction and than is a preposition--which may in fact be followed by a nominal or nominal phrase or clause, as in "easier than to try to do it" and "easier than going that job."  Like the Romance languages, English has compound prepositions like in place of, instead of, on behalf of, for the sake of, out of, outside of, on top of, etc.; cf. out of, off of, etc. (Cf. all of, which in many languages would not usually have anything corresponding to of).  But sometimes a prepositional phrase is embedded in--i.e. serves as the object of--a preposition; e.g. from inside of the house.  An example of a nominal clause functioning as the object of a preposition is "from where we started."
      The symbol = can be used to connect a nominal and its appositive in what in fact is a compound nominal; e.g. <Elizabeth = the Queen> and <Queen = Elizabeth>.  Other compound nouns would be "bread and butter" and "bread or cereal."  
     Subordinating conjunctions created modifier clauses.  But WH-conjunctions (why, when(ever), where(ever), how(ever) can act like whatever, whoever, whichever in forming nominal clauses).  They can also create adverbial clauses like "I'll find it wherever it may be."  In "Whoever wishes to do that," whoever acts like "anyone who"; the relative clauses here are restrictive, whereas the clause beginning with wherever in the foregoing can  either restrictive or (with a preceding comma) on-restrictive.  WH clauses are often indirect questions, as in "He asked/wondered where/why that happened."
      Non-WH subordinating conjunctions are more frequent functors for forming subordinate clauses:  causal because, since, and their many synonyms; concessive (al)though and even though; hypothetical if and its synonyms (on condition that, supposing that) and negative forms (less); nominalizing that and the "(in order) that" which, like "so that" and old-fashioned lest (= in order that . . . not), form (irrealis)  purpose clauses, the that that forms (realis) result clauses, etc.  (Irrealis and realis force are explained in full later.  Past and present are realis; anterior, exochronous [e.g. works, is, has, seems] are irrealis.  The nine posteriors of English  fall into tree groups of marked, doubly marked, and triply marked forms; see Bailey, Essays on time-based linguistic analysis (Oxford University Press, 1996); see "posterior modality" in the index.                       

     The two most important kinds of predictor force in English are those of reality and those of will-and-mind.  They mainly apply to modal verbs, discussed below, and infinitives.  One should think of a scale:

--factual--realis (not factual, not irrealis)--irrealis--

in which irrealis is what is not desiderated or unexpected (dubitative or counterfactual; there are three degrees of irreality in English posteriors in main clause and three degrees of irreality in hypothesis clauses in conditional sentences.  In fact, irrealis can be subdivided into dubious, very unlikely, and counterfactual.  

     Two parameters of reality cross in modal verbs.  One parameter involves the poles of necessity and possibility.    (In modal logic, these are negatively related by a well-known formula of modal logic to the effect that what is necessary is not impossible and what is possible is not unnecessary.)  The modal character of necessity and possibility is each subdivided into what belongs to the will--deontic force, what can be (a command) or what is possible (i.e. permitted)--and th'other belongs to the mind and indeed to being itself (ontology):  Epistemic force involves what must be and what is merely possible in the real world, aside from what is desired or expected.  Negated, epistemic force involves what cannot be or what is not possible. These aspects of modals will be discussed in the latter part of section iv below.
      Force can be illustrated by noting that that deontic may is permissive (or, when negated, prohibitive; these are acts of a will) or epistemic (referring to the possibility of an occurrence or state).  Force will be dealt with in detail in the latter part of section iv below. 

     It is interesting that of the four uses of should, only one is factual rather than irrealis.  When a clause with should depends on an expression that has plus or minus (not neutral, i.e. x) values of the feature [desiderative] or [expectative], should represents a fact:

"We were (not) disgusted that they should be doing that."
"We were (not) surprised that they should be doing that."

    Note the irrealis verb forms in purpose clauses:

"She worked hard so that she might/would be able to afford a long       
     vacation."
"He worked hard in order to be able to afford a long vacation."

      Note the realis verb forms in result clauses:

"We worked so hard that we were able to afford a long vacation."


