Introduction :
Neither in exact nor in social sciences does there exist any recipe guaranteeing discovery. This is what I used to say   to my doctorants   at the beginning of every academic year . Consequently one might wonder how one can TRAIN young researchers in the first place ! All you can do is orientate your " thésards " towards what you consider to be insightful readings and guide them through itineraries of research which have proved fruitful  in the past. Of course your own itinerary can be of help. What I am going to do here is present the genesis and development of my linguistic thinking from the beginning of the seventies until the present day (1999) - almost twenty years.
The question that I shall be trying to answer is the following one : how did the metaoperational approach come to exist ?
As far as I can  remember, there have been six stages  in the genesis of my linguistic thought . Here they are.

"To be sure means that when the right solution is reached, everything
falls into place".

      Hercule Poirot in
The Clocks
(A. Christie).

FIRST STAGE  :  An invariant for BE+ING

The crucial step in my research on " BE+ING in the Grammar of Contemporary English " was the discovery of THE SCOPE OF -ING : the marker -ING  operates on the WHOLE verbal group and not on the verb ALONE. At first I proposed the label " invisible parenthesising " to account for the extraordinary cohesion of VERB + COMPLEMENTATION. This analysis  shows clearly that the minimal pair
I leave tomorrow /  I am leaving tomorrow
is a " trompe-l'œil "  and that it must remain opaque if you deal with it linearly as follows :
-     I                       leave                           tomorrow
2 -    I                    am leaving                      tomorrow
The analysis which sheds light on the functioning of BE+ING must be the following :
3 -     
I                       leave                          tomorrow
4 -     I                        am                       leaving tomorrow
It appears that the status of " tomorrow " is not the same in (4) as in (3) : in (3) " tomorrow " belongs to AN OPEN PARADIGM : one could have said " tonight " or " on Monday " instead. The situation is completely different in (4)  where the choice of " tomorrow " is blocked (CLOSED PARADIGM) because  the scope of -ING is the complex verb LEAVE TOMORROW and not the verb alone . Arithmetically speaking what we have is :
-ING (leave tomorrow)
I claim that this simple reanalysis - which puts an end to three centuries of " progressive form " -  has the following consequences :

  1. The BE+ING utterance is made up of a BINARY RELATION :  I  //  leave tomorrow . This will be the case for ALL BE+ING UTTERANCES. The operator BE links the two members of the underlying predicative relation and permits us to date that relation (its validity) . Thus
  •    I was leaving tomorrow but now I won't.
is perfectly grammatical whereas
  • * I left tomorrow
is ungrammatical.
b) The orientation of the utterance is not the same in (3) and (4) : in (4) the target of the -ING predicate is the grammatical subject (which in this case happens to be the subject of the enunciation) : it is clear that this EXPLICITS the meaning of (4), which can be a way of apologising, as all "  anecdotal grammars  "  of English  will claim. As to (3) it is orientated to the right, that is to say on the date which results from the paradigmatic choice.
  1. One can see the pernicious character of a label such as SVO (Subject-Verb-Object) which occults the analysis of  (4) proposed above. The same applies to Chomsky's tree corresponding to the derivation of (4), which cannot but be identical to that of (3), for how can you signal the status of " tomorrow " in a tree configuration ? All utterances of the SVO type will have to be analysed either as TERNARY (" John plays football but Peter plays tennis "), or as BINARY , that is : S(V+O)  as in " John is playing football ". Not surprisingly I have recently discovered that Turkish displays the same pattern.
  2. The French equivalent of both (3) and (4) is : " je pars demain ". The grammar of English teaches us that this French utterance can be analysed in two different ways, either along the lines of (3) or of (4) :
je       -       pars         -       demain .
(4')   je       -       pars demain.
These two analyses  will help us to understand the difference between:
(3'')  :
Il               partit                 le lendemain (he left the next day).
and (4'') :
Il               partait le lendemain (he was leaving the next day).
  1. My analysis enables us to account for utterances which have resisted the traditional approaches which had -ING bear on the verb alone. Neither Martin Joos (" The English Verb " 1964) nor  Ronald Langacker ("  Foundations of Cognitive Grammar " 1987) have succeeded in explaining  the difference between the two following utterances :                   
Mary resembles her mother               
Mary  is resembling her mother more and more.
In (5) " mother " belongs to an open paradigm : there were other possible choices, for instance " father ", to say the least, whereas in (6) the operator -ING applies to the complex verb " resemble her mother " and it is this nominalized predicate (the  resemblance to the mother)  which is being quantified by " more and more ".
Thirty-five years ago, Martin Joos thought he had found the "single meaning" of be-ing he had been looking for. Here is what he writes in his remarkable review of Akira Ota's doctoral dissertation: "Tense and Aspect in Present-day American English" (Language, Vol. 40, Number 3 (Part 1), July-September 1964, PP. 487-98): "I think I have found (after two decades of trial-and-error) a single meaning for be-ing…My answer is something we have been approaching gropingly, tentatively, almost blindly, for many years..The second-last step gave us Twaddell's "limited duration" or limitation of duration" for a better way of saying this.The last step is the recognition that the duration in question is not duration of the event, the action, the deed; that instead it is duration of
the validity of the predication ".
Well, this is perhaps a step forward  in comparison with the traditional viewpoint which  always refers to  "the action expressed by the verb" but unfortunately  Joos sticks to the notion of "duration" or "limitation of the duration" and so  remains faithful to the classical pattern of explanation, based on intuitive meaning ( moreover no definition of "predication" is to be found). Be it as it may ,  Joos's paper  contains  another interesting point  for researchers in the grammar of BE+ING. This is what he writes immediately after the lines I have quoted above :"Perhaps this is the end of the approach; the only remaining difficulty is offered by a minor idiomatic phenomenon: the fact that we can say "she's always bothering me, where always, forever, eternally  etc. seem to contradict the meaning of "limitation". But Joos sweeps this objection aside and gives the following gloss of the annoying always- utterance :
"At any epoch it is for a limited time true that she bothers me".
Now Joos's "minor idiomatic phenomenon" DOES contradict the limitation hypothesis ! The weakness of the offered gloss is, to say the least, twofold : on the one hand , it concerns an  out-of-context utterance (but this is a common practice at the time); on the other hand, nothing is said about the scope of ALWAYS. The following example  shows more clearly the problem that is at stake:
  1. (nurse to director) : Mrs. Smith says she has seen a ghost.
  2. Oh, well, Mrs. Smith is always seeing ghosts !
My example shows  the crucial importance of context :  here the binary relation "Mrs. Smith / see ghosts" is the target of   the strongly- stressed modal adverb, hence the "irritability" meaning  proposed by traditional grammarians. Here are a few examples of utterances of a similar kind:
I 'm continually forgetting people's names!
He is  perpetually quarrelling with his wife!                                                     
This girl is always knowing something she isn't supposed to !

Biographical Notes     Linguistic Research: Books and Papers     
The Linguistic Credo of Metaoperational Theory     
The Genesis of Metaoperational Grammar
The Double Keyboard of Grammars     
The 10 Components of the Metaoperational Model