Letter-writers Of The Romantic Age

Transcript

1 LETTER-WRITERS OF THE ROMANTIC AGE AND THE MODERNIZATION OF ENGLISH (A quantitative historical survey of the progressive ) René Arnaud Université Paris VII-Denis Diderot AVANT-PROPOS L’étude microdiachronique présentée ici s’appuie sur des relevés qui ont pris une douzaine d’années, pour se terminer en 1994 *. La fin du siècle a vu exploser la technologie et l’usage des ordinateurs, faisant passer les études de corpus de l’artisanat à l’âge industriel *. Il n’est pas certain, cependant, que tout soit bénéfice dans ce progrès matériel: le scanner, par exemple, ne fait qu’enregistrer des signes, alors que le relevé manuel favorise les réflexions sur le pourquoi et le comment de l’écriture. La lenteur n’est pas sans mérites. Les biographes, ou les historiens littéraires, sont enclins à oublier les formes; le linguiste froid risque d’oublier les auteurs, leur vie et leurs émotions, dont le texte est un témoin vivant. Le changement linguistique ne saurait être ni pure inertie ni pure spontanéité: le mouvement psychologique et social s’appuie sur les virtualités de la structure, mais au service de sa propre téléonomie. La présente étude accorde certes une importance particulière à cet extérieur de la langue saussurienne: elle n’exige pas cependant qu’on oublie la part d’autonomie qui définit celle-ci *. Description et analyse d’un changement majeur qui, plus que d’autres sans doute, conduit l’anglais britannique vers ce qu’il est encore aujourd’hui, cette étude s’adresse en premier lieu aux spécialistes du langage; mais ceux qui s’intéressent à la la littérature et au contexte social de l’époque romantique pourront, s’ils veulent bien franchir un instant une frontière disciplinaire, y trouver eux-aussi quelques renseignements et sujets de réflexion: le titre même les y invite. Je tiens à remercier tous ceux qui ont eu leur part dans cette entreprise: Antoine Culioli, pour avoir dirigé ma thèse de 1972, William Labov, pour avoir encouragé cette étude du changement et participé activement à la mise au point finale, Gillian Sankoff pour son aide. Que soient aussi remerciés mes collègues et amis scandinaves, qui ont si largement contribué à promouvoir l’étude diachonique de corpus anglais, pour m’avoir fait l’honneur de m’accueillir au Symposium de 1982, et pour les fructueux contacts qui ont suivi: ceci s’adresse tout particulièrement à Terttu Nevalainen, Mats Rydén, Magnus Ljung et Erik Smitterberg. Je remercie également David Denison pour son aide et ses suggestions, Anthony Warner pour d’utiles remarques, Elizabeth Heliotis, Claude Rivière, Paul Veyriras, pour leur lecture attentive de la version préparatoire et leurs conseils, ainsi que mes collègues et amis de l’Institut d’Anglais Charles V (Université Paris VII-Denis Diderot). Janvier 2003. —————————— * Base de données ROMPROG : environ 22 000 citations codées. *La faible capacité de mon ordinateur individuel m’a contraint pendant longtemps à diviser cette BD en 22 fichiers dont le volume ne dépasse pas 5 Mo : un poids-plume aujourd’hui ! * Sur la complémentarité des approches, cf.: Rissanen, Matti: On Writing a History of English: Variation and Change, 9-12. [Avec une liste des principaux corpora historiques de l’anglais ]. The European English Messenger, VII, Spring 1999. Kastovsky, Dieter: On Writing a History of English: The “ Local ” and the “ Global ”, 13-15. The European English Messenger, VII, Spring 1999. 2 CONTENTS Introduction CHAPTER ONE. OBJECT AND METHODS I.1. General outline I.2. The progressive paradigm I.3.The coding frame I.3.1. Questioning the data I.3.2. Situation of utterance I.3.3. Linguistic factors I.3.3.a. Semantic classes I.3.3.b. Verb form I.3.3.c. Subject I.3.3.d. Object complement I.3.3.e. Time complement I.3.3.f. Duration I.3.3.g. Observations I.3.4. Sociolinguistic factors I.3.4.a. Gender 1.3.4.b. Intimacy 4 7 7 8 12 12 13 14 15 20 20 20 21 22 22 22 24 25 CHAPTER TWO: THE CORPUS 27 II.1. Size and contents II.2. Sampling II.3. Community grammars and idiolects II.4. Identifying the progressive II.5. Data collection 27 30 31 33 34 CHAPTER THREE: RESULTS 36 III.1 Density 36 III.1.1. Overall increase III.1.2. The two Romantic generations III.1.3. Apparent time III.1.4. A study of idiolects III.1.4.a. Individual total densities III.1.4.b. Individual trajectories 36 36 37 39 40 43 III.2. Linguistic factors 48 III.2.1. Semantic classes 49 III.2.1.1. State verbs III.2.1.2. Movement verbs III.2.1.3. Locative verbs III.2.1.4. Processes III.2.2. Verb form III.2.2.a. Overall distribution III.2.2.b. Perfect and processes III.2.3. Development across the lexicon 49 49 49 49 52 52 54 54 3 III.3 Sociolinguistic factors 56 III.3.1. Men and women III.3.1.a. Overall densities III.3.1.b. Extremes III.3.1.c. Diachronic groups III.3.1.d. Pairs 56 III.3.2. Degrees of intimacy III.3.2.a. For the whole corpus III.3.2.b. Diachronic groups 57 CHAPTER FOUR: FURTHER INVESTIGATIONS 60 IV.1. The development of the passive IV.2. The extended present IV.3. Time-frame IV.4. Bounded intervals of time IV.5. Generalizing progressive IV.6 State verbs IV.7. Prospective/retrospective predicates IV.8. Modal auxiliaries IV.9. That-clauses 60 62 63 64 67 74 78 82 84 CHAPTER FIVE: LINGUISTICS AND LITERATURE 83 V.1. Individual works V.2. Trajectories V.3. Chronological groups 90 92 94 GENERAL CONCLUSION 95 BIBLIOGRAPHY OF THE COLLECTIONS OF LETTERS REFERENCES 107 112 ANNEX The progressive across three centuries Binary glimpses List of the letter-writers Specimens of letters Degrees of intimacy The coding frame Verb lists :polysemic verbs; aspectual classes Periods and decades List of figures 120 120 123 127 128 129 132 134 147 148 FIGURES Index 150 176 4 ABSTRACT This quantitative survey of the private letters of great writers of the early and middle nineteenth century aims at assessing the modalities of a major development of English grammar : the considerable increase of the progressive form between 1800 and 1880. Coded in the database ROMPROG (some 22,000 quotations), it reveals that purely linguistic factors had a limited weight on it, whereas a strong connection is found with intimacy. Women are ahead of men in the development. It is suggested that the new existential emphasis associated with Romanticism — in a wide sense — may have helped in the actuation of a long-term trend, favouring a periphrastic verb-form with strong situational expressiveness in spontaneous discourse. This is encouraged by convergent symptoms when biographical data are considered, as well as literature proper. Detailed observations are also given on a variety of linguistic and socio-stylistic points. Introduction The extension of the progressive form ( henceforward “ progressive ” or “ PF ” — this very disputable, but commonly accepted, name will not be discussed here — is certainly the most dramatic change in the grammar of modern English : what used to be an occasional feature of the verb-system has now become an essential part of it. After creeping its way under cover for centuries — there are 10 progressives at most in Tindale’s version of the Gospels and Acts (1534), 49 in King James’s (1611) —, it jumped into modernity at the time of the Industrial Revolution, synchronous with the Romantic Age — 700 occurrences in The Twentieth Century New Testament of 1902 .1 The number of progressives per 100 000 words of text has been found a practical approximation to measure the frequency in a time of rapid change 2 : we will call it density. In the casual style of private letter-writing, the average figure would be a little above 100 at the turn of the century and 350 or more by the eighteen eighties. This rough estimate is at the origin of the present study, taking advantage of the mass of private letters which have been preserved and are still being published, mostly for biographical purposes. Indeed, the century ending with World War II was the golden age of correspondence. Only literate people are considered, then : is it not a disputable choice ? The answer is that it is certainly one of the most practical ways, and possibly the best way, of approaching the everyday language of the past century to assess its history in quantitative terms.3 This survey was carried out along the lines opened by Empirical foundations for a theory of language change (Weinreich, Labov & Herzog 1968 ) and the methodological and substantive developments which followed. The five major issues have been clearly stated again in Labov 1982, 2.1 : the constraints problem : what are the general constraints on change, if any, that determine possible and impossible changes and directions of change ? the transition problem : how (by what route) does language change ? ; a question which can be restated as : “ how can language change from one state to another without interfering with communication among members of the speech community ? ” • the embedding problem : how is a given language embedded in the surrounding system of linguistic and social relationships ? 4 1 The figures are from Mossé 1938. It was first suggested by Mossé 1938. Nehls 1974 and Scheffer 1975 call it the M coefficient. 3 This has now become the major object of the work of the Helsinki school, in connection with early English (Nevalainen, Raumolin-Brunberg 1996.) Other possibilities have been explored recently: plays, dialogue in novels, trials, etc. What we call everyday language is certainly not a homogeneous whole, and the emphasis here is on practical, given the limitations of the research. More will be said on this subject in II.1 on the corpus. As to the connection between spoken and written language, there is an interesting discussion in Romaine 1982, 14-21: I would agree with her that "the function of writing ...is not merely to record the spoken language; writing has an independent existence" p.15. Familiar letter-writing, however, is probably closer to speech than most genres using the same medium. Much remains to be studied in this respect for a meaningful assessment of what community language is, in different societies and historical periods: see e.g. Smitterberg 1998. 4 This is accompanied by the important remark : As linguists we are preeminently conscious of systematic relations : that a given system will change less or more because of its position in a complex web of oppositions and distinctions. But it is quite possible that a linguistic change could be the result of a particular cause isolated from the linguistic or social systems. This means that our common beliefs, justified as they may be, should always be reexamined, at the macro or microdiachronic level. 2 5 • the evaluation problem : how do members of a speech community evaluate a given change, and what is the effect of this evaluation on the change ? • the actuation problem : why did a given linguistic change occur at the particular time and place that it did ? This is well-known to be the most difficult of all problems − accounting for the sporadic nature of change. As Meillet 1921pointed out , the principles developed in the 19th century deal with the way that changes can occur, but not with the way any given change did occur...With the actuation problem we are dealing with cause in the most immediate sense. This last and, in Labov’s terms, most difficult of all problems, is foremost in the present research, dedicated to the linguistic and socio-psychological study of an ongoing developement, after it occurred and with the inevitable limitations due to the distance of more than a century. Second, the transition problem is also at stake, since we are confronted with the rapid development of a central feature of the verb-system. Yet, it is also substantially concerned with the other problems : the constraints problem, because the factors in the linguistic environment which could be hypothesized as effective have been computed, and two extra-linguistic factors (gender and intimacy) have also been observed, suggesting tentative extrapolations; the embedding problem as a more systemic approach of the former. As to the evaluation problem, it has not been specifically studied, but was also kept in mind : the progressive — some of its uses at least — has been the object of stigmatization in Britain. The letters so carefully preserved happen to be those of exceptional people — women and men of letters, as they are aptly called — and their representativity can be disputed. Yet, two observations must be made : 1. Anyone reading their letters will observe the same uncontrolled style they must have shown — with varying degrees, as for most of us (see Chapter II) — in their everyday speech among their family and friends. 2. They certainly lack the anonymity of the “ common speaker ” approach where individualities are lost in the implied homogeneity of a period’s language : their biographies are as well-known as possible, their other works have been preserved. Whatever will be found in their ordinary grammar can be connected with their geographical and social background, the story of their lives and human relations, the books they wrote, etc., for the benefit of both the literary historian and critic and the language specialist, and perhaps even other students of the humanities and of social science. In other words, this is not a study of language as an abstract system of universal rules, but a study of the language used by people, not a study of langue ², but a study of parole , in Saussurean terms. Grammarians have now almost given up the tradition of quoting examples from the best writers, and present-day linguists are suspicious, to a fault, of elaborate literary texts. This was a sensible reaction against the artificiality of “ good language ” as the one and only reference. The modern concept of linguistic communities, with all their diversity, suggests a more balanced consideration of all strata or sub-groups : should the literary microcosm be left aside ? Certainly not. There are at least two reasons for this : 1. At a time when more and more people were enjoying the benefit of education, the great novelists of the Romantic Age, in particular,with their large reading public 5 , exerted on the contemporary language — either directly on their readers or indirectly through them — an influence we cannot easily evaluate, but which must have been considerable. If they were not the only leaders, they were obviously among them. The question of whether literature is an accelerator or a brake is a different question, which studies like the present one may help to answer. This is of interest to the linguist. 2. As individual persons and writers, the study of their language is fundamental if one believes in Mallarmé’s famous statement that literature is made with words, and a parallel between the language of their literary works and their everyday language as evidenced in their letters will always be rewarding. The idea is not new : in books on “ The language of ... ”, private letters have often been considered. Here, however, the approach is quantitative and is not limited to considerations of “ style ”. It concerns a major development of English which it is perhaps not too bold to call its “ modernization ”. How is it that people are gradually led into a new form of expression, consciously or unconsciously, individually or in keeping with the language community (or communities), such is one of the mysteries of language that the present survey is trying to bring to light. 5 The number of copies sold is considerably smaller than the number of actual readers. Circulating libraries (Mudie, SmithErreur ! Source du renvoi introuvable., etc.) bought a large number of them. 6 CHAPTER ONE OBJECT AND METHODS I.1.General outline The present research is supported by a corpus which had to be sufficiently large for a quantitative study of change over the historical period, considering the many factors involved. In such studies, the inevitable gaps in the data cannot be filled by intuitive interpolations or by questioning informants. As to other written data (sometimes found in grammars covering the period), they may hint at virtualities not illustrated by the corpus, but not quantitatively estimated. Even for a comparatively recent time, we are confronted with what Labov1994,21 calls the Historical Paradox : The task of historical linguistics is to explain the differences between the past and the present ; but to the extent that the past was different from the present, there is no way of knowing how different it was. 6 Is it adding a paradox to another to say that the more recent the period studied, the more apt we are to forget the difference ? Few items used in the middle nineteenth century have entirely disappeared from present-day usage (Rydén 1979,16-17), and the language of that recent past seems to us so familiar that the breach of communication is bound to occur without us being aware of it as we are when reading Shakespeare for example. In dealing with historical texts we must never forget this.7 Certainly, it would be a positivist illusion to think that we are dealing with bare data without any theoretical frame to orient the search. Still, while recognizing the value of a more theory-oriented approach, we consider that too strict a theory is in permanent danger of, consciously or not, refusing to consider the facts or data that do not fit it. An observation-oriented theory is of a different nature in the sense that it must be kept open to anything, with the opposite risk of accepting too readily what is most challenging. The attempted theory of the progressive proposed in Arnaud 1973 largely inspires the outline in I.2 below, but the construction of the coding matrix is supported by a large number of converging analyses, from many theoretical viewpoints. The hypothesis governing the present survey is not specific to its object but rather a general methodological principle for the study of change in progress, namely that variation in statistical distributions can suggest which linguistic or extra-linguistic factors are likely to be responsible for the change. However it is not a replication study, since the diachronic trajectories of the individual writers are traced from beginning to end : given the length of time considered, the population has changed, and contributions are diversely weighted. This is far from the requirements of social or linguistic surveys of the usual type, but the fault, if fault there is, is not due to systematic refusal of the immense advantages and progress of statistical science, nor to ignorance of its epistemological foundations, but to deliberate choice. There were two reasons for this : 1. Trying to extract any type of random sample from the corpus implied loss of considerable information and the multiplication of empty cells : even with our approach, there are still too many of them. 2. By studying the whole population — a non-systematic sample, in fact, of the literate population of England at the time — and by collecting the corpus in such a manner that it remains available, the possibility still exists for betterequipped researchers to apply statistical and computational instruments to confirm or contradict our findings, or merely add further precisions. 6 On that point, see also C. Marchello-Nizia 1988. The breach of communication is easier to overcome for vocabulary items, especially with familiar ( from wafer Sellotape ) or slangy words − Keats was fond of them −( from filly to broad ), than for grammar. 7 7 Another serious objection might be raised : in the study of a phenomenon so often presented as a binary choice between simple and progressive forms why not a binary, “ variable rule ” study ? Three answers can be given : 1. A practical excuse : so great is the disproportion between the occurrences of progressives and non-progressives (of the order of 2 per cent at the beginning of the period), that a careful coding of all relevant factors for the whole verbsystem would have implied an amount of work beyond the powers of an individual worker ; 2. The natural counter-reply is that this could have been done, or can still be done, by a team of qualified researchers, or/and that instructive conclusions could be drawn from more limited binary studies (considering for example, only simple past forms, or present+past forms, or perfect forms, etc.). This alternative choice could be quite justified and rewarding : see ANNEX for a few attempts in this direction. 3. More seriously : the binary view8, which we often think obvious, is far from covering the network of syntacticosemantic relationships involving the progressive. The schema presented in the following section gives an idea of it. It has long been noted that the binarism illustrated by the Prague school — Roman Jakobson above all —, originally in connection with the phonological system, must be seen with more diffidence when grammar is in question, especially in the case of a wide-ranging phenomenon. To sum up, this is a quantitative survey of a corpus of 19th century private correspondence with a view to an estimate of the distribution of progressive forms, in connection with a set of linguistic and extra-linguistic factors, over a century — 1780-1880 to give the approximate limits. It is not meant to support a particular theory, either of language or of the history of language, but to provide factual observations with the ultimate object of understanding the conditioning of this change, and possibly to give some lights on the mysteries of change in general. I.2. The progressive paradigm The literature on the progressive is considerable and will not be discussed here. There is no element of the linguistic system which is not involved, but, for the modern period, aspect and aktionsart are commonly referred to. Starting with Aristotle’s famous statements, down to Vendler’s renewal (1957, 1967), and the discussions it originated, they have often interested logicians as well as linguists. The former are even more prone than formal linguists to rely on intuitions and ad hoc examples. To mention just a few disputable considerations, the idea of activity is often referred to... and yet what about such common progressives as standing, waiting or hoping ? Sentences with animate/human subjects are repeatedly quoted ... yet what about everyday expressions like it’s raining, the wind is blowing, the eggs are boiling ? Then the diachronic dimension is commonly overlooked. Without developing a complete theory, we will give a brief survey by considering first the formal auxiliary system and second the semantics of the BE-ING syntagm known as the progressive form. a. The auxiliary system : the following sketch focusses on what appears to have been the auxiliary system, involving aspect and voice, when the progressive developed in 19th century English. The slow semantic change which leads to present-day meanings can be inscribed into it and the quantitative survey may help to understand the direction it has taken and is continuing to take. It seems to be mostly connected with modality and the relation of speaker and speech which French linguists (Benveniste 1961, Culioli 1981, 1995 — see Groussier & Rivière 1996) have studied under the name of “ énonciation ” (cf. also, Wright 1994a, Smitterberg (forthcoming) ). 8 It is almost as old as the discovery of the progressive form as a verb form. Undeniably, this is what language users feel as a rule. In her often-quoted study, Hatcher 1951: 262 writes: "First, since the problem must be the contrast between our two constructions, both constructions must be studied with equal care." Practically all grammarians followed this advice since then. I don't claim that they were wrong, but only that, apart from questions of feasibility, it may obscure some of the problems. 8 The following graph may be helpful for an understanding of the periphrastic forms of the English verb using the auxiliaries BE and HAVE. Prospective viewpoint ----------------------------------------------> 1 2 3 a. BE TO -ING -EN b. HAVE TO -ING -EN <----------------------------------------------Retrospective viewpoint The morphological auxiliary-system connects the two primary auxiliaries (their complementary function in many languages is well-known) and three verb-forms : • TO-infinitive (with its original directional value, in space and time) ; -ING (also found in gerunds, verbal nouns or adjectives) ; -EN, “ past participle ” form, also used for passives, “ passive participles ”9 and adjectives. The table illustrates : • vertically, the complementary use of BE and HAVE. There is an extensive literature on this point : let’s simply say that BE usually refers rather to what is conceptualized as “ internal property ” or “ attributes ” (of an Aristotelian “ substance ”) : she is beautiful and charming ; whereas HAVE concerns “ external property ” : she has beauty and charm: this example to show that, often, the choice, with possible semantic shift, is left to the speaker : cf. also : a wealthy man v. a man of property .10 • horizontally, the sequential system of intervals or phases through which a (perhaps Eleatic) “ body ” in motion or in a process of change (natural or induced) inevitably passes, that is from a “ before ” to an “ after ”. This is relative to the position of the speaker. In our everyday perception, things are certainly more clear when there is a tangible result : the house is built. They are obviously less so, and restricted by modalities, when they are seen in prospect : anything can still happen before the end is reached. The progressive refers to the intermediary stage between project and accomplishment, the inside of the process, so to say, what is “ going on ”, to use the common phrase. We have to remember that this is a metaphor and that intrinsic movement is not necessarily implied : we are staying here. The point is so often forgotten that it might be useful to speak of “ temporary localisation ” or “ temporary occupation ” 11. This last remark reminds us that the various meanings obtained at the different phases must take into account the meaning of the verbs involved, since their aspectual functioning varies accordingly : if we speak of processes, we shall have to ask ourselves whether staying, waiting, living (in its two main senses), and even knowing, thinking, loving... are processes (or activities), and of what sort. The semantic status of participants may consequently be different : agent, cause, etc. for the “ subject ” (first argument), patient, intermediary, goal for the first complement (second argument), etc. Extensive discussion of all this lies outside the present work. This being said, the above system can be used as a practical device to build a variety of sentences : some of them may have been fairly common in 19th century English and are often exceptional or obsolete in present-day English. For example, in column 1 , line a, we will obtain sentences like : 9 The term was fairly common in 19th century grammar : Cf. Lindley Murray (1795), a grammar often reprinted. On this account, ordinary language is not as strict and accurate as grammarians would like, and some uncertainty is part of everyday communication, just as our perception of reality can shift between levels of abstraction according to whether we are engaged, for example, in action or reflection − Cf. Bergson, Matière et mémoire, 1896, p.173-81 − whereas the activity of the linguist is professionally language-centered and reflexive. 11 The term occupation is suggested by Alexander Bain 1863. This famous disciple of Stuart Mill, and the founder of Mind, wrote a treatise of logic and a grammar. He writes: "This peculiarly English form enables us both to confine an action or a fact to the present, instead of extending over all time, and to intimate that the agent is now engrossed, and is barred from other occupation". Like many others, he considers only agent-subjects, and also forgets that a person may have two "occupations" at the same time : and now I am sitting writing in such a pretty dining room (Gaskell, 1848, Letter 18). That was not an unusual expression at the time. 10 9 we are to paint the house-front to morrow / the house-front is to paint ( cf. :we are to kill a pig soon. J. Austen). Line b will generate : we have to paint the house front / we have the house-front to paint. Only when the system has been tried with a variety of items can we really assess its functioning. Some rearrangements occur ; one is well-known,the shifting from : we have a letter to write /we have to write a letter. The same alternation of BE and HAVE can be tested in column 3, where it is perhaps even more clear : the house-front is painted / we have painted the house front. It has often been shown that, in many European languages, there happened a reordering : we have the letter written/ we have written the letter. Just in passing, notice that line b does not seem to be valid for column 2. We do find : the house-front is painting but not normally, in standard British English : ?* we have the house-front painting nor the reordered expression : • we have painting the house-front. Yet in 19th century English we find : Henry Inglis has had my book reading (Carlyle); they have the plan of a Great Council simmering already (G. Eliot, Romola). And also : we have staying with us a Mr Robinson (Wordsworth) the Irishwoman whom I had working for two days (Carlyle) : sentences which, if not entirely excluded today, would be rare. This suggests a potentiality which did not happen to be fully grammaticalized. intermediate stage is more difficult to conceptualize as “ external property ”. The reason might be that this b.Semantics of the progressive : the graphic representation above is meant to illustrate that what we now call the progressive is an ambiguous development of the verb system. The conventional distinction of verb-phrase and nounphrase as separate items, in Indo-European especially, is still a much-debated point. Let’s only recall that English -ING forms, as well as -EN forms — the history of their development is well-documented — have a spectre of meanings between verb and noun : hence the usual names of adjective, verbal substantive or gerund for some of them — or nominalizations in generative grammar —, progressive form of the verb or present participle for others. Some “ adjectives ” in particular, not always easy to distinguish from verb-forms, obviously describe states or occupations seen as temporary : a smiling girl does not always smile, a sleeping horse sometimes wins a race. With the progressive, something similar takes place, which most grammarians have observed. In general terms, leaving aside later developments, the “ simple form ” is synoptic (a term borrowed from Close 1959) ; also quoted by Ota 1963 : 60), that is to say it concerns an act, an event, an accomplishment sometimes, seen as a whole. Many authors have stressed that point, but in terms not always perfectly clear. By using the auxiliary BE (in the sense of exist − the “ existential quantifier ” of formal logic ; cf. also da sein in German, estar in Spanish, tà in Irish, etc.) the progressive is not so much concerned with the act itself as with the “ involvement of the subject ” (Hatcher, 1951), and its (or his/her) temporary condition : the subject IS what it is doing, or experiencing, etc., for the span of time considered. 12 12 In present-day English, it is next to impossible not to use the progressive in the definition, which becomes circular. So that, when trying to devise tests to study the constraints for the use of this form, many have observed that you sometimes have to beg the question. The situation can be the same in the Spanish of bilinguals in the U.S. Cf. Flora Klein 1980 : The question : " Que hacen esos chicos alla afuera ?"(they are waiting at an employment agency) is wrongly interpreted. "It was thus intended to refer to X − the moment of speech itself − and it was so interpreted by Spanish monolinguals, although the question was phrased in the simple present form. Spanish-English bilinguals, however, usually interpreted this question as referring to Y (the general present), in accordance with English usage, and 10 Descartes’s famous cogito ergo sum is usually translated by the simple form : “ I think therefore I am ”. Probably because it dates back from the 17th century, this translation has survived. Philosophers certainly have their opinion about that and even if we might count on probable support from Husserl, let’s not trespass on their ground. However, in contrast to its more synthetic formulation in the Méditations métaphysiques, a translation of the phrase in the Discours de la méthode using the progressive : “ I am thinking, therefore I am ”, would be a clear expression of its existential force. Notice that, in the Méditations, Descartes underlines its temporary iterative meaning, as distinct from a general essentialist sense : “ il faut conclure, et tenir pour constant que cette proposition : Je suis, j’existe, est nécessairement vraie toutes les fois que je la prononce, ou que je la conçois dans mon esprit ” , Méditations métaphysiques..., Paris 1647 ; Paris : Gallimard, 1941 (Bibliothèque de la Pléiade), 167. This would further justify a modern translation by the progressive ; but, at the time, the actualizing, or iterative, sense was normally included, in English, in the ambiguous simple form. Typical in that respect are expressions concerning the weather : it’s raining, snowing, etc., which have the same force as an adjective (windy) or a nominal expression (there’s a storm) — Cf. it’s cold, beside it’s freezing — since they just describe the state of what we call the weather, a sort of mythical entity the Greeks eventually called Zeus. Typical also the binomial constructions mentioned below in section I.3.2. when the two forms are used in succession, the first referring to an act, the second to an action for which the subject (agent or cause) can be held responsible. Notice that this is valid even for non-animate-non-human subjects (the falling roof is clearly a transform of the roof is falling) ; which does not mean that animate and human subjects are not a majority in fact. Notice also that what we call act for want of a better term (event, or “ happening ” being perhaps less obscure) does not imply activity of an agent, or a cause, nor does it imply purpose : if we want the term to cover all the cases where we find a verb-form either “ progressive ” or not, all we can say is that it “ takes place ” in time. As a matter of fact, that common expression “ take place ” is revealing of the epistemological priority of space. This would be the linguistic expression of the existential quantifier of post-kantic formal logicians, who were actually formalizing what is found in ordinary language. This can be connected to an existential ontology. But, as stated above, the present work cannot afford to develop a full theory of the system. When looking for some basic meaning of the progressive — to be adapted to the various contexts and situations — that “ temporary adjectival ” value must be considered as central. Expressions like subjectivity (Mossé, 1938), involvement of the subject (Hatcher, 1951), etc., were largely intuitive analyses for what can be inferred from the logic of the auxiliary system, at least in modern usage. This may well be the reason why the progressive became, unconsciously and by degrees, a more personal, more intimate way of referring to events of one’s or one’s relatives or friends’ existence, easily transferred to other subjects and extending in many directions to give a sort of emotional colouring to one’s prose. History can be a narrative of dry events, where protagonists are not in the foreground ; a letter addressed to close relatives and friends is a picture of one’s existence and feelings.13 Speaking of old theories we can also mention Parmenides. The Ancients had noticed how difficult it is to account logically for movement : hence Zeno’s famous paradox. Aristotle− there have been many discussions concerning aspect based on his developments in the Physics − suggested a partition of time as a sequence of discontinuous periods : touto yar estin o kronos, arithmos kiniseos kata to proteron kai isteron, the number of movement according to the before and the after : Physics 219b, 1,2. Cf. Collobert 1994. A statement using the progressive is a description of the condition of the subject “ for the time being ”, knowing that it won’t last for ever. From there originates the telic-atelic distinction, and the need to adapt the analysis to the semantics of verb-classes. For the space of time considered, it qualifies the subject in the intermediate stage before accomplishment or mere termination. This does not mean that the act itself and its purport is ignored, but it is not so prominent as the “ involvement of the subject ” (Hatcher 1951) : On reading a letter from someone close to our heart, we are more interested in the person than in his or her achievements as such. Perhaps should we say that they are actions rather than acts, though they are both at the same time. It would be begging the question of course to say that the letter replies not to “ what did you do last week ? ”, which might be the object of a business letter, but to “ what were you doing last week ? ” It may interest Mrs Gaskell whether it was a blue or a green dress her daughter Marianne bought : but what really matters to her was whether Marianne enjoyed her shopping, how pleasant or unpleasant were her occupations. so as "What do those kids out there do ?" As a result it often provoked surprise, presumably because applicants at an employment agency are assumed to be unemployed." 13 Cf. the contrast evidenced by Benveniste between "history" and "discourse", with the French simple past (aorist) used for the former and the passé composé for the latter. Even when a speaker relates events comparatively remote, the connection with his own self and the moment of speaking is foremost when the passé composé is used in written French. It is history alive so to say. This periphrastic tense using the present form of ETRE or AVOIR , before it also became used instead of the simple past in the spoken language at least, used to stress "present relevance" in the same way as the English present perfect. In French classical tragedy (Corneille, Racine) it was used to differentiate what had happened during the time-unit of the dramatic action (passé composé) and what had happened at a more remote time (passé simple). 11 This to mention a typical case. The overall orientation is more pervasive and goes beyond individual sentences, not always consistently, with the tentative randomness of what Labov has called the variable rule. This focussing on the subject, mostly felt when it is the sender or the addressee of a letter, but eventually extended to all predications, may be the reason of what we suggest to call spontaneity − effusiveness can be another term −, a dimension responsible for the variations we find in many places in the corpus, either between individuals or at different moments of an individual’s life. The point will developed further : see General conclusion. I.3. The coding frame A fairly large number of factors, of various kinds and weights, are suspected to govern the use of the progressive : they are listed in the coding matrix : see ANNEX. Most of them are well-known ; some, however, are seldom accessible in written texts. They can be listed under three headings : speech situation, linguistic and sociolinguistic constraints. A few points were also included for practical reasons, and the decimal system led to a few breaches in the consistency of the numbering. This is a research frame where the ordering of elements does not claim theoretical significance. I.3.1. Questioning the data Although, as we said, the quantitative research hypotheses must be kept sufficiently loose and open, a certain number of assumptions inevitably lie in the background. Here are some of the legitimate hypotheses which the research was expected to verify. 1. Concerning linguistic factors, it was assumed that the development was differentiated with reference to the syntactico-semantic system of the verb. a. Tenses, voice and modality : it was expected that the increase primarily affected intransitive or unbounded perfect forms (present and past perfect, the second not merely a transfer of the first), probably in parallel with the decrease of the BE-perfect with intransitives (Rydén & Brorström 1987). This does not preclude synchronous development with other tenses, and also voice and modal collocations (the label Ts is just a practical abbreviation). For the least commonly affected of these, it can be enlightening to have more precise data on the rise of the progressive passive and the dwindling of the old equivalent and on the development of the diverse modal forms from very modest origins. b. Predicational verb-classes : the development of the progressive with what are called “ state ” verbs is a 19th century novelty, or almost so. How did it affect the various sub-classes is one of the questions to be raised. But the major question concerns the aspectual semantics of verb-classes. Considering that the original progressive often had a strong connotation of “ middle voice ”, it was hypothezied (on the basis of pilot surveys in Arnaud 1973) that its development would affect primarily unbounded, non-conclusive, predications. Given a reasonably satisfactory assignement of all individual predications to the appropriate class or subclass, the corpus should be sufficiently large to provide confirmation of the trend, and of how it affected the different subclasses. c. A certain number of other suggestions, less central, will be developed below : Jespersen’s “ time-frame ” theory ; the development of present forms (or their transposed equivalents in past forms) referring to events in the (supposedly) near future or past ; the presence of explicit contexts implying duration, whether discontinuous (iterative) or continuous,and even generic statements, contradicting the analysis of the progressive as strictly limited to the interval of the instant. In the same line of thought, the ascription of a terminal boundary is sometimes said to block use of the progressive : instances of it can however be found, requiring further study. Another occasion when it appears difficult, apart from reported speech, to use the progressive, is when a declarative verb is followed by a that-clause or an equivalent object-clause. d. Given the strong association of the progressive with deictic factors (in an enlarged sense), it can be useful to observe how it is connected with “ person ”, and more generally, with determinacy of the subject. e. Apart from their distribution into predicational aspectual classes, lexical verbs have been computed individually, with due consideration given to homonyms (polysemic verbs). This is not without risks on the theoretical level, due to the elusive semantics of individual words. For the present purpose, what counted most was to be consistent for all the extension of the corpus along the periods surveyed. The idea was to test whether the change affected all words (verbs) at the same time, or whether there was some sort of lexical diffusion. This now well-known theory (see Labov, 1981, 1994) has been opposed (mainly by Wang and Chen − Wang 1977 − and the research which followed) to the neogrammarian doctrine of phonetic change affecting, “ mechanically ”, all words at the same time (with the wellknown exceptions of dialect borrowing or analogic change). It is not obvious that the competing theories apply equally well to grammatical change, especially with many parameters involving different levels of analysis. Yet, it was not unreasonable to think that the progressive spread, so to say, “ contagiously ”, over the lexicon of verbs, along lines 12 which can be related to the classes we have discriminated, but possibly also along other dimensions..14 The aim was probably too ambitious, but we might expect the survey to help in unveiling some of the mystery of diffusion over the language system proper. We know, however, that vocabulary (tokens of the lexical types) is strongly linked to the needs of the speaker, and that, especially in the case of the progressive, a situation-bound form, this is largely unpredictable and haphazard. Let’s however note an interesting point, seldom, if ever, noticed by grammarians: new verbs (neologisms) very commonly appear in the progressive as soon as they are introduced. A list will be given in III.2.3. but we can already mention : (I am) calomeling, daguerreotyping (Ruskin), and also the use of the progressive with compounds of the type house-hunting, where nouns first, and then verbal nouns or gerunds have certainly preceded the progressive proper, which looks like a back-formation. The various points mentioned above should not be, if possible, treated separately, and could be subjected to multivariate analysis, or at least to bivariate surveys including, for example, a separate study of the perfect utterances involving unbounded predications. As to the sociolinguistic (or socio-stylistic) factors, the labels speak for themselves. Only part of this ambitious programme could be contemplated within the limits of the present work. I.3.2. Situation of utterance. Most studies have noted that the progressive is sensitive to the speech situation. This point could be studied in the sociolinguistics section, but it is so inescapable that present-day formal linguists have introduced it − sometimes under the heading of pragmatics − in their analyses, so that it seems better to mention it in a separate section. It is indeed one of the tenets of the énonciation approach, partly reflected in some of Jakobson’s views (on shifters, for example), initiated in France by Benveniste, and more recently associated with the name of Culioli 1981, 1995 − Groussier & Rivière 1996. To put it briefly, it emphasizes the insertion of speech in space and time, a topology of reference with the énonciateur (utterer), an abstraction of the traditional notion of speaker, as the origin of the speech process. It differs from the speech situation of the sociolinguistics tradition (Hymes, Gumperz, Fishman, etc.) in that it tries to rely only on whatever trace the operations of the utterance process have left in its product, the utterance proper. Faithful to the doctrine of the autonomy of linguistics heralded by Saussure, it is an attempt to formalize what can be observed in situated utterances, without falling into the traps of philosophy on the one hand or of sociology and psychology on the other. Still, it does not preclude them : the narrow diffidence of some structuralists of the Bloomfieldian tradition definitely hostile to “ mentalism ” is no longer maintained and we can walk across boundaries once tabooed. Even linguists not sociolinguistically-minded, or even those averse to sociolinguistics, cannot conceive of language use completely dissociated from the physical and microsocial conditions, however abstract the picture they eventually give of them. Broadly speaking, in the case of a person writing to a friend or relative, the speech situation, in sociolinguistic terms, is clear and well-defined. A typology of letters − or of sections of them − could be attempted : mere communication of ordinary news, comments of various types about people and events, occasional disquisitions on political or literary matters, more or less passionate outbursts, confidential revelations, expressions of friendship and love, the list could be long and profuse. Here, for practical reasons, it has to be deferred. We will be satisfied to accept that the situation of utterance is that of a normally one-to-one and one-sided written exchange −not forgetting howewer, that letters were often addressed to several people of the same family, or could be read by other people, with or without the consent and knowledge of the writer. This plurality of addressees or readers − perhaps (see below, III.1.4.) responsible for the cooler tone of Macaulay’s letters to Hannah after she married, for example − is one of the inevitable facts that the letterwriter has to accept, just as the ordinary speaker-hearer exchange may be overheard or, to-day, tapped on the telephone. Of this, the sender can be aware or not, and may write accordingly. As a matter of fact, it does not seem likely that our letter-writers were commonly writing consciously for the public or for posterity as Madame de Sévigné did. Some may have suspected that their letters were not only kept − a common practice even for family letters − but shown to other people. Then, when the sender had become famous, letters were even published, and some writers resented that as a breach of confidence. Wordsworth, for example, made it known, in writing, that he absolutely prohibited publication of his correspondence even after his death, especially when he learnt what use had been made of Robert Southey’s letters, of the publication of Coleridge’s and of the proposed publication of Charles Lamb’s letters : it fills 7 volumes in 14 Contagion can also be the word when two or a series of progressives appear in a given context, a fact which may sometimes introduce a bias in the count. This phenomenon has been kept in mind, and on the whole, its incidence is not very significant. With literary texts, it may be part of a "periodic" effect, in argumentative discourse for example, but also of other contrived or unconscious manifestations of "style". Guillemin-Flescher 1981 has carried minute and deepreaching research into such points, which cannot be captured by the wide-mesh net we have to use here 13 Selincourt’s edition !15 But let us appease our conscience by remembering that this prohibition concerns the content rather than the grammar of the letters. As to whether this grammar itself is affected by the awareness of a certain degree of publicity, we are left without a sure guide to give more precision to the notions of privacy, intimacy, familiarity.16 And the only thing we can say is that such niceties of epistolary interchange have to be neutralised for a mass study. Two typical points, fairly easy to distinguish, must however be noted as significant examples of sensitivity to that complex of physical and social factors defining the individual situation of utterance. The first is well-known, though it has only been brought to light in recent times, with occasional mention of the progressive as one of its distinctive features : the performative utterance. Before Austin, Koschmieder 1935 alluded to it under the name of “ coincidence predication ”. As well as it excludes other persons than the first, other tenses than the present, it also excludes the progressive. Perelman 1962 − the analysis is in French − speaks of a soldier refusing to obey : (1) “ I refuse to do this ”. An officer is called, and − this is my personal contribution − may ask (in English if that was the language spoken) : (2) “ Are you refusing to obey ? ”. Whereas (1) is performative, (2) is merely constative, to use Austin’s terms. The officer might have dispensed with the progressive, only saying : (3) ” Do you (really) refuse to obey ? ” However, by using the progressive, he emphasizes the responsibility of the soldier, who becomes definitely a disobedient soldier, someone who deliberately endorses his disobedience, who is not merely responsible for his act, but for an action (see above in the semantic study), for his conduct. A performative is a “ synoptic ” act with aspectotemporal coincidence of initiation and result, with no interval in-between − and actually the clearest case of a punctual event, by definition and without any need of difficult and disputable measurement. It is also something that alters the social order, however minutely, as when we say “ thank you ”. A conduct is less fact-oriented : the word itself refers to the judgments we may pass on a person ; eventually it alters the person : from conduct to character the step is sometimes narrow. Apart from deeds of different kinds from ordinary citizens or people in power, most performative utterances are oral. In private letters, we will of course find expressions like “ I promise to reply soon ”, “ I congratulate you on ... ”, etc. The question here is whether the progressive can occasionally be used for performatives. It’s a matter of definition : exclusion of the progressive can be a defining factor of performatives, a commendable stance if we follow Benveniste’s 1966: 267 advice.to avoid the traps where Austin himself was liable to get caught. Such traps are easier to avoid when one relies on the evidence of the linguistic context. If a policeman says : “ I am arresting you ”, a form which has been heard, though rarely, we may safely assume that he is not strictly following his official guidebook, finding perhaps that the set-phrase is too abrupt, and for various reasons, probably unconscious, prefers a more personal utterance, speaking as a man rather than as a police constable. He might also say : “ I am going to arrest you ”, another milder form for an occasion not usually considered as mild. We could speak of understatements in such cases, and, supposing they were brought to Court, lawyers and judges should have to consider whether some breach of the law had taken place. Another less notorious case where utterance factors are prominent was already explored by Charleston 1941,1955. We find it in sentences of the type : “ When you do A you are (“ actually, in fact ”) doing B ”. There are few instances in the present corpus as clear as, for example, the following from Trollope (Barchester Towers) : The archdeacon perceived that he would be making a false step if he allowed the cathedral clergy to give the bishop just ground for umbrage. But we can mention : 15 "For my own part, I do most earnestly wish that not a single letter I ever wrote shall survive me: and I shall endeavour to make it known to all my correspondents, whether accidental or regular, that such is my wish: and farther that I shd deem a breach of the laws of social intercourse as I wish them to be maintained, between me and my friends and acquaintances, if either they do not destroy my letters, or send them to myself or representatives." Letter to Edward Moxon, his publisher, 10 Dec. 1835. 16 In this order, perhaps: privacy would stress the confidentiality of the letters, intimacy the affective link between addresser and addressee, and familiarity the tone and style of the writing, irrespective of the personal relationship: this is why "degrees of intimacy" has been preferred for an estimate of a social, not a stylistic factor − which would be begging the question. 14 In coming down to Scotland therefore Terry would be leaving a position in which... (Scott, 1828) but I feel that, in writing thus, I am merely flattering myself (Macaulay, 1855); I am not using mere phrases when I say that (Eliot, 1858). Buyssens 1958 called this “ interpretation ”, and Adamczewski 1978 found in such explicit sentences support for his own extended theory of the progressive as “ secondary mention ”, even when the first non-progressive clause is absent. In the same manner as for performatives, whereas the first clause concerns an act, the second takes it up as action, or conduct. The emphasis is then shifted from the event and its consequence to the subject (normally human) who is responsible for it. In logical (and juridical) terms, it is then open to judgment : whether it is the action itself or its author (agent, cause, etc.) which is liable to it is another matter. In Courts of Justice, however, people and not events are tried. The two clear situations above are in fact the only cases when utterance factors can be easily detected in private letterwriting. The first one can be neglected, since performative utterances normally exclude the progressive ; but a separate count was taken of verbs which are candidates to be used in such situations (declarative verbs : Joos 1964 called them “ asseverative ”). The other, the “ interpretative ” binomial sentence, has been noted : the number of instances is small, but they are easy to locate, since the contextual element is so clearly identified that we are verging on syntax proper. It is safe to assume that present-day English developed this explicit or implicit binomial type further.17 Other points normally studied within the framework of formal contextual factors, morphological or syntactic, obviously refer also to the speech situation (person and tense for example) : there is indeed a continuity. Since we are dealing here with writings, it is all the more easy to obey the scholastic dictate of the structuralists that a linguist ought to consider only what is formally present in the utterance. I.3.3. Linguistic factors It must be clearly repeated that the framework of the survey is a research framework, a wide mesh. Many of the points raised are still the object of discussion, and it is hoped that the enrichment of the data provided by this collection will help in solving problems of an often very wide scope. The general semantics of processes, for example, implies, among others, cognitive considerations beyond the limits of an individual language. Inevitably, however, even a tolerant approach cannot fail to reflect some of the convictions of the researcher. One of these is that meaning is a continuous process involving at almost every moment all the components which linguistic analysis obliges us to consider separately, with the inevitable risk of artifacts18 I.3.3.a. Semantic classes For the sake of simplicity, we will speak of “ verbs ” to mean predications. We assume that each occurrence of a verb form involves a basic value which is actualized in the predication.19 A point, not negligible but secondary, is that of homonyms (see ANNEX for a list) : they have been counted separately, of course. As lexicographers know, it is not always easy to say when you have two “ verbs ”, e.g. call : hail or visit ( also said of a ship) or two meanings of the same verb call : hail, name − there are puns on this homonymy− ; here, they have been considered as three separate verbs, as well as labour meaning work or suffer ; but a different decision could be justified. Like lexicographers, we had to be practical, and it was felt that what mattered most was to keep to the same decisions throughout, which is not always so easy. A few illustrations are given for each class : they just give an idea of semantic niceties it was not possible to discuss in detail. Some of them will be considered in the appropriate section. Homonyms can usually be clearly discriminated in context. But specific “ aspectual ” values also pose problems. In the large majority of cases, there was no occasion to doubt, but some contexts can be ambiguous. The semantics of verb predications with respect to use or non-use of the progressive have long been studied by linguists. They give only examples, however, whereas we, like lexicographers, have to include all occurrences in the table for computation. The classification given here reflects a long history of discussions on the semantics of aspect, as old as Aristotle, but, in recent times, dating mainly from Kenny and Vendler’s famous typology : see I,2. We will not discuss it again, and will only give some guidelines. 17 See König 1980. I am indebted to Eric Smitterberg for this reference. Semantics, syntax, morphology are the common names for the main components but we should also remember that, in the case of a typically spoken form, considerations of stress and intonation may intervene; and we usually know little of how they were used at the time in any given utterance. 19 Figures refer to the coding of these classes and subclasses (See ANNEX). 18 15 The main consideration is that of “ telicity ”, a label now in common use, though other terms have been used such as cyclic v. non-cyclic (Bull, 1963), terminative v. non terminative, bounded v. unbounded, or simply state v. action, etc. The labels are not always satisfactory : as exposed above in I.2., for example, the idea of activity is often deceptive (are sleep, glow, starve, freeze activities ?) ; the label “ state verbs ” could apply as well to verbs such as sit, reside, or tarry, or perhaps wait, and Arnaud 1973 had suggested to call “ relative ” (relational) verbs those implying a stable predicational relationship, relying on the distinction epistemologists like Jean Piaget have made between relation and operation.20 To avoid ambiguities, some (Joos 1964 for example) have also suggested “ status verbs ”. Let’s keep the simplest and most usual term. State verbs : In terms of aspectual value, they can be said to form normally unbounded predications, or, more accurately, that the consideration of bounds is irrelevant. If, as many grammarians of old, beginning with the Alexandrians, rightly observed, the verb is an adjective plus BE, the copula is to be taken here in its essential sense, as defining a notional meaning which is not changing so long as the universe of discourse remains stable : when we say my tailor is rich, or he owns a house in Savile Row, we are not concerned with the fact that this will not last eternally. State verbs (or state predications) are an important class inasmuch as they are said, and have long been said, to be averse to the progressive, to the point of being completely immune to it. The question is still open whether this immunity is absolute − if a case is found, then it is traditionally allocated to another class, a very usual pedagogic device, for example with have a good time, a meal, etc., where have is said to be an action verb : take is a common gloss. Since we have no possibility of elicitation to confirm or disconfirm the labelling, it has been found safe to record all instances of such verbs, so that exceptionally for this particular class we have listed verbs (lexical units : for a complete list, see ANNEX) rather than predications ; further analysis will eventually lead to another ascription. Given the sensitivity of this class, it was separated into subclasses for more detailed analysis. Generally speaking, such verbs are so seldom found in the progressive that individual study is always possible. One of the benefits of extensive surveys is serendipity : aside from the general purpose of the search, a few interesting individual discoveries may be made. 01 : Relational/statal : verbs of the be/have type. The relation of inherent or external “ property ” they construct has been extensively studied, and, as a matter of fact, serves also as a semantic or diachronic background for the study of be and have as auxiliaries. Notice however that the stability of the relation can be implicitly limited, for example when we say : Do you have your car ? or even Do you have a car ?21 Among many interesting cases , we will only quote the following : It’s being Papa’s busy day (Gaskell, 1855) we think that we shall be having a peace with a new Ministry (Dorothy Wordsworth, 1804). 02 : Declarative : (asseverative, Joos 1964). It is probably due to the fact that many of them enter into performative utterances of everyday use that they can be said to be state verbs. Any consideration of duration or bounds is irrelevant in this case, since there is coincidence between the statement and the act, just as in performatives. There are borderline cases, however : suggest, for example, can be seen as a process. The following are illustrative of the well-known problems of performative versus constative utterances, especially because they are here at the first person of the present tense (see below in IV,9) : On the whole I am promising to occupy myself more wholesomely (Carlyle, 1833); (of course I am supposing the bare information given and left to the child to work upon) (Eliot,1847). 