     (C) Absolutes like "That being the case, . . . " are not connected to the item they modify by a real conjunction; they could be considered to be constructs beginning with a virtual subordinating conjunction--a virtual since [or "seeing that"} in the foregoing example, represented here by S:  "S that is the case, we will do what we've got to do."  There are sentence adverbs modifying entire clause; they could be placed before the clause with a brace pointing to the clause:  "Incidentally } that is untrue."

     (D) Operators/functors include: 
(i) prepositions and the virtual or real for and to in infinitive constructs; 
(ii) determiners--the articles, demonstratives (this, that) and quantifiers that precede nouns in English; also 's (see above);  
(iii) note that that is the subordinate conjunction that determines a noun clause and an adverbial or object purpose clause (cf. "We did it so/in order that . . ." and "Make sure that . . . ") as well as an adverbial result clause (". . . so much that it turned out right.")
and perhaps (iii) more, most, less, least with modifiers (adjectives, adverbs).  It is omissible in object noun clauses but not in subject noun clauses like "That the facts are does and so won't cause me to alter my course" or in result clauses.

     English nominals do not have cases like Latin and the Germanic and Slavic languages.  This will have been evident in the analysis of 's above.  Though that analysis did not specially involve pronouns, time was when whose was who's and its was it's.  But case analysis will not account for educated usage in all English-speaking lands today.

    For the way in which personal-pronoun forms get reversed in marked (non-default) environments in English--a regularization of what is found in the Romance languages, comparable with calquing adverbial -ment(e) with -like = today's -ly--see the index of Essays on time-based linguistic analysis already cited), under reversals in marked contexts.  The reversals in question affect these pairs of personal pronouns (theoretically it:it, you:you, and y'all:y'all get reversed, and the same is true of French elle:elle, vous:vous, nous:nous):

her:she, him:he, them:they, me:I, us:we

     The marked environments where the reversals take place are when a pronoun is conjoined with another or with a noun (and or or stands between them), or when a pronoun is followed by an appositive or a modifying phrase or clause.  Educated people who would never say "I saw she" say "I saw [or "gave it to"] she and he."  They say "My father and myself [or less-educated "me"] did it," even though they would not say "Me did it."  (But I caught myself saying to my Dalmatian "You and me are buddies.")  Examples from the British Queen and Prime Minister are citable including two of the following::  "to we British," "with we in Europe," "for we who are able," "on they in trouble"
     A double reversal takes place in the marked absolute at the beginning of:  "They being tired, we decided to stop there for lunch."

iii

      Grammar books fail to realize that processes can function like non-processual grammatical phenomena.  Cf. the temporal throw-back in counterfactual hypothesis clauses illustrated in a box later on.  There are at least a dozen such grammatical processes in English.  There can be mentioned here also the inversion of the unmarked (default) order of subject and verb to created normal (not indirect) questions like "Did they do it?"--in which negatives require do-support, i.e. do-insertion.  There is the difference between epistemic may in "They may be on time" (expressing a possibility) and a special form of deontic may in "May they be on time!" (expressing a hope, wish, or prayer; deontic may normally permits something or, if negated, forbids something).  Other processes include deletions that leave virtual operators like deleted to in short infinitives and indirect objects.   There is deletable should represented by $ in "I insisted that he $ not hang around there"; this is allowed but not required in expressions of volition and necessity and in those of importance and propriety.  It offers the only explanation for why hang is not hangs, why there is no do-support for not, and why the sequence of time forms is not preserved in the subordinate clause.  Another process deletes If in "Should they leave; Had they left; Were they present now."   Note that inverted subject-verb order is required here, as in several of the deletion processes.  Speaking of deletions make structures apparent; they are necessary for any rational explication of the phenomena in question.  They are anything but mere artifices.
     Operators/functors can be symbolized as follows:  If nominal words or nomnal/substantive clauses (these begin with real or [if such a clause is a direct object] virtual that) are placed in square boxes the operator can be placed first and set off by a broken line from the rest.  We would have in such a square <the | person> and <that | we are happy>. On the right after such a broken line can be placed 's--which may be attached to a fairly complicated nominal like "the one I spoke to's [home town]."  If modifiers are placed in circles, the same can be done for a preposition.  Likewise for a subordinating conjunction in a subordinating modifier clause.  To the right of a single adverb can be set off -ly when it is appropriate.  If verbs are placed in big diamonds, a small diamond attached on the left point can include real or virtual to for an infinitive and the auxiliaries of finite verbs.  A small diamond attached to the right point can be used for a postverb like out in watch out.   

     Some grammatical phenomena can be explained on the basis of pronunciation phenomena.   The "v"  at the end of unstressed of and 've gets lost:  The weak vowel prevents the total loss of of in lotsa books, but the assimilation of //v// to /b/ in we'b been and the subsequent simplification of /b b/ to one [b] in I been entails the disappearance of 've.  But what of "If I'da been" or "If I hadda been"--the double past-anterior in English hypothesis clauses?  (German and French also have double past-anteriors.)   To say (in normal tempos),  "If we'd been," would--through the assimilatory process heard in "we[ve] been"--become "If we[b] been."  Though this would be tolerable in most contexts, it is hardly tolerable in a (very "marked") counterfactual hypothesis clause.   There is no reason for the double past-anterior in writing, except in drama and novels emulating real usage; a British writer writes "had have"--but this never occurs in my hearing and nullifies the grounds for the development of the double past-anterior.  In connection with constructions caused by pronunciation problems, note "There's three books here"--a kind of expression used by Shakespeare.   There is a problem in pronouncing "there're" (cf. trying to say mirror).  We do not Is there with plural subjects in interrogations.  
     Virtual items are variously symbolized.  The absence of an element can generally be symbolized with zero, usually annotated with a \ (back-slash) through it.  Deleted should (in a context of volition and necessity or of importance or propriety) can be symbolized with $.  Deleted to (with an infinitive or indirect object) can be symbolized as T; cf. also "They went T home."  TH symbolizes a deleted  that, as in "We thought [that] they were ready."  A deleted preposition other than to can be symbolized as P.  In articular adjectives, one or thing can be symbolized with O.  The absence of an object with a transitive verb can be written X, as in "He was the one that punished [or reprimanded] X"; it sounds better here when revised to read "did the punishing/reprimanding."  U stands for a deleted you in imperatives and prohibitions (negative imperatives)  like "Don't U go there"; you is not deleted in emphatic imperatives.   L and M can stand for an absent let  or may in "Be that as it may."    (See below on may in inverted and normal word order.)   In this connection, it may be noted that Let's is now an operator (not a true verb) in "Let's you help up" and "Let's you and me do that."   In the expression in that, (not exactly equivalent to inasmuch as), one could parse the sentence, "He is lucky in that his opponent had to withdraw from that race," as  in/with respect to the fact that"--technically annotated as in R that.  Note that WH is not a symbol for a virtual (deleted) item; it stands for any interrogative or relative pronoun, conjunction, or modifier.  

iv

     Copular verbs include uninflected modal verbs or modals, which are uninflected..  All modals exist in the exochronous ( Copular verbs include uninflected modal verbs or modals, which are uninflected..  All modals exist in the exochronous (timeless) modality:  must, ought, should, will, can, may, and shall (this is rare except in polite  interrogative invitations or suggests like "Shall I/we help?"; also for a strong promise or threat, as in "I shall return").  Some of the foregoing are also paired with a corresponding conditional form used also as past-posterior time-forms in indirect quotations depending on a past predication (and in marked situations outside of indirect discourse).  (see the list in a box near the end):  Will is paired with would, can with could, and may with might; the latter form in each pair is the conditional/past- posterior form.  For modals followed by the short infinitive beginning with have, see below. All modals are found in the exochronous (timeless) modality:  must, ought, should, will, can, may, and shall (this modal is rare except in polite interrogative invitations or suggests like "Shall I/we help?"; also for a strong promise or threat, as in "I shall return").  Some of the foregoing are also paired with a corresponding conditional form used also as past-posterior time-forms in indirect quotations depending on a past predication (and in marked situations outside of indirect discourse).  (see the list in a box near the end):  Will is paired with would, can with could, and may with might; the latter form in each pair is the conditional/past-posterior form.  For modals followed by the short infinitive beginning with have, see below.