03 : Cognitive : this label could be disputed in cases like believe, but it was found necessary to keep the number of subclasses to a minimum. Some linguists (Joos 1964 again, after Hill 1958 ) have included them in the class of “ private verbs ”, meaning that, if they correspond to operations, such operations are not open, not perceptible, and we have to rely on the speaker for their truth-value : in this sense, believe is not different from declare. Here again the existence of bounds is irrelevant, but here again, there are dubious cases, especially among perception verbs : see and hear are a fixture in school grammars where the notion of class-cleavage and the distinction of perception v. activity are traditional. But, apart from the fact that this is an ad hoc, pedagogic, device, are we sure this applies to 19th century use ? I went out for a solitary walk and was finding myself tant soit peu tired of my dear little companions (Thackeray, 1851). 20 Admittedly, a relation results from an operation: even cognition implies discovery, acquaintance implies becoming acquainted, you cannot see something if you don't look at it, etc. 21 Yet the question has been put also for a Manx cat's tail: does a Manx cat have a tail ? speaking of an attribute more permanent and even definitional. 16 If you are seeing that honoured woman will you kindly mention this little fact to her ?(Eliot, 1880). 04 : Affective : the label fits. It is well-known that feelings can change with time, but it is not part and parcel of the definition of love to be transitory, any more than for the less emotionally-loaded like. We will see that love can be a difficult case, too easily dismissed by some grammarians who speak of Don Juanism (who is he loving now ?) and such things : . Two years ago, three men were loving her, as they called it (E. Barrett-Browning, 1846). 05/06 : Prospective/retrospective : the distinction appeared useful, but the two may be fused since there is only a small number of lexical items implying retrospective orientation, which can be studied individually. The names “ prospective ” or “ retrospective ” speak for themselves : they correspond to a disposition of the subject, normally human, whose mind − or attention, or intention − is directed towards an event of the future or the past, sometimes figuratively. There is always a boundary to it, a goal, but this goal is an event not named by the verb itself. Notice that it may seem to be not an event but a thing, a person, an object as it is aptly, with etymological accuracy, called ; but actually when one says : we are expecting Sara to-morrow, it is Sara’s arrival which is expected. When you stop hoping for something, it is not because that something has happened, but because you have lost hope, as we say. So, in the basic meaning of such verbs, the notion of bounds is also irrelevant. In retrospective verbs, the situation is different but in any case bounds are not part of the meaning of the verb proper. Both classes often have an emotional component, positive or negative, so that they could be a subclass of affective verbs. However, prospective verbs, with subjects normally human or metaphorically so, are time-oriented, and are, consequently, a borderline case of state verbs : in ordinary terms, they do not aim at a result or a term, but at a target. This leads to an aspectual contrast (before and after) not altogether very different from many process verbs. They have been ascribed by analysts to the class of state verbs, probably because they cannot easily be identified as processes, and are supposed to refer to “ mental states ” : they are “ private verbs ”(see above). But the distinction between “ mental states ” and “ mental processes ” is a difficult one. They constitute an original class, to be considered separately : a borderline class. Results will be all the more instructive. We are dying to know what that book is (Dickens 1845) ; Now as I know you must be gasping for European intelligence (Scott, 1802) ; for everyday I have been hoping it would be restored to me (Gaskell, 1853) ; The Bearer of this note is meaning to call upon you (Wm. Wordsworth, 1836) ; the ...Tranquillity, for which we had been sighing (Coleridge, 1832) ; We are starving to hear something about our chits (Mary Wordsworth, 1839) ; I was thinking of trying the ponies in the Park (Robert Browning, 1846). 10 : Locative verbs : they are easy to identify and form a comparatively manageable class. In terms of lexical items, they are not very many. Perhaps some verbs like wait (for), especially for etymological reasons, should have been included here : this is again a borderline case, or a case of class-cleavage. Generally speaking, locative verbs signify position, connection of the subject to some point in space, so that they could be said to be just avatars of be when it means be here (da sein, estar, etc.), “ exist ” in space and time. As we said above, they necessarily imply temporary state 22 : any such state will eventually and “ naturally ” end to give place − notice the traditional phrase − to a different position or posture, although this subsequent state is not its “ aim by definition ”, any more than waking up is the aim of sleeping. It is a vicious circle to say that one of the ways of defining the class is its easy compatibility with the progressive so that we will not mention it. To conclude, some members of this class could also be included among unbounded processes : lie and sleep are close to each other, as well as the two meanings of live (be alive and inhabit). However, the emphasis is on location in one case, on activity in the other, if the idea of activity is stretched to include sleep for example. Similarly, there can be some ambiguity between descriptive and locational value, since positions in space, especially for animate beings, are usually accompanied with a descriptive feature : lying, stretching, crouching, etc. Conversely, there are few activities or processes that do not occur in space as we conceive of it. A few illustrations − notice the use of stop in the sense of stay, very common at the time : buried in the same mass of fog and rain in which I had been living for weeks (Macaulay, 1834); I must tell you of one of the young women who were remaining in Prison voluntarily, until we could take them (Dickens, 1847); but we are now roosting for a little while here (Eliot, 1872); Is Fleming stopping at your house ? (Thackeray, 1852). 22 It must always be understood that it is up to the the speaker to define his universe and what he considers as stable in it or not, at the expense of misunderstandings with other people. On some of the puzzling uses of the progressive with locative predications, for example, there are interesting suggestions in Ljung 1980: 128 sq. The classification proposed here concerns referential predicates first, and cannot from the start take interpretation − in a wide sense − into account. Questions connected to it will be discussed in IV.5. 17 20-21 : Movement verbs As a rule, movement here means motion from one place to another, more or less distant. Yet, it also includes what is clearly perceived as movement without change of position. The epistemology of movement is notoriously difficult : can rain, blow or bloom be said to be movement ? 23 We have tried not to extend the notion too much : external activity nearly always implies movement, and even internal activity of the mind or soul has sometimes been treated as movement, not only by poets, but also by philosophers, psychologists, neurologists, etc. But the class is nevertheless one of the clearest, and some movement verbs, particularly common in the progressive, do not call for justification : walk, go or sail in their usual sense for example. Movement verbs, according to the well-known dichotomy, have been divided between bounded and unbounded predications. This is one of the most significant instances when the allocation has to be made in context. A characteristic example is go : it may mean unbounded movement (go on), or movement oriented to a goal even when the latter is not explicitly mentioned (let’s go home ; let’s go). This class of bounded or unbounded predications has long been recognized as particularly receptive to the progressive. Whether it is still recruiting the largest battalions or dwindling in comparative size is one of the objects of the search. 30 to 33 : Processes Movement verbs are actually just a subclass of processes (called activity verbs, in some descriptions)24.Whereas the typology of movement, in spite of some diversity, is relatively clear, the typology of processes is an endless and sophisticated catalogue where confusion of all kinds is apt to occur. Fortunately, the only thing we are concerned with is their aspectual properties. With some reluctance, and only because it is always easier to fuse unnecessary distinctions than to open new classes once the survey is complete, a discrimination has been attempted, within unbounded processes, for a subclass of “ descriptive ” verbs (scarcely deserving the name of processes in its usual sense) : I have no need of an enchanted wax figure to duplicate me for I am melting in my proper person before the fire (Keats, 1820) : Farmers are prospering beyond want (Carlyle, 1855) ; I saw him several times while he was labouring under the disease (Macaulay, 1837). Another has been suggested among bounded processes, distinguishing those which are clearly “ progressive ”, that is to say “ step by step ” : this being in accordance with a common notion of the progressive as “ developing activity ”, which certainly fits with this class. Recover, intransitive, is in this respect not very different from repair, listed here among ordinary bounded process verbs : Add to this that they are repairing the front of St Mark’s, (Ruskin, 1845), yet, a repair can be more or less instantaneous, whereas recovery is seldom so, and, as mentioned above, all operation verbs can be said to be progressive. Walter is recovering from his attack (Scott, 1829) ; Mrs Elliot has been very bad but is mending (Thackeray, 1850) ; We French say Comment vas-tu ? whereas the English expression is How are you ? − cf. also Spanish como esta ? On the use of a "progressive" by Aristotle, Cobb 1973: 84, experienced some difficulty in accounting for the Greek equivalent of walk : "he [Aristotle] argues in this passage that walking is a type of activity, X, which is distinguished by the fact that one cannot say that that something Xs and has Xed at the same time"... "the foregoing interpretation of Aristotle's point in Metaphysics θ 6 only makes sense of what he says in the text by adding a qualification to the predicate "...walks" − viz. "from Founders College to the Scott Library". In French, you don't really need that qualification if you use aller : we cannot say at the same time : je vais et je suis allé..., since aller necessarily implies directionality (apart from the set phrase above and metaphoric meanings : les affaires vont bien) , which walk does not. Yet, in the perfect expression, a place complement is normally needed. Note also the colloquial or dialectal : j'y ai été: cf. I was there (once). It would seem that static or dynamic expressions concerning our existence are sometimes interchangeable: il allait rêvant, les difficultés vont croissant, she went singing, she sat listening. Some light could be shed on the apparently contradictory use of the progressive with both locative and movement verbs by a thorough study of such expressions. 24 Incidentally, this may be due to the fact that we cannot conceive of processes not taking place − note the phrase − in space; and space is clearly included in our spontaneous conception of movement, which has not been disturbed by Einstein. Conversely, space (and time) could justifiably be considered as being, phylogenetically or ontogenetically, derived from movement: this is still an unsolved debate. 23 18 Wheat is expected to be very dear and wool is rising (Hutchinson, 1811) ; He is protracting his stay at Penang as long as he can (Macaulay, 1835). The following : I was fast ripening a plan (Coleridge, 1808) might as well have been included among prospective utterances, which shows once more that classes cannot be tight. Unbounded processes are activities which do not imply a goal or result : they just have a termination, but it is not part of their original meaning. We say that the earth revolves and that the sun shines, and we have some reason to think that this is not really eternal, but yet we do live as if it were so, as well as we accept that water boils in certain conditions as a physical law, hence situated in space but with no conceivable interruption or exception. Interruption, if it occurs, is an accident. This is the largest class, so that the quotations below are just indicative : They were acting Figaro (Carlyle, 1836) ; I was admiring the banks of the Rhine (Macaulay, 1853) ; We have been amusing ourselves with Lady Morgan’s “ Princess ” (Edgeworth, 1835) ; We are living the old life just as if we had never known Paris (E.B-Browning, 1853) ; I am puzzling my way about a poem (Scott, 1812) ; in one town of which (Douglas) the Cholera has been raging - it is now happily abated (Wordsworth, 1833) ; And a slight cough which is teazing me helps to make me think (Eliot, 1863) ; for I am yawning over the parliamentary part of the Register work (Southey, 1809). Bounded processes, on the contrary, have their conclusion built in. It can be positive (build) or negative (destroy). Operations, with identified agent or cause, leading to a clear and tangible result are not the only bounded processes, although it is often a matter of definition : hurt or kiss, e.g., are borderline cases 25, and context will help in solving the uncertainty. Progressive processes (see above) are theoretically bounded processes, but their goal is shifting and can be delayed : learning tends to a result, but, in a large number of cases, who can say that anything is definitely learnt ? grow is more clearly bounded, but not always : one even grows old, after being a grown-up. The decision whether to consider a given utterance as a bounded or an unbounded process is often a difficult one. In : the Tenant had been burning heather (Welsh, 1833), there is a definitely conclusive process, the Tenant was obviously aiming at the result. But in : 4 or 5 candles were burning before us (Mary Wordworth, 1820) whereas we know that the candles will be burnt out sooner or later, people did not light them for that purpose. There are many such cases, so that we must repeat that this is a heuristic research frame, leading to suggestions or directions for further study, rather than to clear-cut solutions. The following quotations are also indicative : while the house is being altered (Eliot, 1871) ; I write at odd minutes when the horses are changing (Edgeworth, 1813) ; Our Dressing-Table is constructing on the spot out of a large Kitchen Table belonging to the House, for doing which we have the permission of Mr.Husket, Lord Lansdown’s Painter (Austen, 1807) ; Margaret is cutting more teeth (Southey, 1803) ; The Picture is of great merit and is now engraving (Wordsworth, 1840) ; In a year...I shall be furnishing my cabin (Macaulay, 1836) ; I am seriously hurting myself in devoting my days...(Coleridge, 1818). Read is a case in point : We have been reading Darwin’s book on the “ Origin of Species ” just now : it makes an epoch (Eliot, 1859); can we be sure that George Eliot and her companion G. H. Lewis had read Darwin’s book throughout ? Yet, they probably had, since G. Eliot ends with an appreciation, and this is why the utterance was included among bounded processes ; but the opposite choice could be justified ; and a verb like read could be treated as referring to an unbounded process in whatever context : we cannot say that a book is “ read ” in the same sense as in a house is built. A general semantics of verbs, or of the notions eventually taking the form of verbs in English, would be a Promethean task. The only thing to offer to the critical reader is to provide lists for each class : they will be found in the ANNEX. 25 A special case could be made when result (effect) and action are in coincidence: she was hurt (affected) by my conduct; isn't this boring you ? aren't you bored ?... 19 A last word to say that metaphoric meanings −they are frequent : “ whither are we going ? ”, would Wordsworth say, speaking of the condition of England − have been treated like non-metaphoric ones. I.3.3.b. Verb form This term covers the morphological verb system (conjugation) inasmuch as it concerns the use of the progressive. Due to the decimal system of coding, it has been found expedient to concentrate on the major points of the system and to dispense with subsections where we know the number of progressives is small at the time of reference : infinitive, passive, modal auxiliary collocations, etc. With a small number of occurrences, more refined analysis will be easy if it is found useful. I.3.3.c. Subject This concerns the determinacy of the subject. The distinction human v. non-human or animate v. inanimate was not coded : it can usually be inferred from the person (except the third) and the meaning of the verb itself, and it is quite safe to assume that the progressive is found mostly with animate subjects and human subjects. This does not mean that inanimates, or perhaps entities outside that dichotomy like the wind or other natural elements have to be overlooked, as they too often are. The coding of persons is interesting and easy : Ota 1963 has noted that it is not indifferent to consider person in studying the progressive. 26 When we come to the third person, it may be useful to note whether it is definite or indefinite. This is a difficult question and it has not been judged necessary to enter into fine distinctions, often out of reach in our texts. Some contexts are ambiguous : one may refer to “ oneself ” or to “ anyone ”, for example. It is usually easier to see whether a definite article refers to an individual or to a class : in both cases it has been treated as definite. In practice, it has not been found useful to develop this point at the present stage of the research. I.3.3.d. Object complement (object) Definiteness of the so-called “ object ” complement is a more difficult matter. “ Object ” is taken in a very general sense : the second argument in the predication, where the first is usually called the “ subject ”. 27 Again, it is not here the place for sophisticated discussions : for example is the “ included object ”, “ object of content ” (Jespersen) a real object (live a life, sleep a sleep, etc...) ? Is mile dans walk a mile an object or a spatial “ circumstance ” ? etc. Interesting as they may be, such questions have little weight in the quantitative analysis : they may be deferred for further investigation, for which the study provides elements. However, it is not always easy to say if an object is determinate or indeterminate, a useful guide for the allocation of a given predication to the right class of bounded or unbounded process. She wrote letters may mean that she was a letterwriter ( e.g. Jane Welsh-Carlyle, in a dictionary of British authors), or that she wrote (was writing in present-day English) an unspecified number of letters at a given moment. Usually the context makes it clear, but not always. For example in : he had been writing letters, ... when he was summoned (Gaskell, 1858), we can assume that there was a definite number of letters, some of them at least completely written ; but in : I’ve been writing Pen all the morning (Thackeray, 1850), Thackeray obviously does not mean that his novel Pendennis was completed, whereas the following quotation is at least ambiguous : I have been writing letters without one moment’s cessation, from breakfast till now (Thackeray, 1856). It is commonly assumed that using the progressive, especially with the perfect tense, implies that the activity may not have been completed : but this is begging the question. A common verb with the same sort of ambiguities is read. 26 Among what are usually called state verbs, he considers a separate subclass of verbs of perception or mental or psychological state, which he calls I-verbs (those which have been called "private verbs" by Joos and Hill: verbs which "occur much more frequently with I/we as subject (both in statement and in question) or you as subject (in question) than with you as subject (in statement) or a third person subject (both in statement and in question)" p.73. This gives formal support to the definition of "private verbs" whose "static meaning is not easily compatible with the dynamics of the "process" indicated by the progressive form" p. 85. 27 This refers to the formalisation given by Culioli of thepredicativesystem: a ternary system of the type XRY, where the relational central element R (in our case, the verb) connects two arguments, X and Y, which only deserve the traditional names of subject and complement when the relation becomes oriented (for ex. active v. passive, etc.). According to the semantics of the notions involved, the nature of the relationship (mere event, processes of various kinds, etc.) may considerably vary. 20 The other items under this heading are sufficiently clear. Some groupings were inevitable. In the list, we also find predicative adjectives or nouns which correspond to a subclass of verbs : look is a clear case (you are looking fine today), differentiated from the original sense of seeing. In this section, thanks to a comfortable capacity of the decimal matrix, some distinctions have been kept which may not be useful for the first general approach, the object of this study, but could shed light on future refinements. I.3.3.e. Time complement Apart from the usual reference to the ternary division of the time-line (past-present-future), three contextual and situational elements have been noted. Iterative and continuous values have been put together because they are often impossible to discriminate − the second is the infinitesimal limit of the first − and above all because they can usually be inferred from the meaning of the verb rather than from the adverbial phrase when there is one. Even always can be iterative when smoke cigars or take medicine is the verb, whereas it is clearly continuous with flow or shine speaking of a river or the sun. To these two values has been added the generic meaning, which deserves special notice since it is normally not congenial with the progressive. Yet it has been found in a number of conspicuous cases, precisely those exceptional occurrences grammarians find precious to the point of exaggerating their significance ; for example : a man, while he is courting a woman (Coleridge, 1819) ; I live under an everlasting restraint − Never relieved except when I am composing (Keats, 1819) ; in any of the thousand ways that fortune is ever offering (Carlyle, 1823) ; Generals, ... and misses and sailors, are taking the air perpetually in its walks (Macaulay, 1814). This point will be developed in IV.5. Jespersen’s description of the progressive as providing a time-frame has long been one of the tenets in the discussions : this is why it was included in this section, to see whether it is so common as has been claimed. Similarly, it has been found interesting to survey instances of extended present, in the sense of “ dilated ” around the point of utterance. The label seems more comprehensive than “ imminence ” or “ immediate future ” by avoiding an idea of proximity which is always debatable, and is sometimes ascribed to psychological time − a valid concept in this case −and also by considering sentences when just, for example, extends to contiguous future as well as to recent past according to circumstance. In any case, the use of the progressive for extended present is a conspicuous development at our period and could be a feasible subject for a binary study of change (see ANNEX for binary glimpses). I.3.3.f. Duration The idea of duration, once commonly used in connection with the progressive, must be treated with diffidence, especially when it is based on the meaning of the verb, from punctual to extended. When there is no explicit context, one has to take into account psychological time, an interesting, but elusive notion. The concept of intervals is basic in the study of aspect, but there is no clear necessity to inquire further into their length, or their internal structure. It happens however that the context, a safe guide in this respect, contains expressions of duration. This is somehow connected with iteration, continuity and generic value by validating a time-space, bounded or unbounded. I.3.3.g Observations This was a sort of “ observations or remarks ” column, always useful in a research of that kind. Individual, remarkable quotations may thus be retrieved, disputable cases eliminated thanks to further reflection. Only one point which could not easily find its place elsewhere has been put here for convenience’s sake : the “ old passive ” forms (the house is building) − to use a practical label. They might have been inserted with ordinary present or past forms since they do not differ in form, but only by the semantics of subject-object orientation : we still accept without reservation that the water is boiling is a normal active form, whereas the dinner is cooking is said to be a relic of the old passive. By noting such outdated forms we hope to bring some light on this problem and to give some information on when and how the new passive was substituted for the old one, so systematic in Macaulay’s historical books. This summary sketch is far from a thorough discussion of all the semantics of the verb and its environment. We have to repeat that an exploratory survey aiming at the discovery of trends cannot be cramped by too strict guidelines. What is valid for the study of individual sentences may not be rewarding for an overview of a large number : in other words, whereas some theoretical views are inevitably orienting the research, the horizon must be kept sufficiently open. In particular, we must beware of notions arising either from our experience of present-day English or from the analyses we have been exposed to. This is not altogether so easy. I.3.4. Sociolinguistic factors 21 The traditional methodology ( D. Sankoff 1978, and later developments) considers a matrix of linguistic and extralinguistic factors, the latter essentially sociological. Yet, some are not strictly so, being rather the reflection in the utterances, spoken or written, of macro- or micro-sociological conditions : this is why we speak of sociolinguistic factors. In a survey of 19th century correspondence, one would like to find some social stratification28. The early Victorian English society happens to be the one which was the subject of Engels’s The Condition of the Working Class in England, published in German in 1845, in English only in 1892, where the picture is not very different from those given by the Unitarian Mrs. Gaskell in North and South. 29 This research is not concerned with such discussions : for a summary overview, see for example Best 1971, Phillips, 1984. They cannot be ignored, however. After due consideration, it was found that, in a society which was far from classless, the social group available for the study was fairly homogeneous as a sample of the larger set of literate people. A sophisticated analysis might discern subtle shades in the general colour : Wordsworth was not an aristocrat like Maria Edgeworth or even Scott, Thackeray haunted circles where Carlyle felt ill at ease, etc. But this would ultimately lead to describing the social or geographical background of the individual writers, hinting at what a finely stratified survey would provide if it were feasible, not to real sub-groupings. Pierre Bourdieu 1982 and passim (see also Encrevé et al. 1983) rightly insists that even in todays’society the language people speak and write is part and parcel of their place − a typically Victorian notion − in society and of their valuation on the “ linguistic market ”. Extending this view to the English 19th century society and to the people who actually wrote letters regularly without the help of an amanuensis, it is not difficult to draw the following picture. Generally speaking, there must have been two sorts of such people : those who wrote with fluency, and those who did not, simply because their schooling had been insufficient. Notice that among the latter were some men and women of the upper class or upper middle class. Desclaire 1973 : 487, n.25 noticed that, so far as expression was concerned, some aristocratic ladies of the 18th century were often far from being as literate as their clerical mentors and were sometimes as illiterate as their servants or Tabitha Bramble and Win Jenkins in Smollett’s Humphrey Clinker : “ L’ignorance, ou le mépris de l’orthographe, procèdent d’ailleurs du peu d’importance que les classes supérieures de la société attachent au savoir et à la culture...Mary Astell remarque que 30 “ as to spelling which they are said to be defective in, if they don’t believe as they are usually told, that its fit for ‘em to be so, and that to write exactly is too Pedantic, they may soon correct that fault... ” . The quotation is from Mary Astell, A Serious Proposal to the Ladies, for the advancement of their true and greatest interest, by a lover of their sex. in 2 parts, London, printed for R. Wilkin, 1697. But aristocrats were so secure about their social rank that, as a rule, they simply did not bother. Not being a historian, the present writer is not qualified to pass judgments on that matter : but it seems plausible that Lord or Lady So and So, or Fielding’s Squire Western for that matter, naturally considered that nice language, and correct writing, were for the parson at the low end of the table, and happily forgotten once they no longer had to be whipped for it. This intuition is endorsed by Phillips 1970 : 14, speaking of the Verney and Wentworth ladies’correspondence at the beginning of the eighteenth century : “ The Verneys and Wentworths wrote a “ prelapsarian ” English, as it were : their sentence structure was not so much incorrect as non-existent. No one then expected a lady, however nobly born, to write other than as she spoke, with, in Henry Tilney’s words [ a character in Northanger Abbey ], “ a general deficiency of subject, a total inattention to stops, and a very frequent ignorance of grammar ” ”. This probably does not apply to all upper class people, but must have been true of a number of them, especially of women. It is then reasonable to expect that their letters should be a valid testimony of the way they spoke, mutatis mutandis. This seems to have been the situation in the early 18th century. But the development of instruction certainly changed such attitudes: Dr Johnson’s lessons had reached a wider audience − Cf. Leonard 1929 −, and the upper class are more and more to include among the fully literate set.31 This source of information, not negligible, would not be considerable and as instructive in the nineteenth century. 28 Recent studies have enriched knowledge on this important point, but mostly with reference to the earlier period: cf. Nevalainen & Raumolin-Brunberg 1996. 29 Disraeli also, a Conservative dandy, gave in Sybil, or the Two Nations, a picture of that divided society ? "Our Queen reigns over the greatest nation that ever existed. "Which nation ? ... "for she reigns over two"..."Two nations; between whom there is no intercourse and no sympathy; who are as ignorant of each other's habits, thoughts, and feelings, as if they were dwellers in different zones, or inhabitants of different planets; who are formed by a different breeding, are fed by a different food, and are not governed by the same laws...'THE RICH AND THE POOR" Sybil, Chapter V. 30 "The ignorance or contempt of spelling result indeed from the limited concern of the upper classes for knowledge and culture. Mary Astell observes that" [ my translation]. 31 Incidentally, when they wrote letters and those letters have been kept, it would be advisable to check whether they were actually written by the person who signed them. When Mary was amanuensis for William Wordsworth, often suffering from his eyes, we have no reason to suspect the authenticity of his expression, except that she may have 22 At the other end of the social scale of the literate, we have what we might call the semi-literate : people who are trying to write as best they can, and will not write many letters. This simply because they don’t have the time, because the post is too expensive even after 1840, and they don’t have correspondents as literate as they are. 32 Going further into the subject would imply a study by qualified historians of early popular letter-writing in Britain. 33 Interesting insights would be obtained, but the corpus would be hard to gather and probably of small size. This is one of the reasons for leaving it aside, at least provisionally. But there is another reason : Labov, with typical sensitivity to people’s reactions and to linguistic insecurity, observed that, when people have to control what they are saying and the manner they are doing it, they are unpredictably and erratically moving between extremes of possible laxity and what is now known as hypercorrection. This means that such data is hardly reliable as a sample of community grammar. A clear case is to be found in the letters addressed by the joiner John Overs to Dickens, kept by the Free Library of Philadelphia, in Dr. Jacques Benoliel Collection in the Rare Book Departement (S.M.Smith 1974.) Here is an extract of the first, a long discussion of Carlyle’s book Chartism. July 20th, 1840 “ My dear Friend Subdued by a fit of spleen − or rather perhaps, I should say usurped by a snarling fiend of discontent and despondency which occasionaly [sic] obtains the mastery I know not why or wherefore, I dare not venture to convey to you certain remarks or rhapsodies which have arisen in my mind partly from your conversation, and partly from the digestion of the book by Thomas Carlyle entitled “ Chartism ” which your kindness has enabled me to peruse. I have therefore waited till the murrain passed away, which I hope you will recieve [sic] as an apology for the all unconscionable time that I have detained your book, and which I herewith return with many thanks... ... And now, my dear Sir, my tediousness is at an end ; I have given you the views, faithfully, of such working men as I have associated with, as recalled to memory by the book of Thomas Carlyle’s, for the perusal whereof I again sincerely thank you. I have told you I do not mix myself with these matters, because they are unprofitable, unsatisfactory, and irritating to my somewhat quiet and studious disposition ; but I can neither be blind nor deaf to what is going on about me ; therefore, if there be any matter connected with, or concerning the working man on which I have cursorily touched, the which you should like to have resolved, I shall have much pleasure in furnishing such information as you demand. Having the honour to be Your very grateful and devoted servant, John Overs. Sometimes the constraint will not be due to incapacity of the writer to write normally and even faultlessly − Overs’ English is reasonably good, with a few misspellings only : he was a lettered man wanting to write a novel − but to the social status of the addressee, often, as in the case of Dickens, felt as intimidating, in a society where submission and condescension were key words − witness Mrs. Gaskell’s Cranford and many other descriptions. So that, with condescending carelessness at one end, and humble hypercorrection at the other, the two extremes of the literate class are deviations from the community grammar we are looking for in the first instance. 34 Then we come to the question of whether Writers in the full sense of the term are to be trusted. Again, one answer would be that we have little or no choice, but the other answer is that, due precisely to their mastery of language, they suggested alterations, etc. But we very well know, and see, that politicians often read the speeches written by others. The magnates of the past also had private secretaries. 32 Apart from the possible influence of manuals of good letter-writing and of what popular letter-writers had learnt of "style" in the literature available, for example the rhetoric of love-letters or of religious effusion, or again of passionate entreaty when asking my Lord for a favour. 33 The situation in France in the 19th century has been studied by Chartier et al. 1991. The French Post-office had the idea of an extensive survey of letter-writing in 1847, before a reform introducing a uniform tax of 20 centimes whatever the distance: among other things, it shows that eight or nine tenth of the letters were business letters. 34 This is not to say that the extremes are negligible, of course. It could be suggested that grammar books take variety more into account by adopting a title in the plural. What there is in common is certainly larger than the variations we may encounter. Yet, if we believe in the consistency of system, such variations justify the plural: we might speak of "Grammar" on the one hand, and of "grammars" on the other, to underline the ordered heterogeneity which Weinreich, Labov and Herzog 1968 consider as the inevitable conditioning of change. After all, this is a less unattainable target than the quest for universal grammar, the Holy Grail of modern linguists. 23 are more apt than others to forget it when they are not writing literature, just as a dancer will walk with natural ease when he is not dancing. 35 This is what intuition as well as the practical experience of reading their letters suggests, anyhow. That it should not be carried too far is also true. From place to place we meet with some more elaborate paragraphs, according to the subject matter, which, for men and women of letters, inevitably included literature or philosophy, politics or religion, art, etc., or also according to the personality and interests of their correspondents : whereas Mrs Gaskell’s letters to her daughters were mostly those of a mother speaking of everyday life (planting flowers, buying dresses, receiving visitors, etc.),whereas Wordsworth’s letters to his publisher Moxon spoke largely of the practical aspects of publishing, some of Macaulay’s early letters to his friend Ellis were full of scholarly considerations, and Carlyle, usually concerned with a variety of everyday occupations (including the provision of clay pipes from Scotland, which, unfortunately, were broken during the transport, or of potatoes and butter in barrels ), did not fail to indulge in some of his customary moral ejaculations on the futility or even sins of the times ; so, for that matter, did Wordsworth, and... Cicero, one of the most ancient letter-writers, speaking of barbatuli juvenes. We cannot abstract the style and language of people from the circle where they live and their various occupations of body and mind. This is true for oral practice as well, and there are micro-situations of speech which ethnomethodologists or other students of social interaction have carefully analysed (Erving Goffman is an outstanding example). Here we are dealing with a large corpus of letters involving a variety of people and circumstances and the neutralization results from the mixture of many speech events which form the flow of ordinary exchange, in writing as well as in speech. To conclude, the written language we are studying is that of a set of educated people who were not extraordinary at every moment of their life. Incidentally, Wordsworth’s idea when he published The Lyrical Ballads, was that of concentrating on the common life of common people, and another Romantic, Victor Hugo, when he said : “ Ah insensé qui crois que je ne suis pas toi ! ” had similar democratic conceptions. How far was this an illusion is another matter, but the Romantics as a group were not averse to spontaneity and naturalness, even in their literary works. Each in his own way, of course. All this means that letter-writing at the time of the penny post, or just before, is to be considered as a fairly homogeneous “ genre ” : individualities, idiosyncrasies sometimes, will emerge, just as in oral conversation, when people stammer of repeat set-phrases − Dickens is celebrated for his keen observation of such tics − , but the overall tendencies of community grammar will be little affected, especially when the item studied is not perceived as a stereotype. Since no class or group discrimination was justified, only two sociological or sociolinguistic factors have been taken into account : gender and intimacy with correspondents. I.3.4.a. Gender : there is no real need to postulate a fundamental difference in mental nature and behaviour between men and women to try to assess whether they differed in language use in a given society. As many works and studies have shown, men and women were brought up differently and their lives were much more separate than they are now : Mrs Gaskell’s novel Wives and daughters, for example, is a vindication of women’s rights to learning. It should however be taken into account that the women whose letters we are reading are even more exceptional than the men : apart possibly from Mary Wordsworth, they are strong personalities, each in her own way. Some of them led a life which met with censure at the time. George Eliot and Elizabeth Barrett-Browning did not behave according to the standards which were to become emblematic of the Victorian age : George Eliot, a non-conformist in religion and in morals, lived with G.H. Lewes, whom the laws and morals of the time forbade to divorce his wife: she insisted on being called “ Mrs. Lewes ” (Haight 1968 : 229), and when he died, she took another companion much younger than herself. Elizabeth Barrett eloped with Robert Browning − creating scandal in London at the time. Whatever we may think of the division of the sexes, of the possibly bisexual origin and inner nature of man, and of the current distinction of sex and gender, it seems no artifact to have a comparative study of their language. This has often been attempted in a broad perspective by some supporters of the feminist movement, but here we are concerned with a question which was raised on several occasions and often in connection with sound change : what is the role of women in linguistic evolution ? In the domain of sound changes, Herman’s study following Gauchat’s at Charmey, in French-speaking Switzerland, gave evidence that when active change is in progress women are clearly ahead of men ; a fact which more recent surveys tend to confirm (Labov 1994, p.72,156; Tagliamonte 1998). Luckily, the corpus chosen offered quite a large amount of women’s contributions (see II.1), fairly evenly distributed along the years so that the diachronic profiles can be compared. I.3.4.b. Intimacy : common sense suggests that the degree of intimacy with the addressee has some influence on the style and, eventually, on the grammar of the letter. Intimacy, familiarity, spontaneity are near synonyms for the personal relationship a letter implies. The first has been retained because it is less circular : it does not refer to style but to the microsociology of the circle of family, friends and more distant epistolary connections of a man or a woman. This means that the question is largely a subject for biographers, who will sometimes hesitate to say whether a given person belongs to the set of intimate or less intimate correspondents. The degrees could be multiplied : anyone among us has an opinion about it, and might find it difficult to say for some of her or his friends which are more intimate than 35 This to recall Paul Valéry's description of poetry as dance and prose as walking: in writing current prose, little attention is paid to the words themselves; it is the message that counts. 24 others. Many will certainly agree that some friends can be more intimate than even close family, and the wider the circle the more difficult it is to draw the concentric limits of intimacy. Then, it is a constant fact that intimacy varies in the course of time : close friendships and even family ties are loosened and are even disrupted, while others develop. The diachrony of love and friendship is not fixed : there are stable attachments, there are also variations in the minds and hearts of both parties. People who read biographies and collections of letters become increasingly aware of such movements. That Coleridge was a changeable person is notorious. Carlyle, on the other hand, was all his life a faithful brother − while he himself confessed, after Jane died and he read her letters to publish them, that he had not always behaved towards his wife as he should have. It is not entirely possible to avoid the circularity alluded to above : it is largely by reading the letters that biographers form an idea of the feelings of the writer. At a purely formal level −not to be trusted absolutely − we can rely on expressions of friendship or affection, as opposed to the conventional phrases at the beginning or the conclusion of a letter. In that respect, English letter-writing gives less freedom than, for example, the French practice of original conclusions, with Voltaire as a typical case. A letter beginning with “ My Lord ” and concluding with “ Your Lordship’s faithful and obedient servant ” is not an intimate letter like one beginning with “ Dear Alick ” (Carlyle), or “ dear Barbara ” (Eliot) : yet, Wordsworth, although not indulging in familiarity with him, had become a friend of Lord Lonsdale and was a friendly adviser to Sir George Beaumont, another aristocrat ; Thackeray, not a member of the nobility, but no doubt a “ gigman ” − to use Carlyle’s sarcastic epithet − enjoyed sufficient freedom with fashionable ladies like Lady Castlereagh to address them as friends. The study of introductory or concluding phrases could be carried further for a comparative study of the relation between form and feeling, of the degrees of reserve and warmth. To conclude on this delicate point, we will say that the final decision results from a confrontation between the formal features found in the letters, the general impression they give and biographical data : this means that we have, as often in biological and human matters, fuzzy sets. Since we are aiming at general and solid results, it has been found sufficient to consider three degrees of intimacy : (1) close, (3) distant, and (2) in-between. In the usual sense of the term, there are no business letters in the collections. Yet, some are not far from it. The most common case is when, for example, at least for part of a letter addressed to a publisher, money matters are raised : this even happens when George Eliot writes to her old publisher and friend John Blackwood. When Wordsworth was campaigning for the protection of writers’s rights, he wrote many letters to Sargeant Balfour, M.P. or even to the Prime Minister Sir Robert Peel, which can be considered as business letters, even when he takes the opportunity to ask for personal news : such letters have not been eliminated as a rule, but they are included in class 3 (more distant). At the intimate end of the scale, we find the family circle. One might object to putting together the type of intimacy between a boy addressing his mother (John Ruskin when he was still a boy or in College), or his father (Macaulay), a grown-up man (Carlyle) writing regularly to his mother or his brothers, and loveletters (those, published in a separate volume, of the Brownings), letters from London or from abroad to the family at home (Maria Edgeworth, Wordsworth, Dickens, Thackeray, etc.) or to brothers, sisters or close friends (Jane Austen, Carlyle, Macaulay, etc.). Still, from the point of view we are adopting here, they share the same freedom of style, casual, newsy, little concerned with form except in its transparency to the feelings, even when it is careful as befits a good writer. For some of the letter-writers, it has not been found useful to have separate degrees : Mary Wordsworth’s letters are addressed to family and close friends only, Jane Austen’s almost entirely to her sister Cassandra, and Keats’s to brothers and friends, in the same tone and erratic spelling. In other cases, there was not enough biographical information available for a scale of distance : except when the content and form of the letter made it clear that it was not addressed to an intimate or a friend, degree 2 was then taken as the neutral class of friendly interchange. When degrees have been established, a list of the correspondents, for degrees 1 and 2 mostly, is given in the ANNEX. Needless to say it could not be complete : only those who were fairly regular addressees have been noted ; nor could it be very strict. As said above, a connection beginning in formal terms sometimes became more and more intimate (e.g. Elma Stuart in the case of G. Eliot) so that all letters to the same person were not always in the same category : but then the moment when the nature of the relation changes is largely subjective, since it is a gradual phenomenon. The lists are largely indicative, and can be confronted with available biographical information. 25 CHAPTER TWO THE CORPUS II.1. Size and contents The corpus of the survey consists of published private letters of 22 people born between the last decades of the 18th century (1767 : Maria Edgeworth), and 1819 (John Ruskin), most of them famous writers of the Romantic Age. The letters were written between 1780 and 1880, approximately, individual extant contributions showing much variation over the years, as could be expected. In a few cases (Scott, Southey, Browning) parts of the collections published at the time were not surveyed. Other letter-writers might also have been included, but as it is, the corpus is substantial : nearly 10 million words, or in other terms, some 30,000 pages or 70 sizeable volumes. Special attention was given to the contribution of women : it is only about half that of men in terms of text-length, but this is more than could have been expected, and, since they use the progressive more than men do, on the total of nearly 22,000 occurrences of the progressive, their share is about 8,000, more than two thirds of the contribution of men. A list of the writers is given in the ANNEX, as well as the bibliography of the collections. The choice of this corpus was not a matter of mere chance or intuition. A comprehensive survey of early and middle nineteenth century English progressives had been made for a doctoral dissertation (Arnaud 1973). The aim then was not a quantitative computation, but it soon became evident that the form was far from evenly distributed in the variety of texts available. Poetry left aside as one type of discourse most distant from ordinary verbal intercourse, it was clear that the frequency varied along some sort of formality scale. The point had already been noted by others, Ota 1963 especially, for contemporary American English, Mossé 1938 more generally. So that if one wanted to find substance for a study relying on attested sentences, it was wiser to concentrate on the documents where density was greatest. The historical linguist cannot rely on elicitations and his own intuitions, even when he has become familiar with the period’s ways, must be followed with diffidence.36 It soon became clear that it needed less extensive reading to find progressives in Punch than in The Annual Register, in lively novels and in their dialogues than in philosophical or historical essays, and in everyday correspondence than in political speeches in the House of Commons37. Even the published scripts of murder trials ( The Trial of William Palmer, 1856, is a good example : see Knott 1952), which seem close to ordinary speech when witnesses from various social classes are crossquestioned, must be considered with precaution : the Assizes is not a place where people speak without control, and the minutes were written by clerks. Definitely, the location for a reasonable density of progressives and the most relaxed type of discourse is the large corpus of private letters kept and collected by the devotees of eminent men and women of the time 38. 36 Is a non-native even less qualified ? This is an old dispute among ethnologists. He is probably less prone than a native speaker to identify the language of the past with that of the present (see above I,1): a French person cannot easily conceive that contemporaries of Flaubert, apart from the novelist himself, did not use the "imparfait" as we do now, an English reader feels quite at home in Dickens's or Mrs Gaskell's novels; but if we exchange books, perhaps the non-native will be less easily deceived. 37 Hansard's transcriptions seem fairly reliable, but reported speech is their rule, and MP's address a very special audience. 38 M.Rydén and S. Brorström 1987, in their extensive survey of the be/have variation in the 18th and 19th century, have also concentrated on private letters; and on comedies, too. It might be worthwhile to survey some 19th century comedies, but there seems to be some artificiality in their style. If a survey of dialogue in novels, which was attempted in Arnaud 1973 and was also suggested by Strang 1982, was not so time-consuming, it would provide a better sample of everyday conversation than comedies. There is always some literary distortion, however. In recent years, explorations of the sociolinguistic variables have been undertaken, and are still continuing. The work of the Helsinki school is considerable in this respect (see e.g. Nevalainen, Raumolin-Brunberg 1996). However, much of what has been done applies to Early English, i.e. before that golden age of private correspondence considered here, which means a different methodological approach: for example, there are occasions when letters by educated ladies follow a conventional pattern. This is no longer the case after 1800. The conclusion should be that there is no "degree zero" of common speech at any time, and that each "genre" deserves to be studied in its own right, with due 26 No doubt manuscript letters from “ ordinary people ” could be found in private or public records : they would seem more authentic witnesses of every man’s or woman’s everyday language. Apart from the disproportionate effort it would require to gather a sufficient amount of material, this is certainly a fallacy. Either the writers were really literate people and the difference between their letters and those studied here would be insignificant, especially concerning a point of grammar not fully conscious, or they were half-literate, and then hypercorrection comes into play : the letters of the joiner John Overs to Dickens ( Smith 1974 ) are typical in this respect (see I.3.4.). Another danger might be that, for some at least, an artificial prose inspired by school-teaching could prevail − young ladies were taught letter-writing manners by teachers or governesses. When you read George Eliot’s early letters to her old affectionate evangelical schoolmistress, Miss Maria Lewis, you are aware of that risk39. There is no need to apologize, then, for the choice of letter-writers who were writers at the same time : accessibility is not the only reason for the choice ; these good writers write ordinary letters most of the time, not monuments of literature. To conclude on this point, we may also recall that the ordinary man is an abstraction : no one exactly fits the model, no Englishman was really John Bull. In sociolinguistic terms, statistical survey of a comprehensive range of speakerwriters is just an inevitable way of smoothing individual idiolects to have a reliable picture of community speech. As to the actual choice of the letter-writers, university library shelves display a long array of volumes of letters from eminent people of the period. The idea was to gather as representative a selection as possible, arranged in chronological order with the best possible contribution of women and it does not claim to be more than a sample of what is already a sample of what was the correspondence of the time. Not all collections of letters have been edited as we would like to find them. There has been steady progress in this respect, and the ongoing publication of the Carlyles’ or of Dickens’s letters are a model of what ought to be done. Many older selections were prepared by family or friends who had their views about what was good, worthless or inappropriate. Usually, they aim at biographers and literary critics, seldom if at all do they think that the text can also be precious for linguists 40. Modern editors record everything, even when they find the handwriting illegible : they simply insert a note to help the reader. They must be congratulated for it. It does not matter very much to us when letters have been omitted or emended for various moral reasons, although it matters a little ; but it seems that in some cases, letters or paragraphs have been suppressed simply because the subjet-matter seemed uninteresting while it is precisely where we could find the informal style we are looking for. It is also important that all forms should be kept without emendations or corrections. John Keats’s bad grammar and spelling is well-known : fortunately his letters have been published, apparently, as they were written. But we cannot always be confident that it was so for others, especially when they were printed in the 19th Century, so that selections which could have been interesting have had to be left aside : Disraeli’s letters for example, because they are scattered and emended extracts in a biographical narrative ; Mary Shelley’s, Landor’s, Clough’s were discarded after inspection, etc. We are fortunate that there is a fair amount of women’s letters. Discontinuous as they are, those of Maria Edgeworth have been included for that reason and for their early date. The Brownings’letters are scattered in several volumes and sorting them was not easy : Robert’s corpus is limited and not very instructive, though he lived till old age. Solid and substantial collections are of course the best, not only because they offer a rich harvest, but also because they allow us to outline the linguistic trajectory of the individual person. George Eliot’s letters, magnificently published by Gordon S. Haight, one of the first to do such work, gave a starting point (Arnaud 1980, 1982) : covering all her adult life with reasonable regularity. But we must be grateful to the people who kept and those who published the extensive collections which form the major part of our corpus : letters were then considered precious, even before their writers became famous. What will be preserved of our telephone conversations ? consideration of the sociolinguistics of the times as the Helsinki researchers rightly emphasized. Explorations into other "genres", comedies for example, can give interesting insights, but on the whole they seem more remote from everyday speech. 39 The fairly abundant literature of what Violette Desclaire 1973 calls "open letters", that is letters addressed to the public, or to a definite public, and often found in magazines, is not what we are looking for, and seems to have been a practice of the previous century, no longer in favour after 1800; even Maria Edgeworth's letters from England are family letters. 40 When Wordsworth, who advised his son to drop his Northern pronunciation when he became a vicar in East Anglia, forgets his spelling and writes nought for note, this may be a hint at what that Northern pronunciation was. 27 II.2. Sampling The sample lends itself to qualitative as well as quantitative studies. Nothing is anonymous: our letter-writers are something like the literary jet-set of the time.41 This means that the selection cannot be said to be a statistical random sample in the technical sense of the term. Population would be a more appropriate term, though it is only a fragment of the overall population. As stated above, this was a deliberate choice, meant to provide maximum information, since it covers practically all the available letters of the authors. It does not preclude further statistical treatment. The question still remains : according to what laws governing the use of language in a given community could sampling be done ? Is every person to be given the same weight as in pre-election surveys ? We know that some people act as leaders in language change, and that the study of diffusion can benefit by consideration of the networks of social relations (Cf. Milroy 1987 ; Labov 2001), but to unravel them in communities long extinct and extend to written texts what can be valid for oral practice would be a very long task indeed. The only point where something like quota sampling will be found here is the substantial contribution of women letter-writers. Even if we submit to the dictate that language is not a statistical concept, nor even immanent in a corpus, quantitative analysis of a sizeable corpus is at least one of the ways of assessing some essential phenomena, especially in the domain of change. To what extent can the observations be valid for the community at large is, theoretically, impossible to demonstrate. In practice, however, that is what corpus-oriented linguists in search of plausible trends always imply. Some other questions cannot be evaded. Quota sampling was mentioned in the case of women versus men, but what about age-groups, generations ? When we put together all the letters written during a given period, we do not separate the younger from the older people : this could possibly be done on the basis of the data provided. It can be another occasion for establishing the difficult relationship between apparent-time and real-time surveys (Labov 1994). The attempted comparison (see below III.1.3.) for the period 1835-40 is indicative in this respect. It seems reasonable to guess that, if only people of the same generation were considered − thus reducing considerably the size of the data − there would be more stability within each group, and a greater gap between groups. Even when people follow the general trend, when they grow old they are sometimes left behind : this question of community change and generational change will be considered in section III.1.3. As was to be expected, the contributions in terms of length of text are variable not only from person to person but for a given individual according to life-circumstances. People write more letters, and longer, when they are away from home : Maria Edgeworth and Wordsworth wrote when they were in London, and many letters were sent from abroad, Macaulay’s from India, George Eliot’s on her tours on the Continent, Ruskin’s when he was painting in Italy, Dickens’s and Thackeray’s from America, etc. This is especially true for letters written to near relatives. Then we have what can be seen as business letters : those addressed to publishers, for example, and the many letters Wordsworth sent to Lord Lonsdale or his son Lord Lowther when he was canvassing for their election, etc. They have been included, in consideration of the fact that publishers, for example, had usually become friends, with ups and downs perhaps, but no more than with other friends : George Eliot’s friendship with John Blackwood − not to speak of John Chapman, with whom she once lived − is a good example of long-standing intimacy. When interpreting the results, this can be taken into account, as well as the cooling of regular friendships— for example, on the occasion of the affair of the Garrick Club,in 1858, between Thackeray and Dickens, with their common friends Yates and Forster involved in the storm —; or even the vicissitudes of family relationships — something happened between Mrs Gaskell and Marianne in 1858, and Macaulay was obviously disturbed when his dear sister Hannah married in 1835 : see III.1.4.b. A thoroughly significant search for such clues would be doing biography in detail : we have to be satisfied here with very general lines, and mere hints. 