     The overall class including modals is that of auxiliary verbs or auxiliaries.  These all take a short infinitive (one not preceded by to).  The auxiliaries be and get form passives, whereas be going to, will, and be to form three of the nine posterior modalities.  With these and conceptual-perceptual verbs (e.g. think, imagine, consider and hear, see, etc.) differ from other verbs in that their default (unmarked) form is exochronous (i.e timeless; e.g. thinks, hears), whereas the default form of other verbs is the present form—is speaking, etc.  So-called performatives like “I declare” or “I pronounce” are only exochronous in form.  But the marked form of the first group of special verbs is present—is seeming, is being, is feeling, is hearing.  Most of these require a hearer to consider a special circumstance; thus, is seeming  may mean “pretending, giving out airs of being”; is being refers to acting or behaving; is feeling may denote a state of mind; and is hearing may be followed by “noises” and refer to a deranged state.   On the other hand, speaks is exochronous and marked, referring to no specific time; e.g.  “speaksof ” may have a generic sense of “refers to.”  Compare these uses:   “In this scene, Hamlet speaks to Ophelia”; “His train arrives at ten tomorrow” or “When he addresses the meeting tomorrow, . . .” 

  

     A systematic pattern offers analogies that can be taken advantage of for getting details right.  For example:
--To understand attributive proven and predicative proved, one can use them where rotten and rotted would respectively be right.  Proven would not be used as a verb (one shouldn't say "they have proven it") or as a participle following a noun (e.g. "a theory proved to be defective").  Only in the position before a noun (e.g. "a proven theory") is proven used as an adjective.
--Instead of preventative, one can think of inventive and get preventive right.
--Instead of violating the //t//-deletion rule in most consonant clusters, one can say often like soften (cf. hasten, moisten, listen, etc.).
--Uneducated confusions of may with might may require more than just comparing them with uses of will : would and can : could.   One may have to explain that after a past, one uses might, whereas both may and might can be used as non-pasts--the difference being that might indicates doubtfulness whereas may entertains a real possibility.

          Foreigners have trouble with will or be gonna in clauses (along with whatever/whoever clauses) subordinate to a main clause having a posterior predication.   where we use the exochronous.  Foreigners' problems with using until for by are legendary (see the analysis in the book cited above, pp. 187-191.  Even native-speakers of English could not do them correctly if reversals in marked (in this instance, negative) situations did not belong to the mechanisms of the human nervous system.  It is easy to see why someone corrected for saying lay intransitively (it is really a contraponent) will then use lie transitively.  Hawai'ian Creole uses no for English not. It is easy to see why a recent development has nat for no--at least in contradicting a statement or in replying to a question.

      Instead of rules, intelligible principles--which serve as explanations--are needing for analysing and teaching grammar--especially English grammar laboring under a the false axioms of Latin and Germanic (Anglo-Saxon) grammar.  (It resembles Anglo-Saxon only by having 28 per cent of its vocabulary from that [Germanic] language.  None of its structures is Anglo-Saxon beyond what is common to languages in general.  Anglo-Saxon words are used to calque (translate) Old French functors, as shown in other pages on this site.  
     Analogies usually make a system more systematic: but the humorous or jocular spice for spouses does, as it happens, the reverse!

THE LIMITED VALUE OF RULES

    A rule that makes no sense and is simply stored as brute memory has little to do with understanding grammar.  Only rarely is a rule-of-thumb helpful.  We have seen a few examples using the analogies of proven with rotten and of soften with often.  But unless the basis of a rule is explained, it won't carry a native-speaker of English far.  If one makes a rule against using a plain past with already or not yet, one should explain why.  The Q of N rule is useful even for native-speakers English:  We say  "a bunch of books are" but "a bunch of grass is," as well as "a lot of books are" and "lotsa sugar comes from sugarcane."  The rule is:  When the quantifier is not one or a single and is followed by a singular or plural noun serving as the subject of a clause, the verb is singular or plural in agreement with the object of the preposition of; when Q is one or a single, the verb is third-person singular.   (Cf. further "Many a person lives alone.")  If with foreigners, we make a rule forbidding be gonna or will in neutral hypothesis, relative, and temporal clauses after a posterior main verb, that can be of help . . . especially if we recognize that "If it arrives tomorrow, . . . " does not use arrives the way we use it in a main clause:  "Her train arrives at noon tomorrow."  

PRINCIPLES GENERALIZED FROM (INDUCED FROM) A PATTERN

     Conditional sentences consist of hypothesis and a contingency that is true when the hypothesis is true; unlike other subordinate clauses, the marked position of a hypothesis clause is to precede the main clause that it is embedded in; of course, the marked order is allowed (without affecting the literal sense), as is the case with other subordinated/embedded clauses.  The basic principles for the form of predicators in subordinate clauses depending on a verb with posterior force (these include imperatives; see the lower cells of this table) using as a modal hypothesis clauses of conditional sentences:

A hypothesis clause (beginning with if, unless, lest, and synonyms of if like Supposing that, etc. can be 
     (a) if past or present or timeless (exochronous):  neutral or counterfactual  
     (b) if posterior (not past or present or timeless):  either neutral or unexpected (dubitative)—or occasionally, expectative.
Past neutral:  “If it happened that way, . . .” 
Past counterfactual:   “If it had happened that way, . . .”   (Many 
     are now saying “If it would’ve happened that way, . . .”)

Models

Present neutral:  “If it is working now, . . .”
Present counterfactual:  “If it were working now, . . .”      

    An intelligent child can deduce for oneself that counterfactuals are made by pushing the time back a degree.  Yet, a common error is to use a past neutral for a past counterfactual:

    “If it was working” or even  “If it were working” for “If it had been ready.”    In a PRESENT counterfactual, was may be changed to were; this is obligatory with if-deletion and a change of word order:  “Were it working now, . . . “  (­If-deletion is restricted to uses with were, should, and had.) 

    Presents and posteriors are more complicated than pasts, since they have a dubitative/unexpectative category, though both obey the time-shift principle of the other hypothesis clauses.  Neutral hypotheses (like temporal and relative clauses in posterior time) use an exochronous or “was going to” form; and dubitative/ unexpectative hypotheses us some kind of past time-form: 

Neutral:   “If it works/is working [“now” or] tomorrow, . . .”
                  “If it was going to (be) work(ing) last week, . . . ”

Dubitative/unexpectative:  
     “If it should (be) work(ing) [“now” or] tomorrow, . . .”
     “If it worked tomorrow, . . . “
     “If it were to work tomorrow, . . . “  (very dubious) 

Posterior hypotheses also have a factual variety:
     “If it’s gonna work tomorrow” or “If it’ll work tomorrow”—where the latter has implications of greater certainty than those of the former form.  

NOTE that, conversely, in main clauses, be gonna is less definite than will, except when the predication is negated or truly interrogated— not in a rhetorical question, an indirect question, or a tag question (which reverses the negatively of what goes before, as in “They aren’t gonna win, are they?” and “They won’t win, will they?”

    All of this can reduced to a few principles by a systematic mind—without which one should not be writing or teaching grammatical subjects.  Trying to set up patterns and rules without the principles is futile—the main reason for the failure of grammars.

     What grammars known to you explain the SYSTEMATICITY of posteriors?
                                Verb                           Conjunction

Neutral                hope, etc.            if, when, etc., although
Unexpectative     question              whether, how 
Expectative        know, promise,
                               expect                 because, even though

     Concerning imperatives (the direct kind--not the "Let's you and I do it" sort, where Let's is an operator rather than a true verb), consider "Make sure that it stays in place" and "Make sure that it will be done on time."   Consider also the dubitative/non-expectative alternative:  "Hasten the process that would bring out the desired goal."   Note here a deviation from the system found in subordinate clauses that depend on non-imperative contingency clause with a posterior predicate:  "If it turned out--or should turn out-- to be possible, they would find out a way to do it"--with would in the main or contingency clause but not in the hypothesis subordinate clause.  Compare what we find in temporal and relative and nominal subordinate clauses to a main clause with posterior predicate of dubitative/unexpectative force:

--I would be there when(ever) you needed me.
--I would be there whenever you should need me.
--I would give it to the one you designated.
--I would give it to whoever you designated/should (might) designate.   