41 It could be interesting to have linguistic data from the jet-set, or even royalty. Incidentally, the published letters of Queen Victoria, even those to her favourite daughter the Princess Royal (Dearest child, 1858-1861) display a very small quota of progressives: the density, evaluated on a corpus of 22,000 words only, is 56; none of our letterwriters shows so small a figure. Apparently, "The Queen was not amused" − to quote one of her favourite ejaculations − with the progressive form, perhaps incongruous in her environment. 28 As could be expected, the corpus is not evenly distributed : the beginning and end periods or decades are less substantial than the central ones : this is due, except in a few cases (Scott, Southey) either to the absence of letters or to incomplete publication at the time of the survey (Carlyles, Dickens). It could be corrected for the last decade by surveying new collections of letters. For the early years, it does not seem that much more can be found, especially in women’s correspondence. Some attempts have been made to compensate for this unevenness, by adding periods or decades, or by computing equivalents “ lots ” in terms of word number : they did not bring significant improvements, so that they are not presented here. However, the letters written before 1800 have been included in the first “ decade ”, and the last “ decade ” covers the period 1865-1880, in order to enrich the data for better approximation. II.3. Community grammars and idiolects So much for the quantitative aspect. In terms of sociolinguistic significance, other questions arise. Chomsky’s homogeneous community is, he owned it, an artifact. Since we are attempting a study of Saussurean parole, a word must be said about the main characteristics of each writer, in terms of naturalness and what we call spontaneity, above all. The reading of that prose from people who have a certain image can provoke surprise. One would expect Wordsworth’s and Carlyle’s style to be somewhat cramped : it is not so. Wordsworth lamented, like Cicero, the disruption of society − he predicted the fall of England at the time when she was becoming the leading world-power − but he did so in the familiar vein, with few oratorical effects. Carlyle wrote in plain terms even at the time when he was translating Goethe or writing Sartor Resartus in a style famous for its ruggedness : how could he have written otherwise when he was writing to his mother, an elderly or old peasant woman smoking her pipe by her Scottish fireside, or to his farmer brothers, or even to his beloved and ill-treated Jeannie, his wife ? George Eliot has been said to be sesquipedalian : what strikes most when you read her letters, especially in the later years, is her everrecurring complaints. She confesses that, when not too well, she is “ writing on her knees ” : how can one be stilted in such a posture ? The champion of cursiveness is certainly Mrs Gaskell : always busy, always in high spirits it seems, she writes with a lively style. Conversely, Macaulay, often writing from the House of Commons, tends to transfer to some of his letters, where we find occasional lengthy dissertations, some of the tone of his Parliamentary eloquence : but not so much as one might expect. As to Scott, taking leave at times of the boring speeches in Court, where he was a Justice (Sheriff-Depute, or Shirra, to use the Scottish word he loved), he seems to enjoy the relaxation he is indulging in. With Dickens and Thackeray, there is some variety. The latter’s numerous brief notes (there was no telephone at the time to suggest or cancel an appointement) have been discarded. At times their letters make one think of their novels, glimpses of humorous stories, portraits, etc. but seldom to the point of being elaborate literary pieces. Jane Welsh-Carlyle is a different case. She is sometimes said to be the British Madame de Sévigné, but whereas it is well-known that the latter wrote for an audience of appreciative people, played the fine writer, carefully calculating her effects, Jane Carlyle is a newsy and marvellous letter-writer simply because it was in her nature. So is it with Jane Austen, who was perhaps, however, one of the first among moderns to take letterwriting as a possible genre − apart from epistolary novels, of course. Yet her simplicity never seems affected. Indeed, letter-writing can be seen as a literary genre. Ordinary conversation could be one as well : Coleridge is said to have been a brilliant talker ; Scott was one too, in a different manner ; we cannot imagine Wordsworth or Carlyle in the same role. Everyone to his own style, this is all we can say. Is there a “ degree zero ” of linguistic expression ? The question remains open, and the answer is probably negative, but, so far as grammar (or pronunciation) is concerned, we must keep to Labov’s rule : “ least controlled most regular ”. Letter-writing answers the requirement in the best possible way. This study is not biography nor literary history or criticism : for a more accurate description, the reader should refer to biographies or to the various commentaries in the collections of letters. It may be helpful, however, to give a short summary of each individual author’s “ epistolary career ”, if we can risk the expression. It stresses also one of the problems we are facing in micro-diachronic analysis : the passing of time implies that everything is moving. A few letters from early youth have been preserved : they are always addressed to parents. Not only those of Ruskin, a much-petted child, but also those of Coleridge − very few − of Macaulay, of Thackeray. Carlyle, as already mentioned, kept faithfully writing long letters to his mother as long as she lived. Then, we have letters to friends at the passionate time of adolescence. Then come love letters : there is a volume of them by the Brownings, but there are also some instances of letters addressed by Scott to his future wife, Miss Carpenter, and of Thomas and Jane Carlyle − were they also love letters in their married life, when one of the couple was on a journey ? Certainly, at least in a way. Even Dickens wrote lovingly to Kate, from whom he was later to become estranged, as to an even 29 greater degree was Thackeray whose wife Isabella had to be sent to an asylum. Then we have all the variations of maturity and middle-age, and, later, the vicissitudes of life mean that some correspondents disappeared, sometimes simply because they had died. The following remarks are presented in the chronological order of the letter-writers. Maria Edgeworth’s main body of letters were written when she left Ireland, her usual abode, to visit friends − personal, political, literary − in the fashionable world of England. We have every reason to think that, while she knew the English spoken in her Celtic surroundings42, even if she possibly could not speak it as well as her friend Walter Scott spoke Scots (Tulloch, 1980), this was not the language she was inclined to write, even to family or friends. There is every reason to suggest that she even kept as much away from it as possible. So that Anglo-Irish influence can hardly be suspected in her grammar. Then we find the group of the Lake Poets, as they have been called. Coleridge was born in Devon, and never lost his local accent. His letters seem to be as erratic as his life, so that it is difficult to have significant guidelines, even to locate the various degrees of intimacy with his correspondents, since he was so elusive and changeable. Robert Southey was born in Bristol. He spent some time in Portugal, and sent many letters from there. But he lived in the Lake District for years, to end his life there, in seclusion and dotage. The others (to whom we have added Sara Hutchinson, since she really belongs there) are a more homogeneous group : people really from the mountains of Cumberland and Westmoreland, and certainly affected by the language spoken there. However, in their writing, they seem to adhere to a fairly controlled Johnsonian English. Unconsciously, they may have carried into it some grammatical features more widespread in the North, the progressive among them. Dorothy Wordsworth became a dotard at a comparatively early age (before 50). Before that time, she does not seem to write very differently from her brother, except that she spoke more willingly of children and domestic events. This is even more perceptible in Mary Wordsworth’s letters, almost all sent to family or close friends, and speaking usually of family matters. Their companion Sara Hutchinson, Mary’s sister, once wooed by Coleridge as his muse “ Asra ”, writes about everyday topics : for the Wordsworth circle, this inevitably includes ongoing literary activities. Coleridge himself, a notoriously versatile man and poet, also usually writes the ordinary news of an extraordinary mind, visionary, often full of passionate ejaculations : he is a borderline case of what we call familiar letter-writing. Very different are Robert Southey’s letters : a literary man and historian, once a close friend of Coleridge, he writes about his works and comments on contemporary events, but there are few pages not written in the usual style of friendly news. No significant difference is perceptible in the language of all these writers, still formed in the eighteenth century tradition of correctness, whether they lived in the North, or the South, or even in Scotland or Ireland. There was some friendly correspondence between Scott and Maria Edgeworth or Wordsworth, both of whom visited him in Scotland. So far as conscious wording is concerned, and since some instances of the progressive may have been already stigmatized as celticisms, as they will certainly be later (Cf. Bain 1863), up to a point they must have resisted the Northern or Celtic influence, in the name of good English. Between that early group and Carlyle or Macaulay there seems to be a gap. The former had lived the times of the French Revolution and the Napoleonic Wars, and the latter were really men of the new age. Between Carlyle and Scott there is also a difference of social class : Scott, a lawyer in Edinburgh, had high connections, such as the Duke and Duchess of Buccleuch, and then King George IV himself − he was in charge of his reception in Scotland − and Queen Caroline, whom he visited in Kensington. The simplicity of his epistolary style is that of a well-born gentleman who was to become the Knight of Abbottsford. Carlyle was a son of a Scottish mason, and most of his family were still farmers, with the exception of his brother Jack, a regular correspondent, who became a doctor. He lived in Scotland for the first forty years of his life, and his letters, often addressed to his mother, are full of questions on the usual occupations and interests of people living in villages : ploughing and harvesting, buying and selling 42 Strang 1982 noted the exceptional frequency of the progressive in Castle Rackrent: although my own figure (360) based on the whole book is lower than hers (500), it remains very high, and is obviously due to the narrator's being an Irish steward. 30 horses, drinking, having love-affairs and illnesses. No hint of gentility in all that : he was a blunt man, later to abhor what he called “ gigmanism ”, the worldly refinement of London gentlefolks his friend Thackeray was so fond of. Beautifully edited by C.R. Sanders and K.J. Fielding, et al.( Duke-Edinburgh Edition), his letters are an exceptional testimony. At the same time, largely due to his Scottish origin, and to his personality, he is a little apart from the community we are exploring. Macaulay, in spite of his connections with Scotland, is a very different letter-writer. His life was centered on Parliamentary life and the writing of historical essays. There are two sides to his correspondence. Addressed to his friends, many letters are about literary matters and are sometimes stodgy, with references to the classics − this is not uncommon, especially with young university men of the time ; but there is a large number of letters to his sisters (Hannah and Margaret) where he is more relaxed in his style as well as in his humour. From India, where he spent many years, he had the satisfaction of earning a lot of money but the litany of his preparations for return tell many correspondents how glad he was to see England again. Carlyle and Macaulay are a sort of transition before we find the great writers we have called “ the later Romantics ”, a practical label which literary historians might not approve. An important historical event must be recalled : the introduction of the Penny Post in 1840. Whereas Wordsworth, a man of small means − he wrote many letters begging employment for his son − often asked Lord Lonsdale to frank his letters, whereas Carlyle, to save postage, kept a system of signals with his family in Mainhill by scoring lines on newspapers, the Penny Post opened new perspectives and the early Victorians became champions of letter-writing. People seem to have been seized with a frenzy of correspondence. Let’s just note at this point that the overall content and manner of that correspondence is fairly homogeneous. The letters vary in length according to needs. Editors have published short notes carried by messengers in London, for example, which we have not included in the count (especially in the case of Thackeray). The longest letters are usually those sent from abroad, from the Continent, or, by Dickens and Thackeray from America. The rythm of existence was accelerating : what with writing books, delivering lectures, travelling abroad, receiving visitors, these early Victorian celebrities sometimes managed to write very long letters. Similarly, in France, George Sand, who fascinated some of them (women like E.Barrett-Browning and, of course, George Eliot), an extraordinarily active woman, found the time, often in the night,to write so many letters that their publication by Georges Lubin consists of 27 thick volumes. No one among the English writers was as prolific as she was, but many have left a good number of letters : nearly 1600, for example, in Thackeray’s case. The letters of John Ruskin are addressed to his parents, including two collections sent from Italy in 1845 and 185152, and this means a certain homogeneity in subject matter and tone. II.4. Identifying the progressive Grammars often refer to expressions that might be confused with “ true ” progressives. Thus it seems appropriate to list briefly some of these, most of them quoted from the letters. Usually, the context was sufficiently clear to resolve the ambiguity. The few dubious cases remaining were omitted from the count. a. Adjectives : Some attributive -ing adjectives of temporary scope, analogous to adjectives such as busy (working), asleep (sleeping) or ill (suffering), can be found in predicative position and thus confused with progressives, whose adjectival force is obvious Even smiling, for example, can be ambiguous in some contexts, such as : it [thecottage ] is smiling in sunset (Thackeray, Pendennis). Also : Mary has been a little peeking I believe (Thackeray 1842 ). But in : An infant is almost always sleeping (Coleridge, 1898 ), we have a verb-form. Sometimes, as in the case of outstanding, forthcoming, the absence of a regular verb (*outstand, *forthcome) excludes the progressive label, but what about : New men, Mr. Harding, ... are now forthcoming in the church (Trollope, Barchester Towers) ? With owing, the context is necessary. Other ambiguous cases include : calculating, complaining, consenting, deserving, existing, forbearing, stirring, and worrying. We also encounter prepositional expressions such as : they are facts which ... are well deserving of attention. In : are you, or is Dr. Mitford, understanding of these things ? (BarrettBrowning to Miss Mitford, 1837 ;15) the preposition confirms the adjectival status (*are you understanding this ?). But at the time, there were still instances of expressions where of follows a progressive, especially in popular or dialectal use (observed in Dickensian characters for example) : they are not present in this corpus. 31 b. Gerundial forms : In : would it be asking too great a favour to beg you... ? the expression is clearly a nominal one ; but in : It would however I fear be outraging the character of antiquary to restore this noble feature (Scott, 1830 ), there may be a little room for hesitation, although it refers to an infinitive. Very few cases are uncertain, however. c. Word-order : Insertion of adverbials between be and -ing is sometimes puzzling : Miss Q. was out riding (Mary Wordsworth, 1839 ) ; while I was in town house hunting (Edgeworth, 1822 ) ; We have been here praying to have the spring put off (Edgeworth, 1898 ) ; and even : I was, a few days after that order reached us, in a rick-yard, looking at the ricks (Cobbett) ; we find it hard to decide if these are real progressives : grammarians sometimes speak of “ dangling participles ”. They were not included, nor such expressions as: I am very near agreeing with Hazlitt that Shakespeare is enough for us (Keats, R, 1817 ), or : that father of mine is “ in trouble again ”. How long he is, growing up to be a man (Dickens, P, 1842 ). d. Near equivalents. The best known is : in the act of, as in: I am just in the act of getting done with that thrice wearisome Legendre (Carlyle, 1821 ) ; we also find: I have been engaged in travelling backwards and forwards to Selkirkshire (Scott, 1804 ). Contrast these examples with in the habit of as in: I am in the habit of taking my papers to Dilke (Keats, 1818 ). Some set-phrases including -ing forms, were very popular at the time ; some of them are still in use. It was ringing ten, striking four, he is rising five (years of age) : [of a horse] or I am going in seventeen, can be seen as progressives, but not : it was locking up (time). Nor can : In two minutes the vessel will be burning hot (Coleridge, 1799 ) ; I was weeping glad (Coleridge, 1824 ), easily be considered as true progressives. An interesting case is that of double progressives such as : I have been sitting thinking of it these 10 minutes (Thackeray, 1849 ) which was counted as two progressive forms. In the careless style of some letters, the well-formedness of a sentence is difficult to judge, as in How sick and sleepless I was lying in bed, when I was told that you were come (Macaulay, 1821 ) : he was not yet, at 21, the eloquent M.P. and historian he later became. As a last word, consider going to, which has been completely grammaticalized in Present Day English. The grammaticalization was not really complete in Early Modern English, where we sometimes find it in competition with coming to. In rare cases, it is impossible to say whether, for example, The President has sent to me twice, and I am going to see him to-morrow (Dickens, 1864 ) is metaphorically referring to an arrrangement (i.e. a new way of stating what was earlier expressed as I am to see the President to-morrow) or to actual movement. II.5. Data collection. Surveying such a huge collection took a long time, and the methods had to be perfected. One of the problems is the lack of homogeneity of the collections, in material form. Volumes vary in size and in print, letters are sometimes numbered, sometimes not, only dated when possible. There are supplements, footnotes contribute to variations in the length of pages of text. This is reflected in the system of references used in the bibliography of the letter-writers. Since the frequency (density here) is calculated on 100 000 words, the way words are counted is important: computers deal with that problem now, never entirely satisfactorily. It would take volumes to define what a word is . I have relied on the common-sense notion of graphic items between spaces (compound words making one unit, as well as abbreviations or figures.). Then random sampling must be attempted, using samples not too long: in this study, the unit was ten times ten complete lines randomly selected in a given book, (by sytematic sampling, to make things easier: i.e. given two random numbers (odd and even, to equalize chances) the pages were selected at a 10 page interval over the whole volume), so that the average number of words per line could be calculated, then to be applied to the whole volume, or volumes. Then, the number of lines for the whole volume was counted as strictly as 32 possible, eliminating pictures, blanks, heads of chapters, etc. At present, the computer certainly allows more precision, provided it is duly fed.43 To conclude, let’s stress again that the study, largely because it is exploratory, relies on a conception of the use of language allowing for some fuzziness in everyday language, even in the case of very clever writers. This lack of rigidity so uncongenial to prescriptivists as well as to formal linguists is probably one of the conditions for language change, in a dialectics of understanding/misunderstanding, functionality/disfunctionality. On this point, often underlined by Labov − cf. for example Labov 1994 : 11 : “ Historical linguistics can then be thought of as the art of making the best use of bad data ” ; see also Arnaud 1983. 43 The trouble with some estimates (Mossé's 1938, for instance) is that they don't tell us how the count was made, and I found it sometimes differed significantly from my own. This is why I have found it necessary to give some precisions, without writing a methodological treatise. 33 CHAPTER THREE RESULTS III.1. Density III.1.1. Overall increase : The first measure of the development for a period of nearly a century between the extreme dates, is that of the density of progressives in the text (frequency, number of occurrences per 100 000 words).44 It was originally computed by five-year periods, but the general trend is more clearly displayed by “ decades ” in Figure 1, where data for the years before 1800 and for 1865-80 respectively − hence not strictly decades − have been fused. However, since the size of the corpus over the years displays a bell-shaped curve, the validity of the graph can be questioned, in the absence of statistical tests (see above I.1). This is why another method has also been used. It consists of dividing the data (words and PF’s) into “ lots ” of equivalent size, arranged in chronological order. This implies a different weighting of people and dates, thus compensating for one of the possibilities of bias, but the even spacing of the successive periods is not preserved and we have only ordinal scaling (Figure 3). We can see that not only has the density increased from approximately 140 in the first decade of the century to reach a peak of some 350 by the 1870s, but the increase is regular over the years. What had been suspected on the basis of test counts is verified with a much larger corpus. Besides this overall estimate and anticipating our study of their possible role, the curve for women shown in Figure 2 displays an increase parallel to that of men. III.1.2. The two romantic generations : Using the word generation in a wide sense, a chronological grouping of the writers has been found useful, evaluating the process, this time discontinuously, by putting together the entire life-production of each age-group. It is a different approach, in that it does not account for possible evolution in a writer’s lifetime : individual trajectories will be studied below III.1.4.b. We do not take it for granted that people do not alter their language manners in the course of their lives, any more than other manners : see below III.1.3. Two generations have been distinguished, with nearly equal contribution in terms of text-length (around 3 600 000 words). The number of men and women in the two groups is equivalent though their contributions are not : women are under-represented. 44 This gives numbers below 400, of which the last figure can usually be neglected.. Mossé 1938: I, 65; II, 30,n.; 271 suggested it; Arnaud 1973 followed suit, as well as Scheffer 1975: 78, independently (he calls it the Mcoefficient). Scheffer also discusses its validity, and suggests another measurement, proposed by Nickel 1966, which he calls the K-coefficient obtained by "dividing the number of progressives in a text by the total number of all verbal forms less those of which it is ascertained that they cannot possibly be replaced by a progressive, and multiplying by 10 000 to get a workable coefficient". This last solution is ideally better than the one I have chosen. Yet it meets with two serious objections: 1. It is definitely difficult, on the one hand, to ascertain which verbal forms cannot be replaced by a progressive especially in a variety of often long non-contemporary texts; this would usually be begging the question. 2. Except if it were limited to part of the verb system (one tense-form, for example), it would render the task really gigantic. 3. It implies a binary view of progressive v. non-progressive forms, which is not entirely legitimate (see I.2). 4. This does not mean that binary studies are valueless. In order to overcome this considerable handicap, Arnaud 1973 used a separate count of non-progressive verb forms in samples of texts to compare their density with that of progressives in other samples of the same text. The comparative distribution of 1000 non-progressives and 1000 progressives gave a measure (Q) of the incidence of the latter. Yet what could be undertaken for a comparatively limited survey was out of the question for a much more extensive one. On this point see Binary glimpses in the ANNEX. 34 Group A : THE EARLY ROMANTICS : Maria EDGEWORTH, Jane AUSTEN, Samuel Taylor COLERIDGE, the WORDSWORTH circle (William, Mary and Sara HUTCHINSON, but excepting Dorothy WORDSWORTH to keep the two groups of equal size), Walter SCOTT, and John KEATS. John RUSKIN’s parents have also been included, with a small contribution. This is the generation of Napoleon Bonaparte and Wellington, both born in 1769, and contemporaries of Maria Edgeworth, of Walter Scott and of the Lake poets. The median of their letter-writing production is situated around 1810-15. Group B : THE LATER ROMANTICS : Thomas Babington MACAULAY, Elizabeth BARRETT-BROWNING, Charles DICKENS, George ELIOT, Elizabeth GASKELL, William Makepeace THACKERAY, Jane WELSH-CARLYLE, John RUSKIN. CARLYLE has been excluded for the same reason as Dorothy Wordsworth above, and because he is in a sense an outsider, due to his Scottish background. The small and atypical contribution of ROBERT BROWNING has also been left out. This is the generation of Disraeli (born in 1804), in a wide sense, since G. Eliot and Ruskin, the youngest, were born in 1819. The median of their production is 1850, that is 35 years later than for group A. Table 1 gives the figures, which speak for themselves : TABLE 1 Density for two generations Group A : 157 Group B : 313 So, in the space of 35 years or so, the figure is doubled. III.1.3. Apparent-time : The main purpose of the study is an evaluation “ in real time ”. “ Apparent-time ” surveys, such as Gauchat’s pioneering work of 1905 at Charmey in Switzerland, compare the output of different generations at a given moment. They have often been just a substitute for the appreciation of current change (see Labov 1994 : 43-112, Bailey et al. 1995, Chambers 1995 : 185-206, Raumolin-Brunberg 1996, for a discussion) on the the assumption that people retain the pre-adolescent system for the rest of their lives. It requires many years to have a neat picture in real time, especially in the field of grammar − Labov’s work in Philadelphia has shown that this is more feasible with sound changes. When the possibility of real-time survey exists, apparent-time might seem useless. However it still remains interesting for two reasons : a. It confirms the results of the real-time survey. The distortions encountered, mostly due to different attitudes to change in a person’s lifetime − there are conservative and progressive people in language as well as in other matters −will help to extrapolate from other similar surveys when real time is not available ; b. It gives insights into the dynamics of change : presumably a mixture of generational change and community change. Only when many such observations have been compiled and confronted shall we know whether the mechanisms are identical or not when linguistic and social parameters have changed. This is why such a study of “ diachrony in synchrony ” has been considered interesting. The central period 1835-40 was chosen. Located near the median of the bell-shaped curve of the overall diachronic corpus, it provided sufficient material for a reliable parallel between two age-groups. In 1835 (see list of writers in the ANNEX), Maria Edgeworth was 68, Mary and Wm Wordsworth 65, whereas Carlyle was 40, Dickens and Thackeray 23. The two groups may reasonably be considered as representing, respectively, the younger generation and the older one by the time when Queen Victoria was crowned (their ages in 1835 are between brackets) : Group a : Edgeworth (68), Mary and Wm Wordsworth (65), J.J.Ruskin (50). 35 Group b : Carlyle (40), Macaulay (35), Jane Welsh-Carlyle (33), E.B.-Browning (29), Thackeray (24), Dickens (23), John Ruskin (15). Table 2 gives the densities for the two groups. TABLE 2 Density for two age groups for the period 1835-1840 a. Age 50-65 : 169 b. Age 15-40 : 260 This would seem to justify the assumption of apparent-time surveys. It gives another perspective than the previous one where the lifelong productions of two generations were compared. It seems honest to say that this is just indicative and that a different sample might give different results. The gap between the younger and the older at that point of time is so clear, however, that it can hardly be disputed. Individual profiles will confirm the trend. Conclusion : Not surprisingly, the year-after-year measurement displays a jagged line. Smoothed by decades, it assumes almost the regularity of a straight ascending line. The arrangement by “ lots ” of equivalent size less evenly distributed over the years and with contributions differently weighted confirms the same regular increase. By grouping the entire contribution of letter-writers into two chronological “ generations ”, we can see that those born after 1800 in the first 20 years of the century used twice as many progressives in the same length of text as those born in the last third of the 18th century, roughly 35 years earlier. At the chosen central point of 1835-40, the picture in apparent-time confirms very clearly the contrast between the age-groups. The extraordinary development of this verb-form in the course of the nineteenth century is not a discovery. Yet the convergence of the results of the multi-faceted approach confirms not only what mere observation of the end points indicated, but the regularity of the ascent, suggesting further analysis of its conditions and modalities. This will be the subject of the next sections. 36 III.1.4.. A study of idiolects : The corpus is not a statistical sample : in fact, we are surveying a group of idiolects. As the support of linguistic theories or analyses, the idiolect, which is too often the idiolect of the analyst, or collected by elicitation, not by recording spontaneous utterances, has been criticized by Weinreich, Labov & Herzog 1968 and other corpus-oriented linguists. The question however remains that la langue, that Saussurean concept referring to the common properties or structures of a given language, is as elusive as ever, to the point of being an article of faith : whether it is a common heritage of the people speaking and writing it now from those who have spoken or written it for centuries, or a more abstract capacity, an avatar of the universal capacity of language, or a combination of the two, is what linguists of the different persuasions are trying to elucidate. But everybody should agree that the common property is not evenly distributed, that idiolects do exist. Merging a large number of idiolects in a random sample reflecting a community grammar was practically impossible for the present research, unfortunately, but, on the other hand, we can have, on the point considered, some idea of individual differences in synchronic and microdiachronic terms. Idiolects are often used and spoken of by linguists, but they are seldom carefully studied for their own sake. Concerning great writers, we find monographs of their language, usually descriptive and stressing the qualitative distinctions of “ style ”, including vocabulary, dialect, etc., and this concerns literary works rather than private expression in letters for example. Before our time, little could be known of the spoken practice, except, occasionally, when early biographers had heard it or heard of it from family or friends. The observations presented below are only tentative and limited in scope. Careful study of private letters should be undertaken for many other features, and even for the whole of a person’s grammar, and would be revealing of the variety usually dissolved in the grey impersonality of community grammars.45 Fortunately, we do not all speak and write exactly alike, and below the ordered heterogeneity of groups we find the consistency and inconsistencies of individual grammars : we will see that they display noticeable differences in the use of the progressive, and in other parameters. This concerns both the overall production of each writer, and their individual diachronic trajectories. The study of such variations is not easy : many may just be random, haphazard or erratic. In some cases, until more sophisticated analyses are tried, all we can do is try to connect phenomena with what we happen to know of a writer’s life and personality. It is easier to raise questions than to give definite answers. Some other points will be considered separately with the different aspects of the study ; but, to begin with, the rough and ready measure of density we have adopted suggests interesting remarks. In their study of the BE/HAVE variation, Rydén and Brorström 1987 have noted that there are conservative and progressive people. This corresponds to our common experience of people’s behaviour in many ways : dress, manners, etc. We have already shown that women as a group in our set are less conservative than men 46 45 Much could certainly be learnt by studying the remarks and character-building in fiction, including observation of individual grammar and usage by the keen observers writers often are. Dickens is probably the outstanding one, but he is not the only one. Linguists and literary people have in common an attention to language where linguistic and metalinguistic processes are in constant interplay; but they do not have the same aims. This will be considered in V. 46 Mossé and Jespersen, in their time, believed in some sort of enrichment in grammar. "l'anglais n'a cessé d'enrichir les procédés qui permettent de mieux traduire la gamme de plus en plus nuancée des émotions et des sensations", "un merveilleux moyen d'expression de l'aspect et de l'affectivité que beaucoup de langues peuvent lui envier"Mossé, 1938, II, 274. This is a linguist's view, but we also have to remember that innovation in language, as often in other domains, is usually judged negatively by most people, especially among the educated, as it is supposed to take us further from the golden age. A critic of the mid-nineteenth century forged the expression: The Great Victoria Bridge has been being built more than two years, as one of the ridiculous expressions the progressive passive could lead to (Marsh, Lectures on the English Language, quoted by Mossé 1938, p.158). By progress, we just mean here development, novelty, without suggesting positive or negative evaluation. 37 III.1.4.a. Individual total densities : Figures 4 and 5 give an idea of how individual densities for women and men respectively, arrayed in chronological order (birth dates), parallel the overall cumulated density displayed on Figure 2. Only the variations between members of the same generation or about can hint at some individual explanation. It seems also justified to compare women with women and men with men, preferably. Concerning the earlier group of people born before 1800, we notice that the highest density (226) is found in Mary Wordsworth’s letters : a native of the Lake District, she wrote only family letters and she had been more deeply immersed in the local dialect, which we suspect to have been rich in progressives. We also see that Jane Austen, only five years younger than Mary Wordsworth, is, with a density of 217, nearly at the same level as the latter, and noticeably above the other women of the period, although she lived in the South (in Hampshire) in comparatively refined surroundings. This may be due to the fact that both have written only to family or close friends : but the same is largely true of the other women, with the exception of Dorothy Wordsworth, an intellectual and a very close literary companion of her brother the poet. On the other hand, the great novelist is generally, in her novels at least, a strict follower of late 18th century prescriptivism (Leonard 1929, Phillipps 1970) : and the progressive was still felt as a novelty at the time, certainly not to be encouraged. Raybould 1957 observed that the use of the progressive in Jane Austen’s novels is connected with a certain idea of gentility : genteel characters use it less frequently than more “ vulgar ” people.47 Are we then to conclude that Jane Austen herself was one of the latter ? We must look for another explanation, and perhaps the very title of one of Jane Austen’s best known novels is a clue : Sense and sensibility. A sensitive woman, she was already a Romantic and the progressive probably attracted her as a more personal and effusive form of expression, even when she was not using it as a writer but simply as a sister. Whereas the simple form was that of the classical expression of sense, the progressive was the new means by which to convey the subtle shades of sensibility. Which does not mean that a conscious decision led to the choice of one or the other : on this point, we can only guess.48 If Maria Edgeworth was far behind Jane Austen in that direction, this may be attributed to another cause. She herself lived in Ireland, where the progressive is known to be a common feature of Anglo-Irish, possibly due to the influence of a Gaelic periphrasis. 49 But she belonged to the gentry and she wrote mostly from England where she visited aristocratic and literary circles at a time when Johnsonian prescriptivism was in favour. It is quite probable that, consciously or unconsciously, she shunned a syntactic form that smelled strongly of celticism and vulgarity : see above in II. 3. In that respect, her practice was not parallel to that of Jane Austen, her junior by eight years, living in different surroundings. Yet, again, we must be cautious with such assumptions when we observe the difference between Mary HutchinsonWordsworth (Density : 226) and her sister Sara Hutchinson (Density: 157) : they had been brought up together, they lived in the same house, they both wrote family letters. Mary often served as secretary to her husband, and as amanuensis when his sight failed in old age. Yet she seems to have been less intellectual than her sister, her letters are more about everyday matters : this may be one of the reasons for the difference, which also may be purely accidental. We also know that two sisters are seldom really alike : what one prefers is often avoided by the other. The corpus for both is comparatively small, another reason to be prudent. The letters written by the women we have just compared are all family letters : degree 1 of intimacy. As far as men of the same group are concerned, the comparison is more significant if only the same degree of close intimacy is taken into account. The difference between Wordsworth (Density: 144) and his friend Southey (Density: 176 ) can hardly be due to their geographical origins, since this would rather contradict the difference : Wordsworth had been brought up in the Lake District, where Southey, born in Bristol, established his residence only after his marriage. It is in Southey’s letters that, until recently (Mossé 1938, p. 149-50 and other sources, the N.E.D. especially) 50, the first instance of the progressive passive was detected, in 1795, when he was only 21 − the second in the present corpus is in one of Coleridge’s letters — who was then 25. This is perhaps a symptom that Southey, then full of revolutionary ideas, was also innovative in grammar, which would explain his comparative addiction to the new form, while 47 Cf. also Stanzel 1957: he found that in the novels of Anthony Trollope, much later (after 1850), the progressive was still more common with the uneducated. His characters often belong to the clergy. 48 See below in V and Figure 34 the parallel between letter-writing and novel-writing. 49 ta se ig baistig ig Corcaigh: it is "at rain" in Cork, where baistig is a verbal noun with "inessive" value, to be translated "it is raining". Cf.Van Hamel 1912; Sjoestedt 1926; Visser 1955. 50 This point will be taken up in IV.1. with reference to Warner 1995, Denison 1998 and Nakamura 1998. 38 Wordsworth would be more conservative. But the overall difference dissappears and is even reversed, though not significantly, when we consider their intimate letters only : Wordsworth (Density: 198), Southey (Density: 189). The two sets of figures have been given just for illustration, although it might suggest a different type of control between intimate and less intimate letters, in the direction alluded to above, with Southey more free and innovative. It seems hardly possible to verify it in the corpus, but this is perhaps a valid notion of more general scope, in the spoken as well as in the written language. 51 Coleridge, on the other hand, with a density of 122 (intimacy 1) against 107 (general), is noticeably behind his two friends. This may be due to the more mixed nature of his correspondence and the changeability of his temperament, but also to the fact that, travelling and changing his residence on many occasions, after a period in the Lake District, he was not exposed so constantly to the influence of the local dialect of the North Western counties. Both born at the end of the eighteenth century (1795), Keats and Carlyle must be considered apart. Keats’s letters were all written during the last four years (1817-1820) of his very short life : he was then a young man of 22-26, and a feverish consumptive. Two suggestions can be offered to explain the exceptionally high density of progressives (322) in his letters, apart from the cursive, very free way in which they were written : the fact that he was brought up in the East End of London in humble or at least mixed surroundings, and may have been influenced by the language of the lower classes, on the one hand, and, on the other, that he was a passionate young man. Thomas Carlyle is another exception. He lived a long life, the second part in London, but the first part in Scotland. Much of his private correspondence is addressed to his family there, but the proportion of other letters increased with time as he became a more popular writer. So that the overall density is not as significant as the trajectory in time, which will be studied below. With a figure of 252, it is not considerably above that of Macaulay 52 , five years younger, and below that of his wife Jane Welsh (303) also Scottish by birth : this latter discrepancy must be due to the tone of his written style, his topics − more abstract discussions − and the greater variety of his correspondents. The women in the second group, born after 1800, display a density above 300, with the exception of Elizabeth Barrett-Browning (205 only). This is probably because she had to live abroad, in Italy, from 1844 until her death from tuberculosis in 1862. If we accept the current notion that the spoken language takes the lead in evolution, the fact of being largely withdrawn from its influence certainly accounts for a more conservative grammar as well as vocabulary. This is a well-known phenomenon. She met visitors from England in Florence, of course, but that could not have the same effect as immersion in the London surroundings. Her companion in Italy was Robert Browning, who is himself idiosyncratic in shunning the progressive. After Elizabeth’s death in 1862, he came back to England where he remained for the rest of his long life. His corpus has been surveyed even for that last long period (18621889), but it is too small to be really significant. However, Browning evinces a density of only 132, the smallest of all in our study except for Coleridge, his senior by 50 years. Some, on account of the intricacies of his introvert personality and style, will be tempted to see here another manifestation of his originality : he is certainly atypical. His refusal of the progressive, even in love letters, may also be partly responsible for the low density in Elizabeth’s letters. Between the two great novelists Thackeray and Dickens, we notice a neat difference of 75 for the overall density : Thackeray 345 , Dickens 271. A comparison of the figures for the more reliable degree 1 of intimacy, Thackeray 378, Dickens 268, confirms and enhances the result. This is rather surprising : as a novelist, Thackeray is usually seen as more conservative in his style, still in the line of 18th century novelists like Fielding. He wrote historical novels located in that century, Henry Esmond , for example, where we find a low density of 186, and the “ Regency 5151 The parallel between grammatical and political or social attitudes has seldom been attempted. However, Denison 1998, 74 and p.c., has suggested that the presence of the progressive passive in Southey and Coleridge's works may be due to their radical iconoclastic ideas. The famous chemist Joseph Priestley, a notorious sympathizer with the French Revolution, who even became a member of the French Convention, would be another interesting case, as well as Cobbett, whose Rural Rides of 1830 − but published earlier in his Weekly Political Register 1803-1835−, a mixture of narrative and reflections, exhibit a fairly high density of progressives for the time (141) The author of A Grammar of the English Language, 1818 and Cobbett's French Grammar, 1823 was however faithful to the old passive: its exciseman can tell it what is doing even in the little odd corner of Binley ( Rural Rides, 7 Nov. 1825). Victor Hugo said : "J'ai mis un bonnet rouge au vieux dictionnaire", and that was true of his grammar as well, even at the time when he was not a very progressive politician. General de Gaulle was known to use coarse language with his family and close companions, whereas he spoke a very controlled and academic French in public or with less intimate people. 52 Not a Scotsman, although a regular contributor to the Edinburgh Review and once a MP for Edinburgh. 39 style ” of his writings has been noted (Phillipps 1978). One hypothesis could be the influence of his wife, who was Irish : but she became insane four years after they married, and they lived separated after that, so that it must have been a very limited influence. Thackeray was often described as a “ swell ”, a “ gigman ”, as Carlyle called such people, fond of aristocratic connections and functions. But he was also a relaxed humorist, whose letters are often illustrated by clever comic sketches, and he does not seem to write differently when he addresses Lady Castlereagh, for example, or anybody else among his friends and connections : there is often little difference between intimate and less intimate letters, and in his case, degrees 1 and 2 were difficult to distinguish. Even with degree 3, he is above the average at the end of his life. No clue can then be given by that consideration. We have seen that Jane Austen seems to have made the progressive a token of ungenteel characters and we have also shown that she used it a lot herself in her letters. Then, whatever the reasons for it, we have just to note that Thackeray was comparatively an addict of the new form, more than Dickens : see also, on Thackeray’s account, Chapter IV on literature. To add to the discussion, Rydén and Brorström 1987 : 201, have observed that Jane Austen “ stands out as remarkably conservative in her handling of the be/have paradigm ”− this had already been observed by Phillipps 1979 − whereas Thackeray is noted as comparatively favourable to the novel forms with have. So that it seems that people may be erratic and even contradictory with respect to new paradigms : innovative with some, reluctant with others. To assess whether this apparent heterogeneity is structured would be the task of studying the grammars of individuals in a systematic manner. The observations above call for further enquiry. Study of present-day behaviour has already shown some of the external constraints, at least for groups. One of the major constraints is consciousness, especially of a stigmatized item. We know that excessive use of the progressive was refused in many circles, mostly because it was supposed to be a celticism, Scottish or Irish, often a synonym of “ low language ”, sometimes also smelling of “ papism ” in the case of the Irish. Among testimonials of this, we have caricatures of Irish paddies, as the immigrants were called : eating potatoes and using progressives or Anglo-Irish expressions were two notorious shibboleths of the paddy, as evidenced for example in George Borrow’s Lavengro, or in Henry Mayhew’s famous interviews of the small people in London in the eighteen-forties and sixties (see Mayhew 1967 ; also Thompson’s The making of the English working class, 1963, 469-85)53. It is quite probable that the stigmatization − a strong word perhaps in this case − was felt more acutely in Scotland and Ireland, not among the lower orders, but among the literati, more than among the aristocracy. The latter, as already said, did not bother : the Duke and Duchess of Buccleuch were fond of the Scottish dialect and certainly felt no shame of their Scottish ascent and accent. What can be identified to a kind of diglossia may explain that Scott displays a density of progressives (157) hardly higher than that of his friend Wordsworth (144), although much higher than that of Coleridge (107) : whereas he spoke Scots quite freely and also a variety of Scottish English (See Aitken & McArthur 1979 ; Tulloch 1980 : 171 sq.), he adhered to the prescriptivism of his fellow-Scotsman Boswell, the celebrated biographer of Doctor Johnson, who “ was in 1761 one of the three hundred Scottish gentlemen who took lessons in the correct English pronunciation ” (Tulloch 1980 : 175). Tulloch notes that “ he [Scott] also seems to have shared the view of many of his contemporaries that, at least since the Scottish king no longer spoke Scots and the old court Scots had died out... English was the “ purer and more classical tongue ” ”. He also mentions use of the progressive, especially with verbs like think and equivalents : I’m thinking meaning I imagine, I’m doubting meaning I’m afraid (p.296) as a feature of Scottish grammar in Scott’s works. At the time of Scott, there was still a feeling of two separate languages, one for local use, the other the only one acceptable in correct writing although, in Aitken’s terms : “ Around 1785 the extreme self-consciousness and the strident note of linguistic insecurity which mark the middle years of the [18th] century die out ”( Aitken & McArthur 1979 : 96. Cf. also Romaine 1982 : 136). With the next generation, that of Carlyle, the change in attitudes and practice becomes more conspicuous. But the question of consciousness, in the absence of objective tests, is difficult to assess. It is probable, as always, that it concentrated on certain stereotypes, not quantitatively significant, such as : someone is wanting you, etc. 53 "The expressions 'the master is calling you', 'he is speaking to you,' 'were you ringing ?' 'I was supposing,' 'he is not intending' are Scotticisms for 'the master calls...'"Alexander Bain, Higher English Grammmar 1863. Bain was professor at the University of Aberdeen. He extends his criticism beyond what were still the usual stereotypes. The trial of Jessie M'Lachlan in 1862 (Roughead 1950) is an interesting piece of evidence on the progressive in Scotland, especially among the poorer classes: Jessie, a poor servant who escaped the gallows only because she was granted a reprieve following a public campaign, says, for example "I had been intending to go...; another witness:" she was needing it, I suppose; she said that I could not be got when she was wanting me"...These were typical expressions. However, the density in the witnesses'statements is only 243, lower than in our letters at the time: but crossquestioning in Court does not encourage familiarity, and the statements may have been emended by the clerks. 40 The new passive progressive is clearly a form which people were conscious of (Mossé 1938 : 155-158), but it accounts very little for the density we have measured. As to the other progressives, except perhaps with some verbs like want or like, or, generally speaking, state or “ private ” verbs, their development, which we find remarkably rapid from the point of view of conventional diachrony, was too slow to be felt as clearly as, for example, the introduction of new words or phrases. It is then quite plausible that the intrusion of this device was mostly felt by those who had sufficient contact with the common people, who were probably the first to have used it extensively, and before all, the immigrates from Scotland and Ireland − an immigration which grew by degrees, but became important after 1840 and the great famine of 1847-8 (Cf. Archer-Jackson 1963 ; Lawton 1959.) So much for groups and communities, but what about individuals ? We know that their attitudes are not uniform, even − and possibly above all − among the intelligentsia, witness the warm debates in France whenever a new attempt is made at spelling reform (the latest was in 1992). The progressive passive was the most conscious innovation and it began to raise diverse reactions as soon as 1822 when, in The Gentleman’s Magazine, one reader censored it, another supported it as solving the ambiguity of the old form : “ the letter is writing ” ; and the debate continued for more than 50 years (Mossé 1938, 155-8). To conclude, the degree of consciousness was probably small and hesitant, and limited to some stereotypes supposedly Irish or Scottish. As to the attitudes, we know that attitudes to language use are not necessarily parallel to those concerning other domains such as fashion, science, or politics, nor are they consistent at different times of history : born under the reign of late 18th century prescriptivism, the 19th century was led to more freedom with romanticism and the contemporary industrial and social revolution. This would be the general frame within which to inscribe individual reactions. Here, we cannot forget we are dealing with artists of the pen who, consciously or unconsciously, may have individually a liking for or an aversion to such and such a form of expression. They may occasionally have justified it, but no mention of this with reference to the progressive was found in our collections of letters. The novels may give a few hints : Romola, by G. Eliot, displays a high density (360), suggesting deliberate choice : see below in V.1. This liking or aversion is found also among ordinary people. Needless to repeat that the argument can be evoked only when the deviation from the average is significant. III.1.4.b. Individual trajectories As a complement to idiolectal sums, the study of individual diachronic profiles may help in the study of two important points : 1. Is there any significant difference between individual trajectories and the general line of evolution ? The most reasonable hypothesis on this account is that people vary in this respect and that the general trend is just the average of their relative speeds. This is connected with the problems summed up as community change v. generational change : Labov 1994, 83. On the whole, however, the deviations must not be great. 2. When the individual line displays peaks or depressions, are they significant and if they are, of what ? Can such deviations be related to what we know of the person’s existence ? This is certainly an ambitious program, requiring fine measurements for which we may not have the requisite tools, and a critical analysis of the observations. Again, we have to stress that this is just a pilot study of limited scope, hoping that it may open the way to others. One of the obstacles is once more the uneven distribution of the corpus along the years. So that, for example, if we compare the figures for any single writer with the overall figure at a given five-year period, we are in fact comparing them with a small set of other individuals whose letters happen to have been preserved for the same period. For that reason, comparing two or more individuals may be more instructive. To sum up, we shall be happy if a few suggestions can be encouraged by the data : students of ancient or medieval history have to fight against much greater odds. The study has been attempted, first by considering individual trajectories and those of sets of people who have something in common. Some of the writers have to be omitted from this inspection : for different reasons, the figures would not be significant. John Ruskin’s parents have left a number of letters too small to offer more than an occasion to check the difference with their son. His own contribution, discontinuous and limited to 200 000 words, displays a decrease possibly due to subject matter, since, in the later letters, he wrote mostly about his experiences as a painter in Venice. Maria Edgeworth’s contribution is uneven and displays no clear profile. Many of her letters were sent from 41 England : unfortunately there is not enough material to study a possible difference between these and the letters she wrote when she lived in Ireland, where she might have been influenced by the local dialect. The letters of the Brownings are scattered in various publications : Robert Browning, as we noted, did not like the progressive, and there is no significant variation around his very low score of 132 during his long existence, even in his love letters, when Elizabeth Barrett appears to have reacted differently. But, with a corpus of only 200 000 words for Robert, half of what Elizabeth wrote during a much shorter period, this estimate may be wrong : it is only highly probable, and could possibly be confirmed on the basis of his literary works ; a doubtful solution, since he wrote practically only verse, and highly elaborate verse at that. Perhaps he ought to have been entirely left out, but after all, nonconformists and eccentrics have a right to be included in a survey. The study will consider successively two diachronic groups or “ generations ”, and the other letter-writers will be seen individually. The Lake poets : The poets commonly known as the Lake poets and their companions form an interesting group of six men and women of the same age who lived in fairly close association, to the point of sharing the same house, temporarily or for a long time − the Wordsworth family, where William lived at Dove Cottage or Rydal Mount with his wife, his sister Dorothy and his sister-in-law Sara Hutchinson. On Figure 6, we can see at once that their densities are close to each other. However, the women display a higher density, as already observed. Two members of the group show some originality : Mary Wordsworth with noticeably higher figures than the rest, Coleridge with smaller figures. This just corresponds to what has been said above concerning the total individual densities. If we look at the curves, we see that they are almost flat, with a slight rise in the case of William Wordsworth whose long life was spent almost entirely in the Lake District. His wife Mary is the only one to display an irregular profile, but the abrupt peak before the end is clearly not significant and probably due to the small size of the relevant corpus. The only certainty, or near-certainty, is that her general density is higher than that of the others. She appears to have begun with a high figure −she was born in the district − but this may also be an artifact of corpus size. To be frank, it is not surprising that they all follow the average tendency, since their contribution to it is considerable. However, as time goes on, there are newcomers − the percentage of the older group passing from 83% to 33 % from period 1 to period 8 − and we must admit that the curve for William Wordsworth, whose correspondence was still fairly voluminous at the end of his life (200,000 words after the age of 70), seems to be rising with the tide : Figure 7. The distinction of intimate letters from the rest is too discontinuous to justify conclusions concerning his whole career, but these intimate letters show a clearly higher density for that last period : this is probably the main reason for his ascending general score. We may also consider that, as he grew old, living in the Lake counties, he was indulging more freely in what had always been a feature of the local speech, no longer contradicted as it used to be by the King’s English : after all, HE was now one of the people of influence on the question of what was good English, just as we recently saw the Beatles be among the modern arbiters. Is this making too much of mere indices ? Looking more closely at the curves, we notice a fairly accented peak for the three friends, Coleridge, Wordsworth and Dorothy Wordsworth, at period 2 (1800-1804). That was indeed the time when they lived and wandered together in the Lake District (and even undertook a tour to Scotland, visiting Scott), a time of intense excitement, physical and moral, one of the peaks of poetic creativity, before Coleridge, always in quest of somewhere else, escaped for his Mediterranean voyage to Malta in 1804. At that time, they had all just passed the age of thirty, and were busy with the publication of the second edition of the Lyrical Ballads, one of the landmarks of romanticism. Each of the three had his own reactions, and we don’t see the same peak in the profile of their other friend Robert Southey, who, however, was somehow apart, even before he left for Portugal. 54 One is tempted to establish some connection between this verbal symptom and the extraordinary excitement of their life in the Lakes. This suggestion is encouraged by the fact that it is not an individual deviation, but that the three protagonists of the Lyrical Ballads’ revolution are sharing it : Figure 8 . 54 He once (Oct. 1798) attacked the Lyrical Ballads, anonymously, in the Critical Review. Southey was a native of Bristol, but came to live in the Lake District. 