Clearly, only the generic clauses with wh . . . ever permit the should alternative; evidently, grammar quite logically regards hypothesis clauses as generic (gnomic).  Like hypothesis clauses, none of the foregoing allows would in the manner of object nominal purpose clauses like "Hasten the process that would bring out the desired goal."

      Is the overall pattern statable in summary form and amenable to an explanation?     If it is, we need a way of distinguishing object purpose clauses from other subordinate clauses in posterior time.  To forward our investigation, it is worth notice that in past time, purpose clauses--which look to the future--can take would or might:

--She worked long hours in order that she might/would have enough for a nice vacation trip.

This leads us to wonder where the object purpose clause--which is adverbial rather than nominal like the immediately preceding purpose clause--might allow might: "Hasten the process that might bring out the desired goal."  The foregoing seems acceptable, though might is less good that would, just as in the adverbial example in past time the converse holds.
     To arrive at a sustainable generalization that correctly predicts the system of educated usage, we can say that it is necessary to separate purpose clauses embedded in a clause with a past redicate looking to the future from other subordinate clauses embedded in higher-level clauses with predicates having future import:

Except for purpose clauses embedded in clauses having a past predicate, subordinate clauses embedded in higher-level clauses with predicates (including imperatives) having posterior import take  exochronous verb forms if neutral.  If dubitative, those that are not purpose clauses allow a past verb and, if generic/gnomic (which includes hypotheses) also should.  If expectative, all subordinate clauses except purpose clauses embedded in a clause with a past predicate allow be gonna or, with stronger expectative assumptions, will.  

This can be tested by seeing whether it correctly predicts that the following is good English:  "Hasten the process that will bring out the desired goal."  This is indeed good English.
     It will take some time for a non-native speaker to internalize this involved set of generalizations--most of which educated native- speakers do not have to be taught, as they have internalized the generalizations from childhood.

If one cannot see the interrelations of the materials in the cells of the above box after they have been pointed out, one has not arrived at an adequate understanding of this vital aspect of  posterior expressions of English.  Such systematic knowledge is much more important that the mainly classificatory differences among count nouns, mass nouns, abstract nouns, and collective nouns--knowledge necessary for knowing when to use few instead of less, a distinction increasingly ignored for linguistically valid reasons--and the determination of which of the four categories is the unmarked use for a given noun--together with the British plural treatment of certain collectives viewed as "teams."  The overall system should be clear in a teacher's mind and presented in such a way that the system can be deduced by a moderately bright student.  To thwart the adventure of discovering the interconnections by laying down a rule goes against the whole concept of becoming educated.

     Major complications for a foreigner learning English (but not so much for a native-speaker) is when a preposition or postverb (the out in work out) or preposition must, can, or cannot be foregrounded to the front of its clause—e.g. “the person I talked to.”  In normal usage, whom would sound strange:  "the person with whom I talked."  There are no few instances where foregrounding the preposition is disallowed, as a page on this website shows.

      Note that English has no genitival “case,” since odor ’s can be attached, respectively, to the front or end of the foregoing, as in “the person I talked to’s book.”  In fact, English (like the Romance) languages lacks pronoun cases, since the forms function differently in these languages from cases in other well-known Indo-European languages.  See the pages on this website concerning form-reversals in marked (non-default) environments.  Whom is rare in educated English.  Many errors occur in using it where it is wrong (even by grammar columnists); but I have never heard an error caused by not using it.