42 We must also note the depression which follows, before the line resumes its ascent, much less pronounced.55 That was a period when the enthusiasm − this is the appropriate word − of the first decade of the century gave place to disquietude, or even Dejection, the title of a famous ode by Coleridge56 : he himself was gone, although he sent letters from Malta. After a period of peace with France, the peace of Amiens in 1802, the French Revolution, in which they still saw the birth of liberty despite the excesses of the Terror, had been confiscated by a tyrant and the Napoleonic wars had begun. It would be ridiculous, of course, to look for a point-to-point correspondence between a linguistic phenomenon and the atmosphere where a small group of people lived, if there were not other instances where the same correlation can be guessed. 57 Walter Scott’s long collection of letters has not been completely surveyed but the line is also ascending in his case, roughly similar to the general curves of Wordsworth and the Lake poets − except Southey and Coleridge as already noted : Figure 9. Jane Austen and John Keats can hardly be compared to the Lake poets. Keats’s letters cover a brief period (18171820) so that he cannot show any clear evolution. Jane Austen was actually 5 years younger than Wordsworth, but she died in 1817 at 42. Her letters, most of them addressed to her sister Cassandra, were written between 1796 and 1817, over a period of 20 years : this is sufficient to justify a brief survey. By grouping the data in order to have enough material for the density counts to be really comparable, we see that the density in her letters passes from about 100 before 1805, to 200 for the period 1805-1809 and to 300 for the last period, 1810-17. This is indeed an abrupt progression, compared to that of the Lake poets and to the general curve : Figure 10. Although a comparison with her novels, all published during the last period, cannot parallel this, we note that in her last one, Emma (1816) progressives reach the stupendous density of above 400 ; see chapter V on literature (Figure 34) and Phillipps 1970, p.111-16. She was certainly a pioneer in using the form. If we consider now the next “ generation ”, the picture is more blurred : Figure 11. Dickens, Thackeray, Mrs Gaskell were of the same age : the three novelists exhibit an ascending diachronic curve, more regular in the case of Thackeray, more jagged for the two others. A close study revealed that the peak in Mrs Gaskell’s profile is connected to her correspondence with her daughters, Marianne (“ Polly ”) especially : Figure 12. Even within the stratum of intimate letters, this correspondence illustrates the weight of the intimacy dimension, and the difficulty to have a strict and stable discrimination. The 55 Anthony Warner (p. c., 1999) offers the following and very interesting suggestions on my paper of 1998: “ there might be some element of age-grading for the lakeland poets, as in Labov’s account of Cedergen’s Panama data. Their ages march together; there is a peak roundabout age 35, then a decline. Could this be a reflex of a hostile evaluation in the 18th C evidenced by the ‘flattening of the curve’ noted by Strang... Also perhaps seen in Hutchinson, Mary W (and consistent with Keats) though distinctly not seen in Southey. The mechanism might be rejection of the hostile evaluation, then acceptance of it, ultimately overtaken by an increase in later life reflecting communal change. It the hostile evaluation weakens or is refocussed in 19C, then such an effect might not be seen among those born after 1800. I appreciate that a change from below is not supposed to show such evaluations, which may be enough to rule out such an account, and though the progressive is being passive is clearly subject to evaluation later, that might be referred to the salience of the sequence. But you clearly suggest that the question of evaluation is not a simple one in this case, and the object of evaluation might not be the progressive per se but the extent of its use, or its presence as an indication of personal involvement or ‘warmth’. ” This seems to me quite plausible, and the validity of the hypothesis should certainly be tested for other cases, including contemporary ones. One of the major difficulties is to assess the degree of consciousness of a given change, especially when it is a change of magnitude. A parallel with habits other than linguistic can be attempted: we are often more indulgent than we claim... 56 As a matter of fact, this Letter to Sara Hutchinson was begun in the spring of 1802, but Richard Holmes, in his magnificient biography of Coleridge, underlines the difference between the first and the final version of the autumn of the same year: "the first version is a passionate declaration of love and renunciation, of almost hysterical intensity; the final version is a cool, beautifully shaped, philosophical Ode on the loss of hope and creative power."( Holmes 1990: 318). Coleridge's visionary romanticism is outside the grasp of a down-to-earth study like ours: but the parallelism with the Wordsworths is clear for that period, nevertheless. 57 I am quite aware that this is returning to a Humboldtian, pre-Saussurean conception of "parole", also found with Anton Marty. A French follower of Freud may also be quoted about this: "Ne pourraît-on caractériser chaque procédé grammatical, chaque syntagme, chaque élément de style, en tant que conduite? Et considérer le choix de telle ou telle expression verbale comme une conduite symbolique bien définie ?" Rostand 1951:10. 43 term spontaneity, because it combines the warmth of the feelings and the cursiveness of the style, finds in her case its most blatant justification : all she writes seems to flow from the heart without restraint, reaching the highest peaks we have found. Part of this may be due to the fact that Elizabeth Gaskell lived in the North : she was born in Cheshire and, after marrying a Unitarian minister, spent part of her life in Manchester, where she was in contact with the the same working class which is described in Engels’s famous book. We remember too that Irish immigration was developing in the area, especially at that time (Archer-Jackson 1963). In her novels, there is also a significant and continuous increase of the density, from a little above 200 in Mary Barton 1847 to almost 400 in Cousin Phillis 1863 or Wives and Daughters (unfinished when she died in 1865). But we realize that the warm tone of her letters to Marianne becomes colder by 1860. As a matter of fact, the year by year graph —not presented here— smoothed by the 5-year periods’curve, displays a clear trough precisely in the year 1862 : from 400 progressives to less than 100. One is tempted to connect it to a serious crisis in the relations between the mother and the daughter : when she was in Rome in 1862, Marianne met Dr. Manning, the Roman Catholic priest who was later to become Cardinal Manning, and she appeared to be on the point of converting to Roman Catholicism. On learning this, the devoted Unitarian Elizabeth Gaskell was serious affected. According to her biographer, she was subject to fits of depression, but in this case it was a severe one : she even had to take to bed (Gérin 1976 : 226). Things finally turned differently since Marianne eventually did not become a Roman Catholic, but the shock must have left some trace.58 If, for the sake of statistical significance, we neglect the ascent corresponding to the period 1848-1853 or about, when letters to the girls were so numerous and passionate, and merely compare the global figures before 1860 and after that date, we find an abrupt decrease from 509 to 391. The overall figure concerning all familiar letters is of course seriously affected so that a comparison would be fruitless, but we have no reason to ascribe this to a general reversal of her language practice, witness the continuing ascent in the novels (see V. 2, Figure 36). No deep insight can be obtained from a study of Dickens’s and Thackeray’s curves, even by considering intimate correspondence only. Concerning Thackeray, the distribution of his letters during the early years and the uneven size of the corpus do not encourage interpretation : the only safe assumption is that, in later periods, between the ages of 30 and 52 (when he died), his curve is steadily ascending for all degrees of intimacy : Figure 13 The same ascending curve is observed with Dickens for the whole of his production : Figure 14 . For the three degrees of intimacy, Macaulay’s curve is ascending, but we are struck by the comparatively abrupt descent at period 10 (1840-4) : Figure 15 ; and we are tempted to ascribe this to a crisis in his personal existence. 1835 was the year when his beloved sister Hannah, whom he used to address as “ my darling ”, married. A bachelor who was then 35, he had kept with her and his other sister Margaret a very intimate correspondence. She appears, even more than Margaret, to have been the focus of his affections, and his biographers have noted that he was seriously disturbed, and possibly depressed, by her marriage. We will again refrain from too bold an assumption concerning the influence of psychological crises on a formal feature of speech, and also remark that, after the marriage, he could no longer write in the same confidential tone : his letters could now be read by his brother-in-law as well ; intimacy was disturbed. A closer examination of the density in the letters to Hannah (in 1832, they were addressed to both Hannah and Margaret) reveals indeed a clear contrast between the period preceding Hannah’s marriage and after that event : before 1835 : 255 after 1835 : 197 Our research had to be interrupted for the Carlyles when only part of their correspondence had been published in the Duke-Edinburgh edition with vol. 11, ending in 1839. For the following 46 years, apart from a selection of letters to his wife −not included − the only corpus was a collection of letters to his brother Alexander. On Figure 16, beside the higher density in family letters already noted, we notice that the curve for Thomas Carlyle is descending : all letters, even the most intimate − the majority − are affected. This may be attributed to his “ education ” : the young son of Scottish peasants was becoming an intellectual, a writer, frequenting other circles where, even in Edinburgh, the influence of dialect was less effective, and finally settling in London in 1834 (period 9). Carlyle was a man with strong family ties : many of his letters are addressed to his mother, who was still living in a Scottish village, and could hardly write, or to his brothers, his sister or his niece. His brother Jack then became a doctor and left Scotland. But the other, Alexander (Alick), was a farmer who left Scotland to establish himself in a farm in Canada. A comparison of the general density for all of Carlyle’s letters, including those to Alick, and of his letters to the latter 58 This is a case where intuition precedes information. I had not had access to the biography before noticing the down slope and its specific reason remained mysterious until I was acquainted with the biographical data. 44 only, after 1839 (the only collection we have reviewed after that date), may be instructive of the influence of the addressee : Overall density : 252 (246 excluding Alick’s after 1839); Density in letters to Alick (1840-76) : 325. When addressing his farmer brother, he was still using the language they spoke in the Scottish village where they were brought up, and where they met again from time to time before Alick left. Among intimate letters as we have computed them, there is some mixture. It might be interesting to study separately the letters addressed to the clan in Dumfries. The Brownings : Figure 17 . The Brownings’ corpus is dispersed between several collections of volumes, especially for Elizabeth Barrett-Browning. One consists of their love letters of 1845-46. It is difficult to establish degrees of intimacy in E.Barrett-Browning’s correspondence, and this would probably bring little benefit : her published letters are to friends (Mrs. David Ogilvy, Miss Mitford, Mr. Boyd, Mr. Kenyon) or to her sister. We see that the total density is 204, whereas Robert’s is only 132, and even 110 (118 for intimate letters), if the letters written after 1880 are included. This is another case of a woman’s using progressives more than a man. It has already been noted that her comparatively low density is probably due to her removal from the evolution in England. Yet, without making too much of it, we have to note that, in her love letters (periods 11 & 12) the density reaches 282 against 211 for the whole period before. No such peak is found in Robert’s love letters : passion did not make him forget his conservative grammar, perhaps. We can also remember, when we compare Robert and Elizabeth, the love letters addressed to Fanny Brawne, 25 years earlier, by another fervid consumptive, John Keats (density 322 for all his letters). Last but not least, we have George Eliot. The study of her correspondence (Arnaud 1979) is partly at the origin of the present work. In that pilot study the result was perhaps too beautiful : the density displayed a regularly ascending line from 257 to 367. The picture presented here (Figure 18) remains practically identical59. So that George Eliot is a typical case of someone who accompanied the microdiachronic evolution, and must even have been among its leaders. As we have seen, when there are individual variations, they can be suspected to have idiolectal reasons related to the vicissitudes of each person’s life. Applied to a very different type of historical quest, and to individual grammars inasmuch as they reflect community grammar, this would have rejoiced the neo-grammarians as supporting their theory that the laws of linguistic evolution are inescapable, mutatis mutandis. See Labov 1979, 1994 and his references ; see also Leskien 1876, who seems to have been the first neo-grammarian, in the context of comparative grammar (Normand et al. 1978). Of course the mutatis mutandis is important, and it is to be hoped that other studies of living and extinct language-use will give precisions about it. Conclusions As usual with social surveys, especially when they relate to the past, there is some heterogeneity in the corpus. The 22 people studied are not strictly parallel : some lived a long life (William and Mary Wordsworth, Carlyle, Browning), others had a brief existence (Keats, Jane Austen). Their surroundings were different, even if they belonged to the same set of literate and even literary people. The letters they have left vary in volume and in their distribution across the years. It is all the more surprising that the discrepancies between them are not greater. a. Total individual densities. We have noted that women show consistently a higher density than men, especially if only intimate letters are considered : many of the men have had more occasions to write to less intimate or even remote people, often on intellectual or business topics. This being said, and if we consider the total lifelong corpora, few people stand really apart from their contemporaries. Among people born before 1800, the only one really exceptional is John Keats : with a figure of 322, he is nearly on the same level as John Ruskin, 23 years younger. Some will certainly remark that they had very different temperaments, the author of Sesame and Lilies on the colder side and the lover of Fanny Brawne on the warmer side. Allowing for the inevitable deviations of sampling, the only one to stand apart as clearly conservative among those born after 1800, is Robert Browning (110) : the writer of often long poems or plays, he was not prolix in his letters. 59 Degree 3 has been omitted, being often absent, and in the general density curve periods 17 and 18 have been fused. Degree 2 shows less regular progression, not presented here, and of dubious significance. 45 b. Individual trajectories. Whenever it appears that a trajectory can be significant of individual progress, our letterwriters usually follow the general trend. We have every reason to think that this was normally unconscious. No one among them has ever openly expressed diffidence or hostility against a form which was not really new but simply increasing − except Macaulay for the new passive. Some even appear to have been leaders : Jane Austen, Elizabeth Gaskell and George Eliot especially, but also, surprisingly, Thackeray. Some of the reasons will perhaps be found in their literary works : yet, this would require a careful, detailed study. Nobody presents a descending line ; at most the line is flat, with occasional variations, probably random. Apart from Browning, whose refusal was constant, only two display no clear rise : Coleridge and Southey. We know that Coleridge was unstable, that he became an opium eater, that the Romantic poet of Christabel and The Ancient Mariner was also a penetrating London critic and lecturer, devoted to abstract prose, philosophical or religious, and that his correspondence is a mixture of discussions of all kinds, rather than everyday news : this may account for the low figure which knew no increase. Southey’s case is certainly different : complete survey of his correspondence would perhaps be more instructive. It is often said that exceptions confirm the rule. Let them at least confirm that we have not tried to force the point and that what was suggested by one case and then another leads to the general notion that such convergence cannot be entirely fortuitous, and that there is some connection between people’s lives and personalities and the language they use, even when unconscious grammar is in question. III.2. Linguistic factors In the conception of the coding matrix, the theoretical tenets of the hypothetical-deductive tradition have been kept reasonably loose. The analysis of the English verb-system by linguists of different persuasions, usually supported by corpora or elicitations from present-day English, has led to some consensus : see I.3. We are all convinced that these factors, with different weights, are the most effective ones : consequently, they ought to be largely responsible for the extension of the progressive, especially at the time when it became so widespread.60 This assumption was supported by a few observations, not clearly quantified : for example, use of the progressive with state verbs have always been said to increase, as well as that with modal auxiliaries, and even occasional occurrence in contexts admittedly blocking it : Ljung 1980 presents original and insightful analyses on some present-day virtualities. Arnaud 1973, a synchronic study for the period 1840-1880, suggested that, among other factors of change, the aspectual properties of predicates played an essential role. Among the hypotheses, we must also mention those involving implicational scaling (De Camp 1973) and “ squish ” (cf. Sag 1973) : a hierarchical ordering of factors can be supposed to govern the stages of the diachronic development. In order to investigate such questions, the coding framework includes also contextual elements of the sentence, sometimes of secondary, informative value for the exploration of what is still a terra incognita. The following does not claim to be the last word on linguistic factors, even within the limits of the corpus considered ; nor does it claim that what has been observed can be extended to other similar domains of microdiachrony without precaution. There are many risks of bias : 1. The corpus itself is uneven ( e.g. the beginning and end periods are under-represented) ; 2. The allotment of verb-predications to semantic classes is notoriously difficult (see I.3 ). More sophisticated justification of it by the use of tests and the computation of a network of syntactic structural or transformational capacities, as was attempted in Arnaud 1973, 616-21, would imply either an impossible amount of work or the limitation of the number of items, forbidding meaningful quantification. III.2.1. Semantic classes Figures 19-24 show the main results of the computation of verb classes. 60 This puts the important question of grammaticalization. Semantic shift accompanies historical development, in grammar as well as in vocabulary. What we call “ progressive forms ” in Old English or even in Renaissance English is obviously different from what the progressive is now. The moment when grammaticalization occurs is still the subject of interesting investigations. It seems plausible to date it from the middle of the 18th century: see General conclusion. 46 1.a. State verbs (Figure 19) : The group of state verbs (Class 1 to class 4) evidences a remarkable increase in density, especially at the end of the period : this will require further study (see IV.6. below), since it is a sensitive domain, where even individual verbs must be considered. But the weight of this group is small on the overall increase : in average, of the order of 2 percent. With so small figures, statistical significance can be disputed (Poisson’s law), but the general orientation seems clear. 1.b. Prospective and retrospective verbs ( Figure 20 ) : these predications have been treated as subclasses of state verbs, in accordance with a tradition which should certainly be reconsidered. They are far from averse to the progressive in our corpus, and this should also increase our diffidence with the notion of activity commonly associated to the progressive. A distinction has often been made between external and internal activity : prospective and retrospective verbs are seen as “ private ” verbs (Ota, 1963 ; Joos, 1964 ; Ljung, 1980) referring to the latter type of activity, whereas, for example, think was usually seen as an ordinary process verb. There is little doubt that there is an aspectual component in them which brings them closer to the other classes in the line of our semantic analysis in I.2. Many of them are also emotionally loaded, a fact which may be largely responsible for their share in a corpus of private letters : feelings are commonly associated to the mere statement of facts. A more detailed development will be found in IV.7. 2. Movement verbs (Figure 21) : Strictly speaking, movement verbs (Cl.20-21) do not form a separate class in terms of aspectual properties : they are rather a subclass of process verbs, with spatial support, and a spatial goal for bounded ones ; this explicitly spatial dimension was the reason for the separation of the subclass from the others, although there are few verbal notions without a spatial component ; but it is less explicit, or sometimes abstract or figurative61. In Figure 21, where classes 20 and 21 have been cumulated by decades, they seem to diverge a little from the main trend (General density). In terms of percentage, a slight decrease was also observed, especially in the last decade. This does not mean that people moved less in the eighteen eighties than a century before, but that this class of verbs, which were the pioneers of the progressive into the verb system, in company with locatives, saw their comparative frequency challenged by other predicates ; unless a complete study of all verb-frequencies showed that, for various reasons, there was also regular decrease in the use of movement verbs at non-progressive forms in the letters, which is unlikely. On this account, a remark seems in point : movement verbs are more versatile than they look ; there are many figurative expressions using them, and, whereas this does not alter their basic meaning, it may affect their frequency according to subject matter or idiosyncratic tendencies. In this respect, a larger number of letter-writers would be an advantage, by limiting the risk of idiolectal bias. On the other hand, figurative use compensates for possible restriction of true movement verbs. Admittedly, ordinary movement verbs are athematic : they appear in a great variety of situations. 3. Locative verbs (Figure 21) : with an average a little over 5 % of all progressives, this class also seems to diverge slightly from the main trend in terms of densities across the decades. In the pilot study (Arnaud 1973), it appeared that these verbs were of high frequency in the progressive, along with movement verbs. The justification for this Eleatic paradox has been given in I.2. In present-day English, the progressive is so frequent with verbs like lie, sit or live, etc. that we underestimate their use with the simple form. Second to movement verbs in their affinity with the progressive, they kept their position better in terms of percentage. The list of lexical items is short, and consists of verbs which are found in a large number of situations : lie, live, sit, stand, stay, etc. See also the lists of glottochronology for basic notions in all languages past or present (Swadesh 1952). There are borderline cases : e.g. wait has been registered here among prospective verbs, and lounge among unbounded descriptive verbs. Some of them, for example kneel, lounge, etc. can be topic-bound expressions of position or posture just as run, stride, march, stroll, etc. are varieties of movement linked to a definite situation. * 4. Processes : the results of the diachronic study of processes are even less significant than those concerning movement and locative verbs. Subclasses have been established for the coding of predications, mainly by discriminating bounded and unbounded processes. The two subclasses have also been divided further, in an attempt to study separately progressive processes (bounded), and descriptive predications (unbounded) : see I.3.3.a. It was not unreasonable to expect some enlightenment from the study of their respective behaviour, within the general framework of an implicational scale, not developed in this study. 61 This is an epistemological question, linked to the conceptualisation of "activity" as movement. Aristotle's conceptualisation of time in the Physics is derived from movement. 47 The observations are presented in Figure 22. Various methods have been tried to get as complete a picture as possible : individual profile of each class, by periods, decades, or equivalent lots in time sequence, cumulated data for several classes or sub-classes to minimize possible bias in the distribution. Only the most significant curves are presented here. There seems to be a hint that the development of bounded processes (Cl. 31+32) diverged from the main trend of density increase in being slightly slower. This is not altogether unexpected : the paragon of bounded processes are operations leading to a concrete, observable result. Building a house, writing a letter are creative processes, whereas breaking a pitcher, or burning a log are destructive, but they all end in an accomplishment which can be described by a perfect : the house is built, the pitcher is broken. This is one of the most ancient foundations of the theories of aspectual values, in connection with voice. What is significant at the period of reference is the appearance of the new progressive passive, the house is being built, not because of its frequency (see IV,1 below) but as a test of the original meaning of all process progressives, suggested by the old form the house is building. They are, so to say, symmetrical with respect to subject − agent or cause − and object − patient. The semantics and the terms adapted to it have raised and still raise endless discussions (see I.2.). It would seem that, at the end of the 18th century at least, the progressive is centered on the process itself, as being the state preceding the result, and leading to it. “ State ” referring to spatio-temporal location, by definition transitory, and also to spatial properties susceptible of transformation ; derived and figurative extensions have complicated the picture. No surprise then that bounded processes should be the first to be affected by the change : their dominance was challenged by other predicates. This is parallel to the comparative decrease of movement and locative verbs. In other words, those predicates which had already conquered the place did not participate as much in the increase as the newcomers. Among the latter, state predications are clearly winners, but with little influence in quantitative terms, as already noted. As to unbounded predications, they seem to have at least kept their percentage, which means that the general increase is mostly affecting them. But these conclusions are only tentative, considering the fragility of the statistics. Subclass 33 (unbounded, descriptive) was hypothezised to be one of the loci of the development. One of the observations the reader of the letters makes is the frequency, at the beginning of the period, of expressions like : it rained, the wind blew when we left, etc. which would seldom be heard or written today. This shows once more how close the progressive is to adjectival phrases, not referring to a process but just to a property ut nunc. Even in the case of the “ activity ” of sleep, if we do not adhere to the strict semantics of “ one word, one meaning ”, we note the equivalence of she is sleeping and she is asleep. There are languages, like Japanese, where a “ verb ” will be used to describe what we conceive as a state, and others like Basque or Irish where the water boils is translated by something like : the water is in ebullition ; cf. Basque ura irakitsen du, where -en is the inessive morpheme. In French l’eau est en ébullition is a quite acceptable translation, in more formal style, of the water is boiling, just another degree of being hot, bouillante. Similarly with les cerisiers sont en fleurs, les cerisiers sont fleuris, to translate : the cherry trees are blooming : cf. in full bloom or blossom. Thanks to the progressive, the adjectival force of the verb is put into prominence. This is not indulging in a simplistic finalist or functionalist notion of progress in language to say that, in this case, the evolution has some logical justification. Process can be a misleading term, just as activity. Paraphrastic gloss reveals at least the ambiguity of verbal sentences where no agentivity can be clearly observed, such as metereological expressions : the sun is shining, the wind is blowing.62 In ancient French, Bluebeard’s wife says to sister Anne : Je ne vois que la route qui poudroie et que l’herbe qui verdoie. It just means that the road is smothered in dust and the grass green : no activity is implied. Often, in logical terms, we have analytic statements, as Alexander Bain 1870 : 14 had already noted, without remarking that this is due to the canonic binary structure of English.63 The fire is burning, the wind is blowing are typical : it just means that the fire exists (for the moment), there is (some) wind, it is windy, etc. Yet, contrary to expectations, no clear precedence of that descriptive subclass has been observed in the counts, whether in terms of densities or percentages : it follows the main course of unbounded predications. It has often been said that the progressive solves the ambiguity of the simple form between momentary and lasting interpretation, especially with indeterminate subjects : snow falls was ambiguous, snow is falling is not. This is particularly true with this type of verbs, and justifies the subclass. As most other classes, it is not tight : the 62 Cf. on this point the discussion on inaccusatives ( Perlmutter 1978, Ruwet 1988). This evokes the question of Indo-european animism and the status of the nominative. In ancient Greek, however, Zeus vei appears later than impersonal vei.( Benveniste 1966: 230). 63 "Fire burns is not a real proposition. It merely repeats, or unfolds, the chief attribute of the subject. Our earliest, and most persistent notion of fire, is the same as is expressed by burning". 48 adjectival force of the progressive is underlined here, but we still can observe it with a more “ active ” subject, even human : he is working is just another way of saying he is busy and not idle. There is a special class of adjectives referring to temporary states or occupations : asleep/awake, angry/merry, sober/ drunk, cloudy/ fair ; ill or sick/fine, etc. are examples, whereas some others are susceptible of less actualized interpretations : happy or sad, for example. For subclass 32 (progressives), the curve, not presented here, does not show any clear increase in density. All bounded predications − except for the so-called punctual events − 64are in fact progressive : this is certainly what suggested that very disputable name of “ progressive form ”. Even meet or burst or jump can imply steps. But there is a class of verbs which are not very difficult to tag as specifically oriented to a goal in the same manner as Achilles running behind Zeno’s tortoise : see list in the ANNEX. The idea is that they are among the predications most likely to use the progressive. We would be tempted to say that, if they do not show significant increase in the period surveyed, it is for the same reason as for locative predications : they had already won the race. To conclude on the chapter of predicational classes, though the figures cannot be considered as indisputably significant, they tend to support ordinary experience and the assumptions of the majority of linguists. All we can say is that the distribution of the different classes and subclasses was not profoundly disturbed in the course of the century. Locative verbs retained their share, movement verbs possibly had to make some room for other process verbs. With reference to the latter, the slight difference observed between bounded processes and unbounded processes may be explained in the following manner. Originally, about the middle of the 18th century, the progressive can be seen as a sort of middle-voice, focussing on the “ event ” itself, not necessarily a process, but only a temporary condition. When the next stage is seen as a result, the momentary stage refers to a temporary property of the subject-agent if the orientation is “ active ” or of the “ patient ”, if the orientation is “ passive ”, although in fact both are involved 65. So that bounded processes were normally leaders in the success of the form. The extension to unbounded events, with just a termination, eventually delayed, came later.66 This is parallel to the development of have with intransitive perfects (Rydén & Brorström 1987). 67 With unbounded intransitives, once the terminal point has been passed, the perfect is one way of looking back on the previous stage at the same time as describing the present one succeeding it (present relevance). Whereas, when the new stage had not been reached, the progressive was a description of the only participant, the subject. It is difficult to give definite conclusions on this extensive study of verb classes : this may be due partly to the nature of the corpus, to the distribution of predications in their respective classes, or even to the semantic analyses behind them. New theoretical developments may permit new insights into the matter : there is some reason to doubt whether they would considerably alter this unconclusive aspect of the survey. The upward movement seems to affect almost evenly the whole system, extending to all types of predications, and more conspicuously to those which had not yet been much affected, the state verbs, and to unbounded processes, an extensive class always adding to its lexicon. III.2.2. Verb form (Tenses) a. Overall distribution Verb form, here, covers the morpho-syntactic variations of the verb according to time-reference, modality or voice. 64 The concept of punctual events is a difficult one, involving a definition of time as duration. The epistemology of duration, in ontogenetic or phylogenetic terms, is even more difficult than that of time-sequence, and is linked with that of movement. Aristotle had also sensed that. Space, time or movement can all be axiomatic. 65 The cook and the dinner are cooking at the same time, but only the dinner will be "cooked", or "have been cooked" although if the chore was too heavy the cook may say: "I am done". 66 This is just a suggestion, until extensive surveys of the previous periods have been made. 67 Instead of saying "I am done", which is ambiguous but stresses the effect on the agent, more clearly now than in early 19th century English, the cook might now use a have perfect. I have been doing it, I have been cooking all morning, are just the new ways of emphasizing the involvement of the agent as well as that of the "patient" in the result just obtained and the effect on the agent as well. Everything we did can be said to have affected us: "L'homme est fils de ses oeuvres"[man is the son of his works]. 49 Whenever the progressive is summarily evoked, present-tense quotations inevitably appear. This involves a risk due to the polysemy, and occasional ambiguity of the present (momentary, extended, timeless, etc.) so that it is certainly advisable to consider the past tense in the first approach.68 Many analysts have also paid special attention to the perfect, the counterpart of the progressive in the aspectual system. It is obvious that the density of progressives varies with the morphological classes of the verb : infinitives, of different forms and meanings ( I expect to be talking with you, Coleridge, 1825 ; I am miserable unless I be soaring in the empyrean, Carlyle, 1823), participles ( I write shortly, having been working my head off , Dickens, 1852), imperatives ( Now, dear child, don't be playing pranks and shocking people, Eliot 185269), passives, and modal expressions ( If he must be reviewing, there is my speech about the West Indies, Macaulay, 1850), are much less common than present, past and perfect progressive forms. The framework for this part of the study is unsophisticated. Classes expected to contain few occurrences have not been subdivided, since the quotations, for passives and even modals, etc. can be studied individually. It was found that percentages were the most significant measures of the comparative development of the main tense forms (Figure 25 ). Considered separately from other factors, they must be examined with a critical eye, for various reasons : 1. The scale is different from one “ tense ” to another, and the validity of the curve varies accordingly, even if we posit a normal distribution ; 2. The distribution (of present and past for example) is closely associated with subject-matter and type of discourse, narrative or argumentative, descriptive, etc. ; 3. As noted above, the present tense is versatile. This again is connected with the topic ; 4. The use of the “ tenses ” will become more significant with at least a bidimensional study, involving in particular the dimension of semantic classes : see (b) below. With those reservations, what can be said of the observed distribution ? Just as for predication classes, some of the small-sized conjugational forms seem to be on the increase. This is particularly true of the passive : Figure 26. The progressive passive was gradually substituted for the ancient form which however was still in use at the end of the period. The point will be further illustrated in IV.1. With the infinitive, the picture is less clear : its use, interesting in itself with its different forms and meanings, is even more exceptional than that of the passive, making a statistical evaluation difficult. The same applies even more to the imperative. The use with all modal auxiliaries, contrary to expectation, does not show very significant increase, which requires further analysis, since they are a hetereogenous group, containing, among other forms, those referring to the future : see IV.8. One of the interesting hints of Figure 25 is that the percentage of the perfect was diminishing. Just as in the case of bounded predications (see also b. below), this suggests that it is due to challenging new extensions. As to the pluperfect, the figures are too small to be really significant. 68 69 Strang 1982 underlines this too. Negative, with the exception of some by Carlyle : . Up Jack ! up and be doing what thou canst !. 1824 Scotticism ? 50 b. Perfect and processes The hypothesis that, parallel to the decline of be perfects with intransitives, or more generally unbounded processes, the perfect progressive behaved differently with those types, has been tested for the two diachronic groups A and B. This is not a truly longitudinal study, but here are the results : Table 3 Processes and perfect Unbounded Perct A B 30 Bounded 33 30/33 7,36% 0,85% 8,54% 0,40% 31 32 31/32 8,21% 2,45% 1,14% 3,59% 8,94% 2,49% 0,60% 3,09% In terms of percentage of the total predications in the progressive, we can see that there is a slight increase of unbounded processes with group B, implying a relative decrease of bounded processes. This is what was expected, although the figures are not considerably different. Considered in term of densities in the two groups, we get the following : Unbounded Classes Density A Density B 30 12 26 33 1 1 30/33 13 28 Bounded 31 4 8 32 2 2 31/32 6 10 This confirms that the increase of perfects with unbounded predications between the two chronological groups is probably greater than with bounded verbs : their density is more than doubled, whereas for the total of bounded predicates it passes only from 6 to 10. II.2.3. Development across the lexicon Nearly 2000 lexical verbs are represented in our corpus. Their frequency of occurrence varies considerably : the most common are go (2000 occurrences approximately), write (1036) come (867). 1766 items are present less than 10 times, and 1034 only once (hapaxes). It is probably better to avoid the term of “ lexical diffusion ” since we have to do here with a type of conditioning very different from what was connected with the neo-grammarian controversy (Wang 1977, Labov, 1981). As we have seen in section I.3, the main problem here is not, contrary to what can be supposed, the listing of verbs with their appropriate distribution into the relevant sets. Given the rough approximation we are aiming at, this is not an insurmountable task. The main difficulty is that we are dealing with a form which is situation-bound, and situations of speech − here, mostly with reference to topics− are of a great variety. Students of lexico-statistics have demonstrated that the vocabulary of a text increases with its length. As has long been known, some words have a high degree of frequency and stability because they are athematic, used in a variety of situations. This is true of state verbs of the different subclasses. A typical case is be itself : it was excluded from the binary study in Arnaud 1973 because it is extremely rare at the progressive whereas its frequency in the simple form implied a considerable wasting of time and effort. But there are other verbs commonly used in many situations and present at the progressive from the very beginning of the period : verbs like do − with a large spectre of meanings and collocations −, or make, or locative and movement verbs. The curve of their density increase (not presented here) can be shown to parallel the general curve for their predication class. But thematic verbs, verbs associated to a given topic, especially with our great writers who mastered a large lexicon, have a tendency to appear erratically, according to need or even fancy. Grouping them into classes duplicates the 51 observations of the semantic study above. The other way of looking at it is to order them in frequency lists, with the perhaps risky hypothesis that, if the verbs of low frequency at a given period are more numerous than verbs of the same frequency level at the previous period, this means that the diffusion affects preferably the class of verbs to which these low frequency tokens belong. It seems sensible not to separate hapaxes (frequency 1 in a given lot) and verbs of frequency 2 or 3. When people use a new verb in a given situation, they are tempted to use it again, sometimes at little distance from the first, by “ contagion ” as statisticians also say. A curve of frequency distributions for individual verb tokens has been drawn for groups A and B (corresponding roughly to the two successive “ generations ” of writers) : it takes the familiar hyperbolic form of vocabulary frequency distributions : Figure 27. Apparently the distribution is very similar in both chronological groups. This tends to show that the choice of the progressive has little or no influence on the selection of vocabulary, or that the choice of vocabulary at a given point is made independently. For the various strata (correspondence, murder trials, dialogue in novels, novels, and selections from Punch (18401870) in the previous binary study (Arnaud 1973) such curves had been drawn for both simple and progressive forms of equal samples of 1000 occurrences : they showed no clear difference between the frequency distributions of the two forms, which would support the suggestion above. For a very significant study of this point, a knowledge of the rules of vocabulary increase applied to the corpus considered would certainly be helpful. We have already noted that new verbs, neologisms, nonce-words or backformations are commonly found at the progressive form, as if two novelties (so long as the progressive was still felt as a novelty) were attracting each other. Here is a short list of back-formations found in our corpus in progressive verb-forms : new roof (Dorothy Wordsworth), house hunt (Thackeray and others), holiday make (Dickens) picture see (Thackeray) quill drive (Thackeray), sight see (Thackeray) wool gather (Thackeray), gaslight (Ruskin), daguerreotype (Ruskin). They seem to be more frequent with the second generation. It does not seem likely that the choice of the participant in the process, giving the first element in the compound (often the “ patient ”, but also “ instrument ” or even something else) which is important for the transformational analysis, accounts for this transfer from gerundial forms to progressives. The rapid introduction of the progressive with those new verbs would rather add a stone to the theories of Gabriel de Tarde 1895 on imitation, or simply to the neo-grammarian secondary rule of analogy. 70 It is well-known that, in modern French, all new verbs take the -er conjugation, the more modern type : zapper, booster, to quote recent anglicisms, from zapping, boosting. Yet, besides the idea of novelty, another motive, probably stronger, is that these new forms (sometimes nonce-words as they are aptly called although many have survived ) have a strong component of immediacy, are usually strongly situated. Provisionally at least, we will then have to admit that the neogrammarian rule is working in this case : the change affects (virtually) the whole lexical system at once. But, whereas only a short length of text, oral or written, is necessary to encounter practically all phonemes, the dictionary of verbs is far from being explored even with 10 million words of written text. Even a few verbs of the small classes 1 to 4 (states) not yet found in the corpus can turn up unexpectedly in a letter : beseech (Dickens), justify (Dickens), exhort (Eliot), prescribe (Gaskell), etc. Their rarity as tokens parallels the small number of lexical items in the class as a whole. III.2.4. Person On a total of 21 777 occurrences, 8937 are at the first person (41 per cent) and 8947 at the second : so that first plus second persons represent nearly half of the total number. This could easily be surmised. If we add to this the probable number of third person utterances referring to human subjects, it is easy to see that, in this type of discourse especially, the progressive is overwhelmingly a form concerning people, and largely in dialogic situations. It is not suprising then that studies have consistently overlooked the other cases. But that the progressive is fundamentally a subjective form is undeniable : Cf. Ota (1963 : 73 sq.) ,who studies what he calls I/you verbs. This is however too vague a concept and it requires further definition : see General conclusion. 70 It does not take a big leap to pass from a gerund or verbal noun to a progressive. However, it is not certain that the verb was introduced first in the progressive: the O.E.D. gives the simple form as the original attested one; but this is only in texts, and often literary texts. Nothing to support this precedence of the simple form has been observed in our corpus, closer to ordinary speech. 52 III.3. Sociolinguistic factors III.3.1. Men and women : The first dimension available for study was the gender dimension, with the traditional query : is there a difference between women and men in receptivity to the progressive ? Several techniques have been used to assess the comparative densities, relying on the convergence of the results. III.3.1.a. Overall densities for the whole period : Women : 247 Men : 211 III.3.1.b. Extremes: Highest density Women : 358 (Gaskell) Men : 334 (Thackeray) Lowest density Women : 157 (Hutchinson) Men : 107 (Coleridge) These figures also point to the extremes of the diachronic curve showing the total output of individual writers (Fig. 4 & 5). III.3.1.c. Diachronic groups : The two diachronic groups already studied for their total densities in III.1.2., can also be referred to for a comparison of men and women. TABLE 4 Group A : THE EARLY ROMANTICS Women Words: 1 435 000 PF: 2621 D: 183 Men Words: 3 023 000 PF: 4430 D: 147 Group B : THE LATER ROMANTICS Women Words: 1 824 000 PF: 5443 D: 298 Men 53 Words: 1 900 000 PF: 5388 D: 281 In both groups, the density for women is higher. The contribution of women to group A, however, is only half that of men for the same age-group. This makes the comparison a little less reliable. III.3.1.d. Pairs : The comparison can be supplemented by selecting people of the same generation supposed to share the same linguistic environment and possibly the same values. 1. EARLY ROMANTICS : Men Wm Wordsworth :144 Coleridge : 107 Women Mary Wordsworth : 266 Sara Hutchinson : 157 2. LATER ROMANTICS : Men Dickens : 274 Thackeray : 334 Women Mrs Gaskell : 358 George Eliot : 344 NB. The pairs of the first generation are indeed couples, since there was a love-affair between Coleridge and Sara Hutchinson. Yet, they met only at intervals and could hardly have influenced each other in this respect. There are two exceptions to the rule : Southey, with a density of 176, is nearly on the same level as the average for women of his age-group, and J.J. Ruskin (179) shows a higher density than his wife Margaret (158) : but for the latter two, the corpus size is too small to be relied on. All attempts to connect the higher density consistently exhibited by women to other factors commonly suggested : − choice of topics (see for ex. Kipers 1987), semantics of verbs, preference for the first or second person, etc. − have failed. The difference between the figures is sometimes not very great, yet it seems difficult not to recognize that, for the population considered, women are definitely ahead of men in their use of the progressive. 71 A possible explanation for that difference will be suggested below. Two remarks may be in point : 1. It is unfortunate that we have so little information about children. The early letters of Thackeray and of John Ruskin which have been preserved are too few to be significant ; yet they show that children were addicted to the new forms − was it due to their being exposed to the language of women, often also of lower-class girls, their nurses, sometimes Irish ? 2. The high densities exhibited by Jane Austen both in her letters and in her novels have been considered in the idiolects section, III.1.4. They may also afford a clue. III.3.2. Degrees of intimacy.72 As exposed in I.3. b, degrees of intimacy are often hard to distinguish. Once more, we rely on the convergence of the results to form a clear idea of the weight of this factor. With a certain number of our letter-writers, it would not be significant to try to differentiate degrees, either because we have no basis for a justified choice, in which cases − Keats, Edgeworth − degree 2 has been taken as the average degree of intimacy. or because nearly all the letters published were addressed to family or close friends − Mary Wordsworth, Jane Austen, the Ruskins, and even Jane Welsh-Carlyle, who had few correspondents outside the intimate circle. The results can be computed for the whole corpus, for groups, and for individuals. III.3.2.a. For the whole corpus 71 This does not mean that being fond of the progressive (with its ambiguous name) is a quality, or that being conservative is a defect. Evaluation is a different sort of problem, first for the general community, and for the sociohistorical student of the resistance to change. 72 See list of correspondents according to intimacy in the ANNEX. 54 TABLE 5 Densities for the 3 degrees of intimacy Degrees Men Women Difference 1 2 3 253 193 136 270 241 191 17 48 55 The results are clear : women are ahead of men for all degrees of intimacy. It must be recalled that degrees were selected before the count was made : they could possibly be disputed in some cases on the basis of biographical evidence, but can the consistency of the results, even when differences are not considerable, be due to chance ? III.3.2.b. Diachronic groups : For group A, where Dorothy Wordsworth has been included, since she had been omitted from the general count in III.1.2. only for the sake of equilibrating the groups, it may be useful to give two sets of figures : • with the elimination of Edgeworth and Keats ( for whom no significant degrees were observed) : General density : 151 Density Intimacy 1 : 182 • with the elimination of writers with only degree 1 of intimacy : General density : 143 Density Intimacy 1 : 177 For group B, eliminating writers whose intimate letters are difficult to distinguish ( Robert and E.B.-Browning) or who were almost always writing to intimate correspondents (Jane Welsh, Mrs Gaskell, Ruskin) we get the following results for the added outputs of Eliot, Macaulay, Thackeray and Dickens : Density Intimacy 1 : 322 Density Intimacy 2 : 310 Density Intimacy 3 : 204 General density : 291 Considering the possibility of bias due to different sample sizes and disputable allocation of intimacy degrees, the differences may not seem considerable. It is however clear that there is a wide gap between Int. 1 and 3 (more than 100), though the corpus for degree 3 is noticeably smaller and consequently less reliable. 73 Between Intimacy 1 and 2 the figures are not much apart. In fact, when we consider individual writers, we see that Macaulay and Dickens are reversing the difference : Macaulay D1 : 235 D2 : 306 Dickens D1 : 268 D2 : 318 In Macaulay’s case, this is possibly due to the comparative restraint noticeable in his letters to his parents, whereas in Dickens’s letters, the ascription of degrees 1 and 2 may be responsible and/or the choice of topic may have more influence than intimacy with the addressee. We don’t claim more than the observation of a tendency. Closer analysis would be necessary to give more precision to the notion of intimacy. Then, the measurement of density is a very rough approximation. It happens however that only when satellites could explore them from a great distance were some features of the earth’s surface seen for the first time. Again what is encouraging is that there are few exceptions to the general trend. Another encouragement 73 On the other hand, it has always been easy to locate the letters addressed to strangers or implying distance. 55 is that, apart from private correspondence, the density of progressives varies in all texts along a formality scale (Arnaud 1973, Smitterberg 1999, 2002).74 The section on literature below will consider some of the reasons for this, just as the study of individual trajectories in III.1.4.b, and the role of women. Convergence is again the key word 74 Formality is no more than a rough and ready approximation: consideration of "genres" (Biber & Finegan 1997, Smitterberg 1999, 2002) is now giving more substantial insights into the variety of language use. 56 CHAPTER FOUR FURTHER INVESTIGATIONS The informations presented under this title concern the linguistic conditionings and constraints of the development of the progressive and can also be seen as a synchronic picture for the period : a collection of individual observations as well as indications of general tendencies. Many of them confirm recognized assumptions, quantifying and illustrating them ; others may occasionally challenge them. Here again, we are relying rather on the convergence of the results and what is commonly admitted than on statistical estimates. It would be beyond the scope of the present work to attempt a detailed linguistic analysis of often-debated questions for which it simply aims at providing material. IV. 1. The development of the passive (Figure 26) : The reasons for this development and the rise of the progressive passive will not be discussed here : see e.g. Warner 1995, 1997 ; Denison, 1998; Nakamura, 1998; Smitterberg, 2002. At the end of the 18th century, the old “ symmetric ” progressive with passive force (the house is building can be seen as a reversal of someone is building the house) was still the only one to be used. It gradually gave way to the new form in be-ing : the house is being built. Some relics are found today : the dinner is cooking, Hamlet is playing. We take it for granted now that : the bell is ringing is not a passive, because the agent is not explicit ; yet, compare : Galatian [the hymn] is singing in the hall with a large troop of boys (Scott, 1829, 88). This could be the case with several verbs used transitively or intransitively. The door was opening, the eggs are boiling, are not so far from 19th century phrases such as : While the horses were changing I happened ... to run up a street (Ruskin, 1845, 13) or : My third boy... is educating expressly for engineers or artillery, ( Dickens, 1862, 3) ; “ O my eye ! Stop thief ! I am strangling ” (Dickens, Our Mutual Friend) ; Lenchen came in with a telegram while I was dressing (Queen Victoria) : a queen does not dress without help, nor is it certain that Jane Austen did her hair alone : I have not seen them yet− as my hair was dressing when the Man and Stockgs came (Austen, 1813, 30). This is largely a matter of the agentive system we have at the back of our minds and anachronisms are always a risk. We are supposed not to be animists and yet many of our agentless expressions suggest some concealed agent. This is especially the case with social phenomena : taxes are increasing, augmenting. A force was collecting at Bridport was a normal expression for Macaulay ; cf. also gathering. All this refers us to the history of the middle-voice and the activepassive development in Indo-European languages. In many cases, the concept of “ squish ”(Cf. Bailey & Shuy 1973 : 62-140) could be helpful here, since there is some variability in the constraints and the passive category is not homogeneous : e.g. the bell was ringing / bells ring ; the novel is selling / novels sell (well) ; a sirloin was roasting / ?*sirloins roast ; our woollens are washing/ ? woollens don’t wash ; baby is washing/ * babies wash. As shown in Figure 26, the new passive was adopted by many of our letter-writers from the beginning, 75 but only occasionally. The figures for densities by decades, adjusted to facilitate comparison with the overall increase, illustrate convergence of the observations.76 They display a crossing of the curves by 1840. Small as the figures are, they undeniably show that our letter-writers accompanied the development. This sudden acceptance after a period of latency when the literate community was more than reticent, sheds some light on the way linguistics barriers are overthrown under the pressure of external forces, and perhaps not differently from other sociological barriers. This must apply to fashion for example such as hemline adjustment or hair-style, but also to more socially important behaviour : the way unmarried couples have become accepted by their own families, etc. Why, for example, so many speeches, in the House of Commons or elsewhere, so many newspaper articles, so many words indeed, had to be said, before the mere idea of democracy was accepted in Britain ? 77 It takes a long time before a 75 It is only recently that instances of the passive have been found earlier than the well-known quotation from Southey's letters ( Warner 1995, Denison 1998, Nakamura 1998). 76 A computation of percentages has also been made: it is so similar to the density curve that it is not presented here. 77 Coleridge, Southey and Wordsworth (cf. his Prelude, begun in 1799) had accepted it in their young years; in 1867, Disraeli was still opposed to the word and the idea: We do not, however, live − and I trust will never live 57 psychosociological current running underground, suddenly erupts, soon to be aknowledged as if it had always been so. We may suggest that linguistic progress happens in two opposite manners, with possibly some intermediate cases : 1. When the change is not conscious, or not clearly so, the development is slow and gradual ; 2. When the item considered is stigmatized, systematically shunned by the leaders of linguistic opinion (see below about Macaulay), the development supported by both internal and external forces, is long contained and suddenly bursts off, as when the gate of a castle is forced − by barbarians as some people will still say ; many such barbarisms became the new rule. Some of the old passives are still in common use at present : sell, cook, play (speaking of a theatre-play ; Dickens also uses act). But even the notorious build is found as late as 1857, in Macaulay’s letters, but also with Mrs Gaskell (1851), Dickens (1844,1854) and J. Ruskin (1845). Among his generation, Dickens seems to have been one of the most faithful to it : in his letters we find 24 occurrences of the old form not normally used today, but there are also 18 progressive passives including one for play (theatre) : Nothing is being played here scarcely that is not founded on my books, 1867. After the fifties, John Ruskin and Mrs Gaskell are the only ones where we find the old forms, occasionally. According to his biographer G. Trevelyan, Macaulay censured the new passive (Mossé 1938 : 157) ; and yet, Mossé found two cases in his early letters; there are 6 in fact, but not a single one after 1834 : youthful sins ! Since Macaulay is a notorious case of faithfulness to the old usage, the passives found in his letters are an interesting example of the development : 1815 I suppose we may expect his Royal Highness the Prince Regent, as the Pavilion is now being repaired and done up. 1819 The Poem is being finally struck off. 1826 But to carry a letter of introduction, to wait in the outer room while it is being read... 1830 The Members of Parliament are being sworn in again. 1834 All the Edinburgh Reviews are being bound. 1852 and my first lunch is preparing. Was ever anything so lucky as Denison’s letter coming just when the cry of Puseyism was setting up against the new Ministry ? (a dubious example perhaps). 1855 I am thinking more of the state of the currency and the bank in 1696 than of what is doing in the City at present. 1857 Lord Panmure has asked me to write an inscription for a column which is building at Scutari. Born in 1800, Macaulay was 34 when he used the modern form for the last time in his letters, and in an expression where the old form still survives today. After that, just as in his historical works, he seems to have deliberately reacted. Only one case has been found when the agent was expressed, in a letter by Jane Austen : our garden is putting in order by a Man who bears a remarkable good character (1807) ; −under a democracy; 18 March 1867, in his speech on the famous night when he and Gladstone were confronted on the issue of the Second Parliamentary Reform. Cf. Raymond Williams, Keywords p. 85. 58 Another, in Mansfield Park, is quoted by Jespersen III 1935 : 352 : the baize was actually forming into a curtain by the housemaids. There are very few cases like that, where the passive force is explicited clearly. Mossé 1938 gives an older example by Walpole : it is there the the search must be making by Manfred and the strangers; and another by Dickens : my cup of tea which was then preparing by Mary Jane. Lindley Murray 1795 suggests the following, probably of his own making : the Indian was burning by the cruelty of his enemies ; the youth was consuming by a slow malady. Two cases where the agentive nature of the subject is dubious : the distinction between agent and cause is notoriously difficult. IV. 2. The extended present : This label covers the use of the progressive when the context makes it clear that the time-reference is situated either after the moment of speaking (with a present or a past tense-form) or before it (with a perfect or pluperfect tenseform). The most common way of stressing this immediacy is just, whatever the orientation. Curiously, grammarians have been mostly interested with the near-future reference, as in the following (with just as a typical adverb, which often looks like a mere expletive) : I am just going down to Lord Aberdeen (Scott, 1807) ; I am just setting off for Sundridge (W. Wordsworth, 1818) ; she is just going to Paris for a week or two (R. Browning, 1870). In the other case, the present relevance is commonly referred to the perfect, not to the progressive itself : Caroline, Anna and I have just been devouring some cold souse (Austen, 1796) ; Mrs Robberds’s son-in-law has just been having the small-pox (Gaskell, 1865) ; Anny and I had been talking about you just before and she had been telling me how my stepfather ... would say to them ...(Thackeray, 1853). In 19th century usage, there is some symmetry between the two, with the utterance point as a sort of magnet. 252 cases of near-future reference have been found, to which can be added 23 cases where the utterance point is in the past. The curve (Figure 28) is ascending first, but then depressed. This may be due to chance, but there may be another reason : it is well-known that, in modern English, three expressions are in competition with the progressive for nearfuture reference, with different connotations. The plain simple form, involving usually an idea of prearranged plan, is reserved to a limited number of “ perfunctory ” cases : it certainly was far more widespread in 1800 than it is now. The be-to form, so common with Jane Austen : it is seldom used nowadays and must have been fairly rapidly on the wane, especially when no idea of constraint or compulsion was associated to it. Then, we have the going to expressions : no count was taken of them for this research. 59 On this account some impressions must be mentioned : . 1. Going to was sometimes in competition with coming to78 2. The going to form seems to have developed after the plain progressive alone had taken that “ intentional ” value. This may be the reason why the latter’s development shows an abrupt decrease by the fifties (periods 12-13), since even today in current practice the two are not always clearly differentiated. We wouldn’t say : we are going to go to the seaside tomorrow, very willingly, so we are going to the seaside will be preferred. Go, come, and other verbs of movement are the most commonly used with extended present meaning. With our letter-writers, we also often find write (with future reference mostly) and read (with past reference mostly). The decrease may also just be due to chance or idiosyncratic preference. IV.3. Time-frame Jespersen’s theory of the progressive as forming a “ time-frame ” for another event is well-known. The prolific and richly documented grammarian has long exerted a considerable influence, and he maintained his views as long as he lived. He claimed he had the idea in 1914, and its expression is fully developed in his Modern English Grammar on Historical Principles of 1931, Part IV : 178. He gives the following diagrams : he is writing he was writing (now) when I entered he is writing (he has begun writing) now (he has not stopped writing) he was writing the moment (he had begun writing) of my entering (he had not stopped writing) And he explains : “ The essential thing is that the action or state denoted by the expanded tense [i.e. the progressive form] is thought of as a temporal frame encompassing something else which as often as not is to be understood from the whole situation. ” This last remark was kept in mind when computing the occurrences of the progressive, but one has to trust the researcher for judging the situation from the context. 1,230 progressives were found, out of some 21,800, satisfying Jespersen’s description more or less, that is between 5 and 6 per cent only. If we remember Bain’s remark that the “ agent is now engrossed, and is barred from other occupation ”, we may have a very general view of the case. Certainly, speaking of people or of things as well, there is always something “ going on ” − “ goings-on ” is the neutral characterisation of the progressive suggested by Leech 1971 : 14, who also mentions its use for a temporal frame, p. 17 −, so that any “ situation, happening or going-on ”, to quote Leech again, inevitably follows something that precedes and precedes what follows 79 , and even what precedes may be going on 77 This metaphoric use of verbs of movement, sometimes with curious redundancy, is well-known and not limited to English: cf. French nous allons partir. When the context implies movement, it can be difficult to discriminate metaphor from ordinary sense. 79 The original aspectual significance of the progressive could be illustrated by the saying, famous in France, concerning Monsieur de la Palisse, a legendary knight, who fought in Italy with king François I : Monsieur de la Palisse est mort Il est mort devant Pavie; Un quart d'heure avant sa mort Il était encore en vie. 60 at the same time as the subject is engaged in the new occupation referred to by the progressive, a point that Bain overlooked. Taken in this sense, Jespersen’s statement can be seen as just another way of emphasizing the existential meaning of the progressive − the condition of the subject in a spatio-temporal frame (see I.2. above) − in the “ stream of events ” paralleling William James’s “ stream of consciousness ”. We can see that a more specific description explicitely implying simultaneity covers only a limited number of situations. 80 IV.4. Bounded intervals of time : One of the perversities of corpus-oriented linguists is to look for cases belying some of the accepted rules of schoolgrammars or even of more sophisticated analysts. Most grammarians have always been cautious with the limitations of the progressive, whether it concerned the lexicon or the contextual or situational factors, but it has sometimes been said or implied that explicitly limited duration would be in contradiction with the imperfective meaning of the progressive81 : one of the common views, mostly founded on present or perfect tense examples, is that the timeinterval when the predication is valid cannot be bounded, at least in present-day English 82 whereas the simple form is synoptic, hence encompassing both ends of the interval. Explicit measurement of the time-interval has been found in about 200 cases in our corpus.83 Does this imply that this interval is bounded ? Certainly not. The usual context where we find such expressions is of the type : I have been writing for two hours. As we know, the perfect in this case is quite in accordance with continuation of the activity, occupation or state of things, and this will be true with the past perfect when it is a present perfect translated into the past. The vast majority of our examples ( nearly two thirds) are plain present or past perfects, to which we should add some indirect cases such as : Your invaluable packet ... having all this time been lying at the post off ( Mary Wordsworth, 1839) . Uncertainties about the real time-span sometimes occur. When Dorothy Wordsworth writes : Mr and Mrs Cooke have been spending a month at Ambleside (D. Wordsworth, 1803), this is uncommittal : the month may or may not be ended. This is a familiar point when one studies the perfect and its “ present relevance ”, further emphasized by the progressive. The remaining 70 cases when the time-limit is explicitely mentioned are more challenging. On this point, Leech 1971 : 16, makes interesting remarks. About : I was reading from 10 p.m. to 11 p.m. / I read from 10 to 11 p.m. M. de la Palisse is dead. He died at Pavia. A quarter of an hour before his death, he was still living (alive). Les vérités de la Palisse are that sort of statements, said to be so incontrovertible that they seem stupid. 80 The same encompassing value has been observed with the French imparfait. Cf. Imbs 1960: 90-1, who speaks of "fond de décor": Le garagiste la rattrapa au vol et la remit sur pied. Pablo criait et pleurait (Sartre); On l'a tuée ? demanda-t-elle, comme Maigret s'asseyait près de la fenêtre (Siménon). 81 E.g. R.L. Allen 1966: 220, who compares what he calls "suffusive" and overlapping predications; Leech 1971: 17; Buyssens 1968: 156, 166. 82 A comprehensive examination of the adverbials used with progressives can be found in Scheffer 1975: 50-53. In his corpus, adverbials denoting restricted duration, answering the question "how long?", represent 18.7 % of all adverbials. Scheffer refers us to Diver 1963: 148-9 for a criticism of the ambiguities of the term "limitation". Cf. also Joos 1964 and his "limitation of the validity of the predication", supposed to be diminishing gradually as time goes on (p. 107, 112). He calls the progressive the "temporary aspect", and says, about examples like "Am I really hearing what you are saying ?": I propose that this means, or rather its use of the temporary aspect means: Assuming that the predication is valid for the time principally referred to, then it is 99 per cent probably valid. for certain slightly earlier and late times, it is 96 per cent probably valid for times earlier and later by somewhat more than that, and so on until the probability of its validity has diminished to zero. 83 Buyssens 1968: 156 sq. gives a long list of modern quotations on this point. He concludes that the progressive unsually underlines that the span of time was longer than normally expected and that its end is coming. 61 he writes : The Simple Past tells us that the speaker started to read at 10 o’clock and finished at 11 o’clock. The Progressive, however, does not specify either the time of beginning or the time of completing the activity : all we know is that reading was in progress for that hour. Hence it would be a fitting answer from a suspect being interrogated by a detective. The detective would ask What WERE you DOING between 10 p.m. and 11 p.m. ? being uninterested in whether the activity persisted after that period or not ; and the suspect would reply in kind. This can be referred to the analysis attempted in I.2. of the semantics of the progressive and to I.3.2. on the situation of utterance, at least when animate-human subjects are considered. The aspectual value in terms of boundaries is superseded or overshadowed by the adjectival value of temporary occupation. Boundaries are not entirely irrelevant as when the synoptic simple form is used, but they are overlooked, or implicitly flexible.84 The analogy with temporary state adjectives would be instructive in this case : She was asleep from 10 p.m. till 10 a.m. ; Her telephone was busy for two hours ; We were all glad of his success until we learned that he had been disqualified, etc. But : ? We were all enjoying his success until... Until and till may reasonably be said to set a time-limit to the predication. Sometimes, it can be said that until opens a new time-span instead of closing the previous one85 : Your book, I imagine, is being reserved until the right time for advertising school books. [ i-e. comes ]( Eliot ,1863). But even this example, in the passive, defines a clear and “ punctual ” time-limit and most if not all seem to do so, thus closing the prospect. The example given by Leech was in the past tense, but many in our corpus are in the present . They mostly concern locative verbs : We are staying at Boulogne until the 10th of October( Dickens, 1853) ; Do you understand that I am stopping here till Saturday morning 11 o’clock (Gaskell, 1860). Yet we also find : but till then I am delivering a course of extempore lectures which (Carlyle, 1838) ; I am saving up the perusal of Vanity Fair until I have done Dombey. (Dickens, 1848). In the past, more common, we find : I was gambling until 2 o’clock this morning ( Thackeray, 1851). There are many cases when only the length of the time-interval suggests its limitation. Again, locative verbs are the most common : Mrs Somerville and I were sitting on the opposite sofa all night (Edgeworth, 1822) ; but we also find : I was for several hours passing through a succession of spots which ... (Macaulay, 1834). On Wednesday I was riding for 12 hours canvassing rather a feat for me (Thackeray, 1832). In the following : From the moment we came till we left them Mr & Mrs Fowler were contriving and executing sthg for our amusement (Edgeworth, 1821), the time-limits are clearly defined. 84 85 Or Culioli's "open boundaries" (Cf. Groussier & Rivière 1996, s.v.) In oral speech, there may be a pause before until and/or an intonation and stress signal. 62 The present tense is not unusual : I am writing by the river-side for a few days (Dickens, 1857). I am writing at the Reform Club until 4 o’clock when I have an ingaygement [sic] with such a charming person (Thackeray, 1849). In some cases, the sentence has iterative meaning, so that it can be said that the time-limits are included in the predication, not just a secondary piece of information : They are paving, repairing, gaslighting, drumming, from morning till night (John Ruskin, 1845). What Ruskin is complaining of is not that they are paving etc., but that they do it all day long. This type of sentence seems to be perfectly acceptable to-day. But perhaps less so the following, also iterative : Father is driving at fine speed from 8 o c till 12 at night (Mary Wordsworth, 1839). Remember that we have no way of checking whether Mary Wordsworth’s sentence was entirely acceptable in her time. But this is one instance of the usual type : What are you doing these days ? Where one may reply : I am idling from morning to night. Obviously, the question of time-boundaries requires special attention when bounded movement predications (class 21) or bounded processes (class 31) are considered. This has been the subject of discussions, especially by Bennett and Partee 1972, Dowty 1977 and Vlach 1981a and b, in attempts to formalize tense and aspect in English along the lines of Montague’s Formal philosophy 1974 , also called Montague grammar. The logical concept of truth-value is not entirely adequate for the task, since for example it led Dowty to assume that there are no progressives of achievement sentences. Again, we will be satisfied with examining the few cases when the whole predication reads like an achievement. Still, I am getting worried a little, because I was all day long getting from Milan to Baveno (John Ruskin). This is definitely a terminative movement predication. It was first tentatively included among progressives in our corpus probably because of the adverbial all day LONG , instead of simply all day. Yet it is at least a borderline case, and for want of the support of stress, pause and intonation signals it would have in oral speech, there is some ground for not considering it as a true progressive. The sense is clear : it took me a whole day to go from Milan to Baveno ; getting from M. to B. took me a whole day . And the reordering : • I was getting from Milan to Baveno all day long would have been, probably, an unacceptable sentence, supposing that the one written by Ruskin was entirely acceptable. We have here one of those intermediate structures which, according to some historians, led to the development of the modern form. 63 From the moment we came till we left them Mr & Mrs Fowler were contriving and executing sthg for our amusement (Edgeworth, 1821) ; On Saturday I was correcting proofs literally from morning till night (Eliot, 1853). The two sentences above are not strictly identical. In the second, the predicate is clearly to be analyzed as correctproof (cf. the nominalization proof-correcting). One could say that the progressive solves the ambiguity of the determinacy of proofs, by giving it a notional, or generic value (Cf. Leech, 1969 : 151). With Maria Edgeworth’s sentence, the situation is more embarrassing, but it is not unreasonable to say that something suggests indeterminacy and leads to a similar interpretation of the predicate as contrive-something, execute-something. There is another girl ...to whom we are giving a few shillings now and then ... until we are ready to take her (Dickens, 1847). Give is a bounded process, but repetition, in spite of the mention of the − somewhat indefinite − sum, tends to iterative-generic meaning, thus allowing use of the progressive (see below IV.5). Your book, I imagine, is being reserved until the right time for advertising school books (Eliot, 1863). Aside from the fact, mentioned above, that until is less of an enclosure than it seems, this passive form may be one of George Eliot’s idiosyncracies. The old form of the passive (*is reserving) was not acceptable ; 86 so instead of using an “ impersonal ” active form : they are reserving your book, she may have used this curious passive as a sort of understatement, although not being herself the publisher who was delaying the issue. Our conclusion, then, must be that the supporters of Montague grammar are right in this respect : explicitely bounded achievement sentences were as difficult to use in 19th century English as at present. Time-limits were possible with unbounded processes only, or iterated or generalized bounded ones. 86 No inventory of the limitations of that older passive form seems to have been made. 64 IV.5.Generalizing progressive A number of approximately 900 quotations (out of almost 21 800) can be ascribed to such uses of the progressive : they have always embarrassed grammarians, who didn’t always distinguish 19th century from present-day usage. The progressive is rightly considered as a truly actualized form, so that its use with adverbs like always for example, seems contradictory. Recently, some logicians or linguists have noted the parallel, in this respect, of verb and noun phrases where the definite article, for example, can assume a generalizing sense : the child is father of the man , a famous line by Wordsworth is typically puzzling. This shifting from particular to general is sufficiently fascinating to justify closer examination. Such generalizing uses are ordinarily presented as secondary, and even Leech 1971, who dedicates to the point two interesting pages, puts them under the heading “ other uses ”. Comrie,1976,37 the author of an enlightening study on Aspect, speaks, like many others, of “ emotive effect ”. Consideration of what happened in 19th century English may help to understand the apparent contradiction. Two contrasting expressions, one actualized the other iterative, of fairly frequent occurrence − but probably more in the oral than in the written language, witness the minutes of murder trials at the Assizes (Notable British Trials Series published by William Hodge & Co.) − are worth mentioning in this respect : in the act of, and in the habit of. (see II.4). I am just in the act of getting done with that thrice wearisome Legendre (Carlyle 1821) ; I am in the habit of taking my papers to Dilke (Keats 1818). It has been found convenient to consider separately the following types : IV.5.1. Extended situation (present or past) : The French people, as usual, are making as much noise as possible about everything that is of no importance (Dickens, 1850) ; He is taking a violent medicine (Dorothy Wordsworth, 1803) ; Tell me that you are sleeping naturally and well (Eliot, 1875). “ In these cases, the concept of “ limited duration ” is applied not to the individual events that make up the series, but to the series as a whole. The meaning is ‘HABIT IN EXISTENCE OVER A LIMITED PERIOD’ ” : Leech 1971 : 27-8. This can also be applied to : I am not going out again (Eliot, 1862) ; I am seeing nobody (Eliot, 1879). George Eliot is not intimating a prohibition as when a politician says to journalists : I am not answering questions 87 she is just saying that she has not resumed her old habit of going out, and the expression could be affirmative as well, since it is constative, not at all performative. A small number of such examples (10) has been found in the corpus. IV.5.2. Iterated situations Not very different from the above are subordinate clauses, usually introduced by when, in the sense of whenever In fact, there is a gradient, leading from the particular to the universal. 87 Incidentally, this well-known use of the progressive in present-day English has not been found in our corpus. It is not far from a performative, and the problem is whether the negation is, in logical terms, within or without the predicate proper. Whereas: I disown you as my friend; I deny you the right to do this, are clearly performative utterances, what about : I do not include you among my friends and I don't allow you to do this ? In: I am not answering questions, the politician above is just setting himself a momentary line of conduct, a law applying only to himself, not altogether very different from the "principle" followed by George Eliot. 65 When you go into her room she is lying generally as if she were in a state of insensibility ( Hutchinson, 1835) ; she looks so pretty when she is playing on the harp ( Edgeworth, 1830) ; some of the pleasantest time is talking with Mrs E. when we are going to bed ( Edgeworth, 1813) ; I seem to lose twenty of my years when we are chatting together (Carlyle, 1823) ; when I am not suffering for vicious beastliness I am the greater part of the week in spirits (Keats, 1817) ; The reference is to iterated individual situations, whatever their number. But the meaning becomes generalized in : I never feel more contemptible than when I am sitting by a good looking coachman ( Keats, 1819) ; Referring to what was said in I.3.2. about the transition from punctual action to behaviour, conduct and character, we realize here that, from the repetition of such situations, Keats comes to an appreciation of himself − he seems to have been a handsome young man, but was at the same time insecure, and suffered from being small-sized. With a generic subject the situation becomes a sort of prototype : I by no means accede to the doctrine that ladies cannot attend to anything else when they are working (Edgeworth, 1805). In Keats’s remark : when a poor devil is drowning it is said he comes thrice to the surface (Keats, 1818), due allowance must be made to the fact that drowning is not an instant process ; but we also find : I cannot but feel that the Right man is always in the Right Place when he is pleasing his Par. or his Mar. (Thackeray, 1855) ; a man, when he is courting a woman (Coleridge, 1819) ; never to interrupt your elders while they are speaking ( Coleridge, 1807). Even in the sense of whenever, when can also be said to provide a generalized time-frame, by iteration of individual situations. Actually, this may correspond to different stages in the process of generalisation. (1) Any time when you are writing, 2s/3 worth of stamps would be most acceptable (Eliot 1879) ; (2) but mind I won’t have you to come in a whirl just when people are coming in and out of the house ( Gaskell 1859) ; (3) remember all his advice etc. when you are practising (Gaskell 1851) ; (4) and when you are sitting in the evening solitary and silent... then imagine...(Ruskin 1836). There is some uncertainty concerning the above examples. Especially in (1), an interpretation is suggested of reference to a single occasion in the (near) future ; (2) may suggest a more general recommendation, but taking its origin from an occasional event. ; (3), since practising refers by definition to repetitive action, leads to a generalized meaning, whereas (4) may definitely allude to regular behaviour : whenever. No other case of reference to a single future act has been found. When normally refers to recurrent facts, which can be seen as habits, since there is only a small step between repeated and habitual, especially when animate subjects, or moods or states of mind are concerned. This is the prevailing use : some 70 cases have been counted, out of 130 when-sentences, that is more than 50 %. 66 IV.5.3. Generalization The next stage − a short step further − is generalization of an act or event considered as typical. Repetition becomes a rule. Some 25 cases have been counted. for it happens often that consumptive people seem to ail nothing when they are moving about in carriages (Dorothy Wordsworth 1808) ; You know Kerry when you are walking with a lady you should always consider...(Edgeworth 1818) ; when any woman has been scolding, beating, or otherwise abusing one of the other sex, ...she is made to ride stang ( Gaskell 1838) ; if men will preach just when turtle soup is coming in, what can they expect...(Macaulay 1831) ; when a man is drawing or working in any way, for other people, he has always a hope that... (Ruskin 1852) ; he asked whether dead people ever dreamed while they were lying in the ground (Dickens 1846) ; It is so much a custom with me to try to understand what people “ think ” when they are talking to me, that... (Dickens 1841) ; if this be so when we are merely putting together words or colours, how much more ( Wm Wordsworth 1805) ;. In : when the blue of heaven is growing dark (Ruskin 1836), the context made it clear that this was a general observation. A good example is : at this moment the SW wind is ...whirling the leaves and blossoms about in a way that reminds me of the tricks it is playing with the surf (Wordsworth, 1823). Obviously, when he writes this, Wordsworth thinks of his experience (reminds), which goes beyond the actual landscape he is looking at. The wind’s way of playing tricks with the surf is, so to say, a common feature of the wind’s nature observed by a poet, animated perhaps, but nevertheless a stable physical law : just as water flows or boils, it is natural for the wind to blow leaves and blossoms and the surf. Instead of a physical law suggested by induction, it can be just a mental phenomenon presented as normal : we believe our dreams to be real when we are dreaming (Coleridge 1816). And then what is worded as a mere indicative can in fact have the force of an aphorism : the Right man is always in the Right Place when he is pleasing his Par or his Mar. (Thackeray 1855). Other examples are on the margin of habit and rule, which is not surprising since generalization is a gradual process. About this point − and possibly about many others − we are confronted with a well-known linguistic option. According to advocates of the conventional approach “ one word one meaning, and if you change the words you alter the meaning ”, the sentences above cannot be strictly identified to the same sentences at the simple form, which, at least in many cases, can be substituted for the progressive. In Wordsworth’s phrase : which reminds me of the tricks it is playing with the surf, we have supplied the “ generic ” meaning, because the context obliges us to do so. But Wordsworth could have written : it /always, sometimes/ plays with the surf. So that the sentence may just be a slip of the pen, or, perhaps, hint at local use of the progressive in the Lake District. To someone used to the variationist approach, this may rather be seen as an instance of the hesitancy characteristic of intermediate historical stages.88 It would indeed be hard to supply an actualized meaning in this case. 88 Or perhaps only of the uncertainties of language in action. 67 But in Coleridge’s we believe our dreams to be real when we are dreaming, it seems quite possible to use the simple form to refer to an habitual activity. Jespersen’s time-frame theory hardly applies here, unless it is taken in a very vague sense. It is probably in such cases that the deep difference between generalization and truly generic value is felt. The statement is not a ready-made rule or law, but rather the conclusion drawn by an empiricist from observation of what happens. At the same time, as usual, the progressive stresses not the fact of dreaming, but the “ involvement of the subject ” and the -ing form is very close to an adjective (cf. adreamed, already obsolete at the time, although a-dreaming was not), especially when intransitive. Nearly the same could be said of the following sentences, some of which were quoted above : I by no means accede to the doctrine that ladies cannot attend to anything else when they are working (Edgeworth 1805) ; Cf. “ occupation ; what one is engrossed in ”, suggested by Bain 1863 (cf. also the equivalents at work, busy). In the following : when a man is doing his business (Coleridge 1812) ; a man, while he is courting a woman (Coleridge 1819) ; like a horse when he is passing the door of his stable (Coleridge 1825) ; the name of a marabout appealed to by a man when the sword is hanging over him will save his life (Eliot 1865), the nature of the generalization is underlined by the indefinite article a/an, which is obtained, by indefinite summation of individuals.89 In logical terms, the set is extensional, not intensional. In other cases, we have such terms as any, people, men, also implying indefiniteness. Eventually, we come to something which is not far from a truly generic meaning : It is the madness and folly of man to disregard his immortal part even while it is entering its eternal state (E.BBrowning 1828) ; and yet how hard, how impossible, to remember or to reason when the heart is breaking ( E.B-Browning 1831). These extended uses of the progressive can readily be explained by an existential process which is not limited to grammar. An act, strictly speaking, is something effected in the present, more or less extended in time according to the type of process. When we say that it is actualized, we are just giving a modern equivalent of the 19th century expression in the act of − where the spatial, inessive, preposition is revealing. Then the act may be repeated, and we have the different stages of iteration. If iteration is more and more frequent, it becomes a habit. The term normally applies to animates, mostly men, but can easily be extended to animals − a cat has habits of her own − and even, by metaphor, to inanimates, provided they can pass by different “ moods ” : the sea, the atmosphere, the seasons for example. So that in the end, the habit becomes a regular attribute : cf. “ smiling Somerset ”, a blushing girl, but often represented by a different adjectival form : talkative, smiley. An occupation which was only momentary when your first engaged into it and which has become habitual can be mentioned on your identity card under the heading “ profession or occupation ” : I used to paint for my pleasure from time to time and now I have become a painter. Repeated practice becomes profession : What is John doing ? may get the reply : he is a musician, not unjustifiably, since occupations may change with time. The different stages have been illustrated by the above developments. They can be summed up as follows : act > habit (conduct) > attribute (character). Those terms are just for commodity’s sake. As already exposed about Bain’s term of occupation, they seldom apply indifferently to men and things : can the wind have a “ habit ” of blowing ? Applied to the physical world, a word like rule is normally used, according to the inductive approach. Notice that we can “ make it a rule ” to abstain from visiting at night; this is what Mrs Gaskell did: I am not visiting in an evening (Gaskell, 1852), but she probably did not mean that she would never alter that principle. So that, even when it becomes so generalized, the progressive is never entirely deprived of its original existential force. 89 In Culioli's terminology, this would be called "extraction-échantillonnage"; cf. Groussier & Rivière 1996, s.v. 68 IV.5.4. Appreciatory expressions Under the subtitle interpretation of what he calls “ aspect total ”, Buyssens 1968 : 135-156 studies a number of binomial sentences where a progressive is used, usually in a subordinate clause, to refer to the event already presented before in the simple form. The verb may be the same, or it may introduce a different idea connected to it, as a sort of commentary on the first. This type of anaphora, already alluded to by Arvid Smith 1917 , Charleston 1955, and Ota 1963, was examined in Arnaud 1973 : 552-559. Adamczewski 1978 gives it a considerable significance. The binomial, contrastive, sentence can take different forms and if we study it in connection with when, it is only because it is the most common and typical conjunction in present-day English for that type of subordination. We know that when (or while ) may encompass various relationships, passing from temporal value at a given point of time to iterative meaning (whenever) and from that to a more abstract one not very far from if and indeed leading to it. This cannot be developed here, and is connected to the drift from simple past to modal preterit, for instance. When a traveller of the Romantic Age was speaking of going to Rome, it was almost customary to say : When you go to Rome, you will see /you will be seeing/ the Pope, or : If you go to Rome... in which case if does not imply absolute uncertainty about the journey. Here are some typical examples borrowed from Buyssens 1968 : When Elizabeth put Ballard to death, she was not persecuting (Macaulay, in Buyssens 1968 :136) ; In a sense, when one of you votes Labour, you will also be voting Democrat ; and when one of you votes Conservative, you will also be voting Republican (The Listener 17/11/1949, in Buyssens 1968 : 140). Cases like the above, when the event reported is neither serial −although the second suggests it − nor modalized, are exceptional ; but we find : If I should go to one of the tea-parties in a dressing-gown and slippers, I should be insulting society, and eating peas with my knife ( Thackeray, quoted by Jespersen, 1931 : IV, 287 ; cf. also Ljung, 1980 : 70). If, especially when the situation is seen as somehow typical, can be substituted for when, as we see here, and the assertion becomes hypothetical. This seems to be essentially a dialectical or simply dialogic or even rhetorical formula, and it is certainly more common in the written than in the spoken language. At the same time, that sort of equivalence relationship of simple and progressive forms deserves attention in a broader perspective. Adamczewski 1978 believes that it illustrates the basic value of the progressive as “ anaphoric ” in a wide sense : what was (simple form) an instance of doing becomes an instance of saying. The utterer (énonciateur) has passed the predication through a “ filter ” and it is, so to say, modalized. In other terms, the action of the subject is explicitely qualified by the speaker instead of just reported. Similar views are expressed by Ljung 1980 :69-96 in a very enlightening chapter under the title The interpretative progressive. This reminds us of the discrimination of “ narrative-history ” and “ discourse ” which is the basic difference between simple past and “ passé composé ” in French, according to Benveniste 1966, 238 sq. and passim. The schematic remarks below are just elements of what would justify long developments. In the light of what was said in I.2. this interpretative function should not surprise us. By qualifying the subject of the sentence for the interval of time of his act (or condition) the progressive opens the way to judgment and appreciation of the actor. 90 It does not report an act (or merely an event) but describes a behaviour, which can be estimated, hence the label “ appreciatory ” used here. Apart from possible metaphors, there seem to be only animatehuman subjects for that type of sentence, and declarative verbs have been quite rightly pinpointed by Buyssens 1968 : 90 As already said in I.3.2, in theory, people and not acts can be subject to appreciation or judgement. However, for various reasons which could justify a chapter on ethical pragmatics, the judgement is often presented as bearing on the act itself. By underlining the existential involvement of the actor-subject, the progressive has a modal connotation, evokes the possibility of a judgement, which can be positive or negative according to the situation (objective and subjective). In the normally dialogic situation of speech (obvious in the case of letter-writing), it is left to the addressees to give the interpretation they like. 69 I am not reciting these facts for the purpose of recrimination... I recite them in order to explain why it was that we did not have...between twelve and fourteen divisions fighting...(W.Churchill, in Buyssens 1968 :147) ; I’m not saying that it isn’t attractive ; I merely say that if you were my wife I should be in a perpetual state of agitation (Woman’s Own 29/7/1954, in Buyssens 1968 : 147). The negations in those latter examples show the argumentative significance, the adversative force of the binomial structure. It has often been said that the progressive adds emphasis to a statement. This is probably due to its being firmly anchored in the situation of speech, to the existential force of be. In our corpus, some twenty sentences can be quoted. In most of them, the first part of the binomial formula is of the form of in -ing, or by -ing : He used always to tell me that in giving me a good education he was leaving me the greatest good (Welsh 1820) ; In coming down to Scotland therefore Terry would be leaving a position in which... (Scott 1828) ; Perhaps I am doing a superfluous thing in writing all this to you (Eliot 1859) ; you would be doing something towards this by having a carefully written review of a single book (Eliot 1858) ; By giving occasion for more of this frivolous /.../ kind of comment/.../ I should be stepping out of my proper function and acting for what I think an evil result (Eliot 1876) ; But if you are acting right, I should be acting right in imitating you (Coleridge 1795). Incidentally, the proximity of some of the -ing forms above to ordinary progressives is certainly significant and supports the approach suggested here. But we also find other forms of subordination such as, for example : I should be ridiculously sinning against my own Law while I was propounding... (Coleridge 1804) ; At the same time I should be dealing disingenuously with you were I not to tell you... (Mary Wordsworth 1821) ; I should only be deceiving myself and misleading you, were I to encourage a hope... (Wordsworth 1838) ; There is only one case when a plain indicative was found : I am not using mere phrases when I say that (Eliot 1858). IV. 6. State verbs In the coding frame, predications of classes 1 to 5 have been called state verbs. Two points must be kept in mind about this : 1. the coding matrix is a research frame : the labels for specific verbs or predications can always be revised according to the observations they have provisionally supported. The quantitative approach assumes that individual corrections of limited scope will not alter the comprehensive figures and overall conclusions ; 2. in the case of state verbs, seldom used with the progressive (prospective verbs excepted), it has already been said that all occurrences were noted, to permit further study. This is an old debate. Scheffer 1975, 61-4, lists the verbs which have been considered as state verbs by 10 linguists : they are seldom unanimous, and often based on intuitions. 91 He is not the first to have noticed that such verbs are sometimes found in the progressive. When this happens, it is left to the researcher to decide how to account for it with reference to the basic interpretation of the verb form. The usual solution is to say that the predication is not really stative, but refers to an activity of the subject. It has been observed that the subjects are usually human subjects, or metaphorically so. 91 Some have tried to find stativity criteria and even tests. Remarkable in this respect is Sag 1973, who suggested a "progressive squish" (in the line of J.R.Ross) i.e. a matrix correlating degrees of stativity to a scale of less to more receptive environments, among which "futurate progressive", "process progressive", "habitual progressive" and perfect progressive. A comprehensive list is also given in R.L. Allen 1966, Appendix A. 70 Considering that many grammarians are also teachers, it is neither surprising nor reprehensible that they should be looking for an explanation which will be easy to sell to students of the different grades. A fairly common way of doing it is to look for equivalents : in have a meal or have a look, for example, it has been suggested that take can be be a substitute for have, since it is clearly an activity verb. This may be a device in practical teaching, but fails to be as satisfactory for the theoretical linguist. We all have an intuitive notion of activity ; but yet we are often at pains to justify such a term. 92 It is well-known that Aristotle’s concept of time was derived from what has been translated by “ movement ” ; yet it seems that for the Greek philosopher what was meant by movement was fundamentally change [ metaboli ] not yet kinesis : Physics IV, 11 (Cf. Collobert 1994 : 58 ). It is only at a second stage that the notion of continuity, especially in the succession of “ transport ” [pheromenon ] IV, 11,219b , leads to the numbering or measurement of time, parallel to the measurement of linear space, and in connection with speed. But the way time is conceptualized in the system of aspect, at least in English, is better illustrated by Eleatic philosophers. Things will probably become more clear if we avoid the concepts of duration, speed, and terms like dynamic (versus static) in the study of English aspectual forms. A term like “ dynamic ” suggests input of energy, and this has often been seen as an essential feature of the verb as opposed to the noun’s “ inertia ”. Discussing this would mean a long debate, and I will give, briefly, only my own views about it (see also I.2). If, when accounting for the use of the progressive, we want to cover predications as different as walking and lying (not lying down, of course !), working and sleeping or idling, talking and believing or (more usual) thinking or hoping, we will be happier if we dismiss altogether the notion of activity to consider only the successive “ situations, “ occupations ” or “ happenings ”, “ goings-on ” (Leech, 1971 : 14, is not the first to find it difficult to select a covering term) affecting a subject whether animate or not, human or not. To sum up, we are dealing with temporary, contingent states. In another respect, and referring again to Aristotle’s discussions, it would seem that the idea of “ aspect ” epistemologically antedates that of time93, in the sense that it does not consider the measurement which normally underlies the concept of duration : “ normally ”, since duration may merely imply just an interval, meaning that the event takes time, between a beginning and an end, remote from each other or not ; but the measurement of the length of time (be it psychological time) or the mere consideration of length are irrelevant. This is why the term “ progressive ” is inadequate and may be responsible for the above misconception. Referring again to the Eleatic vision of time, we might say that “ events ” progress in the way Achilles does when pursuing his tortoise, i.e. there is a successivity of intervals of space-time, external to each other. Historians of scientific thought have shown that it is only later that a mathematical revolution will solve Zeno’s famous paradox and introduce infinitesimal numbering. To speak less philosophically, we will say that “ progress ” is all right if by this we mean passing from one state to another, whatever the length of time between the two. Change is a better term than movement to name this. It would seem that what is obvious for processes has been extended to all verb predications, following the old fallacy that the verb is dedicated to actions. It should be stressed now and again that verbs can be used for many “ events ” which do not correspond to activities, unless we are animists and think that shine, bloom, sound, etc. are activities of inanimates, as well as wait, lounge or fast for animates. 92 Quirk & al. 1985 are apparently the first to have observed that lie, sit,etc. cannot be called activities, but are static postures; they introduce the label of "stance". However, this looks like a sort of evasion. Arnaud 1973 suggested to call "relative verbs" all the verbs of classes 1 to 6, with reference to the cognitive distinction between operations and processes on the one hand, and relations on the other. The relation (or connection) with space, when no movement or change ut nunc is implied, would then be characteristic of locative predications. 93 Whether this is historically true is still a debated question. 71 Grammarians as old as the Alexandrians have repeated that the verb is an adjective plus the verb to be (in its existential sense, should we say now 94 ), so that it basically refers to what “ takes place ” at a given moment, thus introducing the notion of successivity before that of progression. At the same time, we know that the brothers Lumière have given us the illusion of movement by mere succession of static shots at the requisite speed. The short-term diachronic study of state verbs in connection with progressive use may help in discriminating them and suggesting some of the reasons why they accept or resist the progressive. State predications, or rather state verbs in this particular case, have been grouped in classes 1 to 4 of the coding frame, according to semantic distinctions which are not really tight but can be helpful for practical purposes. The overall evolution for those classes is clearly shown in Figure 19, a cumulated histogram where the contribution of the subclasses by periods can be estimated. Considering the small figures for the individual subclasses, it may be difficult to draw clear conclusions from a detailed review. It seems, however, that class 1 (relational-statal) plays a dominant role in the progression, and among such verbs, that have is largely responsible. This justifies looking at this verb a little more closely. HAVE 2 occurrences of have to can be set apart from the 104 cases. The 102 remaining have been distributed into the following empirical categories : 1.Causatives (undoubtedly the most clear-cut category) : I am having engraved on a brass plate to be fixed up over the street door... the following words (Dickens, 1846) : 13, including 2 ambiguous cases ; 2.“ Experience ” : 2.a. Weather (or climate) : I hope that you are having as much sunshine (Eliot, 1876) : 27 occurrences. 2.b : “ Time-span ”, “ period ” (good or bad) : we have been having much musical pleasure of late (Eliot : 1878,) : 12 occurrences. 2.c : Health : She has been having the old headaches (Gaskell, 1864) : 10 occurrences 2.d : Food : Henry and Daisy had been running and having luncheon on the hills (E.B-Browning 1831): 6 occurrences. The remaining cases can less easily be grouped under these categories, except by metaphor or extension of the sense : cf. I like to think that you are having all sorts of delices (Eliot , 1843) : food ? good time ? ; do you write to Nice and say what a great suxess [sic] I am having ? (good time ?) (Thackeray, 1856). But what about : we shall be having a Box of Tea from Mr. Twining’s very soon (Mary Wordsworth, 1815) ? Indeed, the attempt to find a satisfactory cover-term seems artificial and unrewarding. The term “ experience ” itself cover all instances of use of the progressive. The progressive is the form that describes best human experience or the condition of things at given moments, what is “ going-on ” : this includes having or being, in the momentary, transitory sense. Internal (be) or external (have) property is usually permanent, and this may be the principal reason for the rarity of the progressive with the two verbs ; whereas an activity or a posture is not always habitual. But there is no intrinsic reason to exclude the progressive in such cases. The list above shows, however, that the properties in question are usually not physical objects, but something a little more abstract, implying interiorisation : the famous “ subjective sense ” of the progressive. 94 We find it still in Destutt de Tracy 1825 : "Un verbe n'est autre chose qu'un adjectif uni à l'adjectif étant, qu'un adjectif renfermant l'idée d'existence, et par cela même pouvant avoir des modes et des temps". 72 This being said, the authors display uneven tendencies with this verb : the champion is undoubtedly George Eliot, this casting perhaps some more light on her exceptionally sensitive nature. Figure 29 has been established by decades. It shows that the curve is resolutely rising by the forties. The contribution of George Eliot is obviously considerable, since, for decade 8 (1865-1880), she supplies 38 of the 45 occurrences. But the slope of the curve is not affected if we subtract her contribution. For decades 5 to 8, the other major contributors are Mrs Gaskell (17) and Thackeray (11). One cannot make too much of such small figures, yet we notice that two women are ahead of the group ; and also that Thackeray (who is absent from the last period since he had died earlier) displayed a clearly “ modernist ” tendency on this account. Perhas a more extensive survey would confirm the trends the present one merely suggests. IV.7. Prospective/retrospective predicates (Figure 20). Under that label have been collected a hybrid numer of predicates with the following two characteristics : 1.They refer to mental states or conditions, “ private ”, non-overt “ activities ” − to repeat the disputable term of several grammarians : “ attitudes ” seems more justifiable. This is why they have commonly been included among state verbs. 2.They are supposed to have a certain “ dynamic ” force, since they are oriented to a prospect ; but − and in this respect they differ from process verbs − what they are aiming at, their goal, is not a result in the sense of being a completion or achievement of the “ activity ” they represent (see 1.3.3.a.). It is difficult to use completely unambiguous terms in this case: prospect seems less ambiguous than goal. In terms of “ truth value ”, of “ being the case that ”, what such predications are aiming at is the realization of another event than the mental event they designate. What obscures it is that this outcome is not always represented by a verb-phrase but fairly often by a noun-phrase. In terms of boundaries, this means that the right-side boundary is open, since it opens the following interval rather than closing the previous one. This is why prospective verbs are at the same time analogous to locative verbs and different from them. When you are sitting you are not expecting to stand up or to lie down, although you know that in the natural course of events you will certainly have to. The same will be seen if we refer to unbounded movement verbs : a horse may canter and then run, but running is not the prospect of his canter. A verb like wait is significant in this respect. Etymologically, it is connected to watch, and does not imply the directionality which becomes explicit in wait for, and has also become associated with the verb used alone. It could have been included among locatives, were it not that the prospective sense is now the most usual. So that prospective verbs are a class of their own, for research purposes at least. The mental states they normally refer to are dynamic, oriented towards a future event, for the realization of which they are not responsible. By definition also, these states are transitory, and could very well be expressed by temporary adjectives. When a mother is “ expecting a baby ” she is said to be “ expectant ” or “ pregnant ”, and when a hen is “ hatching ”, the verbal status of the term can be disputed. The inessive force of the -ing form had often been underlined, and was still sometimes represented in 19th century English by the prefix a-. Just as one can be asleep instead of sleeping, one can be awaiting instead of just waiting, to mention two of the cases when the inessive preposition-prefix has survived. Another point to be underlined concerning these predicates, is that they accept only animate subjects, essentially human ones, except, as usual, by metaphor : some supper, which was waiting for us there (Dickens, 1845). Also, they are what Ota 1964 calls “ I-verbs ”, meaning they are mostly found at the first or second persons, a fact which is connected with the previous remark. Even mere intuition − such as we find in Hatcher’s article of 1951 − suggests that the class is not homogeneous with respect to the use of the progressive. We are quite familiarized with it with expect, look for, plan, and wait (for) or even watch (for) ; but verbs like want, wish, have always been embarrassing, and think (of) alluding to a project is one that raises the question of the “ activity ” component of verbs which Vendler has classified among state verbs. In a sense, we accept more readily that sleep, suffer or even die may be seen as active than such verbs of purely mental reference. • It is not possible here to try to solve the problems. Perhaps, if we accept that the activity of language, even when it does not look like it, is always dialogic (which is obvious in the case of letter-writing), considerations of “ observability ” or “ evidence ” suggested by Ljung 1980 will help us greatly. For such “ activities of the mind ”, the speaker (or the letter-writer here) is the only warrant : they are subjective, as we say, and the dichotomy between 73 “ referring and non-referring predicates ” which, according to Russell, Searle and Ljung, applies to performatives, may also be applied in the case of some prospectives. In this respect, promise would be a good introduction, since it is a well-known performative, but one where the truth-value or “ being the case ” of the “ thing ” promised is deferred. So that : • • *I am promising to write soon is not acceptable, whereas : He has been promising to write soon, She is promising to write soon, though not very natural perhaps, are more acceptable. Wish can also be a performative : I wish you luck. Yet we find : I am wishing Margaret many happy returns of the day (G. Eliot, 1879). where the performative force can hardly be disputed : another example of the audacity in George Eliot’s use of the progressive. One cannot speak of understatement in this case, but rather of overstatement. This emphatic use of the progressive has often been noted, and we have here a typical example. One way of accounting for it is to say that by using a constative rather than a performative utterance, George Eliot is less authoritative : instead of the usual, non-referring speech-act relying on her social right to utter a wish, she actually refers to it, objectifies it, just as a wishing-bone is used to give some magic strength to a promise of marrying soon. This is a sophisticated justification for a statement of everyday life, where wishing does not cost much, but this might be the path leading to the extension of the progressive to subjective verbs, among which prospective predicates are included. This also might help us to account for the degrees of acceptability of the progressive between, say, expect and hope. Expect is just a factual statement, whereas hope implies stronger affective involvement. In the dialectics of interpersonal influence or even authority, they will have a different status. A person in authority, a father for example, can say : 1.I expect you to pass your exam ; 2.I hope you will pass. Which of the two is more likely to be taken as a demand ? That prospective verbs should be included among “ state verbs ”, hence not easily found with the progressive, is contradicted by their comparative frequency in our corpus : some 1330 occurrences have been counted, for the list of verbs below (in alphabetical order) : aim, anticipate, apprehend, await,cast about, clamour for, contemplate, crave after, desire, die (for, to), dread, drive (special), dream, entertain, expect,fear ; forget, frame wish, gape for, gird off for, groan for, hanker, hatch, hope, hunger (for) hunt (for), intend, itch, languish, long (for), look (for, forward, to, mean, meditate, muse, need, pant (for), pine (for), plan, project, promise, purpose,rave (for), raven (for), scheme (for), search, seek,sigh (after, for) starve (for), strive (after, for) tend (to), think (of), thirst (for), threaten, tire (for), try (for) vow, wait,want,watch, weary (for, to), wish, yearn (after, for). As already said, the group is not homogeneous. Tentatively, the following subgroups can be suggested (the number of occurrences is between brackets) : 1.Expectation : besides expect (96), wait (279), − and await (23) −, is the most typical of “ objective ” orientation to a future happening, with comparatively small affective involvement. Such predications could easily be assimilated 74 to mere locatives, although the latter are more neutral towards the inevitable end of the situation they describe. Watch (for)(1), anticipate (6) can also be included here. 2.Plan : in common terms, planning is intellectual anticipation, implying some activity, so that the following verbs might have been included among process verbs, with e.g. prepare. The subgroup would include : aim (9), cast about (4), contemplate (6), intend (29), plan ( 21), project (4), promise (1), purpose (13), scheme (for) (8), search (for) (2), seek (20), tend (to) (5), think (of) (86), try (for) (1). Clearly, such intellectual anticipations may also imply more or less positive or negative involvement, since, generally speaking, consideration of the future seldom leaves people entirely indifferent, hence the well-known link of futurity and modality. In this respect promise could be seen in some contexts as the opposite of threaten which is included in the next group, due to its intrinsic semantic content. 3.Emotion : apprehend (1), clamour for (1), contemplate (6), crave after (1), desire (4), die (for, to) ( (14), dread (2), drive (1), dream (5), entertain (1), fear (6) ; frame wish (1), gape for (1), gast for (2), gird off for (1), groan for (1), hanker (after) (3), hatch (1), hope (63), hunger (for) (3), hunt (for) (9), intend (19), itch (1), languish (1), long (for) (66), look (for (134), look forward to ( 52), look (on) to (9), mean (25), meditate (57), muse (1), need (6), pant (for) (3), pine (for) (9), rave(n) (for) (1) , raven (for) (1), sigh (after) (for) (3) starve (for) (1), strive (after, for) (4) tend (to) (5),thirst (for) (3), threaten (6), tire (for) (2), try (for) (1), vow (1), want (56), weary (for, to) (17), wish (39), yearn (after, for) (8). With different degrees, these verbs imply valuation of the expected event, and affective involvement, positive (hope) or negative (dread, threaten). Between the subgroups, the limits are not strict, and must be appreciated in context. But, generally speaking, the distinctions may lead to different meanings of the progressive with reference to human subjects, which is normally the case. With subclass 1, the quite natural use of the progressive is equivalent to what we find in locative predications, with the important difference that these prospective ones are transitive and that their “ object ” − what is anticipated − lies in the (usually near) future ; a future which is not part of the meaning of locatives. Subclass 2 stresses activity − usually mental − and it can be assimilated to ordinary processes, with the major reservation (see above) that the outcome is not a result of that activity in the usual sense. If we consider a plan as the result of planning it is only a transitory one95, not the real future objective, and if it can be said that the “ result ” of promising is a promise, this is only valid when it constitutes a performative utterance (present tense, first person) : in all other contexts, it refers to a future accomplishment relying on the word of the subject, normally human. We can see that subclass 3 is totally different, even though there is a certain amount of interference. In terms of the number of verbs, it is the richest with our romantic writers, probably because it embodies degrees of emotional involvement and suggests stylistic effects. This involvement will be stressed by using the progressive. At the same time the attention is shifting from the expected event to the expectant people and their feelings. The frequency of the progressive with subclass 3 deserves special consideration. Two of these verbs have always attracted the attention, above all of prescriptive grammarians. Want is a case in point, which was usually ill-treated, its use in the progressive being considered as a scotticism − an instance of the common propensity of taking foreigners for scape-goats. 46 cases, presented below, were found, of which one or two can be taken as meaning lack, with possibly adjectival status. Among the people responsible we find such notoriously Scottish people as Jane Austen, Thackeray, George Eliot and Mrs Gaskell !... To tell the truth, we find it also in the letters of Walter Scott, Carlyle and his wife, but far less than in Mrs Gaskell’s. Remembering that Bain, who censored such scotticisms in his Higher English Grammar of 1863, was himself a Scotsman, born at Aberdeen in a rather low class of society, we will refer this to the well-known self-criticism and social insecurity of educated people in everyday contact with a local dialect considered as substandard. Sociolinguists of the Labovian persuasion will describe such cases as stereotypes, the surface phenomena on which people focus their critical eye for prescription and proscription. 95 An interesting case in the typology of "objects", a very deceitful term: see for example, Jespersen 1931, III:232 sq.). Plan would be the cognate object of planning. 