The only use of rules, once insight into structures and functors has been acquired, is to avoid commonplace errors; e.g. “Use proven only in a context where you would use rotten”; “Use might only where would fit—for a past-posterior, a doubtful posterior, or a doubtful present” . . . or “Often rhymes with soften because both have the same sound change.”  A RULE WHOSE RATIONALE IS NOT EVIDENT IS WORSE THAN NOTHING.  A rationale is necessary; e.g. the backward time-shift for indicating irrealis (dubitative or counterfactual) force.  For grammar is neither a set of pre­scrip­tions (rules) nor “what sounds right” but structures that exist in the brains of speakers of the type of English (or any other language) that one aims to teach.

It has been noted earlier that English distinguish the "features" of [desideration] and [expectation] (where non-expectation can be subdivided into dubitative, very unlikely, and counterfactual values) relevant to analysing English grammar.  Force is an aspect of modal verbs neglected in many school grammars. There are two kinds of force in English ....... and reality force, which will now be discussed.

     Reality force is relevant to analysing modal verbs.   “It must be finished on time” (a command) or “It must be there somewhere” (a logical deduction).  Should and ought express a weaker force of deontic force.  Some uses of must, should, and ought can express epistemic likelihood, as in "It must be in there; it's gotta be there"; "Well, it ought to be there."  There are two kinds of necessity and possibility--epistemic,, which has nothing to do with a will; and deontic, which involves a will that commands or permits.  There are likewise two kinds of possibility; deontic possibility, when not negated, is expressed by deontic can or epistemic may.  (Note scope of not with may: it negates what follows deontic may; it negates epistemic may itself; cf. “They may not stay with me, as I'm too busy to put up with them” vs. “There may or may not be staying with me; they haven't told me.”  Can has, in addition to the two uses, another, so-called “root” (the term has hardly any meaning; it’s just a tag) use. 

Root can = “be able”:   “We can win if we work hard.”
Deontic (permissive) can:  “You can sit here.”
Epistemic (logical) can”:  “It can be true.”

     A quirk or two occurs when can is negated negation (see p. 234, note, on the book by Bailey cited above):   F. R. Palmer observed that, since there is a difference is the scope of “-n’t” in can’t (the negative negates can:  can't = "be unable") and permissive may not, there ensues an odd consequence.  Although permissive "Can I leave now?" and May I leave now?" are similar except in the latter's  being more deferential, their similarity disappears when they are negated and epistemic:  "John may not be in his office" does not rule out the possibility of John's being in his office the way "John can't be ion his office" does.  As Palmer says, the former allows the possibility of John's being in his office, whereas the latter is not open to such a possibility.  This is another part of English grammar that adult native-speakers and children of any sort do not need to be taught.  But a teacher of grammar ought to know about it.  When not is stressed in can not (written as two words), the sense is that "is able not to . . ."

     The modal verbs (“modals”) are:  must, ought, will/would, can/could, may/might, should—also shall, now used only for and polite suggestions in interrogative form (“Shall I/we do it now?”) or for insistent promises (“I/we shall return!”) or insistent admonitions (“You [or “they”] shall finish it by tomorrow--or else . . . ”).  

  Note that none of the four uses of should are no longer conditional or other variants of shall.  A good advanced grammar will show when irrealis would and could can be used as pasts.  (A good class exercise would be for the class to work these out on the pupils’ own.) 

     Would can be a past-posterior in indirect quotations; but in a main clause, was/were going to is used.  Could is used in indirect quotations but is preferred to was/were able in independent (not embedded) clauses when could is negated or truly interrogated.  The different kinds of “non true” interrogation are discussed on a sub-page of this website.   Since modal verbs generally lack past forms, other expressions like had to or was obligated to for must are used; must have, should have, and ought to have are also available where appropriate. 

Will/would have three non-posterior--…--and should has four uses, of which one is factual (see above), one is equivalent to ought (to), and the others are conditional and dubitative (see the box dealing with conditionals above). The three non-posterior uses of will and would are:

     Volitional:  “be willing, insist on, persist in”; e.g. “They will/would keep arguing.”  Cf. the more metaphorical “That key won’t/wouldn’t work in my lock.”

     Characterizing; e.g. “He will always answer like that” and “She would sit there for hours like that”—which is more marked than the unmarked “She useta sit there for hours like that” (also consuetudinary, but not characterizing).  