75 Wish was probably less stigmatized. Yet it is also supposed not to accept the progressive easily : 39 cases have been found with a variety of letter-writers. In the following, we may easily admit that it is the outward expression which is described, so that it becomes an overt activity similar to asking, or calling : Mr Bowles has been wishing to take some extracts from Field’s Church (Coleridge 1815) ; Anny is always wishing for you (Thackeray 1848) ; Mamma has been wishing for a wind to blow away the cholera (Ruskin 1832). Ruskin was only 13 when he wrote this. But the majority of cases definitely refer to the usual psychological meaning, e.g. : Mr. J.W. frightens me. He will have you. I see you at the altar... He must be wishing to attach you (.Austen 1817) ; I have been wishing for your advice (Macaulay 1842) ; I...am always wishing I was at home in New York (Thackeray 1853). Thinking No verb can better refer to mental activity than think, and it has been suggested in I.2. that Descartes’ cogito could be translated by a progressive in modern English. In our corpus, many instances of think or think of and even think to have been found with prospective force : the 86 examples mentioned above. Discussing the point at large would carry us too far, since “ thinking ” is of great complexity — a commonplace: 1.In terms of the time-sequence − the basic dimension of the progressive − it is a Janus-concept. The speaker’s “ thoughts ” can be directed, besides the actual present, backward in the direction of the past (sometimes with postposed back), or forward in the direction of the future. But in all cases it can be treated as agentive mental activity : witness Rodin’s famous Thinker, whose static posture nevertheless suggests that “ something is going on inside ”, invisibly (Cf. Martin Joos’s notion of “ private verbs ”). We will leave to psychologists the task of accounting for this activity of the mind which the gerundial form “ thinking ” actually suggests : as we know, the neurological processes it implies can be measured by encephalography. The common representation of thinking as an activity makes it all the easier for this verb to be used with the progressive. 2.In a theory of speech-acts, it may also be ambiguously interpreted in the two ways suggested by Ljung 1980, 45 sq. and passim, either as a quasi-performative, a non-referring predicate, or as a “ constative ” (expressive) predicate referring to what is passing within the mind of the speaker. By using the progressive, the speaker may wish to give his statement a less authoritative effect, something one had to guess when the simple form was the one normally available. I think is a (mild) performative, since it “ juridically ” forces you to take the assertion seriously −even if you know that people sometimes lie − ; I am thinking openly describes a “ state of mind ” with no outward manifestation except what the speaker says, but in such a way that you are not obliged, whatever your confidence in her/him, to accept what she/he thinks as definitely reflecting the judgement it introduces. This is of course valid at the first person and when followed by a proposition, but may also help to understand the many facets of other cases. Examples illustrating this are presented below in IV.9 on that-clauses (object clauses). It becomes obvious that it is only in this sort of utterances that think can be included among state verbs, in the subclass of declaratives. 76 IV.8. Modal auxiliaries It seemed reasonable to expect the use of the progressive with modal auxiliaries to spread. For the whole corpus, the number (831) is less than 4 % of the total. Broken into the different classes of modals, 96 the figures are too small to give more than indications: Figures 30 to 33. There seems to be some progression affecting all modals, with the possible exception of must, but, compared to the general curve (Figure 31), the curve appears slightly less sloping. The graphs give an idea of how each modal (or pair) is affected. They have been computed by decades to provide a more legible picture. This point certainly ought to be studied further in connection with semantic variations in a domain notoriously subtle. Only a list of quotations is given below, by individual authors and in chronological order, with a few brief comments. Dorothy WORDSWORTH 1787 but I shall be making you as melancholy as myself 1793 but enough, he is my Brother, why should I describe him ? I shall be launching again into panegyric 1795 it will place me in such a situation that I shall be doing something 1796 It was singular enough that they should be assisting at a fire 1811 should you be going away before the end of next week (...) we set off, ... 1801 We then concluded that you must have been coming yesterday 1805 but I find that he must have been standing there pleading to be admitted, for at least ten minutes 1815 for people who might be relied upon for the journey must be constantly coming from Paris 1820 How you must be enjoying yourselves on this beautiful day ! Below, a few other pioneering attempts. The above would not be exceptional today, but they certainly were at the time, stressing as they do Dorothys’emotional involvement. 1800 or perhaps Harriet may be writing to Elizabeth The context makes it clear that this is a suggestion, not a guess. 1822 but we are afraid you will be getting married Notice how this all-important event in Dorothy’s world as well as in her contemporary Jane Austen’s affecting the subject − a girl − in submission to a course of events already ordained. 96 97 is seen as Putting together shall and will, may and might, can and could is not really justified for all cases, but may be valid for a rough estimate of the development. 97 Cf.her letter of 1817 to Cassandra : Mr. J.W. frightens me. He will have you. I see you at the altar. I know it must be so. He must be wishing to attach you. 77 1817 I may have heard these particulars from someone who had visited her, and I must still be turning to this subject It is unusual, even today, to use this form to express compulsion or necessity. .I ...hope that ... it [your illness] may be taking a resolution to leave you altogether. The expression, stressing immediacy, is quite normal today, but was rare at the time. Mary WORDSWORTH 1821 At the same time I should be dealing disingenuously with you were I not to tell you... A clear case of appreciative-adversative comment (see IV.5.4.). SOUTHEY 1807 Your nameless daughter must be now advancing towards weeks of kissability and your raw hands must be getting more useful every day They are so interwoven ...that I must either relate all myself − or be continually stopping to explain and refer the reader to... These are quoted only because they seem to be among the earliest examples. In the latter two, must clearly expresses necessity. William WORDSWORTH 1818 Either they must surpass us in zeal, ... or some other powers must be acting on their side Necessity/obligation again, not probability : the context makes it clear. 1830 Attend particularly to John’s pronunciation at Church – his unaty, his charaty, his inexpressable, and the Northernisms that must unavoidably be creeping upon him. 1834 Should I be silent upon the part you have taken as the Public Leader of the 62 or 63 Petitioners, I should not be treating you with sincerity. 1838 I should only be deceiving myself and misleading you, were I to encourage a hope Two cases of appreciatory binomial. COLERIDGE 1795 But if you are acting right, I should be acting right in imitating you. 1804 I should be ridiculously sinning against my own Law while I was propounding Binomial appreciatory sentences again. 78 DICKENS 1826 Her answer was... so that she could be working on one while the other was drying. I am inclined to think you must have been anathematizing me pretty considerably, for not keeping my promise. Early examples of something not unusual today. THACKERAY 1836 Thou knowest well as I, Beowulf, we must be hammering at the head of the spear Again a must of obligation. 1849 I shall be forgetting my own text 1850 and until I had got up my courage at 12 o’clock when I thought the Xning could be taking place and said Amen there IV.9.That-clauses Subordinate clauses introduced by THAT or WHETHER, etc., and other such patterns with a declarative or cognitive verb in the main clause, are usually called object-clauses . The two types of verbs can be treated separately. IV.9.1. Declarative verbs : A declarative verb is often used in an utterance which is not very different from a performative. Some transformationalists (McCawley, for example) even suggested that a performative ought to be introduced at the early stage of derivation of all sentences, even when it does not appear on the surface. What interests us here is the occurrence of the progressive on the left side of a final pattern such as : I declare /am declaring /that Sir Robert is a fool The theory of performatives, beginning with Austin himself, has underlined that a true performative is in the present indicative, first person. It has also been remarked that it excludes the progressive. There is an interesting discussion of this point in Ljung 1980 : 97 sq. According to him, declarative verbs, which he calls illocutionary, with reference to Austin’s famous analyses (Joos 1964 calls them asseverative), usually enter in non-referring predicates. Ljung also alludes to Bertrand Russell’s discussions, anticipating even Austin’s. Although Russell and logicians of the post-kantian tradition make use of the concept of truth-value, which may be regarded with some diffidence by linguists, the views expressed by Ljung are enlightening. Illocutionary utterances, he says, in the same way as performative utterances “ do not refer because they are used to express a relation between the speaker’s mind and a proposition...I suggest that we refer to such uses of predicates as expressions ” 1980 : 104. In Russell’s terms, such assertions have no truth-value. Ljung goes on to study a certain number of cases, taken from present-day usage, and encounters situations when a progressive can be found in utterances which, on first inspection, may appear as similar. The consideration of tense and person is essential, just as with performatives, and, according again to Ljung : 1.In some cases, declarative verbs are used in a “ reportive ” way, referring to events : this is normally what happens with other persons than the first, other tenses than the present. 2.In other cases, they are “ expressive ” of the relationship of the speaker and his predication : this is normally the case with the first person in the present tense. But we would then have to account for statements of the following type : ? I am suggesting that we leave. 79 Though unusual, it is not unacceptable. It means that the speaker is actually reporting on his own speech-act, just as a police constable might possibly say : I am arresting you (see above I.3.2), probably by some sort of understatement. The above remarks will help us in the following brief survey of that-clauses in 19th century English. you will understand that I am stating what I know (Dickens 1847) ; if I come to Italy en garçon (as I am already swearing I will) (Dickens 1847). Obviously, Dickens is not uttering an oath, but reporting it. This is very similar to the sentences studied above in IV.5. 3 on generic use and when clauses. These are the only two quotations (and the first is not a that-clause) in our corpus when the first person present is used, to quote Ljung 1980, 106, in a “ split ego ” situation, where the speaker refers... but to himself and his own speech-act. The following is more difficult to accept : I am presuming that the fact of the 2 vols. having been lettered “ The Works of G E ”...( Eliot 1862). The only justification seems to be, just as for the fictional constable alluded to, that this is an understatement. By presenting her statement − the illocutionary force of which is somehow dubious, but which is definitely to include among declarative verbs − as a constative referential predication, she wants to soften what might be taken, by the publisher she is addressing, as a reproof. The other that-clauses are more clearly of the constative (“ reportive ”) type. While Opposition is asserting that the Catholic Association is eminently beneficial, Mr Canning as positively asserts that it is a pest (Wordsworth 1825). This sentence illustrates the adversative-interpretative meaning we have already noticed. At a superficial level this is similar to a “ time-frame ” effect, but in fact the relationship is of a dialectical type. Which of the two assertions is most emphasized is however difficult to tell : the emphasis is rather in the contrast between the two. The following can easily be seen, due to the use of past tenses, as merely narrative-descriptive. I was only begging and entreating that some one of them should keep watch (Coleridge 1832) ; I was confessing the other day that I felt myself (...) growing credulous of joy (E. B.- Browning 1841) ; I was laughing and maintaining that mine shd be as like as his [portrait] (E.B.- Browning 1846) ; I hope that you had been conjecturing that there must have been some mistake to account for my silence to your note ( Eliot 1863) ; Henry and I were agreeing yesterday that... (Eliot 1880) ; I was remarking that here one was allowed to be anything one liked except a fool, or a bore − to which he answered... (Welsh 1841) ; A lady was suggesting to me the other day that...(Welsh 1837) ; I had just been propounding to Forster if it is not a wonderful testimony to the homely force of truth, that one of the most popular books on earth has nothing in it to make anyone laugh or cry [Robinson Crusoe] ( Dickens 1856) ; a certain young Irishman has been insisting occasionally these two years that I...must absolutely come out and pay him a visit (Carlyle 1838) ; Twice, I think, just after I had been declaring that he was evidently getting better, he came upon us next night (Carlyle 1834) ; I had been holding forth and asserting that I thought I could exercise the influence (Dickens 1842) ; and Mrs. Boyd had been confessing that she had lost all her taste for poetry (E. B.Browning 1831) ; My pen has been declaring itself independent this half hour (Scott 1829). The following : 80 ...saying that Ld Lonsdale has been naming that next week was the time that...(Mary Wordsworth 1814), is the transposition of a clearly performative (even official) speech-act : reported speech. As to the examples below, they seem to concern not the mental facts (which Joos 1964 called “ private ”), but their outward expression. Searle 1969 and Ljung 1980 suggest discussion of the various types of such declarative predicates in terms of “ sincerity conditions ”. Booksellers are every year finding it more and more their interest to sell an increased number of copies at a low price rather than... (Wordsworth 1838) ; everbody at Chawton will be hoping and predicting that they cannot be good for anything [gloves] (Austen 1813) ; I hope that you had been conjecturing that there must have been some mistake to account for my silence to your note (Eliot 1863) ; when we were regretting to Ld Sligo that we had missed seeing so many persons on our tour...(Edgeworth 1834) ; Papa was regretting the other day that he did not question him (E. B.-Browning 1831). This is perhaps too simple a way of accounting for the transposition from “ private ” thought − when the subject is the only warrant of the truth-value of his statement − to public expression. If we see the question in terms of truthvalue, we can say that the person reporting about it is actually endorsing that truth-value. Prospective verbs, too, are often “ private ” verbs, corresponding to states of mind ; but they may also be declarative, in which case the same principle as above applies : when the tense is not present and first person, the statement becomes constative : everybody at Chawton will be hoping and predicting that they cannot be good for anything else (Austen 1813). Think As exposed in IV.7. on prospective predications, think is a typical case of multivalence. Its presence in the headclause introducing an “ object ” clause (usually with that) is the clearest circumstance when it can be included among state predications. It can be considered as a quasi-performative and the syntactical and pragmatic constraints deserve to be studied further for the period surveyed here. Only a few glimpses are given here. Just as for classical performatives, the typical context of declarative force is the present tense, first person. Consequently, many of the quotations below would probably be found odd in present-day English of the controlled type. and if summer were come I am thinking it will be well to secure quarters about Kirkcaldy (Carlyle 1821 ) ; Things are better ordered considerably as they are − I’m thinking (Welsh 1826) ; We are thinking that if the Kitchen grate is still to set, it might be as well to...(Carlyle 1828). The three quotations above are from the Carlyles. They seem to be clear cases of scotticism − notwithstanding our reservations about this label ; at the same time the declarative force is undisputable. and I am thinking with delight that you will learn something of Mrs. Burne-Jones (Eliot 1870) ; I am writing a book and thinking that I will never finish it (Eliot 1871) ; I am thinking whether it would not be wise to retire from the world and live here (Eliot 1852) ; But I am thinking that I may well be dédommagée by going back with Mrs. Boyd (E. B.-Browning 1831). In the above, despite the presence of that, we may accept that the utterance is rather of the constative-descriptive type, hence referential. The reference is to a state of the mind modalizing the assertion : with delight in the first example, underlines this subjectivity. In the last sentence, by E.B.-Browning, this is less obvious, so that the expression may just be an understatement. Another point would concern the interval of time often necessary for passing from intention to decision. Suppose a king saying to a minister : “ I think the army should be sent there ”, this would be taken as a decision of the king : but not everyone is as authoritative. Human relations include degrees and a subtle interplay between the parties : language is dialogic. I daresay I’m thinking that I stay on, not to meet the Bishop...(Thackeray 1850). 81 Thackeray begins with “ I daresay ” : already, whatever the interpretation we give, this is a precaution. He definitely −we understand it as well as his correspondent − does not wish to meet the Bishop, but he still hesitates, and possibly expects advice. as I often sit a long time without saying anything - when I am thinking, or when I am thinking I am thinking (Dickens 1845) ; I am often thinking I ought still to try ( Carlyle 1870). These two examples show that that is not enough to signal a declarative predication : often underlines the descriptive meaning. Are you thinking you are never to see my sweet face anymore ? (Welsh 1826) ; It is better, you are thinking, as it is (Welsh 1826) ; you are thinking I presume,that I might have done your bidding with less delay (Welsh, 1826). The second person does not always entirely eliminate the declarative force, as the first sentence shows. Yet, the utterance becomes normally referential descriptive. It must have been difficult at the time, however, to use the progressive in this connection, except possibly in Scotland. IV.9.2. Cognitive verbs “ Object ” clauses are also found after cognitive verbs. We will not undertake a typology of such patterns, but they are mentioned merely because they are supposed to be averse to the progressive in the main clause . In the three sentences which follow, we notice a special use of the cognitive verbs which usually, in school grammars, is referred to their activity component. another had been seeing his horse’s coat cut and his tail docked [reported speech] (Thackeray 1829) ; Notice that, in this case, the sense may be causative, close to have the coat cut and the tail docked. Hearn wants you to be spying out and seeing if you cannot see a little roughbear coat (Gaskell 1852) ; I have been hearing from Miss Marshall how long suffering and kind Mr. Bray is towards him (Eliot 1865) ; Did you ever hear, or are you hearing now for the first time, that Mr Irving dreams dreams and sees visions ? (E. B.-Browning 1831). In the next : and even in Edinr one is constantly hearing : such and such people “ are not in my way ” (Welsh 1837), there is not need to refer to Scottish usage. It has been included here because of the “ object ” expression reported, but it is not substantially different from other cases when hear has a noun object. Verbs such as find and believe have been included among cognitive verbs in a wide sense. They can also be followed by an “ object ” clause : I went out for a solitary walk and was finding myself tant soit peu tired of my dear little companions (Thackeray 1851) ; All day I had been finding it the greatest possible hardship to have to sit upright (Welsh 1842). Whereas in Wordsworth’s third person statement already quoted : Booksellers are every year finding it more and more their interest to sell an increased number of copies at a low price rather than... (Wordsworth 1838), we may assume that he is reporting on booksellers’complaints. The first person tells us of a state of mind the truthvalue of which is definitely endorsed by Thackeray and Welsh above, so that, even in past tenses, the progressive must be somewhat exceptional. However, we may recognize again a split-ego situation. I know... and that you will be believing me a dead woman if I delay any longer (Welsh 1823). 82 This time, Scottish usage may well be responsible. practical purposes, to an “ object ” that-clause. 98 98 The accusative construction has been identified here, for Some reflection should certainly be applied to the reasons for that "dialectal" use, as well as for the use of the progressive in India, even today. 83 CHAPTER FIVE LINGUISTICS AND LITERATURE While claiming that our letter-writers showed as much spontaneity as any other literate people, we cannot, however, entirely forget who they were. There is a sociology of literature as old at least as Hippolyte Taine, the author of a once popular Histoire de la littérature anglaise 1864-1869: “ Race, milieu, moment ” − race, environment, moment − was the framework within which he and other contemporaries justified the flowering of literature at given places and historical periods, classical ages especially. Such views have been elaborated by modern sociologists. They do not deny the rights of the individual to have her or his own style. Should we consider the use of the progressive form as a significant element in the style of a writer ? For the present limited excursion into a controversial domain, we will take style in a very general sense, i.e. a writer’s original manner of writing, with the only ambition of providing data and suggestions. Scholars, especially in France, have attempted to shed some light on the practice of writers, with sophisticated statistical studies of their vocabulary and chosen grammatical features. 99 That Pythagorean approach can be frustrating : between Mallarmé’s dictum of “ La littérature ” as “ le hasard vaincu mot après mot ” (Variations sur un sujet : Le mystère dans les lettres), and a less idealized view, there is a wide margin where we find for example “ period style ” (Regency: Jane Austen and even Thackeray), or local style (Scott’s use of Scots in his novels), Kiltartanese (Yeats, Synge and the Celtic revival), social style (popular speech with Dickens and many others), and the more or less laborious attempts at imitating the language of the past in historical novels (Kingsley’s Westward Ho, Thackeray’s Henry Esmond, Flaubert’s Salammbô, etc.). Ultimately, we have the original features a reader can recognize in the individual writer, the daily bread of university courses. So long as literature is made of words, 100 we all know what style,or “ écriture ”, can be, without always being able to give a clear definition of it. The main question for the linguist is that of consciousness, i.e. the deliberate use of a given item, even in novels or other fiction, supposedly less constraining than poetry for example. Then, we will contradict the somewhat paradoxical view we have adopted of literary writers as common people with another dictum : “ once a writer, always a writer ”, just as comedians or politicians, for example, are supposed never to forget their role. There are few occasions to suspect the correspondence of the 22 people in the sample, most of them writers, to justify this dictum to a large extent. However it may be a clue to some idiosyncrasies in the corpus and to the independent use of the progressive by some of the letter-writers. Browning and Coleridge on the negative side, Jane Austen and Thackeray on the positive side, are cases in point, but we find some digressions in many letters reminding us for example that Macaulay was a parliamentary orator, a remarkable historian and essay-writer, Carlyle an indefatigable polemist, and George Eliot sometimes a sesquipedalian philosopher. This is in fact related to topic rather than to style and can be taken as an accident since such digressions occur, just as they do in private conversations, in practically all collections of letters ; but for example, they are absent − not to mention Catherine Ruskin, the essayist’s mother, whose English is erratic − from Mary Wordsworth’s correspondence : she was not a writer nor even very intellectual, one of the possible reasons for the high frequency of a familiar item in her letters. Then, less disputably, a person who is busy writing a long novel can be expected to convey to her or his other writing activities some features of the literary work in progress. The place where we find some common ground with letter-writing in literature proper is obviously the novel. Poetry has to be left out, even in long narrative poems such as Wordsworth’s Prelude, or Browning’s The Ring and the Book101 : the constraints of verse are such that a grammatical item cannot be said to occur as freely in poetry as in 99 Some of the promoters were English, however: Yule 1944, Herdan 1964,1966. Let's only mention, among others: Muller 1964,1968; Thoiron 1979; Juillard 1983; see also Sven Jacobson (ed.) 1983. 100 Mallarmé was probably the first to stress it so strongly" 101 Even if Oscar Wilde is known to have said: "Meredith is a prose Browning... and so was Browning"... 84 prose.102 When Coleridge or Wordsworth published some prose, that was not in novels, which excludes them from this line of study. When a comparison between novels and letters can be attempted, there are two ways of doing it : 1. by considering individual novels, trying to discover whether the progressive can be a token of something characteristic, in itself or in association with other items. Needless to say, this requires a qualitative study to complement the quantitative observations. Notice, however, that the density of linguistic items could very well be used for expertise, along with other facts when possible, either to attribute a text to so and so, or to locate it in so and so’s career. Let’s suppose that a novel should be unearthed and thought to have beenwritten by Jane Austen at the end of her life, a statistics considerably different from that presented below could be invoked to dispute the attribution ; with Jane Austen more than with many others, because her novels have a notable homogeneity of style.103 2. The second approach is diachronic : given the rising curve of a writers’use of the progressive in the letters, to what extent is there a parallel in the novels ? Only very few insights are attempted here. V.1. Individual works : The density of progressives in George Eliot’s Romola is remarkable (370). It is known that she spent a lot of time doing research in view of it, while her publisher John Blackwood was desperately waiting. He wrote to his wife in 1861 : “ Her great difficulty seems to be that she, as she describes it, hears her characters talking, and there is a weight upon her mind as if Savonarola and friends ought to be speaking Italian instead of English ” (Haight 1968 : 349). Notice the modal progressive used by Blackwood (1818-1879), a Scotsman.104 We know what pains Romantic or post-Romantic writers of historical novels took to convey the sense of historical distance which was absent from classical literature, trying to give at least an outlandish ring to their sentences, were it only by the spelling.105 They are seldom rewarded, George Eliot not any more than Gustave Flaubert, by the appreciation of present-day readers ; but, at the time, Romola was a success : the Queen presented it to Disraeli. Was this intensive use of the progressive, for which there exists a replica in Italian − but much less common − one of the ways of giving to the verbose novel, in the narrative and the dialogues, a sort of Florentine ring ? The frequent recurrence of certain grammatical items, less clearly perceived by the reader than, for example, names of people (Baldassare, the Frati), or of Italian place names, was perhaps thought apt to carry the imagination away to the Florence of the Medici, instead of to Cimmerian mists, as could be the case in some Scottish tales106. In many stories or comic plays, the progressive was commonly used as a shibboleth of the Irish. So is it, apparently, in Samuel Lover’s Handy Andy: Lover was an Irishman himself, but he left his native Dublin at 36. The density found in this famous Irish humorous story of 1842 is fairly high for the time (270), but not above that in Dickens’s Curiosity Shop, published in 1841. On looking more closely at it, we see that the progressives are concentrated in the dialogues of the Irish characters. It is even possible that Lover, who had become a literary figure in London − he was with Dickens among the founders of Bentley’s Miscellany − may have, consciously or not, avoided what he knew was a 102 Goux's 1971 original statistical study of the progressive in Pop-music shows that it was comparatively a common item in Tin-pan-alley at the time of the Beatles. On this point, see also Scheffer 1975, p.103-5. Why is it so ? The rhyme or the length of the verbal syntagm may be one of the reasons in Good day sunshine where we find : I love her and she is loving me . 103 The same could be said of Hemingway, for example, who appears to have shunned the progressive. 104 As a matter of fact, density in his own letters attains the level of 374, a higher figure than for any of our writers, for a sizeable corpus of nearly 100 000 words. His son William, for a much smaller corpus, shows a density of 418. In spite of censure, the progressive seems to have prospered in the middle nineteenth century in Scotland, even among the literati. Boswell had died long ago. 105 Leconte de Lisle wrote Qain for Cain, and Charles Péguy, in Victor Marie Comte Hugo, tells the amusing story of how Victor Hugo invented Jerimadeth (j'ai rime à -dait − rhyming with Ruth se demandait) for his poem Booz endormi in La Légende des Siècles, the name of a place not found in atlases, even German ! 106 Tulloch 1980: 296 did not attempt a thorough study of the progressive in Scott's novels. He mentions a few cases, looking like ordinary shibboleths of the time. 85 distinctive celticism, often laughed at : the title of his burlesque opera Il Paddy whack in Italia is an indication of his attitude to the speech of the Irish and to their boisterous and pugnatious manners. Grammatical shibboleths are certainly less common than those concerning pronunciation or vocabulary, and, when they consist of phrases, they are not very different. In they are after killing your pigs, neighbour, the Irish construction, often quoted, becomes a set-phrase. It is much more difficult to have systematic dispersion of a feature over the whole narrative, especially, when the shibboleth is a shibboleth of magnitude : the progressive was already common enough in England so that it was not noticed most of the time ; but the Irish or the Scotch − as they were still called − simply made more extensive use of it. Conversely, we find a refusal of the progressive as a modern feature in Thackeray’s Henry Esmond. A chapter in The language of Thackeray (Phillips 1978) is dedicated to it. Commenting on a sharp criticism of Fitzedward Hall in Modern English 1873 giving a list of Thackeray’s anachronisms, Phillipps admits that Hall is right to censure a sentence like was being battered down : even now, nobody has been able to find an occurrence of progressive passive before the second half of the 18th century (Denison 1998 ; Nakamura 1998), whereas Esmond takes place during the reign of Queen Anne (1702-1714). The question of what is valid and what is not in Thackeray’s novel is beyond the present project, but I found, with the same method of measurement, a density-difference of 100 between two novels written at few years’distance Pendennis (1850 : D : 287) and Henry Esmond (1852 : D : 186) (Arnaud 1973 : 603) : when we remember Thackeray’s fairly extensive use of the progressive in his correspondence, we must admit that he deliberately abstained in his historical novels from the modernization this form implied. Historical stories or folkloric plays, etc., even by people who really know the language of the common folks , but are not professional linguists, are inevitably bound to be distorted : observations on the negro dialect in Mark Twain ‘s Huckleberry Finn, on Bernard Shaw’s cockney in Pygmalion, and other imitations, are instructive of the refraction through the prism of literature ( Labov, Lecture at the Maison des Sciences Humaines in Paris, 20-2-1986). 107 This is partly due to ignorance and false notions of language, even from highly intelligent people.108 There is always an element of parody in local colour, which explains why such attempts are seldom successful, except when explicitly presented as parodies, as in the case of cocoliche (Italianized Spanish) plays once popular in Buenos Aires.109 This is not the only situation when something else than pronunciation or vocabulary can be used for literary effects. Guillemin-Flescher 1977, 1981 was mostly concerned with a comparison of the French and the English approach to syntax and semantics when she studied the translations into English of some of Flaubert’s works. But, admittedly, his use of the imparfait − something as exotic to the English-speaking readers as the progressive to the French − is highly elaborated. 107 In French, we can mention George Sand's novels (La Mare au Diable, François le Champi, Le Meunier d'Angibault), where country folks from Berry, in central France, speak a somewhat fictional dialect. 108 Jean d'Ormesson, de l'Académie française, and a brilliant writer, in his attacks of the suggested spelling reform, used an extraordinary argument: "people won't be able to read Corneille in the text". Who ever does ? There is practically no edition of the Cid with the original spelling, and what about Rabelais or François Villon? Publishers know what the sales would be. 109 They are no more accurate than stage-Irish. 86 Raybould 1957 studied Jane Austen’s progressives in some detail, but not their frequency compared to that of contemporaries ; nor did Phillipps 1970 : 111 sq., in his long and penetrating comments on Raybould’s article. The present research adds this dimension to the debate. It seems likely that Jane Austen had found in the modern form something worth her attention.110 The most obvious fact noted by Raybould and Phillipps is the number of progressive to-infinitives : for a total number of 292 progressives in her letters, there are 11 infinitives. This corresponds to a density of 8.16 against a general density for the same periods (up to 1819) of only 2.11 in which she is included. Even allowing for the difficult rejection of the null hypothesis with small figures, there is no doubt that she was particularly addicted to progressive infinitives, as the studies covering also her novels had noted. On the other hand, there is no occurrence in the letters of present participle progressives like the concert being just opening, or the extraordinary their being going to be married (in Sense and Sensibility) − in fact a nominalization, which could nevertheless be included in a “ squish ” next to the progressive. Next, we have what has been called the “ extended present ” in IV.2., which, in the language of the time more than now, looks both directions like Janus : towards the future or towards the past. In fact, there are four ways of referring to the near future, even in our time :1. the simple present (we leave to-morrow) ; 2. the progressive present (we are leaving to-morrow) ; 3. the going to expressions (we are going to take a house) ; 4. the be to form, which was very common at the beginning of the 19th century, especially with Jane Austen, and seems now obsolescent. Did she also use 2, now very common ? The answer is no. She refers to the extended present in the direction of the past by using the perfect associated with just, but to refer to the immediate future she usually makes use of the be to phrase. As Phillipps very aptly says : “ These constructions, of the type, ‘John is to come home at once’, are common, of course, in present English ; but she uses them more widely. It is easy to find parallels today for the use in the sense of what is preordained or arranged. ” Phillips 1970 : 137. But he goes on to say that there are other cases when the best gloss would be : is about to or is likely to. This interesting discussion cannot be developed here, so the important remark is that we find no example in Jane Austen’s Letters of the progressive with immediate future reference. For the same period, from the end of the 18th century up to 1817, 21 progressives of this type have been found in the letters of many writers, from Maria Edgeworth to Keats − the latter will increase his contribution noticeably in the following years −, Scott giving the highest number. In this respect, Jane Austen was still writing in the language of the 18th century, and, when future reference was meant without the modal overtones of be to, she relied on the context or the situation, especially with terminative verbs, to make it clear, and used the simple form : On Thursday Mr and Mrs Bridges return to Danbury, Miss Harriet Hales accompanies them to London ; We dine today at Goodwestone ; Martha comes, and a ball there must be ; Mrs Busby drinks tea and plays at cribbage here to morrow ; if the weather permits, Eliza and I walk into London this morning, etc. The now common going to forms were not part of this research, regrettably perhaps. They don’t seem to have been used by Jane Austen. Their development becomes conspicuous only later, with some ambiguities, and the occasional use of come instead of go. Another point which both Raybould and Phillipps have noted is her using the progressive (“ expanded tenses ”) when the “ durative aspect of the verb, now normally conveyed by the expanded tense, is severely limited, implicitly or explicitly, by some circumscribing of the time in the context ” Phillipps 1970 : 111. This is what we call explicitly bounded duration, sometimes considered as incompatible with the progressive (see above IV.4). 26 cases have been found in the corpus for the same period when Jane Austen wrote, 4 of them in her letters. Such small figures hardly allow generalization but still, the compared densities, 1.05 and 2.97 tend to show that the observations of the two analysts are founded. As a matter of fact, such use seems to increase later on, since there are 204 occurrences in the whole of our corpus. This corresponds to a density of 2.11, clearly above the general figure for periods 1-5 (up to 1819), but still below Jane Austen’s. This again can be a clue to her experimenting with the new form consciously − which is likely in her case − or unconsciously. The question whether this was a transfer into her letters from her novels, or, in the opposite direction, a transfer from her own speech, oral or written, is a difficult one and beyond proof. The most likely is that there was feedback both ways, with the restriction that, careful as she was to differentiate her characters by their language, not one among them really spoke her own. This − and a supplementary reason for the above development − is the clearest case in our study when reality and fiction can sometimes be fused and confused. 110 On this question, see also Strang 1982. 87 V.2. Trajectories Jane Austen’s letters, as well as her novels, cover a brief period, 20 years including all, but in fact only 10 with an acceptable corpus size for the letters ; much less, 6 years, for the novels (1811-1817), which on the contrary provide a comfortable size. Two observations can be made about Figure 34 : 1. The overall density figure for the novels (from 216 to 400) is much greater than for the letters (217). Only Pride and Prejudice (1813), Mansfield Park (1814) and Emma (1816) have been surveyed, with the following results : P&P, 235 ; MP, 390 ; E, 404. Not only is there a remarkable increase over three years only, but we reach the stupendous figure of 400, never found elsewhere. 2. Due to corpus size, the longitudinal values for the letters may want significance. They nevertheless show an increase from approximately 160 to around 300. If we rely on this token, we may explain this less acute rise in two ways : 1. Jane Austen’s novels contain a lot of dialogue, which encourages use of the progressive, especially for character-building as said above. It could be studied more precisely in the direction opened by Raybould 1957 and Strang 1982 ; 2. the novelist was satisfied with the effects she had obtained with that form, which was a novelty for her and her readers in such situations, and she was encouraged to enhance it. This remark would be stressing the amount of conscious or partly conscious elaboration which separates the literary works from the everyday prose of private letters. The comparison between letters and literary works should be more rewarding with many others who have had a long career. Only a few glimpses have been found possible within the limits of the present study. With reference to the first romantic “ generation ”, the attempt could be worthwhile for Walter Scott, since his letters are fairly abundant, and his novels as well ; but with the important reservation that, being often historical novels, they cannot be a mere reflection of his own usage : cf. Tulloch 1980. The second period, that of people born after 1800, is a better field, with the great early-Victorian novelists and their abundant production in both fiction and letter-writing. To them could be added Carlyle and Macaulay, with prose works of a different type, but where progressives can find place. Macaulay is an interesting case. After a volume of Essays, critical and historical (1843), his famous History of England from the accession of James the Second was written and published in successive volumes between 1848 and 1861 (after his death in 1859). Figure 35 is a comparison of the densities in letters and historical works. It shows clearly that the development of the progressive in the narrative historical works is parallel to the development in Macaulay’s Letters, but at a lower level, which is not surprising. It is important to stress that Macaulay’s works are homogeneous and have never been considered as particularly modern in tone ; he was one of the strong opponents of the new passive, and makes extensive use of the old one : while innocent blood was shedding under the forms of justice, the new Parliament met ”( History I) ; “ while the foulest judicial murder ... was perpetrating, a tempest burst forth ” (History II ), etc. Carlyle’s French Revolution, published in 1837, but actually written before 1835 and rewritten after that date 111 shows a density of 180, a high figure for the time, and considerably higher than in Macaulay’s Essays seven years later. There must be two reasons for this : 1. Carlyle had even begun writing it in Scotland, under the influence of his native Scots dialect ; 2. It is much more than a plain narrative in Macaulay’s fashion ; it reads indeed like a novel, at least as full of emotion as Dickens’s Tale of two cities also about the French Revolution. Nevertheless, the historical narrative displays a much smaller density than the letters : 180 against 286 for the latter at the same stage of the author’s life. Perhaps a parallel could be tried for at least some of the other works of the famous essayist : but, whereas his correspondence reads like that of any man of the time, his literary style is far from straightforward and fluent. 111 One of the most extraordinary cases of a precious MS entirely destroyed, at a time when the only way to keep a duplicate was to have it hand-copied. Carlyle had lent it to Stuart Mill, whose servant inadvertently threw it into the fire. Refusing a generous offer of money from his friend Mill, a rich man, Carlyle had to write it again from memory. This may account for the exceptional spirit of this work, full of historical errors, but the best "poem" on the French Revolution ever written. 88 V.3. Chronological groups We have seen that, in the case of Jane Austen and despite her short career, there is some parallel between the density increase in novels and in letters. This was an occasion to remark, however, the distortion between works of art and actual everyday practice, oral or written. Even the dialogue in novels, seen by Strang 1982 as the best reference, cannot be entirely relied on : yet, its inspection, possibly with the modern facilities offered by scanning, could be rewarding. A pilot study (Arnaud 1973 : 604) of 1000 progressives in a random sample of dialogue in novels revealed densities very similar to those of correspondence or slightly higher112. Ideally, some insight could be obtained by comparison either with works of fiction or other narrative works, randomly selected along the century, or with a corpus of such works by the letter-writers themselves. Unfortunately, this is not feasible, for want of suitable material. In the first “ generation ” surveyed, there are only three writers of narrative fiction : Maria Edgeworth, Walter Scott and Jane Austen. We have seen that the latter’s novels mirror everyday speech in a very special, Austenian, way. In the case of Maria Edgeworth, it’s the corpus of letters which is comparatively thin. Walter Scott could provide a substantial corpus, the long line of his Waverley Novels on library shelves ; but most are historical novels, which means that it is difficult to control what is early nineteenth century language, or the supposed language of the time of Ivanhoe for example, and what is anachronic confusion (on such points, see Tulloch 1980). This is not to say that the study is uninteresting with reference to the techniques of novel-writing. Then the bulk of our letter-writers for that period are poets, who, with the exception of Southey with his History of the Peninsular War, etc. wrote little narrative prose likely to contain a number of occurrences sufficient for a statistical survey. The second chronological group, on the contrary, lived in the golden age of the novel. Studying either random samples chronologically ordered, or well-chosen individual works, could be rewarding for an overall view confirming or supplementing, or perhaps contradicting what is found in the letters. This would be a huge task, for which the present one can provide a pattern. Arnaud 1973 gives some elements for a chronology of novels published between 1841 (Dickens’s The Curiosity Shop) and 1868 (William Collins’s The Moonstone) but, for that interval of 27 years only, it does not show significant increase, even when historical novels are excluded from the list. Another form of assessment is to compare the chronology of the letters and the chronology of the other prose works. We have seen that Jane Austen displays an increase in novels paralleled in her letters ; which is unlikely to be due to chance. Carlyle, Macaulay, Mrs. Gaskell, Thackeray, Dickens and George Eliot have all published volumes of prose at successive points of their existence. A parallel is then possible to confirm or disconfirm whether the increasing use of the progressive in the correspondence is found also in the other prose works. Again, this would be an immense task, and only a few examples will be given here. Among the numerous novels of Charles Dickens, Arnaud 1973 studied only two landmarks : The Curiosity Shop 1841 and Our Mutual Friend 1865, that is about the two extremes of the novelist’s career. Here is the comparison of the densities with that in Dickens’ Letters : 1841 The Curiosity Shop : 266 ; Letters (1840-44) :266 [sic] ; 1865 Our Mutual Friend : 261 ; Letters (1865-69) : 379. Whereas there is conspicuous increase in the letters, the density in the two novels is stable. A longitudinal study including a larger number of works might perhaps suggest the reasons for this, with due allowance to the variety of parameters and to the statistical significance of the figures. A Tale of two cities (1859), a historical picture of the French Revolution, with a density of 190, betrays once more that the modernization of style brought about by the progressive was deliberately excluded from historical narratives (with the notable exception of G.Eliot’s Romola : see above ). As a matter of fact, this is also supported by the low figure for Charles Kingsley’s historical novel Westward Ho (1855) : 173. Within the limits of the present work, apart from Jane Austen, the most interesting parallel between letters and literature proper is probably to be found with Mrs. Gaskell. Using a count made by Evelyne Tessier 1980 (Université Paris VII dissertation, unpublished), the densities for 5 novels (Mary Barton, 1848 ; Cranford, 1853 ; North and South 1855 ; Cousin Phillis, Wives and Daughters, 1865) have been put side by side with those for the 112 At the time, it was based on separate samples of progressive and non-progressive tokens. 89 letters to her daughter Marianne at their respective periods : Figure 36. Two novels belong to the same period and their density values have been fused as well as those for the letters. The curve for the novels shows a regularly ascending line from 220 to 375 between 1848 and 1865, without the depression for the early sixties which, as we have suggested, may be due to a coldness in her relations with her daughter, also reflected in the fact that her letters were less frequent and/or shorter. That difference could support the hypothesis (see above in III.1.4.b and Figure 12), since there is little probability that her private feelings should be reflected in the novels as clearly as in the letters. Should such comparative observations be confirmed for other writers, George Eliot for example, their credibility would be enhanced. Statistical tests only tell us when the null hypothesis cannot be rejected, not that it must be accepted. Linguists non longer rely only on the language of “ good writers ” for the description, aside from prescriptive rules, of community grammars113. Socially-oriented linguists would fain write “ communities ” in the plural (Georgian and Early Victorian grammars for example, to account for social strata and linguistic geography, etc.). At the same time, it would be ridiculous to discard the lessons of literary works, for a variety of reasons, including the admiration we have for them and the influence they have had. The above remarks and observations point to directions which could be explored further, and contribute to bridge the unfortunate gap between the social and language sciences and what still belongs with the humanities. GENERAL CONCLUSION The present survey was undertaken with the hope of providing insights on the progressive in Romantic times, and more generally on historical developments : yet, trying to conclude with more than suggestions would certainly be overambitious. What can be done in “ the art of making the best use of bad data ” − the task of historical linguistics according to Labov − can be summarized under three heads : observations, generalizations, principles (Labov 1994 : 12). Observations are already a construction of facts on the basis of the quantitative data ; generalizations are an attempt to go beyond them to unveil some of the particular rules they may illustrate ; as to principles, they go further beyond by questioning some accepted premises of linguistic theory. The further we go from observations and facts, the less assured is our progress and the more likely our suggestions to be challenged. As already said, our observations cannot claim to be entirely innocent of preliminary theorizing, open or covert. It was also emphasized that this theorizing should not be too rigid, and was supposed to be acceptable by a large number of the linguists who have studied the progressive and its history : hence the comparatively unsophisticated outline in I.2. As stated in the introduction, the guidelines of the research were those of Empirical foundations for a theory of language change (Weinreich et al. 1968), a critical overview of the major theories of diachronic evolution. The five problems (constraints, transition, embedding, evaluation, actuation), which the authors kept separate for the sake of Cartesian methodology, are often linked or overlapping : in the field of grammar for example, it is sometimes difficult to distinguish constraints and embedding, transition and actuation ; and evaluation can generate new social constraints. Many discussions have taken place since that landmark was set, especially around the actuation-transition problems, and grammaticalization has recently been on the foreground (Traugott & Heine (eds.) 1991 ; Pagliuca (ed.) 1994). In 113 This view of good usage is typically represented for French grammars by Vaugelas's Remarques sur la langue française (1647) and the Belgian Maurice Grevisse, whose Le Bon Usage (1936) has been reprinted many times. Grammars of English have commonly followed the same practice, less avowedly; Jespersen's A Modern English Grammar..., for example, refers to an extensive bibliography of writings. 90 connection with the progressive, this raises interesting questions, which the present survey may help to answer, as to the nature of grammaticalization : is it merely an internal process or should we also take external factors into account, were it only that of learnability, implying psychological considerations ? Even when internal parameters only are considered, how should the process be described ? Finally, what can be said about the time of its development, its duration and stages, and the local and historical conditions (internal or external). No more than a summary discussion will be attempted here. The classical type of grammaticalization involves lexical items becoming grammatical. It can be applied to the progressive as part of a study of auxiliaries, since they all derive from main verbs : yet, what can be comparatively clear for modals can send us to very remote periods indeed with reference to primary auxiliaries. However, if a limited historical period is considered, the question can be seen as mostly syntactical, beginning with description of the movement from one type of organization to another If the description is correct, does it mean that the explanation will follow, or can simply be dispensed with ? Among recent efforts to shed some light on the development of the auxiliary system in the Modern English period (18th and early 19th century), Warner’s (1993, 1995, 1997) are noteworthy.114 He sees the development of the progressive passive − the clearest case of innovation in our survey − as a predictable element of the overall system of auxiliaries undergoing complete transformation in the course of the 18th century. Since predictability is one of the challenges often presented to linguists, this is a bold stand. He does not take into account external constraints which may have either encouraged or censured the move. We know that the new passive form was repeatedly stigmatized (see above IV.4.1) : this certainly explains why it took so long to become settled ; the present survey shows clearly how the two forms went on side by side. It also shows that the logic of everyday users − of the enlightened class, including professional grammarians − is not necessarily that which governs the investigations of modern linguists, nor the formal systems they apply to language. Warner’s arguments are strong enough, however, at the same time as cautious, and we may admit that the conditionings he describes were sufficient to justify the, to his mind inevitable, development. 115 Social factors would only have affected the rate of its diffusion. One of his points is that the change affecting the auxiliaries was connected : 1. with the decline of thou,especially in addressing children − although he confesses that we have too little information on the spoken language to warrant this − with consequent loss of inflectional categories in auxiliaries ; 2. in “ the virtual going to completion of periphrastic DO”. On that second point, Nurmi 1996 concluded that, for the period 1590-1620 at least, no evidence can be found to support it. Yet the question remains open and should not be discarded. This calls for further quantitative research in extended corpora: on this point, see also Kroch 1989 and Tieken-Boon van Ostade 1987. Then, the connection between the developments affecting the progressive and the BE/HAVE variation with intransitives (Rydén and Brörstrom 1987) certainly requires new attention. Finally, the history of the progressive from its beginnings, the subject of Mossé’s description, could probably shed some light even on its modern “ not solely aspectual meanings ” ( Hancil 2000; Smitterberg 2002). 114 However, the lexicalist approach is perhaps too apt to generate ad hoc formulations : subcategorizations (e.g. of -ing as progressive or not, (+PRD) or (-PRD)) etc. are justified in terms of distributional properties only. Then the concept of learnability (connected to Lightfoot' s “transparency principle”)does not seem to be thoroughly grounded on psycholinguistic observations, and evokes the old “parent-child” model of language change, even if some assumptions seem reasonable. This could open a long discussion. 115 Another reservation would concern his corpus, mostly relying on Visser and also, often, on Jane Austen. The former is certainly a reliable source, but it could be extended; the latter, as I have found, (see above in V) is probably suspect: she was an experimenter, and comparison of her novels and her letters shows that both are somehow idiosyncratic, interesting in any case, but to what extent representative of community use, we don't know. Rydèn & Brörstrom 1987 have also shown that writers (and this probably applies, more generally, to people) can be, often erratically, either conservative or progressive; or even innovative. Idiolects need further exploration: Jane Austen's original usage can be a good starting point, illustrating the possible role of individual women in language innovation and change. 91 Such points were not the object of the present research: yet the data provided may eventually help in their study. Heine 1994 studied the grammaticalization of the progressive in Ewe, a language of West Africa. Although we have to be cautious in applying the label “ progressive ” to a number of analogous forms in many languages, such comparisons can be illuminating (Arnaud 2000 ). He notes, to the satisfaction of the present writer, that “ the grammaticalization of locative constructions as verbal aspect categories is not confined to Ewe, rather it is observed world-wide, it forms in fact the most common pattern of evolving new progressive aspects ”, “ ... It is well known that in many, if not all, languages, spatial expressions are metaphorically employed to conceptualize temporal notions...location in space also serves to express temporary states, contingent situations and progressivity (cf. Comrie 1976 : 98ff) ”. He gives a table of the most common schemas serving as the source for progressives, the first of which is quite in agreement with the suggestions in I.2. (Heine 1994 : 268-269) : “ X is at Y ” ( Location schema) gloss : “ he is at/in/on eat-ing ”. Forgetting for a moment the very concept of grammaticalization we have to ask ourselves what was exactly the type of change evidenced by the present survey. A complete coverage of the question implies a wider approach. Among the queries, one is fairly classical : the language or the language users, which is foremost ? This is another way of formulating the dichotomy of langue v. parole in Saussure’s Cours. The survey indisputably puts the users in front, an approach which does not preclude the formal approach, with the potentialities of change present in the user’s minds under the once popular name of competence. Among the difficulties we have the fact, which entails a serious hitch along the road to predictability, that the potentialities suggested can lead to diverse routes : “ spaghetti junctions ” for example, etc. So that the linguistic field can be compared to a gigantic web, where the future generations will surf, no one dares to say which way. A limited idea of such a web was suggested by Lloyd Anderson 1975 for the semantic space of get, have and be : the latter two are connected to the progressive, as we know. To put things differently, linguistic structures have some capacities of assuming new shapes (but they are more like the physics of soft materials than that of stones), and we are often in a position to say what these shapes cannot be, but not what they will necessarily be. This is possibly where Meillet comes in, when he links linguistic change to social change. This − and the above remarks − certainly sounds like a commonplace to many linguists, yet it seems to have been frequently overlooked or even censured. The case of the passive progressive discussed above is perhaps a case in point : was it as inevitable as Warner thinks ? It would be an interesting task to list predictable changes which eventually did not happen : the blind alleys of historical linguistics. If the transition problem is difficult, but can be cleared by careful observations, the actuation problem which is connected to it is far more formidable, since it puts the question of causality. In the perspectives of Empirical foundations, the causality of language change is not seen as purely systemic but is referred to structured heterogeneity within the socio-cultural matrix. This emphasis on sociology, usually stressing the communicative function of language, was already, especially in France, typical of the early 20th century which saw the birth of modern sociology. Other possible causes, often favoured before, were entirely discarded, such as psychology in general : Cf. Hugo Schuchardt’s criticism of Saussure’s Cours (Normand & al. 1978, 174-181), particularly concerning Saussure’s dichotomies of langue /parole and synchrony/diachrony which Schuchardt thinks are in fact only one ; or imitation, so dear to Gabriel de Tarde 1895 ; or even climate, etc. As a matter of fact, sociology itself was taken in a very general sense : Meillet speaks of “ social change ” without qualification. After a debate lasting for almost a century, and the victory of structuralism, are we not entitled to adopt a more tolerant view ? Why should we exclude any cause or factor a priori ? The constraints problem and the embedding problem are more familiar to linguists, insofar at least as only the linguistic environment is concerned, and familiar to society-oriented linguists, when external factors are taken into account. Clearly, the research frame presented in I.3 was conceived for the purpose of measuring the comparative influence of both groups of factors, on the basis of experience and the observations of a great many people who have studied the progressive, mostly with reference to present-day British or American English. We shall now try to sum up the results of the quest, observing what can be reasonably taken for certainty and what is still subject to various degrees of uncertainty. 1. .A considerable increase of the density. 92 All the figures confirm that the frequency of the progressive in the type of discourse studied increased considerably and steadily. There is some reason to think that it paralleled the increase in spoken English, and that it also extended − with possible different rates and modalities, worth surveying (Smitterberg 1998, 2002) − to other strata, where density, being smaller, requires a larger volume of material ; this, except, perhaps, for the study of tokens of oral speech, would not considerably alter the general picture. It is also quite probable that this development originated in the language of the lower orders of society, but little is known of the way it spread across the various geographical dialects, apart from the accepted fact that the Celtic dialects of Ireland and Scotland were noted for their addiction to it.116 It is also likely that their influence was felt in the North West counties in country dialects, such as in the Lake district as we have assumed, and that the large number of Irish and Scottish immigrates, many of them, especially the women, introduced into the Victorian homes as servants, contributed to its development in city dialects. The reasons for the development of the progressive in Anglo-Celtic dialects are beyond the scope of the present work. 117 It is also difficult to say whether this influence originated the movement in modern English, or merely reinforced it. This is part of the actuation problem discussed below. Let’s just remember that Irish and Scottish immigration became significant only by the eighteen forties, when, as we can see, the movement had already been launched : so that the hypothesis of reinforcement seems more likely to be right. 2. Density and degrees of formality The progressive has been shown in a previous survey (Arnaud 1973) to occur at different frequencies across the strata or genres arrayed in order of formality, and was more common in the less formal strata. Within the more homogeneous domain of epistolary writings, we notice a subordinate order of formality, uniting the less formal language of women in general and the degrees of intimacy for both men and women. This confirms that the less formal a discourse, the more likely it is that it contains a high density of progressives : another instance of parallelism between the conventional social scale of speech and what is studied under the label of style or register. 3. The weight of linguistic factors The study of the purely linguistic factors required considerable effort, especially in distinguishing predication classes. Some aspects of the change which could be hypothesized, and have been long documented in most cases, are confirmed : from the increasing use with state verbs to the nearly complete substitution of the new passive to the old one, we can only enrich the existing knowledge with significant figures. But the predication classes do not display distributional differences sufficient to reject the null hypothesis : the conclusion must be that this influence is a slower process, too slow to be measured accurately with the methods used here. We have to be satisfied with plausible hints which a more extensive review over a longer period should confirm and qualify. 118 116 So far as I know, historical dialectology (cf. Görlach, 1988: 218) does not give us much information on this account, mostly because of lack of evidence on the spoken usage and familiar writings. Mossé (1938:35-36) suggests that the progressive, between the 13th and the 15th century, spread from the North to the South through the Central Midlands. Mustanoja (1960) and Nehls (1974) also noticed that it was more common in the North in the Middle-English period. Cf. also Nurmi (1996: 157; 1999: 65). 117 They must not be very different from the virtualities existing in English proper, except that verbal nouns are a prominent feature in Gaelic, thus encouraging the periphrastic development. Mossé, however, did not think that the influence of Gaelic was at the origin of the "periphrastic form" in English. He also believed that the English phrases due to the influence of Celtic in the English of Celtic areas (Scotland, the "Pale" in Ireland) did not penetrate into English, and that the influence of Celtic dialects was not necessary to reinforce the development : Mossé 1938, p.66. Cf. also Arnaud, 1998, 142-43. 118 Not all people will accept the semantic classification given here, nor the ascription of individual items to a given class or subclass: it seems unlikely that a different choice would lead to more convincing results. 93 4. Lexical diffusion. Does the development affect all the lexical system at once or is there some sort of lexical diffusion ?(III, 2.3.) This is another point about which no firm conclusion can be drawn : the study of vocabulary is full of thorns. Vocabulary increases with length of text : new words (new verbs in the present case) are added each time you extend it. A binary study of verbs that take or do not take the progressive at the successive periods would therefore be tempting, but the task is formidable and elusive : new verbs occur erratically according to needs, so that it is only by considering classes as we did that quantitative measurement is possible. One of the observations, however (already made in the computing for Arnaud 1973) is that the new verbs in a new section of text usually belong to the class of processes and that the enrichment may consist of any item of the corresponding lexicon, with a preference for unbounded processes, which refers us again to the distribution of classes. The addition of new state verbs to the list of those accepting the progressive is a well-known fact, but there are too few occurrences of each for a significant quantitative study. 5. Explanations and principles The lessons which can be drawn from this extensive corpus are only tentative with respect to causality. Many, if not all, of the current hypotheses concerning the linguistic conditioning and modalities of the development in modern English are undoubtedly supported by the quantitative observations. The progressive has been gaining ground over the verb paradigm and its lexical and semantic constituents, in the direction which apparently tends to eliminate many of the original barriers. It would not be taking a great risk to say that, in 21st century English, even performative utterances will be affected, even verbs like know or belong − there are already some symptoms ; this, of course, implying modernized semantics. Unless the trend be reversed, which is always possible, though unlikely for the short term. In quantitative terms, it is however difficult to measure with precision the influence of purely linguistic factors, as we have seen even with the verb classes covering a large number of items. Unfortunately, it was not found possible for this exploration to take into account all contextual or situational factors beyond the limits of the sentence and even of the main clause : they obviously play a role, sometimes considerable, but it is not easily submitted to quantitative analysis .119 The typology of clauses is a complicated matter, and a term like situation refers to a number of occasions which can better be apprehended in oral contemporary discourse, or even in modern texts with the help of contemporary speakers’ elicitations and reflections. The case is slightly better if we rely on the formal parts of the sentence proper, but it probably requires a larger corpus, and a longer period, to attain a satisfactory degree of reliability. With reference now to the sociostylistic concept of degrees of formality, the short term development is shown with convincing evidence to be strongly connected to the two factors investigated : gender and intimacy. Our reflection on this with respect to principles and to causality can take two directions, one suggested by what we may know of diachronic processes in general, the other has to be so to say “ invented ”, since it carries us to littleexplored paths. a. Long term and short term Concerning diachronic processes, it is perhaps artificial but practical at least to separate long-term trends and shortterm developments. The discussions about long-term “ drift ” are as old as historical linguistics, but fairly well documented, and we may reasonably admit that the birth and development of the progressive (including semantic change) are not outside the grasp of diachronicians, provided they use the appropriate tools for it. This is a general change that affects the very foundations of the verb system, and which can be apprehended only at a high level of 107 Strang 1982, Wright 1994a and Smitterberg 1998, among others, have rightly taken clausal distribution into account. The subtleties of the arrangement of discourse have been studied by Guillemin-Flescher 1977. Often, the reference system is only indirectly present in the sentence, since no utterance is entirely independent of what was said or thought before (presupposition or Culioli's préconstruit). In some of the points developed in IV, the analysis extends clearly beyond the clause or the sentence proper. 94 abstraction. 120 The discussions around Aristotle’s views are in the right direction if they take care not to abstract from the actual use of the language too beautiful logical models, which sometimes happens even with ordinary language philosophy. In this particular case, analysts have a problem of equilibrium, since the gap is very wide between the abstract and sophisticated discussion on the concepts related to aspect and the careful study of actually attested utterances. b. Internal and external explanations In a clear overview, Heine 1994 : 254-259 opposed two types of explanations, internal and external : “ According to the former, phenomena are explained exclusively with reference to and within the domain to which they belong, while according to the latter, explanations are sought outside the relevant domain, i.e., they are derived from independently motivated principles. ” He adds that “ There is a growing consensus that internal explanations are weaker than external ones... and indeed, according to some, internal explanations are not explanations at all, that is, true explanations are by nature external ones ”. Among many who took sides in the debate, Givon 1979 :3-4 is specifically mentioned as providing “ perhaps the mots exhaustive and relevant catalogue of “ strongly explanatory parameters ” :... “ 1. propositional content, 2. discourse pragmatics, 3. the processor, 4. cognitive structure, 5. world-view pragmatics, 6. ontogenetic development, 7. diachronic change, 8. phylogenetic evolution. ” While it is not certain that the catalogue is complete, one can see that the discussion here takes some of these points into account, possibly adding new perspectives. In this debate, which reminds one of the distinction of analytic (deductive, syllogistic) and synthetic (inductive) judgements in classical logic, one will not be surprised to see a socially-oriented linguist side with Heine. But it is perhaps just an article of faith, and it is reasonable to admit that the two types can be complementary, whatever we mean by explanation, and the strength of it. c. Some possible external causes One of the founders of structural linguistics, Ferdinand de Saussure, never proposed a satisfactory solution to his own paradox : la langue, a social entity, was conceived as a completely autonomous system. It is not sufficiently known that, at the time when, following Durkheim’s discovery of the specificity of the fait social, Saussure insisted on the specificity of la langue, he was also exposed to the ideas of another very influential social scientist, Gabriel de Tarde. Tarde accounted for social phenomena in terms of interpsychologie, with imitation as the main vehicle of diffusion in many domains, including language. So that linguistic innovations could be propagated in the same way as fashion in dress, hairdo or furniture. Tarde’s suggestion was promptly ostracized by the early sociologists ; yet it was at least an attempt to solve the mystery of the interface between the individual and the group, especially in trying to account for historical developments. La langue as a social entity was seen by Saussure as a product : nothing was said of its production. This is why it has been said that Saussure’s parole must be the locus of diachrony. Dorozewski, 1933 (see Norman et al. 1978, 105) believed that the concept owed a lot to Tarde’s views. What is its mode of existence we don’t really know, but it can only be of a dialogic nature. If external influences are not excluded as a matter of principle, several explanations can then be suggested for the type of development we are surveying. The simplest is Hermann Paul’s avalanche effect. The movement, once started, gains momentum, and we find the typical S curve of diffusion. Nobody has been completely satisfied with this since it looks like just description, and can hardly be justified or falsified. Then we have societal change (from Meillet 1921 to Guy 1990). Undoubtedly, the period we have surveyed (17801880) was one of considerable societal change (the French Revolution, the Napoleonic wars, the Industrial revolution, etc.) and, without further precision we could leave it at that, just as a qualification of the avalanche effect. A side-aspect but an important one of societal change is when it induces contact phenomena. In this respect it is not impossible that Anglo-Celtic dialects, Irish English especially, had an influence (see above) . Massive Irish immigration began only by the 1840s (the railways navvies were often Irish), but even before that time, there were 120 In terms of generative grammar it is not a low-level rule, but among the first, involving primary auxiliaries. 95 many Irish nannies and Scottish gardeners in the homes : they may have sold some of their speech habits to the children and their mothers : cf. Labov 1990 : 219, quoting Gauchat 1905. Yet, again, why should avalanche and societal change so strongly affect some variables and not others ? The same could be said of Tarde’s imitation and fashion in language habits : it could however justify a renewed approach to style, applied to all types of discourse in a diachronic perspective. Even in everyday conversation, extensive use of progressives might well be one of the symptoms or indices of the change from Augustan and Regency English, to what would become Victorian English : it was felt , at variable degrees of consciousness, to be more modern. Notice that in this respect, valuation tends to contradict stigmatization in the everlasting conflict between generations, the younger people adoring what the seniors burnt. We will then have to look for the reasons of this new, often perhaps unconscious or at least implicit, valuation. d. Psycho-historical considerations. It is certainly more daring to suggest some connection between the semantics of the progressive and the sociohistorical conditioning of the change. This can be done by tentatively correlating the two concepts of womanhood and intimacy on the basis of accepted psychological and sociological attributes. It has been shown that the women in the survey consistently used progressive forms more frequently than the men : wasn’t this perhaps a matter of the topics they favoured as suggested by a present-day survey of conversation in Kipers (1987) ? The point would require closer examination but, at first sight, it does not seem to be the case Intimacy − the opposite of distance − was a practical label to sort the letters into three classes. Renewing the intuitive considerations of many older students of the progressive, perhaps it provides a clue to explain the growing success of the progressive in Romantic times, especially with women. This would definitely contradict Eckert’s (1989) views on the irrelevance of expressiveness to account for women’s share in language variation, but by relating it to gender, not to sex, i-e. by referring it to more comprehensive characteristics. Intimacy and distance are microsociological concepts, whereas other terms such as effusiveness and spontaneity run the risk of circularity, since they refer to features that pertain to the text proper. Intimacy is a biographical fact, although, as we said, it is often revealed by the letters : in historical research, we cannot ask the writers what their real feelings are. The other labels refer more specifically to the reflection in the text of the relationships involved. Undoubtedly, psychology intervenes. When we speak of intimacy between two people we mean more than the mere fact of sharing the same house, the same bedroom or the same bed ; hence intimacy is ambiguous, both physical and affective. Although it is not recommended, in scientific research, to deal with ambiguous concepts, the timehonoured metaphor in Saussure’s Cours suggests a possible resolution. The two sides of the ambiguity, the sociological and the psychological, are somehow similar to those of his sheet of paper, and the still unsolved puzzle of langue v. parole. We are not dealing with the semantics of current lexical items − the meaning-shifts of a term like Romantic, for example − , but with a formal morphosyntactic system of deep significance, all the more so because it is largely unconscious. Its existential force has been emphasized in I.2.. Applied to the situation of letter-writing, it certainly connotes the opposite of distance : if the present tense is used, it is a super-present, so to speak . From this derive suggestions of warmth, sensibility, expressiveness, and all the subtle shades a writer like Jane Austen masterfully introduced in her novels, but also found in the works of many other writers. On the other hand, is there some connection between our analysis of the auxiliary system in I.2. and the extralinguistic modalities of its increase ? At a superficial level, a term like spontaneity has been suggested. It may help to encompass the two sections of the sociolinguistic study, which is, indeed, suffused with popular psychology. The link is, one has to confess it, that women are supposed to be more spontaneous than men 121, more sentimental perhaps, people say, to provoke feminists who will have it that the difference has been forced upon women. This extreme feminist view is clearly opposed to what is known of the image of the Victorian woman, an emblematic socio-historical avatar of the condition of woman across the ages : (Basch 1979). We have to acknowledge that, at 121 Possibly, it is rather the aptitude to express their emotions, in speech or in writing, which differentiates women, as recently suggested and documented by Braconnier 1996. This concerns the social psychology of the period: the supposed coolness of British males, which the French called "le flegme", is sometimes felt in the written exchange. Yet our Romantics are not very typical in this respect. 96 the very time when the feminist movement had already begun in Britain with Mary Wollstonecraft and later John Stuart Mill, Harriet Martineau, etc., women as independent and militant as George Eliot and Elizabeth Gaskell were far from refusing the traditional psychological attributes of femininity, and would not have found insulting the picture illustrated in Ruskin’s Sesame and Lilies, for example. It would be anachronic to project on most Victorian women our own ideology. If the suggestion of a link between gender (not sex : this is not just a concession to feminists) and what is called spontaneity is judged irrelevant and even provoking, then we can take things the other way round and say that the spontaneity which is also present in the intimate letters of men is an expression of feminine gender in them. Spontaneity is a term that somehow resists definition ; 122 it might be measured by the speed at which the pen ran on the paper : writers often allude to scribbling with bad pens, or to the trickeries of ink, and editors note the erased words, the blots, etc., characteristic of informal writing, at a time when writing tools were still very imperfect, and paper easily scratched. Unfortunately, we can only judge of it indirectly. There are formal signals : erratic punctuation, incomplete sentences, misspellings (Keats is famous for them), erased words, the desultory shifting from one topic to another, exclamations, hesitations ; all tokens of that minimal distance defining intimacy. Obviously, this is not a strict concept, especially when degrees have to be established, and its application is certainly loose and questionable, as we said in the related section. Spontaneity is clearly connected with intimacy, although it can be found with less intimate correspondents, and could certainly be divided into varieties. It is, in traditional terms, a stylistic concept, and we may easily admit that “ every one to his own style ” : Macaulay’s spontaneity is of a different kind from that of Thackeray, who was often joking, and George Eliot’s different from Mrs Gaskell’s, who, in her everyday correspondence, was less intellectual, more newsy ; the two Scotsmen, Scott and Carlyle, also had a different manner of addressing their intimate correspondents, the former usually relaxed, the latter − who so often complained of bowels trouble −always ready to lecture his correspondents. It is an interesting journey into human nature to read the letters of so remarkable people. To conclude, spontaneity is something that can rely on internal and external evidence : formal indices or symptoms, on the one hand, and what we learn from biographies and from the content of the letters themselves concerning human relations, on the other. The research presented here clearly shows the process of the spread. What would be its responses to the questions put forth by Labov (1982) ? For the short term, the constraints (or the motives) are psycho-sociological rather than purely formal. It is quite possible that long-term drift and short-term change obey different rules ; or that a shorttime spurt is just surfing on the long wave of syntactic and semantic renewal. This would describe the actuation of the change in this case : a reinforcement to which avalanche, overall societal change, dialectal and popular influences may have contributed. The whole study illustrates the transition. Even though we should not exaggerate its incidence (it never affects now more than 8 to 10 per cent of the total verb tokens), the progressive has become part and parcel of the verb system, no longer an exceptional idiomatic phrase. Is this not a sort of revolution, a modernization ? As to embedding, we see that, as usual, sociological factors converge with stylistic factors, with the use of the progressive increasing in the lower strata. The evaluation problem was considered only incidentally, but it is clear that prescriptive statements, negative as usual, did not inhibit the development, because it was an insidious and quantitative novelty. The form had existed long before, only its spread could be censured, apart from stereotypes. We may also ask ourselves if our illustrious writers were just carried along by the tide, or if they actually led the movement, and, if so, to what degree. The reading public was growing, and the influence may also have been indirectly transferred through various channels. This point should certainly be studied more closely, not only in the field of literature − many writers have insisted that they took their language lessons with the common people 123 − but also in the more general way of reciprocal influence between the community and the intellectual and social élite. Trying to discover such channels in the diffusion of the less conscious syntactic changes would be difficult, but certainly as rewarding as in the better-known field of vocabulary. 122 Under the label spontaneous speech, Labov 1994, 157-8 includes casual and careful speech, as opposed to the controlled styles used in reading. In letter-writing, this seems to apply as well, if we consider that there is a gradient, with individual variations and degrees of intimacy. Writing − especially at the time and with such writers − does not easily allow for the same casualness as speech, so that it is something between casual and careful expression. 123123 “One is not bound to respect the lazy obtuseness or snobbish ignorance of people who do not care to know more of their native tongue than the vocabulary of the drawing-room and the newspaper”: George Eliot to Walter W. Skeat, the famous dialectologist, in defence of her sometimes inaccurate use of the Midland dialect in Adam Bede (North Staffordshire and Derbyshire) and Silas Marner (North Warwickshire)(1872). 97 The progressive and the romantic world-view Are the actuation and diffusion observed just a contingent phenomenon, or is there a deeper connection with what the progressive periphrasis embodies into the verb-system ? Answering this is going from mere description to the study of causality. The following suggestions, in terms of external causality (cf. point 5 of Givon in b above), are then to be taken with a pinch of salt, and some will certainly find them too bold. Is it just a coincidence that the comparatively abrupt increase of the progressive is synchronous with the birth and development of Romanticism ? This analytic form, we have seen, stresses existence with its concreteness in the ever-changing, transitory picture of the world, just as John Robert Cozens, Constable and Turner’s landscapes are dedicated to passing lights and clouds. In the classical age, time and nature seemed immutable, and events were usually described by a verb-form sometimes ambiguous, but essentially synoptic. With the 18th century, a new picture emerges, leading to a greater interest on the individual, whether this means things, events, or people. When the speaker or the addressee are also the subjects of the sentence − the majority of cases − it goes with the new emphasis on self. The coincidence is probably not fortuitous, and it can be at least surmised that the romantic atmosphere − or should we say ideology ? After all, Kierkegaard was a Romantic − is not without influence on the actuation of a sleeping virtuality of the system of English grammar. The French philosopher Georges Gusdorf 19821993 : 394, in his extensive and masterly study of Romanticism, dedicates long developments to the spatio-temporal universe of the new age, as opposed to the epistemology of the classical age : “ la vérité intime de la conscience humaine sert de modèle à la vérité des choses extérieures... A cet espace [l’espace kantien, toujours euclidien ], le romantisme oppose un espace qualitatif, l’espace vécu de la présence ou de l’absence... Le temps est, comme l’espace, un exposant de la présence humaine, en fonction de la succession des événements. ” [ the intimate truth of human consciousness serves as the model to the truth of external things... In constrast to the kantian space, still a euclidean space, the space of romanticism is qualitative, is the space of actual experience, of presence or absence... Time, like space, is an exponent of the human presence, according to the succession of events ; my translation ] 124 This does not mean that it was limited to enlightened literate people : the reverse could be true. Enlightenment was an intellectual phenomenon, the sentimental impulse was not. Wordsworth, the great novelists, and many others of the contemporary intellectual elite, were interested in the common people. The fact that the population surveyed consists of writers may have led to this “ literary ” interpretation, but it must be stressed that romantic is understood here in a very comprehensive sense, as an atmosphere involving even the illiterate masses : encompassing for example the sentimental current of religion (Evangelism, non-conformism of various denominations including Methodists, Unitarians, Quakers, etc.) and the popular appeal of writers like Dickens. Warm seems an appropriate metaphor to qualify the popular new form. Stressing this expressive function, somehow neglected in the recent past, and referring to psychology is, obviously, indulging in neo-Humboldtian linguistics. Wilhelm von Humboldt was a friend of Schiller, and also a product of the Romantic age. He used the term Weltansicht (Humboldt 2000: 180), implying a less intellectualized view than Weltanschauung — cf. B. L. Whorf’s famous theory of linguistic relativism (Whorf 1956: 263: A change in language can transform our appreciation of the Cosmos.) Humboldt’s considerations have occasionally suggested some ridiculous interpretations on the national character of individual languages. Yet, in the narrower line of historical development which is ours — and was one of the major lines of Humboldt’s thought—, several students of the progressive, as we have seen, often referring to not strictly present-day usage, have underlined its “ subjective force ” ( Curme 1930 : 374 ; Mossé 1938 : 274 ; Granville-Hatcher 1951 ; Charleston, 1960 ; Storms,1964 ) and several others have more or less sensed it, at least as a by-product of its aspectual function (Scheffer, 1975 ; Strang, 1982 :449125). Studying the foregrounding function of the progressive, Wright 1994126 mentions some of the 124 Gusdorf remarks that British romanticism is in many ways specific, but this mostly applies to the literary sphere. With reference to the universe of the British landscape painters, here is another typical observation from a very different view-point : "both the Cotman and the Cox [watercolours] are typical products of the romantic period, which showed people becoming more aware of their place in the natural world". C. Nugent, 1993, 63 (my italics). To what extent are language forms and world view mutually conditioned is an old and fascinating question: cf. Benjamin Lee Whorf 's studies of the Hopi , Whorf 1956, 263. 125 She suggests that the development of the progressive may explain the transition from the epistolary novel to the modern novel: "what changes is the participation of the reader, who is made to feel as if he is seeing rather than being told". 126 She believes that the subjectivity found in writings, even in letter-writing , may well be just a reflection of the subjectivity always expressed in spoken form. Unfortunately, our historical investigations have to be concentrated 98 adjectives describing it in the extensive literature : emotive (Zandvoort), vivid (Jespersen), interpretative (Ljung 1977), phenomenal (Goldsmith & Woisetschlaeger 1976, 1982). These originally more or less intuitive judgments can certainly be applied to the 19th century rise of the progressive, provided that the term subjective is taken to refer not only to the subject of the sentence but to the speaker as well. The rate of the development of the progressive in preromantic, romantic and industrial times is so high that it certainly required a strong drive. The psychosociological forces of the romantic mind may have provided at least some of it. To resume a more professional discourse, we can repeat that traditional historical linguistics is certainly more concerned with the long term than with the short term ; but is in fact most of the time concerned with the middleterm, that is of the order of several centuries. But, while avoiding the traps where some of the earlier comparativists were caught, we are still aware that there are very long term trends of a general kind, diachronic universals : either formal (for example the development of analytic after synthetic forms ; Cf. Robin Lakoff 1972 ), or of a more abstract nature, such as semantic drifts affecting grammar (L. Anderson 1975 ; Traugott 1982). Boulle 1988,1996, proposed a “ theory of the volcano ”. According to it, deictic markers flow down historically like a lava from “ warm ” values − near the crater − to cold or cooler values, that is from situated forms, in close connection with the speaker (“ énonciateur ”) and the moment of speech, to aoristic meanings.127 The hypothesis is documented in many languages and is somehow parallel to the drift from marked forms to unmarked forms, resulting from the generalisation of the former, in a cyclic movement, proposed by the “ naturalness ” theories : cf. for example, Ch.J.N.Bailey 1973. Obviously, the progressive is situated and the simple form is usually aoristic. No surprise then if the progressive was first introduced in oral speech, where it still prevails as we are entitled to think. Both diachronic trends are possibly at work in the development of the progressive in the 19th century : the tendency to develop analytical, periphrastic forms 128 and the renewal of illocutionary force by the adoption of what is felt to be more definite, more concrete, more close to the situation of speech and the speaker, who is often the subject of the sentence as well. This, however, raises another thorny question. When she spoke of “ involvement of the subject ”, in her paper of 1951, an important landmark, Anna Granville Hatcher was taking the word subject in the usual sense of “ subject of the sentence ”. She was not alluding to the speaker, who is the subject of the utterance process ; nor to another subject, the analyst. Three levels have to be considered, in fact : the level of reference, the level of utterance, and the metalinguistic level. At the level of reference, speaking of 19th century usage, we can largely accept Hatcher’s definition of “ involvement of the subject ” : “ (1) the effect of the activity on the subject, (2a) his absorption in activity, or (2b) the results or aims of this activity. ” ; except that activity is a disputable term, as we have seen, in many common cases, and that we must also, especially in 19th century English, consider “ symmetric ” passives of the old type, where the subject of the sentence is actually the “ object ” (the patient) of the predication : our garden is putting in order (J. Austen). In many cases, the progressive also played the part of a “ middle voice ”. This is the first level of analysis. The second level is that of the speaker (or “ énonciateur ” in Culioli’s terms, to stress that it is a formal entity) : one cannot say that it is not involved in any statement “ it ” makes, since this is what speaking, or writing, implies. The third level is that of the analyst : the grammarian, for example. The second and third “ subjects ” are necessarily people ; only the first can be non-human and even non-animate. on the written documents, from which we can only draw inferences concerning the spoken language. Vast questions then arise as to mutual influences in modern times between the two. Letter-writing seems, of course, to encourage subjectivity more perhaps than any other situation. Linguistic change may well affect differently the different "genres", but it would seem that it is in the same direction. 127 There is a spatial, geographical counterpart to this: Paul Kay 1977 suggested that there are languages, mostly oral, which can do without forms disconnected from the actual situation of speech: the aorist is one of such forms. The notion of "primitive" languages can be renewed by this observation, simply emphasizing the enlargement of man's universe in terms of space and historical time. 128 The periphrastic perfect (passé composé), originally mostly a "present relevance" form in French gradually took the place of the aoristic simple past (passé simple) which was still common even in ordinary speech in the nineteenth century and more recently in some dialects. Was the modern world-view altered from this shift? This would be a parallel suggestion to that concerning the English progressive. 99 It would seem that those three levels are not always as separated as we should wish them to be.129 For convenience’s sake, let’s consider the third level subject, the subject of metalinguistic discourse. It is now well-accepted − Bloomfield 1944 already spoke of “ secondary and tertiary responses to language ” − that any speaker is also a linguist at times and more often than we believe. This involves a fairly frequent confusion between language and metalanguage, once responsible, for example, for the famous allusions to the “ life of language ” (“ la langue fait ceci ou cela” ) which Schuchardt denounced as well as Saussure : in words, but they did not easily dispense in practice with that sort of transfer, forgetting the activity of language users. So when someone says : “ You have been drinking ! ” it is not “ you ”, the subject of the sentence (S1), who is emphasizing the effect of drink. The reply may be : “ Yes, I drank six Martinis ” ; it is the speaker(S2) who is alluding to the condition of S1 instead of just referring to the act that originated it. Perhaps S1 will reply : “ You’re telling me ! ”, which would be reacting to the speaker’s statement rather than to the facts themselves : in which case, both parties would assume the part of S3, the metalinguistic subject.130 This is particularly true when human agents are subjects of the sentence, but can easily be extended by some sort of spontaneous animism − rather than mere grammatical analogy − to animals, and even what are called inanimates (tools, carriages, food, meteorological entities, etc.). It must also be remembered that speech is always dialogic, and this is still true in the absence of the interlocutor : letter-writing is an outstanding case of the absence-presence of the other party. The confusion, or rather the unpredictable shifting between levels, is more frequent between the subject of the sentence and the speaker, subject of the utterance process. First because the two are often the same (first person). Second because there is a often a transfer, which we can simply describe by a word like “ interest ”, the opposite of “ detachment ”. This may be how intimacy is correlated with what we call spontaneity. In other terms, we are led to use once more what has long been considered by linguists as almost an obscene word : expression. The emphasis on communication (even when it is not reduced to the cognitive component, still a common fallacy) sometimes forgets that language (parole) is not limited to it. 131 In his often-quoted paper of 1958, Roman Jakobson (in Sebeok 1960), after Buehler, gave a description of the different functions of speech (cf. also Labov 1982). Besides phatic function, which certainly accounts for much of letter-writing, yesterday, and telephoning today, we cannot deny the importance of expression in it. So that a term like spontaneity will stress the expressive value, whereas intimacy refers to the fact that it is also communication, but often not purely cognitive. When we are studying parole instead of langue, we must consider that speakers are human beings who have not been trained to Cartesian methods and are inclined to syncretism and transfer. The tentative remarks above are perhaps a way of considering the statement by Saussure himself that all that is in la langue has first been in la parole, and one of the often-quoted remarks by Schuchardt that, whereas synchronic linguistics is the linguistics of la 129 On this point I am indebted to Jean-Blaise Grize and his approach to everyday natural logic with the concept of "schematization": see for example Grize 1982,1990. It must also be noted that in the corpus studied, 41 % of the occurrences have first person subjects. 130 This might be further formalized. Some generativists once postulated an original performative of "saying" at the origin of derivations. In any case, trying to develop the hidden movements of the mind leading to an utterance is an endless task. Much of the intricate network of notions and mental operations has been called "préconstruit" by Culioli, something in which physical and also psychosociological factors obviously play a large part, not easily formalized with the tools of conventional logic. And it seems that ordinary logic is far from Aristotelian: on this again, see J.-B. Grize, passim. 131 This was typical of the structuralist school (Hockett, etc.) and and has been successfully challenged since then. 100 langue, diachronic linguistics is the linguistics of la parole, another way of summing up the suggestions of Empirical foundations. This is perhaps going too far and confusing the short term and the middle and long term. Our purpose was not to cover the long-term drift where the virtualities of system have to be described in more abstract terms, but to throw light on the actuation, or simply on the acceleration ; and the above suggestions are just an attempt at connecting the two. 101 BIBLIOGRAPHY of the collections of letters AUSTEN : Chapman, R.W. (1932). Jane Austen’s Letters to her sister Cassandra and others. Oxford : Clarendon Press. [One vol. ed. : 1952]. BROWNING, Robert & BROWNING, Elizabeth Barrett : Wise, T.J. (1933). Letters of Robert Browning. New Haven : Yale University Press. XX,389. [ 1973 reprint : New York, London : Kennikat Press.] Berridge, Elizabeth. (1974). The Barrets at Hope End. The Early Diary of Elizabeth Barrett-Browning. London : John Murray : XIII, 276. Kintner, Elvan. (1969). The Letters of Robert Browning and Elizabeth Barrett-Barrett, 1845-1846. Cambridge, Mass. : The Belktrap Press of Harvard University Press. I : XVIV,568 ; II : VII, 1119. Huxley, Leonard. (1929). Elizabeth Barrett-Browning : Letters to her Sister, 1846-1859. London : John Murray : XXV,334. Miller, Betty. (1954). Elizabeth Barrett to Miss Mitford. London : John Murray : XVII, 284. Heydon, Peter.N. ; Kelley, Philip. (1974). Elizabeth Browning’s Letters to Mrs. David Ogilvy, 1849-1861. London : John Murray : 174 p. McCarthy, Barbara. P. (1955). Elizabeth Barrett to Mr. Boyd. London : John Murray : XXIX, 299. CARLYLE : Sanders, Charles R. & Fielding, Kenneth J. . (1970-). The Collected Letters of Thomas and Jane Welsh Carlyle. Duke Edinburgh Edition. Durham, North Carolina : Duke University Press. Vol. 1 : 1812-1821 : IX, 421 ; Vol.2 : 1822-1823 : XIII, 497 ; Vol. 3 : 1824-1825 : 446p ; Vol. 4 : 18261828 : XV, 493 ; Vol. 5 : 1829-Sept.1831 : XII,448 ; Vol. 6 : Oct. 1831-Sept. 1833 : 446p. ; Vol. 7 :Oct. 1833-Dec. 1834 : XV,442 ; Vol. 8 : Jan.1835-June 1836, 365 p. ; Vol. 9 : July 1836-1837, 434 p. ; Vol.10 : 1838-1840, 258 p. Marrs, Edwin. W. (1968). The Letters of Thomas Carlyle to his Brother Alexander, with related family letters. The Belknap Press of Harvard University Press : Cambridge, Mass. XII, 830. COLERIDGE : Griggs, Earl Leslie. (1956-71). Collected Letters of Samuel Taylor Coleridge. Oxford : Clarendon Press. 6 vols. Vol. 1 : 1785-1800, xxxix, 659 ; Vol. 2 : 1801-1806, viii, 661-1219 ; Vol.3 : 1807-1814, lii, 544 ; Vol. 4 : 18151819, xi, 545-1000 ; Vol. 5 : 1820-1825, lxi, 529 ; Vol. 6 : 1826-1834, x, 531-1089. DICKENS : House, M. ; Storey, G.(1965-) The Letters of Charles Dickens. The Pilgrim Edition. Oxford : Clarendon Press.I. : 1820-1839, XLV,744 ; II. 1840-41, 547 p. : ; III : 1842-1843, 692 p. ; IV : 1844-1846, 771 p. ; V : 1847-1849, 753.p. Anonymous.(1980) The Letters of Charles Dickens, edited by his sister-in-law and his eldest daughter.London : Chapman & Hall. [Second edition]. 2 vols. EDGEWORTH : Hare, A.J.C. (1894). The Life and Letters of Maria Edgeworth.[Reprinted 1971] : Freeport, N. Y. : Books for Libraries Press ; 2 vols . I. 339 p. ; II. 341-704. 102 Colvin, Christian (1971). Maria Edgeworth : Letters from England (1813-1844). Oxford : Clarendon Press. 103 ELIOT : Haight, G. S.(1954). The G. Eliot Letters. New Haven : Yale University Press. 9 vols.Vol 1 : 1836-1851 : XXVII, 378 ; Vol. 2 : 1852-1858 : 513p. ; Vol. 3 : 1859-1861 : 475p. ; Vol. 4 : 1862-1868 : 502p. ; Vol. 5 : 18691873 : 475p. ; Vol. 6 : 1874-1877 : 440p ; Vol. 7 : 1878-1880 : Vol. 7, 523 ; Vols. 8-9 : Supplements. GASKELL : Chapple, J. A. V. ; Pollard, Arthur. (1966). The Letters of Mrs. Gaskell. Manchester : Manchester University Press, XXIX, 1010. HUTCHINSON : Coburn, K. (1954).The Letters of Sara Hutchinson from 1800 to 1835. London : Routledge & Kegan Paul ; XXXVIII, 474. KEATS : Rollins, Hyder Edward. (1958). The Letters of John Keats. 2 vols : I. 1814-1818, XXII,442 ; II. 1819-1821, XIII, 440. Cambridge : Cambridge University Press. Gittings, R. (1970). Letters of John Keats. A New Selection. Oxford : Oxford University Press. [Reprinted with corrections, 1975]. MACAULAY : Pinney, Thomas. (1974-1981) The Letters of Thomas Babington Macaulay. The Claremont College, Claremont, California : Cambridge University Press ; 6 vols. I : 1807-Feb. 1831, 320p. ; II : March 1831-1833, 377 p. ; III : 1834-1840, 387 p. ; IV : 1841-1848, 403 p. ; V :1849-Dec. 1855, XIV, 984 ; VI : 1856-1859, XII, 484. RUSKIN : Van Akin Burd, J. (1973).The Ruskin Family Letters. The Correspondence of John James Ruskin, his Wife and their Son. 2 vols. I. lviii, 417 ; II. 421-792. Ithaca & London : Cornell University Press. Shapiro, Harold E. (1972). Ruskin in Italy. Letters to his Parents, 1845. XXII, 263. Oxford : Clarendon Press. / Bradley, J.L. (1955). Ruskin’s Letters from Venice, 1851-1852. New Haven : Yale University Press, XX, 330. SCOTT : Grierson, H. J. C., assisted by Davidson Cook, W.M. Parker and others (1932). The Letters of Sir Walter Scott. London : Constable & Co. [Centenary Edition ; 12 vols.] : Letters from different periods have sometimes been included in a given volume. Vol. I (1790-1807) : XCV, 531 ; Vol. III (1811-1814) : XV, 542 ; Vol. VII (1821-1823) : XXII,511 ; Vol. XI (1828-1831) : XVIII, 496 ; Vol. XII (1831-1832) : XII, 520. SOUTHEY : Curry, Kenneth. (1965). New Letters of Robert Southey. New York & London : Columbia University Press ; 2 vols. I : 1792-1810: XIX, 551 ; II : 1811-1838 : XIX, 552. 104 THACKERAY : Ray, Gordon N. .(1945). The Letters and Private Papers of William Makepeace Thackeray. London : Oxford University Press. 4 vols.I : 1817-1840 ; II : 1841-1851 ; III : 1852-1856 ; IV : 1857-1863. WELSH : see Carlyle. WORDSWORTH, Mary Burton, M. A. (1958). The Letters of Mary Wordsworth, 1800-1855 Oxford : Clarendon Press : XXIX, 363. WORDSWORTH, William & Dorothy : Selincourt, Ernest. de (1967). The Letters of William and Dorothy Wordsworth. Oxford : Clarendon Press. Vol.I : The Early Years 1787-1805 : XXXIII, 729. [Revised by Chester. L. Shaver]. Vol. II, 1. The Middle Years : 1806-1811 : XXVI, 546.[Revised by Mary Moorman]. Vol. II, 2 . The Middle Years : 1812-1820 : XX, 691. [Revised by Mary Moorman and Allan G. Hill]. Vol. III, 1. The Later Years : 1821-1828 : XXXII, 730. [Revised, arranged and edited by Allan G. Hill]. Vol. V, 2. The Later Years : 1829-1834 : XXII, 789. [Revised, arranged and edited by Allan G. Hill]. Selincourt, Ernest de (1939-1988). The Letters of William and Dorothy Wordsworth. The Later Years, Vol. 2 : 1831-1840. Oxford, Clarendon Press : XIV, 545-1059. The Letters of William and Dorothy Wordsworth.VI. The Later Years. Part III : 1835-1839, XXV, 794.VII. The Later Years. Part IV : 1840-1853, XVII, 951. [Revised, arranged and edited by Allan G. Hill from the first edition by the late Ernest de Selincourt]. Oxford: Clarendon Press. 105 REFERENCES La philosophie analytique, Cahiers de Royaumont 1962. Paris : Editions de Minuit Aitchison, J. 1981. Language change : progress or decay ? London : Fontana. Aitken, A.J. ; McArthur, Tom 1979. Languages of Scotland. Edinburgh : Chambers. Adamczewski, Henri 1978. Be + -ing dans la grammaire de l’anglais contemporain.Paris : Champion (Thèse de doctorat d’Etat 1976). Akerlund, A. 1911 On the History of the Definite Tenses in English. Cambridge & Lund. Anderson, J.M. 1982. Language form and linguistic variation. Papers dedicated to Angus McIntosh. Amsterdam : John Benjamins. Anderson, Lloyd 1975. Uniting three traditions : the classical study of grammatical-category space, and lexical decomposition in generative semantics, in R.W. Fasold & R.W. Shuy, (eds.) Analyzing Variation in Language, Washington. D.C. : Georgetown University Press, 241-68. Archer-Jackson, John 1963. The Irish in Britain. London : Routledge and Kegan. Aristotle : Metaphysics 6/1048b 18-35 ___________ : Nicomachean ethics 1174 a 29 b 5 ___________ 1994. Traité du temps. Introduction, traduction et commentaires par Catherine Collobert. Paris : Kimé. Arnaud, René 1973. La forme progressive en anglais du XIXe siècle. Lille : service de reproduction des thèses (Doctorat d’Etat, Université Paris VII). [Out of print.] ________________ 1979. Regard sur la variation syntaxique : la forme progressive dans la correspondance de George Eliot (1838-1880). Cahiers Charles V, 1, 105-18. ________________ 1980. Quelques observations quantitatives “ en temps réel ” sur un changement : l’accroissement d’emploi de la forme progressive dans la 1ère moitié du XIXe siècle. Actes du Congrès de Poitiers de la Société des Anglicistes de l’Enseignement supérieur. Etudes Anglaises N° 90. Paris : Didier Erudition, 67-75. ________________ 1983. On the progress of the progressive in the private correspondence of famous British people (18001880), in : Sven Jacobson, (ed.) Papers from the Second Scandinavian Symposium on syntactic variation (Stockholm, May 15-16, 1982), 83-94. _______________ 1983. Les règles, le jeu et les règles du jeu. Quelques réflexions méthodologiques sur l’étude sociohistorique de la variation. Modèles linguistiques, V,1, 17-62. _______________ 1998. The development of the progressive in 19th century English: A quantitative survey. Language Variation and Change, 10, 123-152. _______________ 2000. A propos du schéma ETRE+PARTICIPE PRESENT dans quelques versions du Nouveau Testament. Modèles et traductions diverses. Cycnos 17. 203-12 No spécial 2000. (Journées Charles V sur les propositions relatives et l’aspect be+ing). Université de Nice. Bailey, Ch.-James N. 1973. Variation and linguistic theory. Arlington, Virginia : Center for Applied linguistics. Bailey, Charles-James N. ; Shuy, Roger W. (eds.). 1973. Washington, D.C. : Georgetown University Press. New Ways of Analyzing Variation in English. Bailey, Guy ; Winkle, Tom ; Tillery, Jane ; & Sand, Lori. 1991. The apparent-time construct. Language Variation and Change 3 : 241-264 Bain, Alexander 1863. Higher English Grammar. London: Longmans, Green and Co. ———————1870. Logic. London : Longmans, Green and Co. Basch, Françoise 1974. Relative Creatures. Women in Society and the Novel. London : Longman (Penguin). 106 ___________________ 1979. Les femmes victoriennes. Paris : Payot Bennett, Michael ; Partee, Barbara 1972. Toward the logic of tense and aspect in English. System Development Corporation, Santa Monica, Calif.; Bloomington : Indiana University Linguistic Club. Benveniste, Emile 1966. Problèmes de linguistique générale. Paris : Gallimard. Bergson, Henri 1896. Matière et mémoire. Paris : Presses Universitaires de France. (1941 edition). Best, Geoffrey, 1971. Mid-Victorian Britain 1851-75. Bungay, Suffolk : Weidenfield and Nicolson. (Revised ed.: Panther Books, 1973) Biber, Douglas 1995. Dimensions of register variation : A cross-linguistic comparison. Cambridge : Cambridge University Press. Bloomfield, Leonard 1944. Secondary and tertiary responses to language. Language 20, 45-55. Bolinger, Dwight 1971. The Nominal in the Progressive. Linguistic Inquiry II, 2, 246-250. Boulle, Jacques 1988. La théorie du volcan; in : M.-L. Groussier (ed.). Pourquoi la linguistique diachronique ? Paris Université Paris VII-Denis Diderot Boulle, Jacques 1996 (forthcoming). Le renouvellement des systèmes aspectuels. Thèse de doctorat, Université Paris VII-Denis Diderot. Bourdieu, Pierre 1982. Ce que parler veut dire. Paris : Fayard. Braconnier, Alain 1996. Le sexe des émotions. Paris : Odile Jacob. Brinton, Laurel J. 1988. The development of English aspectual systems. Cambridge Studies in Linguistics 49. Cambridge : Cambridge University Press. Bull, William E. 1968. Time, tense and the verb. Berkeley : University of California Press. Buyssens, Eric 1958. Les deux aspectifs de la conjugaison anglaise au XXe siècle. Etude de l’expression de l’aspect. Bruxelles, Paris : Presses Universitaires de Bruxelles; Presses Universitaires de France. Bybee, J. ; R. Perkins & W. Pagliuca 1994. The evolution of grammar : Tense,Aspect and Modality in the Languages of the World. Chicago & London : The University of Chicago Press. Cerquiglini, Bernard 1991. La naissance du français. Paris : PUF (Collection Que Sais-Je ?). Chambers, J. K. 1995. Sociolinguistic theory. Linguistic variation and its social significance. Oxford, U.K ; Cambridge, Mass. : Blackwell. Charleston, Britta Marian 1941. Studies on the syntax of the English verb. Swiss Studies in English II, X, 209. _______________________________ 1955. A reconsideration of the problem of time, tense and aspect in English. English Studies XXXVI, 1, 263-78. _______________________________ 1960. Studies on the emotional and affective means of expression in modern English. Berne : Francke Verlag. Chartier, Roger (ed.) 1991. La correspondance. Les usages de la lettre au XIXe siècle. Paris : Fayard. Close, R.A. 1959. Concerning the Present Tense. English Language Teaching XIII, 57-66. Cobb, Allan 1973. The present progressive periphrasis and the metaphysics of Aristotle. Phronesis, A Journal for Ancient Philosophy. Assen : Van Gorcum 18, 80-90. Collobert, Catherine 1994. Aristote, Traité du temps. Physique, Livre IV. Introduction, traduction et commentaire. Paris : Kimé. Comrie, Bernard 1976. Aspect. Cambridge : Cambridge University Press. Culioli, Antoine 1981. Pour une linguistique de l’énonciation. Opérations et représentations. Paris : Ophrys. ___________________ 1995. Cognition and Representation in Linguistic Theory. Texts, selected, edited and introduced by Michael Liddle, translated with the assistance of John T. Stonham. Amsterdam, Philadelphia : John Benjamins. Curme, George O. 1931. A Grammar of the English Language. Vol. III. Syntax. Boston : D.C. Heath & Co. 107 Dahl, Östen 1981. On the definition of the telic-atelic (bounded-nonbounded) distinction. Syntax and semantics 14, 79-90. DeCamp, David 1972. What do implicational scales imply ? in Ch.-J.N. Bailey ; Roger Shuy (eds.). New ways of analyzing variation in English. Washington, D.C. : Georgetown University Press. Denison, David 1993. English historical syntax: Verbal constructions. London and New York: Longman. —————— 1998. Syntax ; in Suzanne Romaine (ed.). Cambridge History of the English Language IV. Cambridge : Cambridge University Press. Desclaire, Violette 1973. Langue et société. Etude méthodologique : le style des correspondances féminines sous les deux derniers Stuarts (1688-1714). Thèse de doctorat d’Etat, Université Paris VII-Denis Diderot. Diver, William 1963. The Chronological System of the English Verb. Word 19-2 ,141-181. Dowty, David 1972. Studies in the logic of verb aspect and time reference in English. Studies in Linguistics 1, Department of Linguistics, University of Texas at Austin. _________________ 1977. Towards a semantic analysis of verb aspect and the English imperfective progressive. Linguistics and philosophy, 1, 45-77. Eckert, Penelope 1989. The whole woman : Sex and gender differences in variation. Language Variation and Change 1 ,245-268. Ellegard, Alvar 1953. The auxiliary do. The establishment and regulation of its use in English. Stockholm : Almqvist and Wiksell. Encrevé, Pierre ; de Fornel, Michel ; et al. 1983. L’usage de la parole. Actes de la recherche en sciences sociales 46. Paris : Editions de Minuit. François, Jacques 1983. La résurgence des catégories aristotéliciennes energeia/kinesis dans les discussions actuelles sur le mode d’action télique. Suivi d’une bibliographie commentée. DRLAV 28, 79-96. Fuchs, Catherine (ed.) 1991. Les typologies de procès. Paris : Klincksieck. Gauchat, Louis 1905. L’unité phonétique dans le patois d’une commune. in Aus Romanischen Sprachen und Literaturen : Festschrift Heinrich Morf, 175-232. Gérin, Winifred 1976. Elizabeth Gaskell. A biography. Oxford : Clarendon Press. Givon, T. 1989. Mind, Code and Context. Essays in Pragmatics. Hillsdale, N.J.-London : Lawrence Erlbaum Associates. Goldsmith, John ; Woisetschlaeger, Erich 1976. University Linguistics Club. The Logic of the Progressive Aspect. Bloomington : Indiana _________________________________________________ 1976. The logic of the English progressive. Linguistic Inquiry 13 , 79-89 Goux, Henri-Pierre 1971. A statistical study of the progressive form in today’s English pop-music. Mémoire de maîtrise, Université Paris VII-Denis Diderot. (Unpublished). Granville Hatcher, Anna 1951. The use of the progressive form in English. A new approach. Language 27,3, 25480. Grize, Jean-Blaise 1982. De la logique à l’argumentation. Genève-Paris : Droz. ______________________ 1990. Logique et langage. Paris : Ophrys Groussier, Marie-Line (ed.) 1988. Pourquoi la linguistique diachronique ? (Table ronde 24 Oct. 1987). Paris : Publications de l’association des médiévistes anglicistes de l’enseignement supérieur 13. Groussier, M.L. ; Rivière, C. 1996. Les mots de la linguistique. Lexique de linguistique énonciative. Paris : Ophrys. Guillemin-Flescher, Jacqueline 1977. Analyse comparée des catégories grammaticales et des opérations de référence en français et en anglais. Thèse de Doctorat d’Etat, Université Paris VII-Denis Diderot. 108 ____________________________________ 1981. Syntaxe comparée du français et de l’anglais. Problèmes de traduction. Paris : Ophrys Guy, Gregory 1990. The sociolinguistic types of linguistic change. Diachronica 7 , 47-67. Haight, Gordon S. 1968. George Eliot. A biography. London, New York : Oxford University Press ; (Reprinted in Penguin Books 1992). Hancil, Sylvie 2000. Beon + ende en vieil-anglais. Cycnos 17. 91-104. No spécial 2000. (Journées Charles V sur les propositions relatives et l’aspect be+ing). Université de Nice. Hatcher, Anna Granville 1951. The use of the progressive form in English. A new approach. Language 27,3, 25480. Heine, Bernd. 1994. Grammaticalization as an Explanatory Parameter, in W. Pagliuca (ed.) Perspectives on Grammaticalization. Amsterdam/Philadelphia : Benjamins : 255-287. Herdan, Gustav 1964. Quantitative linguistics. London : Butterworths. ___________________ 1966. The advanced theory of language as choice and chance. Berlin, Heidelberg, New York : Springer. Hill, Archibald 1958. Introduction to linguistic structures. New York : Harcourt, Brace. Holmes, Richard 1989. Coleridge. Early Visions. London :Hodder & Stoughton . ( Penguin Books 1990). Humboldt, Wilhelm Von 2000. Sur le caractère national des langues et autres écrits sur le langage. Présentés, traduits et commentés par Denis Thouard. Bilingue. Paris: Editions du Seuil. Imbs, Paul 1960. L’emploi des temps verbaux en français moderne. Paris : Klincksieck. Jacobson, Sven (ed.) 1983. Papers from the second Scandinavian symposium on syntactic variation. Stockholm : Almqvist & Wiksell. Jakobson, Roman 1958. Linguistics and poetics, in : Ths. Sebeok (ed.). Style in language. Cambridge, Mass. : M.I.T. Press, 350-77. Jespersen, Otto 1931. A modern English grammar on historical principles, Part IV. London : Allen and Unwin (Reprinted 1961). Joos, Martin 1964. The English verb. Madison : The University of Wisconsin Press. Juillard, Michel 1983. L’expression poétique chez Cecil Day Lewis. Vocabulaire,syntaxe, métaphore. Etude stylostatistique. Genève : Slatkine. Kastovsky, Dieter. 1994. Studies in Early Modern English. Berlin ; New York : Mouton de Gruyter. Kay, Paul 1977. Language evolution and speech style, in : B.G.Blount & M. Sanches (eds.). Sociocultural dimensions of language change. New York : Academic Press, 255-68. Kenny, Anthony 1963. Action, emotion and will. London : Routledge and Kegan Paul. König, E. 1980. On the Context-Dependence of the Progressive in English, in : Roher, C. (ed.) Time, Tense and Quantifiers : Proceedings of the Stuttgart Conference on the Logic of Tense and Quantification, 269291.(Linguistische Arbeiten 83.) Tübingen : Max Niemeyer Verlag. Kipers, Pamela 1987. Gender and topic. Language in Society 16,4 , 543-557. Kytö, Merja ; Rissanen, Matti ; Wright, Susan 1994. Corpora across the centuries. Proceedings of the First International Colloquium on English Diachronic Corpora : St Catharin’s College, Cambridge, 25-27 March 1993. Amsterdam-Atlanta, GA : Rodopi. Klein, Flora 1980. A quantitative study of syntactic and pragmatic indications of change in the Spanish of bilinguals in the U.S., in : Wm Labov (ed.). Locating language in time and space. New York, London : Academic Press. Knott, G.H. (ed.) 1952. Trial of William Palmer. Edinburgh & London : William Hodge & Co. (Notable British Trials Series). 109 Koschmieder 1935. Zu den Grundfragen der Aspecttheorie. Indogermanische Forschungen 53, 280-300. Kroch, Anthony 1989. Function and grammar in the history of periphrastic “ do ”, in R. Fasold and R. Schiffrin (eds.). Language Variation and Change. Orlando : Harcourt Brace Jovanovich.133-72. Labov, William 1980 (ed.) Locating language in time and space. New York, London : Academic Press. __________________ 1981. Resolving the neogrammarian controversy. Language 57 , 267-309 (1981). 1982. Building on empirical foundations, in W.P.Lehmann; Y. Malkiel, (eds.). Perspectives on historical linguistics. Amsterdam : Benjamins. __________________ _________________ 1994. Principles of linguistic change. Vol. 1 : Internal factors. Oxford, U.K. ; Cambridge, U.S.A. : Blackwell. _________________ 2001. Principles of linguistic change. Vol. 2 : Social factors. Oxford, U.K. ; Cambridge, U.S.A. : Blackwell. Lakoff, Robin 1973. Language and woman’s place. Language in Society 2,1, April 1973. Lawton, R. 1959. The Irish Community in Liverpool in 1851. Irish Geography, IV, No.1. Leech, Geoffrey N. 1971. Meaning and the English verb. London : Longman. Leonard, Sterling A. 1929. The doctrine of correctness in English usage 1700-1800. New York : Russell and Russell, (1962 reprint). Leskien, A. 1876. Die declination im slavisch-litautschen und germanischen. Leipzig. Ljung, Magnus 1980. Reflections on the English Progressive. Göteborg : Acta Universitatis Gothonburgensis (Gothenburg Studies in English 46). Marchello-Nizia, C. 1988. L’argumentation sur une langue de corpus, in M.L. Groussier,(ed.). : Pourquoi la linguistique diachronique ? Paris : Publications de l’association des médiévistes anglicistes de l’enseignement supérieur13, 78-97. Mayhew, Henry 1862. London labour and the London poor. London : Frank Cass. (New edition. 1967). Milroy, Lesley 1987. Language and social networks. London and New York : Blackwell. Mossé, Fernand 1938. Histoire de la forme périphrastique être + participe présent en germanique :I. 1ère partie : Ancien germanique ; II. 2e partie : Moyen Anglais et Anglais Moderne. Paris : Klincksieck (Collection linguistique publiée par la Société de linguistique de Paris. XLIII). Mourelatos, Alexander P. D. 1981. Events, processes, and states. Syntax and semantics 14, 191-212. Muller, Charles 1964. Essai de statistique lexicale; L’illusion comique de P. Corneille. Paris : Klincksieck. __________________ 1968. Initiation à la statistique linguistique Paris : Larousse. Murray, Lindley 1795. An English grammar comprehending the Principles and Rules of the Language illustrated by Appropriate Exercises and a Key to the Exercises. York. (5th ed. 1824) Mustanoja, Tauno 1960. A Middle English syntax. Part I. Parts of speech. Helsinki : Société néo-philologique. Nakamura, Fujio.1998 A word on the history of the English progressive passive. Tenth International Conference on English Historical Linguistics, Manchester, 26 August. Nehls, Dietrich 1974. Syncron-diachrone Untersuchungen zur Expanded Form im Englischen. München : Max Hueber. Nevalainen, Terttu ; Raumolin-Brunberg, Helena (eds.). 