     Expectative or assumptive or epistemic use in “They will’ve arrived by now” and “They would’ve been missed by then.”  Would occurs in this usage only in indirect quotations; “We thought that he would be finished by then.”

      It needs to be observed in somewhat more advance studies of modal verbs that transvestite modals exist.  Under negation or when truly interrogated (not in rhetorical, indirect, tag, and other questions), need, dare, and partially ought (and even help) can be quasi-modals and act like modals, rather than taking a long infinitive (one preceded by to) and having the negative placed before them with do-support:
-- “They needn’t/daren’t do that.” and “They didn’t need/dare to do that.”
-- “Need/dare we go there?”  and “Do we need/dare to go there?”
(Note the different SCOPE of –n’t in mustn’t and needn’t.)
    In addition, one should be aware that “has (got) to” is a non-modal that may be used for must or should.  Cf. “be able” for can and others like be sposeta, be liketa, be bounda.  Logically—and especially in books for foreigners—one should write useta, hafta/hadda/hasta (and ‘ve gotta and wanna) for their (non-clause-final) quasi-modal senses.  This not only distinguishes the pronunciations from non-modal-like usages; it would also prevent ungrammati­cal separations of these units by foreigners like “They used in those days to swim in this creek” and “She had yesterday to go to his office.”  British but not American speakers can have a past of the full form:  “had got to.”  There is not corresponding infinitive, participle, or gerund.  We avoid the contractions clause finally; we say “(didn’t) use to,” “(don’t) have to,” “have got to” and “(don’t) want to” at the end of a clause.

    Consuetudinary-be (as in “I always be myself in such circumstances,” etc.) is dealt with on this website.  

    And what of punctuation?  I know that some of the best publishers have given up on it, but it is confusing to leave a space in a compound of object+verbal noun like tool-using/er or demon-worship(er).  It is confusing to leave a space between double modifiers of a noun (or of anything else) like sweet-scented oil and nice-smelling dogfood--though we don't do it with proper names like North American habitat.  Perhaps a hyphen can be allowed in Eastern-Orthodox mentality.

EXCURSUS ON MORPHOLOGY

     Some of the latter distinguish attributive and predicative forms—e.g.   drunken : drunk, rotten or swollen : swelled : rotted, burnt : burned, blessëd : bless(e)d.  Isolated differences occur in comments below on distraught :  dis­tracted, fraught : freighted, wrought : worked and wreak (which more often has the form wreaked) and bereft : be­reaved.  (Cf. cloven and cleft as participles of cleave; and stricken and struck as semantically differentiated participles of strike.  All shook up doesn’t mean what all shaken up means.)   Gerunds and attributive present participles of the structures illustrated in off(e)ring, lab(e)ling, soft(e)ning, and foll(o)wing lose the middle syllable when pronounced at normal tempos; the middle syllable is retained in predicative uses.  Gamboling is usually trisyllabic.  Note how   -ling without a preceding “e” (as in tabling) exhibits a “u” when the “l” is followed by a vowel (other than in suffixed -ing), as in tabulating, tabular, etc.  It would be more logical to write tabul, fabul, particul, vehicul, muscul (cf. muscular), articul (cf. articulate), etc.  Contrast circul : circular with cyle : cyclic and syllable : syllabic.  (Tabloid is an ignorant invention that violates the pattern heard tuberculoid.)

     With a few exceptions (dealt with by me elsewhere), we postpose predicative adjectives.  Consider the theory proved by Einstein:   We cannot say “the proved by Einstein theory.”   Where German would say “the by two metres longer table,” French and English say “the table longer by two metres.”  All three languages can say “a much longer table,” where longer is attributive . . . like rotten in “a rotten apple” and proven in “a proven theory.”

The other kind morphology—word-formation—is normally neglected in school grammars.  This explains the malformations (cf. tabloid for tabuloid, like celluloid) in our time.  A good example is systemic for systematic and mispronunciations galore like pre-sentation for pres~entation (where ~ represents a syllable boundary).  The correct pattern is usually hear in rec~reation and re-creation.

Hits on this website from 19981122
 till 20050917:  426,797