1996. Sociolinguistics and language history Studies based on the Corpus of Early English Correspondence. Amsterdam-Atlanta, GA. : Rodopi. Nickel, Gerhard (1966). Die “ Expanded Form ” im Altenglischen. Neumunster. Normand, C. ; Caussat, P. ; Chiss, J.-L. ; Médina, J. ; Puech, C. ; Radzinski, A. 1978. Avant Saussure. Choix de textes (1875-1924). Bruxelles : Editions Complexe. 110 Nugent, Charles 1993. From View to Vision. British watercolours from Sandby to Turner in the Whitworth Art Gallery. Manchester : The Whitworth Art Gallery. Nurmi, Arja 1996. Periphrastic DO and BE+ING : Interconnected developments ? in Nevalainen & RaumolinBrunberg (eds.). 1996. Sociolinguistics and language history. Studies based on the Corpus of Early English Correspondence. Amsterdam-Atlanta, GA. : Rodopi. _______________ 1999. A social History of Periphrastic DO. Mémoires de la Société Néophilologique de Helsinki LVI. Ota, Akira 1963. Tense and aspect of Present-day American English. Tokyo : Kenkyusha. Page, Norman 1972. The language of Jane Austen. Oxford : Blackwell. Pagliuca, William (ed.). 1994. Perspectives on Grammaticalization. Amsterdam/Philadelphia : Benjamins. Palmer, F. R. 1974. The English verb. London : Longman. Perelman, Ch. 1962. In : La Philosophie analytique, p.294. Perlmutter, David 1978. Impersonal passives and the inaccusative hypothesis. Proceedings of the Fourth Annual Meeting of the Berkeley Linguistics Society : The University of California at Berkeley. Phillipps, K.C. 1970. Jane Austen’s English. London : Andre Deutsch. ——————1978. The Language of Thackeray. London : Andre Deutsch. ——————1984. Language and Class in Victorian England. Oxford : Blackwell. Raybould, Edith 1957. On Jane Austen’s use of expanded verbal forms. One more method of approach to the problems presented by these forms. Festschhrift Brunner, Wiener Beitrage zur Englishchen Philologie, 175-90. Raumolin-Brunberg 1996. Apparent time, in : Nevalainen & Raumolin-Brunberg (eds.). Sociolinguistics and Language History. Amsterdam-Atlanta : Rodopi, 93-109. Röhrer, C. (ed.) 1980. Time, Tense and Quantifiers : Proceedings of the Stuttgart Conference on the Logic of Tense and Quantification, 269-291. (Linguistische Arbeiten 83.). Tübingen : Max Niemeyer Verlag. Romaine, Suzanne 1982. University Press. Socio-historical linguistics. Its status and methodology. Cambridge : Cambridge _________________ (ed.) 1998. Cambridge History of the English Language IV. Cambridge : Cambridge University Press. Rostand, François 1951. Grammaire et affectivité. Paris : Vrin. Roughead, William 1950. Trial of Jessie M’Lachlan. Edinburgh & London : William Hodge & Co. Ruwet, Nicolas 1988. Les verbes météorologiques et l’hypothèse inaccusative, in Hommage à la mémoire de Jean Stéfanini. Aix en Provence : Publication de l’Université de Provence Rydén, Mats ; Brorström, Sverker 1987. The Be/Have variation with intransitives in English. Stockholm : Almqvist & Wiksell. Rydén, Mats 1979. An introduction to the historical study of English syntax. Stockholm : Almqvist & Wiksell International. Sag, Ivan A. 1972. On the state of progress on progressives and statives, in Ch.-J. N. Bailey ; Roger Shuy (eds.). New ways of analyszing variation in English. Washington, D.C. : Georgetown University Press. Samuels, M.L. 1972. Linguistic evolution with special reference to English. Cambridge : Cambridge University Press. Sankoff, David (ed.) 1978 Linguistic variation. Models and methods. New York : Academic Press. Sankoff ; D. ; Labov, W. 1979. On the uses of variable rules. Language in Society 8, 2, 189-222. Sankoff, Gillian 1973. Above and beyond phonology in variable rules, in : Bailey, C-J.N. & Shuy R. (eds). New ways of analyzing variation in English. Washington, D.C. : Georgetown University Press. Scheffer, Johannes 1975. The progressive in English. Amsterdam : North Holland ; Oxford : American Elsevier (North Holland Linguistic Series). Searle, John 1969. Speech Acts. Cambridge : Cambridge University Press. Sebeok, Thomas 1960. Style in language. Cambridge, Mass. : M.I.T. Press. 111 Sjoestedt, M.L. 1926. L’aspect verbal et les formations à affixe nasal en celtique. Paris : Champion (Collection Linguistique de la Société de Linguistique de Paris). Smith, Arvid 1917. Ueber eine bisher unbeachtete funktion des progressiven form, Beiblatt zur Anglia 28. 244-51 Smith, Carlota S. 1983. A theory of aspectual choice. Language 59,3, 479-501. Smith, Sheila M. 1974. John Overs to Charles Dickens : a working-man’s letter and its implications. Victorian studies, dec. 1974, 195-217. Smitterberg, Erik 1998. “ The sun is shining, the air is delicious ! I like the climate of Manchester ! ! ! the progressive form and genre variation during the 19th Century. 10th International Conference on English Historical Linguistics, Manchester, 21-26 August 1998. _____________________ 2000. Pragmatic Functions of the Progressive Form in the 19th Century. 11th Conference on English Historical Linguistics, Santiago de Compostela, 7-11 Sept. 2000. ______________________ (2002). The Progressive in 19th Century English. Dissertation, Uppsala University. A Process of Integration. Doctoral Stanzel, Franz 1957. Die Erzänsituation und die Umschriebenen Zeitformen. Festschift Brunner, Wiener Beitrage zur Englischen Philologie, 220-31. Strang, B.M. 1982. Some aspects of the history of the be+ing construction, in J. Anderson, (ed.). Language form and linguistic variation. 427-74. Swadesh, Morris 1952. Lexicostatistic dating of prehistoric ethnic contacts, in Proceedings of the American Philosophical Society, 96, 462-463. Tagliamonte, Sali 1998. Was/were variation across the generations: View from the city of York. Language Variation and Change 10, 153-191. Taine, Hippolyte 1863. Histoire de la littérature anglaise. Paris. Tarde, Gabriel de 1895. Les lois de l’imitation. Paris, Genève : Slatkine Reprints 1979. Tieken-Boon van Ostade, Ingrid. 1987. approach. Dordrecht : Foris. : The auxiliary Do in Eighteenth-Century English. A sociohistorical Thoiron, Philippe 1979. Index et concordance pour “ Alice’s Adventures in Wonderland ”. Problèmes, méthodes, analyse statistique de quelques données. Thèse de doctorat d’Etat, Université Paris VII-Denis Diderot. Thompson, E.P. 1963. The making of the English working class. Harmondsworth, England (Penguin Books). Traugott, Elizabeth Closs 1982. From propositional to textual and expressive meanings : Some semantic aspects of grammaticalization, in W.P. Lehmann & Y. Malkiel (eds.). Perspectives on historical linguistics. Amsterdam : Benjamins : 245-271. Traugott, Elizabeth Closs & Heine, Bernard (eds.) (1991). Approaches to Grammaticalization. Amsterdam Benjamins. Van Hamel, A.G. 1912. On Anglo-Irish Syntax. Englische Studien Bd. 45 Vendler, Zeno 1957. Times and tenses, Philosophical Review. Vendler, Zeno 1967. Verbs and times, in Z.Vendler, Linguistics in philosophy. Ithaca : Cornell University Press. Whorf, Benjamin Lee 1956. Language, thought and reality. (Selected writings). Cambridge, Mas.: The MIT Press. Visser, G.J. 1955. Celtic influences in English. Neophilologus XXXIX, 276-89. Vlach, Frank 1981. The semantics of the progressive. Syntax and semantics 14, 271-292. Wang, William S. (ed.) 1977. The lexicon in phonological change. The Hague : Mouton. Warner, Anthony 1993. English auxiliaries. Structure and history. Cambridge : Cambridge University Press. _____________________ 1995. Predicting the progressive passive : parametric change within a lexicalist framework. Language 71, 3 : 533-557. 112 Wartburg, Walther v. 1963. Problèmes et méthodes de la linguistique. Paris : Presses Universitaires de France. (2nd ed.) Whorf, Benjamin Lee 1956. Language, thought and reality. (Selected writings). Cambridge, Massachusetts : MIT Press. Williams, Raymond 1976. Keywords. A Vocabulary of Culture and Society. Glasgow : Fontana. Woisetschlaeger, Erich F. 1977. University Linguistics Club. A semantic theory of the English auxiliary system. Bloomington : Indiana Wright, Susan 1994 a. The mystery of the modal progressive, in Dieter Kastovsky (ed.). Studies in Early Modern English. Berlin ; New York : Mouton de Gruyter. 467-85. __________________ 1994 b The place of genre in the corpus, in : Merja Kytö, Matti Rissanen & Susan Wright (eds.). Corpora across the centuries. Amsterdam-Atlanta, GA : Rodopi : 101-107. Yule, G.U. 1944. The statistical study of literary vocabulary. Cambridge : Cambridge University Press. 113 ANNEX The progressive across three centuries Density of progressive forms for 100 000 words in a selection of fiction works A number of studies dedicated to the progressive and its history contain at least allusions to its increasing frequency. The table below covers only the early modern and modern periods — when semantic change was not too considerable —, and the corpus of fiction, the only one where figures have been computed for a long period in a sufficient number of works to give a useful picture: Figure 37 below.. However, it can only be indicative for many reasons: 1. The item studied is not clearly and consistently defined. The points to consider for that identification have been listed in II. 4. Yet, apart from the inclusion or exclusion of the going to form, the distortions should not be considerable. 2. More serious are those introduced by the statistical methods. Beginning with Mossé 1938, few scholars describe them, even summarily. Strang 1982: 432 is an exception: she said she selected what she called “ chunks ” of 30 000 words, including the beginning and end of the novels she surveyed –why should beginning and end be especially significant ? She may be supposed to have selected the chunks either randomly or in a systematic manner. It can be seen that her counts differed notably from my own in the case of Jane Austen’s Emma or Maria Edgeworth’s Castle Rackrent. As exposed in II.5, many aspects of sampling are liable to make comparison difficult, as will be seen by comparing the results for some given books in the table below: 1. Since the frequency (density here) is calculated on 100 000 words, the way words are counted is important: see on this point II.5. 2. Counting progressive occurrences is easier, once you have pencilled them in the book: a hand counter is useful in this respect. Errors are inevitable, as in any statistical survey: the point is to minimize them. The discrepancies have not been disguised, on purpose. In preparing averages for the graph, by periods of approximately 20 years, not evenly filled, a selection was made: historical novels (Westward Ho ! Henry Esmond, Waverley Novels, etc.) and fiction works likely to contain too much dialect (Castle Rackrent, Handy Andy) have been left out to avoid bias: they are in italics. Margaret Drabble’s Middle Ground too, since it seems to betray deliberate avoidance of the progressive. When several counts exist, I have chosen that which seems to me the most accurate: usually my own, for the reasons given above. There is still some ground for further refinement. Sources: Arnaud, René (1973). La forme progressive en anglais du XIXe siècle. Doctorat d'Etat, Université Paris VII 1972. Lille: Service de reproduction des thèses. Mossé, Fernand (1938). Histoire de la forme périphrastique être + participe présent en germanique. 2e partie: Moyen anglais et anglais moderne. Paris: Klincksieck. Nickel, Gerhard (1966). Die “ Expanded Form ” im Altenglischen. Neumunster. Scheffer, Johannes (1975). The Progressive in English. Amsterdam: North Holland; Oxford: American Elsevier (North Holland Linguistic Series). Strang, Barbara (1982). Some aspects of the history of the BE+ING construction, in John Anderson (ed.). Language Form and Linguistic Variation. Papers Dedicated to Angus McIntosh. Amsterdam: Benjamins. 114 Date Density Author Title Source 1720-40 1719 1726 61 Defoe 80 Swift 54 " Robinson Crusoe Gulliver " Mossé Strang Mossé 1740-60 1740 1742 1748 1749 1751 186 80 92 50 80 30 Richardson Fielding " Smollett Fielding Smollett Pamela Joseph Andrews " Roderick Random Tom Jones Peregrine Pickle Strang Strang Mossé Strang Strang Strang 1759 1766 1768 1778 119 169 190 128 Johnson Goldsmith Sterne Burney Rasselas Vicar of Wakefield Sentimental Journey Evelina (Ist part) Mossé Mossé Mossé Mossé 500 360 190 123 111 140 235 213 390 353 404 283 56 133 156 63 110 250 423 313 93 223 100 Edgeworth Castle Rackrent Edgeworth Austen " Belinda Sense & Sensibility " Austen " Austen " Austen " Scott Scott Scott Peacock Peacock Austen Austen Hogg Scott Hogg Scott Pride & Prejudice " Mansfield Park " Emma " Waverley Guy Mannering Old Mortality Headlong Hall Melincourt Northanger Abbey Persuasion The Brownie of Bodsbeck A Legend of Montrose Justified Sinner Count Robert of Paris Strang Arnaud Strang Arnaud Mossé Strang Arnaud Strang Arnaud Strang Arnaud Strang Strang Strang Strang Strang Strang Strang Strang Strang Strang Strang Strang 266 269 277 213 241 245 280 Dickens Lover Dickens Disraeli Marryat C. Bronte " Curiosity Shop Handy Andy The Chimes Sybil Children of New Forest Jane Eyre " Arnaud Arnaud Mossé Arnaud Arnaud Arnaud Strang 1800-1820 1800 1801 1811 1813 1814 1816 1814 1814 1816 1816 1818 1818 1818 1818 1819 1824 1832 1840-1860 1841 1842 1844 1845 1847 1847 115 1847 1848 1850 1851 1851 1852 1855 1855 1856 1856 1857 1859 1859 270 211 287 263 266 186 173 205 352 223 195 190 385 543 307 277 Thackeray E. Bronte Thackeray Borrow Gaskell Thackeray Kingsley Trollope Hughes Mulock Trollope Dickens Eliot " Meredith 1860-1880 1860 1861 1865 1868 1872 1897 309 323 251 230 285 379 Eliot Eliot Dickens Collins Hardy Wells 1900-1920 1911 1914 1920-1940 1920 1928 1930 1859 " Vanity Fair Wuthering Heights Pendennis Lavengro Cranford Henry Esmond Westward Ho ! The Warden Tom Brown's Schooldays John Halifax Barchester Towers Tale of Two Cities Adam Bede " Richard Feverel " Strang Arnaud Arnaud Arnaud Arnaud Arnaud Arnaud Arnaud Arnaud Arnaud Arnaud Arnaud Arnaud Strang Arnaud Mossé The Mill on the Floss Silas Marner Our Mutual Friend The Moonstone Under the Greenwood Tree The Invisible Man Arnaud Strang Arnaud Arnaud Mossé Mossé 520 Lawrence 414 Joyce The White Peacock Dubliners Strang Mossé 512 Galsworthy 676 Waugh 713 Waugh In Chancery Decline & Fall Vile Bodies Mossé Strang Strang 730 513 781 630 703 557 648 854 280 Room at the Top Unconditional Surrender The Pumpkin Eater The Girls of Slender Means The Old Boys The Red and the Green The Man w. the Golden Gun Life at the Top The Middle Ground Nickel Strang Scheffer Scheffer Scheffer Scheffer Scheffer Scheffer Arnaud 19601957 1961 1962 1963 1964 1965 1966 1966 1980 Braine Waugh Mortimer Spark Trevor Murdoch Fleming Braine Drabble 116 Progressives in novels since 1700 By 20-year spans 700 600 D e n s i t i e s 500 400 300 200 100 0 1740 1760 1780 1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 Dates (ends of spans) Densities BINARY GLIMPSES In studies of the progressive form, the most common assumption is that, basically, we must account for the choice between progressive and simple forms. As stated in I.2., this is only partly true. Nevertheless, when feasible, the binomial study can be helpful and revealing. However, the proportion of progressives to simple forms is very small : of the order of 2 per cent at the beginning of the 19th century, all types of discourse (genres) considered, to possibly some 8 or 10 per cent in present-day English. This means that parallel survey of the two forms in a given corpus implies disproportionate effort : to collect the same number found in one page for the non-progressive, 50 pages or about are necessary for progressive occurrences. There are several ways of remedying to that. One is to concentrate on tenses of major significance : the past and the perfect would be the most rewarding, undoubtedly. Conversely, we can eliminate some verb-forms such as the infinitive, the participle, or even the passive, etc., where the progressive is rare, but at the expense of missing interesting phenomena. Finally, BE, a very common verb, exceptionally found with the progressive, can be excluded with little bias : but this must be clearly stated. In Arnaud 1973, where the statistical survey was not in prominence, it had been found better to cover the whole verb-system. To minimize the impact of the unbalanced distribution of the two forms, the following method had been adopted : separate random samples of 1000 tokens were taken of the progressive and of the simple form in each of the various strata and frequencies (densities) were then compared in the following way : P/S = Q P : frequency of the item in the progressive sample ; S : frequency of the item in the simple form sample. The Q coefficient thus mirrors the comparative frequencies, but does not give their actual values. These can be estimated in a rough way with reference to the overall frequencies of the two forms, but that Q coefficient is as good a measure so long as comparison is aimed at. 117 The corpora On the basis of preliminary surveys, 5 “ strata ”132 were selected as likely to provide a substantial harvest of progressives : 1. Punch, the well-known satirical magazine (P) 133 ; 2. a random selection of private correspondence, of the type used in the present work; 3. a choice of the reports of criminal trials of the time (T) ; 4. a random selection of contemporary novels (N) ; 5. finally, within the latter, a random selection of passages of dialogue (D). Table A Distribution of Q in the verb forms P C T N D Present 2.27 1.69 1.34 1.72 1.94 Past 1.73 1.06 1.60 1.36 1.90 Perfect 2.35 2.74 1.06 2.51 2.72 Pluperf 1.18 2.00 2.24 0.79 1.28 Infinitive 0.14 0.07 0.06 0.15 0.10 Modal 0.15 0.01 0.19 0.16 0.19 Mod pfct 0.75 0.25 0.50 0.07 0.25 Pass 0.16 0.12 0.08 0.15 0.10 Total 8.78 7.81 5.58 6.89 8.51 Apart from Punch, selected because its style seems close to ordinary spoken English, especially in the numerous cartoons, but which could also be treated as a very specific type of discourse, it can be seen that the strata where the frequency of the progressive is greatest are the dialogue in novels and private correspondence. The comparative frequency with the perfect tenses is also noteworthy. Table B Distribution of Q in the verb classes Classes P C State 0.12 0.16 Locative 3.10 1.53 Movt 2.65 2.14 Ubdd proc. 1.38 1.47 Bdd proc. 0.96 1.07 T 0.12 6.10 2.00 1.26 0.74 N 0.07 1.47 2.19 1.33 0.70 D 0.09 1.17 2.86 1.45 0.81 Total 0.12 2.67 2.37 1.38 0.76 Nb vbs 25 8 16 32 27 % prog 3.6 10 31 31 23 % simple 31 5 13 22 27 Nb vbs refers to the number of lexical items (verb bases). It is clear that, although BE was omitted, the percentage of state verbs is great at the non-progressive form, in comparison with their low frequency in progressive utterances.. Conversely, locative verbs have a high rank as well as movement verbs. Situated utterances Another feasible survey consists in measuring the choice of progressive v. simple form in situations where it is not constrained by internal factors (the now or then context). Two such situations were surveyed in the letters of George Eliot and Mrs. Gaskell. Contrary to what one might think, it is not so often that people say what they do at the very time when they are doing it, apart from the well-known case of the conjurer or the practical scientist in front of an audience : they don’t need 132 Actually this is not the appropriate term in the statistical sense, in the absence of solid distributional laws. It is just hinting at what has been considerably developed since, the multi-genre approach: cf. Smitterberg 1998, 2000 and forthcoming. 133 Only the collections for 1841 and 1870. 118 to. However, when writing letters, it is fairly common that people do so, witness the occurrences of write (more than 1000 in our corpus). Linguists will not easily be convinced that there are two ways of saying the same thing : “ one word one meaning ” is almost an article of faith among us. My own belief is that there are situations, fairly common, when some uncertainty or hesitancy happens. In such situations where minimum attention is paid to words, simply the difference does not matter, whereas we, as linguists, are commonly paying a lot of attention to language. A philosophical reference to this view can be found in Henri Bergson’s Matière et mémoire (1896, 1941 : 174-81). Bergson insists that our everyday experience is at the level of action, i. e. somewhere between the extremes of generalisation and individuation. We can reasonably assume that when G. Eliot or Mrs Gaskell, writing to a near relative or friend, use the expression “ I write ” or “ I am writing ”, we have no way of proving that they mean different things... nor can we prove the opposite. This, in my opinion, is one of the clearest cases of variable rule. The results are given below, after Arnaud 1980. Table B I write/ I am writing in the letters of G. Eliot and Mrs. Gaskell, by five-year periods George Eliot : I write 9 3 1 5 4 7 8 5 Total : 42 ____________________________________________________________________________________ I am writing 6 5 3 4 9 7 5 12 Total : 51 Mrs Gaskell I write 0 2 14 9 0 Total : 25 ____________________________________________________________________________________ I am writing 1 3 10 12 3 Total : 28 This tends to show that, in similar situations, the letter-writers hesitated between the two forms they considered equivalent for practical purposes. What used to be called free-variation has been given renewed status by Labov, in the micro-diachronic perspective we are adopting here. This seems more reasonable than trying to suppose conscious or unconscious semantic subtleties at the back of the writers’minds in this type of situations. Notice that, especially in the case of George Eliot, there is a tendency to favour the progressive phrase with the passing of time. Table C Go, come and other verbs with future reference (extended present) in Mrs Gaskell’s letters by five-year periods GO 1832-45 1846-50 1851-55 1856-60 1861-65 Totals S 4 5 41 15 11 75 P 4 9 34 29 19 95 COME S P 1 4 0 6 21 32 9 10 8 6 39 58 Other verbs S P 3 1 7 1 22 4 13 8 8 1 53 15 Total S P 8 9 12 16 84 70 37 47 27 26 168 168 Cumulated 17 28 154 84 53 336 Percentage S P 47 53 43 57 54.5 45.5 44 56 51 49 50 50 In this case, with more significant figures, the same conclusion seems justified, since contexts did not give any indication of difference of perspective, between, say, mere expectation and more personal committment. 119 120 LIST OF THE LETTER-WRITERS Chronological order. Dates between slashes : first and last letters surveyed. Women-writers are in italics. Nb PF Density EDG : Maria Edgeworth (1767-1849)/1791-1840/ MAW : Mary Wordsworth (1770-1855)/1800-1854/ 606 344 185 266 WMW : William Wordsworth (1770-1850)/1790-1850/ 1234 144 DOW : Dorothy Wordsworth (1771-1832)/1787-1834/ 902 177 HUT : Sara Hutchinson (1771-1835)/1800-1835/ 316 157 SCO : Walter Scott (1771-1832)/1790-1832/ 818 157 COL : Samuel Taylor Coleridge (1772-1834)/1791-1834/ 1192 107 SOU : Robert Southey (1774-1843)/1792-1838/ 656 176 AUS : Jane Austen (1775-1817)/1796-1817/ 292 217 RUM : Margaret Ruskin (1781-1843)/1814-1842/ 136 158 RUJ : John James Ruskin (1785-1842)/1808-1842/ 101 179 CAR : Thomas Carlyle (1795-1881)/1813-1876/ 3722 259 KEA : John Keats (1795-1821)/1817-1820/ 425 322 MAC : Thomas Babington Macaulay (1800-59)/1807-1859/ 1135 225 WEL : Jane Welsh-Carlyle (1802-1866)/1819-1842/ 815 303 EBB : Elizabeth Barrett-Browning (1806-61)/1827-1861/ 1128 205 GAS : Elizabeth Cleghorn Gaskell (1810-1865)/1832-1865/ 1172 358 THA : William Makepeace Thackeray (1811-1863)/1820-1863/ 1679 345 DIC : Charles Dickens (1812-1870)/1832-1870/ 1813 271 BRO : Robert Browning(1812-1889)/1838-1889/ 184 132 ELI : George Eliot (1819-1880) (Mary Ann Evans)/1836-1880/ 2327 320 RUS : John Ruskin (1819-1900)/1827-1851/ 761 326 Total PF 21759 Total density 222 121 SPECIMENS OF LETTERS (extracts) Due to the wide dispersion of progressives ,they are very few in the extracts below which have been selected only to give an idea of tone and subject matter. Grasmere, 16 April, 1802 My dear Coleridge I parted with Mary on Monday afternoon about six oclock, a little on this side of Rushy-ford. Poor Creature ! she would have an ugly storm of sleet and snow to encounter and I am anxious to hear how she reached home. Soon after I missed my road in the midst of the storn, some people at a house where I called directed me how to regain the road through the fields, and alas ! as you may guess I fared worse and worse. With the loss of half an hour’s time, and with no little anxiety I regained the road. Unfortunately, not far from St Helen’s Auckland the Horse came down with me on his knees, but not so to fall overhead himself or to throw me. Poor beast it was no fault of his ! a Chaisedriver of whom I inquired the next day told me it was a wonder he could travel at all, he wanted shoeing so sadly, and his hoofs cleaning and paring... Yesterday after dinner we set off on foot meaning to sleep at Paterdale...at Ambleside we called on the Luffs to see how Luff was but learning that the Boddingtons were upstairs we did not see either Luff or his Wife. He has been dangerously ill but is now recovering fast. We reached home at dusk : so ends my story. Wm Wordsworth Plymouth Grove, Monday evng [15 Nov. 1852]. My dearest Polly, [Marianne, her favourite daughter]. You did not read my last letter carefully I conclude, or else your letter was a long time on the way, I don’t know which, for it had to be forwarded to me from Fox-How, and reached here yesterday. About the night gowns, they are at Knutsford... if I had know [sic] here at Midsummer how very much you wanted them, I would have had them put in hand sooner - as it is, and what with carriage etc they would come to as much as ready-made ; so if you are really dead-pressed, and can’t wait till Xmas get three two ready made (calico) from Silvers’Corn-Hill. O yes ! We’ll send a p-order for your cloak...There I think that’s all about dress,- except/.../ Moreover Hearn wants you to be spying out and seeing if you cannot see 2 little rough bear coats or paletots for F E and Julia. Be gracious, civil and handsome to servants and write me a particular account, not so much of the Duke’s funeral, as of Mr Mrs Chapman and their ménage and children. Every body here is going into mourning...Mrs Rich and Snow Wedgewood come here tomorrow Your own affec Mammy. (Elizabeth Gaskell). To Hannah Macaulay [M’s sister]. London, November 30 1833 My love, I enclose you a letter from the Temple − I have not time to write much to day, nor have I much to tell you if I had time. Things stand as they stood, except that the report of my appointment [to a high magistracy in India, where Hannah will accompany him] is every day spreading more widely, and that I am beset by advertising dealers begging leave to make up a hundred cotton shirts for me, and fifty muslin gowns for you, and by clerks out of place begging to be my secretaries. I am not in very high spirits to day, as I have just received a letter from poor Ellis, who is wandering among the corrupt Corporations of Kent, and to whom I had not communicated my intentions till yesterday. He writes so affectionately and so plaintively that he quite cuts me to the heart [...] He is to return to London on the 12th or 13th of next month. On the 11th - I forget whether I told you or not,- there is to be a dinner given to Lushington by the electors of the Tower Hamlets [...] Ever yours dearest, T B Macaulay 122 Degrees of intimacy Some names appear in two columns, for various reasons, e.g. the passage from distant early acquaintance to intimacy, and the nature of topics (Macaulay’s letters to Napier are often about classics). 1 AUS all letters BRO BRO BRO BRO Blagden, Isa : “ dearest Isa ” Elizabeth Barrett (Mrs. Browning) Pen (son) & F. Sarianna CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR CAR Carlyle (mother and father) Carlyle, Alexander (brother) : “ Alick ” Carlyle, Jean (sister) Carlyle, John (brother) : “ Jack ” Welsh, Jane (Mrs. Carlyle) COL COL COL COL COL COL COL Coleridge, George (brother) Cottle, Joseph Montagu, Basil Poole, Ths. Southey, Robert Wade, Josiah Wordsworth, Wm DIC DIC DIC DIC Dickens, Ch. C. (son) Dickens, Mary (daughter) Hogarth, Kate (wife) Hogarth, Georgina DOW DOW DOW DOW DOW DOW DOW Clarkson, Catherine Hutchinson, Sara Marshall, Jane Mary Hutchinson (Mrs.W.) William W’s brothers W’s children EBB neutralized EDG ELI ELI ELI ELI ELI ELI ELI ELI ELI ELI ELI 2 3 Murray De Quincey, Ths. Badams Montague, Anna Graham Hope, David Hunt, Leigh Inglis, Henry Johnston, James Mill, Stuart Mrs.Welsh Wilson, John Napier Bull, John Hunt, Leigh Tait, C. Wilson, Jane Emerson, R.W. Taylor, John Boyd Hessey Tait, Wm. Fraser Robinson, H.C. Julius Goethe Eckermann Aitken Gleig Fraser, J. D’Eichthal all other letters few Southey, R. Pearson, Wm. Robinson, H.C. all letters Mrs. Bodichon : “ Barbara ” Bray (Mrs. née Hennell) : “ Cara ” Hennell, Sara Sophia “ Sara ” Houghton, Mrs. Henry Lewes, Charles Lee Lewes, Mrs. Charles Lee Parkes, Bessie Raynier Stuart, Elma Blackwood, John Blackwood, William Chapman, John 123 ELI ELI ELI ELI ELI GAS Combe, George Congreve (Mrs. ) D’Albert Durade, François Pattison (Mrs. Mark) Spencer, Herbert Daughters (Marianne: Polly) nil HUT all letters KEA all letters MAC MAC MAC MAC MAC MAC MAC Empson Macaulay, Frances Macaulay, Hannah Macaulay, Margaret Macaulay, Mrs. (mother) Macaulay, Selina Macaulay, Zachary (father) MAW RUJ all letters all letters RUM RUS all letters all letters SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SCO SOU SOU SOU SOU SOU SOU SOU SOU SOU SOU SOU SOU SOU SOU SOU Lewis, Maria All others (few) Ellis Napier Baillie, Joanna Carpenter, Charles Erskine, Charles Erskine, William Lockhart, J.G.(Sophia’s husband) Morrit Richardson, John Scott, Charles (son) Scott, Lady (wife) Scott, Sophia (daughter) Scott, Thomas (brother) & his wife Scott, Walter (son) & his wife Sharpe, Charles Clerk Clephane, Miss & Mrs. Constable Croker Rutherford, Miss Buccleuch Cadell Baillie, Joanna Kerr Abercorn, Lady Ballantyne, John Ballantyne, James Ellis, George Erskine, Charles, William Heber Hughes, Mrs. Laidlaw Mackensie, Colin Montagu, Lord Morritt Richardson Rose Scott, Mrs (mother) Sharpe, Ch.K. Skene Smith, Miss Southey Stuart, Lady, Terry Melville, Lord Wordsworth. Smith, Miss. Bedford, G.Ch. Bedford, H.W. Danvers, Ch. Fricker, E. Southey, Bertha Southey, Isabel Southey, Mrs. Robert (wife) Southey, Ths. Williams Wynn, C.W. Cottle, Joseph May, John Southey, H.H. Rickman, John Seward, Anna Duppa, Richard King, John Coleridge, S.T. Poole, Ths. Betha, Matilda Morgan, J-J. Biddlecombe Standert, H.C. Senhouse, H. Landor, W.S. Robinson, H.C. Britton, John Grahame, J. Murray, John Crocker, John Lockhart, J.G. Ainsworth, W. Hodson, Mrs. Tonna, Mrs. C.. Cunningham. Edgeworth, M. Surtees, Robert Hamilton, Lady Haydon Byron, Lord Hartstonge Polwhele Laing 124 SOU SOU SOU SOU SOU SOU Hone, William Laing, David Kenyon, Lord Moxon, E. Farr Rosqe Pickering, Wm. Reade, John E. THA THA THA THA THA THA Brookfield, Mrs. Carmichael-Smith (his mother) Daughters Fitzgerald Perry, Kate Shawe, Isabella (wife) WEL WEL WEL WEL Carlyle, Mrs (Thomas’s mother) Carlyle, Thomas Carlyle,Jean Stoddart, Eliza WMW WMW WMW WMW WMW WMW WMW WMW WMW WMW WMW WMW WMW WMW WMW Clarkson, Catherine Fenwick, Isabella Hutchinson, Sara (Mary’s sister) Mary Hutchinson (Mrs Wordsworth) Wordsworth Dorothy (sister) W’s brothers W’s children Stanley, Lady Procter, Mrs. Baxters Perry, Kate Welsh, Helen Welsh, Mrs. George Montagu,Anna Hunter, Susan Sharp, Richard Haydon Lowther, Lord Lonsdale, Lord Scott, Walter Jewsbury Stuart, Daniel Moxon Gordon, J.H. Scott, John Landor, W.S. Watts, Alaric Taylor, John Cunningham. Lockhart, J.G. 125 THE CODING FRAME VERB VERB (lexical item) Cl SEMANTIC CLASS (predication type) : 0 State : 01 Relational/statal (be/have type) : belong ; owe... 02 Declarative : agree, beg, suggest, refuse, name... 03 Cognitive : find ; seem ; hear ; believe... 04 Affective : love, like, hate, admire... 05 Prospective : aim, wish, mean, pine(for), expect, fear, wait ( for)... 06 Retrospective : recall, forget, repent... 10 Locative : sit, reside, live (inhabit), hover, tarry... 20 Movement (unbounded) : ascend, walk, stroll ; go (on)... 21 ” (bounded) : go (=leave) ; arrive, draw (near), sail (for)... 30 Processes (unbounded) : bark, dine, help, happen,strive, worry ... 31 Processes (bounded) : build, kill, cook, make(up) ; give, deliver, obtain ; cause ; marry... 32 Processes (bounded, progressive) : grow, decline ; become ; recover ; learn... 33 Processes (unbounded, descriptive - adjectival) : bloom, glow, bleed, flourish, sleep... N.B. Verb meaning is ascribed in context : some lexical items can belong to more than one class : go (leave), go (progress). Metaphoric meaning is not considered apart, but real homonyms are differentiated (own : possess/ confess). 32 and 33 are indicative subclasses of 31 and 30 respectively Ts VERB FORM (“ tense ”, etc.) 0 Bare stem (infinitive, imperative, subjunctive) 1 2 3 4 5 6 7 8 9 Present Past Perfect Pluperfect To infinitive With modal auxiliary Passive (modern) : is being tried Participle (being writing) “ Modal ” perfect or pluperfect ( I wish I were...) ; also reported speech, etc... N.B. Some (0,8,9 and even 6) are unfrequent. They can easily be retrieved for further analysis. S SUBJECT 0 Zero subject 1 1st person 2 2nd person 3 3rd person definite (Mary, he, a visitor, the door) 4 3rd person indefinite (people,anyone, some) 5 it 8. what 126 C “ OBJECT ” COMPLEMENT N.B. : direct or prepositional ; “ accusative ” for vbs with a “ dative ” complement ; “ goal ” for directional verbs. 0 1 2 3 Zero Definite Indefinite Object clause (that,whether...) or nominalisation or equivalent 4 5 6 8 -ing complement : this linen is wanting washing Infinitive Predicative adj. or noun, including well (it is looking nice, she is looking well) what, where T TIME COMPLEMENT 0 1 2 3 4 5 6 Zero Present Past Future (including subordinate when clause ) Iterative (often, whenever...), continuous (always, never) or generic “ Time-frame ” (Jespersen). N.B. : 1,2 or 3 are then implied. “ Extended ” present (today, presently...) D DURATION and DEGREE 0 Zero (or other than below) 1 Degree : fast, very much, enough ; slowly, little... 2 Unbounded duration (atelic) : for years, long, ever since, still... 4 Bounded duration (telic) : for two years, till now... Obs OBSERVATIONS 4 Exceptional insertions between be and -ing 5 Remarkable examples 6 Dubious cases 7 Old passives : preparations are making I 1 2 3 INTIMACY Close (family circle, close friends) Friends and relations More distant connections AUT AUTHOR YEAR YEAR No REFERENCE NUMBER TEXT TEXT with references (page or letter number, according to the source). N.B. 1. In the ROMPROG database, for practical reasons (decimal system ; non-exclusiveness of distributions, etc.), some groupings under a given label (e.g. time complement and duration or H) are not strictly coherent. 2. Some points are just supplementary information, sometimes redundant. 3. This research framework was not completely exploited and does not claim theoretical significance. VERB LISTS 127 Polysemic verbs See also : list of verbs belonging to several classes. Verb Class ABSORB1 ABSORB2 ASSIST1 ASSIST2 BEAR1 BEAR2 BEG1 BEG2 BURN1 BURN2 CALL1 CALL2 CALL3 CROSS1 CROSS2 DECLINE1 DECLINE2 DESIRE1 DESIRE2 DIE1 DIE2 DRAW1 DRAW2 DRAW3 DRAW4 DRIVE1 DRIVE2 ENLARGE1 ENLARGE2 FEEL1 FEEL2 FIND1 FIND2 FLIT1 FLIT2 FLY1 FLY2 GET1 GET2 GET3 HANG1 HANG2 HOUSE1 HOUSE2 INSIST1 INSIST2 KEEP1 KEEP2 LABOUR1 LABOUR2 31 30 30 30 30 21 2 30 31 33 30 30 2 21 21 32 2 5 2 31 5 30/31 21 30/31 20/21 30 31 30 30 3 31 3 20/21 31 20/21 30 31/32 32 21 10 31 30 31 30 2 30 10 30 33 Meaning let in, imbibe interest, engross help be present hold, carry sail ask mendicant set fire to ;trans. be on fire hail visit name walk, etc. across cancel, delete decrease, dwindle (intr.) refuse wish ask, urge expire long, desire pull come, approach sketch, paint 31 write a cheque progress on some vehicle push, hit with, etc. make larger expatiate ; speak at length touch have the feeling of discover be aware of move quickly... remove : transitive birds, etc. flags, etc. obtain become advance be suspended suspend contain store in emphasize, stress urge retain stay (home,etc.) work suffer 128 LIE1 LIE2 LIVE1 LIVE2 LOOK1 LOOK2 MEDITATE1 MEDITATE2 PASS1 PASS2 PASS3 PASS4 RESOLVE1 RESOLVE2 ROLL1 ROLL2 ROLL3 ROLL4 SET1 SET2 SHARE SHARE 31 SMOKE1 SMOKE2 SPEND1 SPEND2 STAMP1 STAMP2 STEAL1 STEAL2 STIR1 STIR2 STOP1 STOP2 TEAR1 TEAR2 THREATEN THREATEN1 THINK1 THINK2 TRUST1 TRUST2 WEAR1 WEAR2 10 30 10 30 30 30 30 5 20/21 30 30 31 31 31 20 30 30 30 31 21 30 divide 33 30 30 30 31 30 31 21 30 20 31 10 31 20 5 30 30 5 2 31 30 32 in bed, etc. tell lies dwell be alive gaze have the appearance of ponder, think arrange in advance move past spend (time, etc.) happen enact turn into, dissolve decide move like a ball ship thunder cause to revolve install, set up go, go down(sun) have in common emit smoke a cigar, etc. pass time use up (money, etc.) put a stamp on on the ground rob go away shake ; transitive move ; be up arrest ; transitive stay rip, cut run fast be imminent menace have in the mind anticipate have confidence entrust clothes become worn Aspectual classes State Cl. 1 BE BELONG DETAIN HAVE Cl. 2 (ctd) Cl.2 (ctd) Cl.3 BID BLESS BLESS CALL3 PRESCRIBE PRESUME PROMISE PRONOUNCE BELIEVE DOUBT HEAR MISUNDERSTd 129 OWE PROVE Cl.2 ACCUSE ACKNOWLEDGE ACQUIESCE ADVERT ADVISE AFFIRM AGREE ALLEGE ALLUDE ANNOUNCE APOLOGIZE ASSERT BEG1 BEG2 BESEECH CONFESS CONFIRM CONGRATULATE CONJECTURE COUNSEL DECLARE DECLINE2 DESIRE ENJOIN ENTREAT EXHORT FORGIVE IMPLORE IMPLY INSIST MAINTAIN NAME OWN PERSUADE PREDICT PROPHESY PROPOSE PROPOUND PROPOUND RECOMMEND REFUSE REGRET REMARK REMARK REMIND REQUEST STATE SUGGEST SUPPOSE SWEAR THANK TRUST URGE VOW Prospective-retrospective CL. 5 AIM ANTICIPATE APPREHEND AWAIT BEAT UP FOR BREW CAST ABOUT CLAMOUR FOR CONTEMPLATE CRAVE DESIRE DIE TO/FOR DREAD ENTERTAIN EXPECT FEAR FERMENT FRAME WISHES GAPE FOR GIRD OFF FOR GROAN FOR Locative (Cl. 10) HANG HOVER KNEEL LAY LEAN SEE SEEM Cl.4 GRIEVE HATE LIKE LOVE PREFER RESENT CL.6 HANKER HATCH HOPE HOPE HUNGER TO HUNT FOR INTEND LANGUISH FOR LONG LOOK FOR LOOK FWD TO LOOK ON LOOK OUT FOR LOOK TO MEAN MEDITATE2 MUSE OF NEED PANT FOR PINE FOR PLAN LIE1 LINGER LIVE1 LODGE PAUSE PROJECT PROMISE PURPOSE RAVEN FOR SCHEME SEEK SIGH AFTER SIGH FOR STARVE TO STRIVE AFTER STRIVE FOR THINK2 THIRST TO/FOR THREATEN TO TIRE FOR TRY FOR TURN OVER WAIT WANT WEARY FOR/TO WISH YEARN REMAIN RESIDE REST ROOST SIT SIT UP SOJOURN SQUAT STAY STOP2 FORGET LOOK BACK RECALL RECALL RECALL REGRET REPENT TARRY 130 131 Movement Cl. 20 ADVANCE ASCEND BEAT ABOUT CANTER CAPER CARRY ON CIRCUIT CIRCULATE CRAWL CREEP CRUISE CUT ACROSS DANCE DRAW2 DRIFT FLEET GALLOP GAMBOL GET ON GLIDE HOBBLE HOP JAUNT JIG JOG JOURNEY LEAP MAKE WAY MARCH MOUNT PACE PADDLE PATROL PERAMBULATE PLOD PLY PROGRESS PROWL RAISE RANGE REPASS REVOLVE RIDE ROAM ROVE ROW RUSH SAUNTER Cl. 21 SCALE SCAMPER SCRAMBLE SHIFT SKATE SKIM SLAP ALONG SLEIGH SLIDE SLIP SOAR SPANK AWAY SPIN ON SPRAWL STAGGER STALK STAMP STRAGGLE STROLL STRUT STUMP ABOUT SWIM TACK TEAR2 THREAD AWAY TOIL TOUR TRAVERSE TROT TRUDGE TRUDGE TUMBLE VAGABOND VEER WADE WAG WARP WAVER WHEEL WHIRL WIN WAY WIND WING WORK UP WORK UP ZIGZAG ZIG-ZAG-ZIG APPROACH ARISE ARRIVE BEAR2 BEND O’S STEPS CLAMBER CLEAR THE BAR COME COME BACK COME DOWN COME IN COME OFF COME OUT COME OUT COME ROUND COME UP CROSS1 DEPART DESCEND ENTER ESCAPE FLEE GET1 GET3 GO AWAY GO BACK GO DOWN GO OFF GO OUT GO ROUND GO TO BED KNEEL DOWN LAUNCH LEAVE NIGH OVERSTEP PENETRATE PRESS AWAY PUT OFF QUIT REACH REASCEND RECUR REMOVE RETREAT RETURN RISE SET SET FOR SET IN SET OFF SET OUT SIT DOWN STAND UP START STEAL STEP TEND TIVY OFF TRIP OFF 132 Double-class verbs (unbounded/bounded predications) Cl. 20-21 CLIMB COME ON DRIVE1 FLIT1 FLOW STARVE GAD GLIDE GO GO ON SURVEY HASTEN SWALLOW HURRY JUMP MOVE PASS PLUNGE THRESH PLY PROCEED THRUST RAMBLE TRANSLATE ROLL1 TRANSPORT RUN SAIL TUMBLE STAMP STEER STEP STRIDE TRAVEL TURN WALK WANDER WHIRL Cl.30-31 ADMINISTER ARRANGE BEGIN BESTOW BLOW DO DRAW1 DRAW3 DRINK DRIVE2 MOW NET NURSE PANT PAINT SOW SPEND2 SPIN STAMP BURN1 BURST BUY CALCULATE EAT ENCLOSE ENGAGE ENLARGE PASS2 PASS3 PASS4 PAVE STRIKE SUBMIT SUCK CARRY EXCITE PAY CAST CATCH CAUSE CHANGE CLEAN FALL FORM FUME FERMENT GAD PICK PLY PRESUME PRINT PROCEED SWEEP SWING TAKE TEAR1 COBBLE COIN GATHER GET1 PROMISE PROVIDE THROW COLLECT GET2 PUBLISH COMPOSE GIVE PUSH CONTRIVE COOK HAMMER HOUSE PUT RAISE TRIM COPY CORRECT COVER CRY CUT DEFER DELIVER DEVISE DEVOUR DIG DIGEST DISSECT DIVIDE HURRY ISSUE ITCH KNIT KNOCK LAY LEAD LET LOSE LUG MAKE MEASURE MOVE READ REAP RECEIVE REVISE ROAST RUN SCOUR SCRAPE SCRIBBLE SEND SET SEW SHARE TURN USE WARM WASH WASTE WATER WEED WHILE WHIRL WORK WRING WRITE 133 Process verbs Unbounded predications Cl. 30 ABET ABSORB2 ABUSE ACCELERATE ACCOMPANY ACT PROCRASTINATE ADDLE ADDRESS ADMIRE ADVERTISE AFFECT AFFLICT AFFORD AGITATE AID AIR ALARM AMUSE ANATHEMATIZE ANGLE ANNOY ANSWER APOSTATIZE APPLY ARGUE ASCRIBE ASK ASK FOR ASSAIL ASSAULT ASSIST ASSIST AT ASSUME ATTACK ATTEMPT ATTEND ATTITUDINIZE ATTRACT AVOID BABBLE BACKBITE BACKSLIDE BAIT BALANCE BANG BARGAIN BARK DELIGHT DELVE DEMONSTRATE DEMUR DEPLORE DEPRECATE HUG HUGGER MUGGER HUM HUMOUR HUNT HURL PREVAIL PREY PRICK PRIZE PROCEED DEPRECIATE DESCRIBE DESECRATE DESERVE DETECT DEVISE DEVOTE DICTATE DIG DIGEST DIGRESS DILATE DIN DINE DIP DIRECT DIRECT DISCHARGE DISCOURSE DISCUSS DISPARAGE DISPERSE DISPUTE DISSECT DISTINGUISH DISTRUST DISTURB DO ABOUT DOCTOR DOG DOSE DOTE DOZE DRAG DRAWL DREAM DRILL DRIP DROOP DRUDGE DRUM HURRY HURT HYBERNATE HYMN AWAY HYPERBOLIZE IDLE ILL TREAT IMAGINE IMBIBE IMITATE IMPART OSF IMPASSION IMPORTUNE IMPOSE IMPUTE INCLINE INCONVENIENCE INCROACH INDICATE INDULGE INFLICT INFLICT INHABIT INJURE INQUIRE INSPECT INTEREST INTERFERE INTRIGUE INTRUDE INVEIGH INVESTIGATE INVITE INVOKE IRON ISSUE ITCH JAMMER JAR JARGON JAW AWAY PROFER PROFIT PROMPT PROPAGATE PROSE PROSECUTE PROTECT PROTEST PRY PUDDLE PUFF PULL PUMP PURR PURSUE PUT UP PUZZLE QUACK QUADRILLE QUARREL QUAVER QUESTION QUILL DRIVE QUIZ QUOTE RACK RACKET RADIATE RAGE RAKE RALLY2 RANSACK RANT RAP RASP RATTLE RAVE REACT REAP REAR REASON 134 BASK BATHE BATTER BATTLE BAWL BEAR WITNESS BEAT BEAR1 BEFRIEND BEG2 BEGUILE BEHAVE BELABOUR BELEAGUE BELIE BELLOW BEMOAN BEND BENEFIT BESET BESIEGE BESTIR OSF BEWAIL BILE BITE BLACKBALL BLAME BLAME BLASPHEME BLEND BLUBBER BLUNDER BLUSTER BOARD BOAST BOB BORE1 BORROW BOTANIZE BOTHER BRAG BRAVE BRAWL BRAY BREAKFAST BREATHE BREED BREW BRIDGE BRIGHTEN BROOD BUFFET BUMP BUOY UP BURBLE BUSTLE ABOUT DUN DWELL EARN EARTHQUAKE EAT THEIR TERMS ELECTIONEER EMBITTER EMBLAZON EMBRACE EMPLOY ENACT ENCORE ENCOURAGE ENDEAVOUR ENDURE ENFORCE ENJOY ENLIVEN ENNUY ENQUIRE ENTERTAIN ENUMERATE ENVY ESCORT EULOGIZE EXAGGERATE EXAMINE EXCEED EXCHANGE EXCITE EXCLAIM EXEMPLIFY EXERCISE EXERT EXHIBIT EXIST EXPATIATE EXPERIENCE EXPERIMENT EXPERIMENTALIZE EXPIATE EXPLAIN EXPLORE EXPOSE EXPOUND EXPRESS EXTOLL EXULT EYE FAG FAINT FALL BEHIND FAN FANCY FARE FAST JEST JINGLE JOKE JOLLYFY JOSTLE JUDGE JUGGLE JUGULATE JUMBLE ABOUT JUSTIFY JUSTIFY KEEP KICK KISS LABOUR1 LAKE LAMENT LAUGH LEAN LECTURE LEGACY-HUNT LIBEL LICK LIE2 LIGHTEN LIONIZE LISTEN LIVE2 LOATHE LOITER LOLL LOOK AFTER LOOK1 LOOK2 LOUNGE LUG LUNCH LURK MAINTAIN MAKE BELIEVE MAKE UP FOR MAKE UP TO MANAGE MANIFEST MANOEUVRE MANT MANTLE MARK MARVEL MEDDLE MEDIATE MEDITATE1 MEET MENTION METAPHYSICIZE MILK REBEL REBUKE RECITE RECKON RECLINE RECONNOITRE RECREATE RECRUIT REEL REFER REFLECT REFRAIN REGALE REJOICE RELATE RELAX RELENT RELY REMONSTRATE REPEAT REPEL RE-PERUSE REPORT REPOSE REPROACH REPROBATE REQUIRE REREAD RE-READ RESIST REST REVEL REVERIE REVERSE REVIEW REVILE REVOLVE RIDDLE RIFLE-SHOOT RING RIOT RISK ROAR ROAST ROB ROCK OSF ROLL2 ROLL3 ROLL4 ROLLICK ROMANCE ROMP ROUT ROW RUB RULE 135 BUSY OSF BUZZ BUZZ CACKLE CALL1 CALL2 CALOMEL CANT CANTABRIGIANIZE CANVASS CARE CARE CARESS CARP AT CARRY ON CARRY OUT CAST O S EYE CAST UP CAUSE CELEBRATE CHAFE CHALLENGE SENTIMENTALIZE CHANT CHASE CHAT CHATTER CHEAT CHEAT CHERISH CHEW CHIDE CHIME CHIRP CHOP CHRISTMAS CHUCKLE CHURN CLAIM CLANG CLASH CLATTER CLAW CLING COAST COGITATE COIN COLLOQUE COMBAT COME IN FOR COMFORT COMMUNE COMPARE COMPARE COMPLAIN COMPLIMENT FATIGUE FEAST FEED FEEL FEEL FOR FENCE FERMENT FIDGET FIGHT FIGURE FILL FIND2 FINESSE FINGER FIRE FISH FLAG FLATTER FLAUNT FLIT2 FLOAT FLOUNDER MIND MINE AT MISCHARGE MISHAPPEN MISJUDGE MISLEAD MISLUCK MISS MISTAKE MIX MOAN MOIL MOLEST MONOPOLIZE MOPE MOUNT GUARD MOURN MOVE MOW MUDDLE MURMUR MUSE RUMBLE RUMINATE RUMMAGE RUN RUSTLE SAY SCAN SCATTER SCHEME SCOLD SCOOP SCOUR SCRAPE SCRATCH SCRAWL SCREAM SCREECH SCREW UP SCRUB SEARCH SEIZE FLUCTUATE FLURRY FLUSTER FLUTTER FLY FOLLOW FONDLE FOOL FOOT FORAGE FORBEAR FORCE FORK FREQUENT FRET FROLICK FRONT FROWN FUME GAD GAFFAY GALL GALLANT GALLIVANT GALLOP-SCRAWL GAMBLE GAPE GARGLE GASLIGHT GASP GAZE GET ON GET3 MYSTIFY NAME NEGLECT NEGOCIATE NET NIBBLE NOURISH OBEY OBSERVE OCCUPY OFFEND OFFER OFFICIATE OOZE OPERATE OPPOSE OPPRESS ORDER OVERFLOWER OVERGUARD OVERLOOK OVERPOWER OVERSTRAIN OVERWHELM OVERWORK PANT PAPER PARADE PARLEY PATRONIZE PAY VISIT PEAL PECK SERMONIZE SERVE SET UP FOR SHAKE SHAKE HANDS SHAM SHED SHELL SHELTER SHEW SHILLY SHALLY SHOCK SHOOT SHOP SHOUT SHOW SHRIEK SHRINK SHUFFLE SHY SIGH SIGHTSEE SIMMER SIN SKETCH SKIM SLAM SLASH SLAVE SLIP SMELL SMILE SMOKE2 136 COMPLY CON OVER CONDUCT CONFINE OSF CONFOUND CONFUSE CONSIDER CONSOLE CONSTERNATE CONSULT CONSULT CONTEMN CONTEMPLATE CONTEND CONTEST CONTINUE CONTRADICT SPEC.CONSTABLE CONTRAST CONTRIBUTE CONVERSE CONVEY COQUETTE CORRESPOND COUGH COUNT COUNT ON COURT COVET COWER CRACK ON CRACKLE CRAVE CREAK CRIMP CRITICIZE CROSS EXAMINE CROW CRUSADE CUDDLE CUDGEL CULTIVATE CURL CURSE DABBLE DAGUERREOTYPE DALLY DANGLE DARN DASH DAWDLE DEAL DEBATE DECLAIM DEDICATE DEER HUNT GIVE EAR GIVE ONE S MIND GIVE OSF TO GIVE OUT GIVE VENT TO GLANCE GLAR GLEAN GLORY GNAW GO WRONG GOGGLE GOVERN GRAPPLE GREET GRIN GRIND PEEP PEER PEG AWAY PELT PEN PERFORM PERK PERORATE PERPLEX PERSECUTE PERSEVERE PERUSE PESTER PET PHOTOGRAPH PICK PICKLE SNEER SNEEZE SNIFF SNORE SNORT SOAK SOAP SOB SOLACE SOLICIT SOLILOQUIZE SOOTHE SOUND SPAR SPARE SPEAK GROAN GROPE GROWL GRUMBLE GRUNT GUESS GUIDE GUZZLE HAIL HALLOO HALT HAPPEN HARANGUE HARASS HARP HARROW HARVEST HAUL HAUNT HAWL HEAL HEALTHIFY OSF HEAVE HELP HESITATE HIDE HIGGLE HINDER HINT HISS HISTORIFY HIT HOE HOLD HOLIDAY MAKE HOMOEOPATHIZE HONE HOUSE PICTURE-SEE PINE PIONEER PIPE PIQUE OSF PIRATE PITCH PITY PLAGUE PLASH PLASH PLASTER PLAY PLEAD PLEASE PLOT PLOUGH PLUISTER ALONG PLY POACH POETIZE POINT POKE POLK PONDER POP PORE POSTURE POUR POWDER AWAY PRACTISE PRAISE PRATTLE PRAY PREACH PRESENT OSF PRESIDE PRESS SPECULATE SPEND1 SPEND2 SPIT SPLASH SPORT SPOUT SPRAWL SPURN SPURT SPY SQUABBLE SQUEEZE SQUIRE SQUIRT STAGNATE STAMMER STAND STAND OUT STAND TO STARE START UP¨ STICK STIMULATE STING STIR1 STIR2 STOOP STORM STRAIN STREAM STRETCH STRIVE STRUGGLE STUDY STUMBLE STUN SUBJECT 137 DEFEND DELIBERATE HOUSE HUNT HOWL PRESUME PRETEND SUCKLE SUCK-SUCK 138 Process verbs Bounded predications Cl. 31 ABOLISH ABRIDGE ABSTRACT ACCEPT ACCOMPLISH ADD ADOPT ALIGHT ALLOW AMPLIFY ANCHOR APPEAR ARRAY ASSIMILATE ATTAIN ATTRIBUTE AWAKE AWAKEN BAPTIZE BEGIN BIND BLISTER BORE2 BRAID BRING BROIL BUILD BURK BURY CALCINE CAST ANCHOR CEASE CEMENT CHOOSE CLEAR CLEAVE CLOSE COIN COMBINE COMMENCE COMMISSION COMMIT COMPILE COMPLETE COMPOSE OSF COMPOUND COMPROMISE COMPUTE DRIFT DROP DROWN DRY DYKE EDIT EDUCATE EMBARK EMBEZZLE EMBROIDER EMPTY END ENDANGER ENGRAVE ENRICH ENTER ERECT ESTABLISH ETCH EVADE EXCLUDE EXECUTE EXPEL EXTERMINATE EXTRACT EXTRICATE FABRICATE FALL IN WITH FALL IN WITH FASHION FATTEN FEATHER FETE FIND1 FINISH FIT FIX FLEDGE FLING FOLD FORGE FORSAKE FORTIFY FORWARD FOUND FRAME FRANK FREE OSF LAND LEAVE LEVEL LIGHT UP LOAD MAKE UP O S MIND MANUFACTURE MARRY MASTER MATURE MEDICINIZE MEND MISDATE MODEL MODIFY MOULD MOULDER MURDER NAIL UP NERVE NEW ROOF NOTE NUMBER OBTAIN OCCASION OMIT OPEN ORGANIZE ORIGINATE OVERCOME PACK PART PASS1 PEEL PERISH PERVERT PILE PLACE PLANT PLUCK POLISH POLLUTE POST POSTPONE PRECLUDE PREPARE PRESENT PREVENT ROT ROUND OFF RUB UP RUIN RUMMAGE UP RUN MAD RUN OVER SACRIFICE SATISFY SAVE SAW SCLAW SCREEN SEAT SECURE SEE SEEK UP SELL SEPARATE SET OFF SET TO SETTLE SHAKE OFF SHAPE SHAVE SHIFT SHUT SKIP SLAUGHTER SLING SMASH SMELT SMOOTHE SNIP SORT SPIER OUT SPILL SPLIT SPOIL SPUD UP SQUANDER SQUEAL OFF START START STEAL STITCH STOP1 STOW IN 139 CONCEIVE CONCOCT CONFER CONFIRM CONNECT CONSTRUCT CONSUME COVENANT CRAM CREATE CROSS2 CROWN CURE DAD DARKEN DEAFEN DECEIVE DECIDE DEFER DEGRADE DELAY DELINEATE DELUDE DEPRIVE DESERT DESPATCH DESPOIL DESTROY DETAIN DETERMINE DIE DIGEST DISAPPEAR DISAPPOINT DISCOMPOSE DISCOVER DISENGAGE DISMANTLE DISMISS DISPATCH DISPOSE DISSIPATE DISSOLVE DISTRESS DISTRIBUTE DO AWAY WITH DOUBLE DRAIN DRAW4 DRESS FREEZE FRY FULFIL FULFILL FUME FURBISH FURNISH GAIN GET IN GET UP GIRD ON GIRDLE IN GIVE BIRTH TO GIVE RISE TO GIVE UP GO OFF GRANT GRAVEL HALF OPEN HAND HANG HARNESS HEAP HEW HOARD HUNGER ILLUSTRATE IN INCREASE INCULCATE INDITE INFORM INITIATE INSCRIBE INSERT INSTIGATE INSTITUTE INTER INTERPRET INTERRUPT INTERWEAVE INTRODUCE INVENT INVEST INVOLVE JOIN JOT DOWN KICK OFF KILL KICK OFF PROCURE PRODUCE PROMULGATE PULL DOWN PUNISH PURCHASE PUT OFF PUT OUT PUT TO QUIET RAVAGE REACH UP REALIZE REAPPEAR REARRANGE RECOPY RECORD REDUCE REFORM REFRESH REGILD REKINDLE RELAPSE RELIEVE REMAKE REMEDY REMOVE RENDER RENEW REPAINT REPAIR REPAY REPRINT REPRODUCE REPUBLISH RESCUE RESELL RESERVE RESOLVE1 RETAIN RETARD RETIRE RETOUCH RETRANSCRIBE RETRIM RETURN REWARD RIP OPEN ROOF ROOT OUT STRANGLE STRIP STUFF SUBDUE SUBMIT SUBORN SUBSCRIBE SUBSTITUTE SUCCEED SUFFOCATE SUM UP SUPPLY SWEAR IN THATCH TIDY TIE TRAIN TRANSCRIBE TRANSFER TRANSFORM TRANSMIT TUNE TWITCH UNBURTHEN UNCOVER UNDERMINE UNDO UNDRESS UNLOAD UNPACK UNTIE VICTUAL WAKE WARP WEAN WEAR2 WEAVE WET WHET WIN WIND UP WINK OUT WITHDRAW WREATHE WRIGGLE OFF WRONG YIELD ZINCIFY 140 Verbs belonging to several classes Only by reference to the context can this class cleavage be understood : the implied difference of meaning may vary ; sometimes two or several meanings are subsumed by a common lexical label : e.g. contemplate : consider, plan ; name : appoint a day, call ; rest :stay, relax ; stand : “ posture ”, tolerate ; etc. Classes 32 and 33 are not to be taken as very strict. This list of homonyms covers some of the main cases. BREAK BREW BRIGHTEN BROIL CHANGE CONFIRM CONTEMPLATE COVER CRAVE DARKEN DESIRE DETAIN DISMISS DRAG DRAIN DROP DRY EMPTY ENTER ENTERTAIN FALL FILL FLY FORM FREEZE FUME GAIN GATHER GET ON GET UP GET1 GET2 GET3 HANG HATCH INCREASE LAY LEAN LEAVE 30 5 30 31 30 2 5 31 5 31 2 1 31 30 31 31 31 31 21 5 30 30 20 30 31 30 31 31 20 31 21 30 21 10 5 31 10 10 21 31 32 LOOK UP 30 LOSE 32 MAINTAIN 33 MELT 31 32 MEND 31 NAME 30 PASS1 33 PROMISE 30 PROTRACT 32 QUIET 5 RAISE 31 REGRET 32 RELAX 32 REST 32 RETREAT 32 RETURN 32 SCHEME 32 SET 31 SHIFT 30 SKIM 31 32 SLIP 32 SPRAWL 21 30 33 STAND 31 32 START 31 33 STARVE 31 STEAL 32 STICK 32 STRETCH 30 32 SWELL 32 SWIM 30 31 32 TOIL 31 32 TUMBLE 30 32 TURN 31 WARP 31 WASTE 32 WAVER 30 31 WEAR2 30 WHEEL 31 WIN WRITHE 30 31 2 32 31 2 20 2 30 31 20 2 30 10 21 21 5 21 20 20 20 20 10 21 31 21 30 30 30 20 20 20 21 20 30 20 31 20 31 30 32 32 30 33 32 30 21 31 5 32 32 30 31 6 32 30 32 31 30 30 31 31 30 30 30 30 31 33 31 33 32 32 30 30 30 30 3132 31 31 32 30 32 30 32 33 141 Periods and decades The whole time-span (exactly 1787-1880) was originally divided into periods of five years, but, the letters for the end periods being in small number, 1787-1799 was counted for one period, and the few letters dated after 1880 have been added to period 17. In many cases, data have been computed by decades according to the table below, where the three last periods have been fused, to compensate for smaller corpus size. Periods Dates 1. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1787-1799 1800-1804 1805-1809 1810-1814 1815-1819 1820-1824 1825-1829 1830-1834 1835-1839 1840-1844 1845-1849 1850-1854 1855-1859 1860-1864 1865-1869 1870-1874 1875-1880 Decades I II III IV V VI VII VIII LIST OF FIGURES Figure 1. General density, by decades. Figure 2. Density for men and women, by decades. Figure 3. Densities, by equivalent chronological lots. Figure 4. Individual densities for women, in order of birth Figure 5. Individual densities for men, in order of birth. Figure 6. Trajectories : Lake group. Figure 7. Trajectory of William Wordsworth : 1800-1850. Figure 8. Trajectories of the Lake Poets : 1800-1835. Figure 9. Trajectories of Scott and Wordsworth Figure 10. Trajectories of Jane Austen and the Lake poets. Figure 11. Trajectories of Dickens, Thackeray and Gaskell. Figure 12. Trajectory of Mrs Gaskell. Figure 13. Trajectory of Thackeray. Figure 14. Trajectory of Dickens. Figure 15. Trajectory of Macaulay. Figure 16. Trajectory of Carlyle (early periods). Figure 17. Trajectories of the Brownings. Figure 18. Trajectory of George Eliot. Figure 19. State verbs : cumulated densities, by decades. 142 Figure 20. Prospective & retrospective verbs : densities by decades. Figure 21. Locative & movement verbs and general density (adjusted), by decades. Figure 22. Process verbs and general density, by decades. Figure 23. Density of major classes, by decades. Figure 24. Percentage of major classes, by decades. Figure 25. Percentage of major tenses, by decades. Figure 26. The two passives, by decades. Figure 27. Number of lexical units, by frequencies, for groups A and B. Figure 28. Extended present : adjusted densities. Figure 29. Have. Figure 30. Modal auxiliaries. Figure 31. Shall & will. Figure 32. May & might. Figure 33. Must. Figure 34. Density in Jane Austen’s letters and novels. Figure 35. Density in Macaulay’s Letters and History. Figure 36. Mrs Gaskell’s letters to Marianne and novels. Figure 37. The progressive across three centuries 143 FIGURES 400 343 305 300 278 D e n s i t i e s 219 200 209 190 158 140 100 0 1 2 3 4 5 Decades 6 7 8 144 Figure 1. General density, by decades. 400 359 299 300 272 D e n s i t i e s 236 236 234 218 200 193 172 183 177 173 147 117 100 97 0 Decades Men Women 169 145 Figure 2. Density for men and women, by decades. 400 D e n s i t i e s 300 200 100 <1805 1815 1825 1835 1845 1855 1880 Chronology Density all D women D men Figure 3. Densities, by equivalent chronological lots. 400 358 350 D e n s i t i e s 320 303 300 266 250 217 200 185 177 158 157 150 205 100 50 0 EDG MAW DOW HUT AUS RUM Densities WEL EBB GAS ELI 146 Figure 4. Individual densities for women, in order of birth. 345 350 326 322 300 271 259 D e n s i t i e s 250 225 200 150 176 144 179 157 132 107 100 50 0 WMW SCO COL SOU RUJ CAR KEA Densities MAC THA DIC BRO RUS 147 Figure 5. Individual densities for men, in order of birth. 600 500 D e n s i t i e s 400 300 200 100 0 1 2 3 4 5 MAW WMW 6 7 Periods (1800-50) HUT COL 8 9 DOW SOU 10 11 12 148 Figure 6. Trajectories : Lake group. 400 D e n s i t i e s 300 200 100 0 1 2 3 4 5 6 Decades Figure 7. Trajectory of William Wordsworth : 1800-1850. 300 250 D e n s i t i e s 200 150 100 50 0 1 2 3 4 5 6 Periods (1800-1835) WMW COL DOW 7 8 149 Figure 8. Trajectories of the Lake Poets : 1800-1835. 200 180 160 D e n s i t i e s 140 120 100 80 60 40 20 0 1 2 3 4 D. Sco 5 Periods 6 D. WmW 7 8 9 150 Figure 9. Trajectories of Scott and Wordsworth. 400 D e 300 n s i 200 t i e 100 s 0 1 2 3 4 5 Periods (1790-1820) D. Aus D. Col D. WmW D. Sou Figure 10. Trajectories of Jane Austen and the Lake poets. 600 D e n s i t i e s 500 400 300 200 4 5 6 7 Decades (1825-70) GAS DIC THA General Density 8 9 151 Figure 11. Trajectories of Dickens, Thackeray and Gaskell. 500 400 D e n s i t i e s 300 200 100 0 8 9 10 11 12 Periods (1830-65) 13 14 15 10 11 Periods (1825-65) 12 13 14 Figure 12. Trajectory of Mrs Gaskell. 500 400 D e n 300 s i t i 200 e s 100 0 7 8 9 152 Figure 13. Trajectory of Thackeray. 500 400 D e n 300 s i t i 200 e s 100 0 9 10 11 Figure 14. Trajectory of Dickens. 12 13 Periods (1835-70) 14 15 16 153 300 250 D e n s i t i e s 200 150 100 50 0 4 5 6 7 8 9 Periods (1810-60) 10 11 12 13 Figure 15. Trajectory of Macaulay. 300 250 D e 200 n s i 150 t i e 100 s 50 0 5 6 7 Periods (1815-40) 8 9 154 Figure 16. Trajectory of Carlyle (early periods). 300 250 D 200 e n s i 150 t i e 100 s 50 0 10 11 12 Periods (1840-65) EBB 13 BRO 14 155 Figure 17. Trajectories of the Brownings. 400 D 300 e n s i 200 t i e s 100 0 9 10 11 12 13 14 Periods (1835-1880) Figure 18. Trajectory of George Eliot. 15 16 17 156 16,00 14,00 12,00 D e n s i t i e s 10,00 8,00 6,00 4,00 2,00 0,00 1 2 3 4 5 6 7 8 Decades Cl 1 Cl 3 Cl 2 Cl 4 Figure 19. State verbs : cumulated densities, by decades. 24,95 25,00 22,62 20,00 D e n s i t i e s 16,11 15,54 15,00 13,75 12,43 10,00 8,45 8,06 5,00 0,00 1 2 3 4 5 Decades 6 7 8 157 Figure 20. Prospective and retrospective verbs : densities, by decades. 120,00 100,00 D e n s i t i e s 80,00 60,00 40,00 20,00 0,00 1 2 3 4 5 6 7 Decades D 10 D 20+21 Gl D : 3 Figure 21. Locative & movement verbs and general density (adjusted), by decades. Values for the general density have been divided by 3 to facilitate comparison. 8 158 400 D e n s i t i e s 300 200 100 0 1 2 3 4 5 6 7 8 Decades D 31+32 D 30+33 Gl D Figure 22 : Process verbs and general density, by decades. 160 140 120 D e n s i t i e s 100 80 60 40 20 0 1 2 3 4 5 6 7 Decades State Mvt Prosp-retr Bd process Loc Unb. process 8 159 Figure 23. Density of major classes, by decades. 0,50 0,40 D e n s i t i e s 0,30 0,20 0,10 0,00 1 2 3 4 5 6 Decades 1-4 20-21 5-6 31-32 10 30-33 7 8 160 Figure 24. Percentage of major classes, by decades 0,60 0,50 0,40 P e r c e n t 0,30 0,20 0,10 0,00 1 2 3 4 5 6 Decades Pt Plpfct Past Inf Figure 25. Percentage of major tenses, by decades. Pfct Modals 7 8 161 6,00 24 5,00 D e n s i t i e s 4,00 3,00 38 35 2,00 17 24 10 9 20 22 7 1,00 3 0,00 17 0 1 2 8 5 2 3 4 5 6 Decades D old D new Figure 26. The two passives, by decades. The figures on the curves refer to the actual number of passives for each decade. 7 8 162 1000 N b o f 900 800 700 600 500 u n i t s 400 300 200 100 0 Freq Frequencies Gr.A Gr.B Figure 27. Number of lexical units, by frequencies, for groups A and B. 900 800 D e n s i t i e s 700 600 500 400 300 200 100 0 1 2 3 4 5 6 Decades Gl D D extpt 7 8 163 Figure 28. Extended present : adjusted densities. To facilitate comparison of the curves, figures for extended present have been multiplied by 10. 1200 D e n s i t i e s 1000 800 600 400 200 0 1 2 3 4 5 6 Decades Gl Density D have*100 Figure 29. Have. To facilitate comparison of the curves, figures for have have been multiplied by 100. 7 8 164 400 D 300 e n s i 200 t i e 100 s 0 1 2 3 4 5 6 7 8 7 8 Decades D mod*10 Gl d Figure 30. Modal auxiliaries. To facilitate comparison of the curves, figures for the modals have been multiplied by 10. 6,00 5,00 D e n s i t y 4,00 3,00 2,00 1,00 0,00 1 2 3 4 5 Decades 6 165 Figure 31. Shall & will. 4,00 D e n s i t y 3,00 2,00 1,00 0,00 1 2 3 4 5 6 7 8 Decades D ma/mi Figure 32. May & might. 1,80 1,60 1,40 D e n s i t y 1,20 1,00 0,80 0,60 0,40 0,20 0,00 1 2 3 4 5 Decades 6 7 8 166 Figure 33. Must. 500 400 MP E 1814 1815-17 D e 300 n P&P s i 200 t y 100 0 1813 Dates Novels Letters 167 Figure 34. Density in Jane Austen’s letters and novels. P & P : Pride and Prejudice (1813) ; MP : Mansfield Park (1814) ; E : Emma (1816). 300 250 D e 200 n s 150 i t 100 y 50 0 Periode 1840-44 1845-49 Dates D Letters Figure 35. Density in Macaulay’s Letters and History. D History 1850-54 168 600 518 509 500 459 D & s i z e 400 C. Phill 334 300 N &S 305 Cranf. 234 MB 190 200 100 39 0 1850 1855 1860 1865 Periods D Marianne Wds(thds) D. novels Figure 36. Mrs Gaskell’s letters to Marianne and novels. MB : Mary Barton (1848) ; Cranf. : Cranford : 1853 ; N & S. : North and South (1855) ; C. Phillis and other Tales (1860, posthumous). 169 Progressives in novels since 1700 By 20-year spans 700 600 D e n s i t i e s 500 400 300 200 100 0 1740 1760 1780 1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 Dates (ends of spans) Densities Figure 37. The progressive across three centuries. — —INDEX (Text) Adamczewski 14; 69 Aitken & McArthur 39 Anderson 92; 99 Archer-Jackson 40; 43 Aristotle 10; 15 Arnaud 1; 6; 11; 15; 25; 26; 32; 44; 45; 46; 50; 51; 55; 69; 86; 89; 92; 93; 94 Austin 13; 79 Bailey 34; 56; 99 Bain 29; 48; 60; 67; 75 Basch 96 Bennett and Partee 62 Benveniste 7; 12; 69 Best 21 Bloomfield 100 Boulle 99 Bourdieu 21 Brorström 36; 39; 48 Buehler 100 Bull 15 Buyssens 14; 69; 70 Chambers 34 Charleston 14; 69; 98 Close 9 Collobert 10; 71 Comrie 64; 92 Culioli 7; 12 Curme 98 De Camp 45 Denison 56; 86 Descartes 10; 76 Desclaire 21 Dorozewski 95 Dowty 62 Eckert 96 Encrevé 21 Gauchat 96 Gérin 43 Givon 98 Goldsmith & Woisetschlaeger 99 Groussier 7; 12 Guillemin-Flescher 86 Gusdorf 98 Guy 95 Haight 23; 26; 85 Hatcher 9; 10; 98; 99 Heine 90; 92; 95 Herzog 4; 36 Hill 15 Humboldt 98 Jakobson 100 Jespersen 19; 58; 59; 69; 99 Joos 14; 15; 46; 76; 79; 81 Kipers 53; 96 Knott 25 Labov 4; 5; 6; 11; 21; 22; 24; 27; 28; 32; 34; 36; 40; 44; 50; 86; 90; 96; 97; 100 Lawton 40 Leech 59; 60; 61; 63; 64; 71 170 Leonard 22; 37 Leskien 44 Ljung 45; 46; 69; 73; 76; 79; 80; 81; 99 Mayhew 39 Meillet 5; 92; 95 Milroy 27 Montague 62; 63 Mossé 10; 25; 37; 40; 57; 58; 98 Murray 58 Nakamura 86 Norman et al. 95 Normand & al 92 Normand et al 44 Nurmi 91 Ota 9; 19; 25; 46; 52; 69; 73 Pagliuca 90 Phillipps 37; 39; 42; 86; 87 Phillips 21; 86; 87 Piaget 15 Raumolin-Brunberg 34 Raybould 37; 87; 88 Romaine 40 Rydén & Brorström 11 Rydén 6; 36; 39; 48 Sag 45 Sankoff 21 Saussure 12; 95; 100 Scheffer 70; 98 Schuchardt 92; 100 Searle 74; 81 Smith 22; 26; 69 Smitterberg 55; 93 Strang 88; 89; 98 Swadesh 46 Tarde 51; 92; 95 Thompson 39 Traugott 90; 99 Tulloch 29; 39; 88; 89 Vendler 7; 73 Vlach 62 Wang 12; 50 Warner 56; 91; 92 Weinreich 4; 36; 90 Weinreich, Labov & Herzog 4 Wright 7; 99 INDEX (Footnotes) Buyssens 60 —C— Chartier et al 22 Cobb 17 Culioli 19; 61; 67; 94; 100 —D— Denison 37; 38; 56 Desclaire 26 Destutt de Tracy 72 Diver 60 Durkheim 95 —G— Gauchat 23; 34 Görlach 93 Grevisse 90 Grize 100 Groussier 61; 67 Guillemin-Flescher 12; 94 Gusdorf 98 —H— Hancil 91 Hatcher 7; 73 Herdan 84 Herzog 23 Hill 19 Holmes 42 —I— Imbs 60 —J— Jacobson 84 Jespersen 36; 75; 90 Joos 19; 60 Juillard 84 —K— —A— Allen 60; 70 Aristotle 17; 46; 48 Arnaud 25; 33; 71; 93; 114 —B— Bain 8; 39 Benveniste 10; 47 Bergson 8 Biber & Finegan 55 Braconnier 96 Brörstrom 91 Brorström 25 Kay 99 Kenny 15 Klein 9 König 14 Koschmieder 13 —L— Labov 23; 34; 97 Leech 60 Ljung 16; 74 171 —M— Mallarmé 84 Marchello-Nizia 6 Mossé 4; 32; 33; 36; 93; 114 Muller 84 Murray 8 Mustanoja 93 —S— Sag 70 Scheffer 33; 60; 85; 114 Shuy 56 Sjoestedt 37 Smitterberg 4; 7; 14; 55; 56; 91; 94; 118 Stanzel 37 Strang 25; 29; 49; 87; 94; 114 —N— Nakamura 37; 56 Nehls 4; 93 Nevalainen 1; 4; 21; 25; 110; 111 Nickel 33; 114 Nugent 98 Nurmi 93 —P— Paul 95 Perelman 13 Perlmutter 47 Phillipps 39 —R— Raumolin-Brunberg 4; 25 Rivière 61; 67 Romaine 4 Rostand 42 Russell 74; 79; 110 Ruwet 47 Rydén 25; 91 Rydèn & Brörstrom 91 —T— Tagliamonte 24 Taine 84 Thoiron 84 Tulloch 85 —V— Van Hamel 37 Vaugelas 90 Vendler 15 Visser 37 —W— Warner 1; 37; 42; 56; 91; 92; 113 Weinreich 23 Whorf 98 Williams 57 Wright 94 —Y— Yule 84