Skip to navigation – Site map

Notes on peripeteic when clauses

Graham Ranger
p. 113-133

Abstract

On considère généralement que la catégorie des subordonnées adverbiales possède un fonctionnement selon lequel une subordonnée repère permet de localiser une principale repérée. Certains emplois de when, cependant (il s’agit du ‘when de péripétie’) semblent remettre en cause cette représentation, dans la mesure où les rapports repère-repéré habituels sont à première vue inversés. Une description précise des propriétés aspectuo-modales des procès en cause permet d’effectuer un classement bipartite parmi ces énoncés, distinguant ceux qui marquent une succession immédiate de procès, et ceux qui marquent l’interruption d’un premier procès par un deuxième. On s’aperçoit également que le fonctionnement de ce type d’énoncé peut être rapproché à celui des constructions corrélatives. A partir de la problématisation et de la description, on mettra à l’épreuve trois hypothèses d’explication. On trouvera, enfin, que le fonctionnement du when de péripétie relève d’un processus de réanalyse selon laquelle l’instabilité quantitative de la lexis marquée comme la subordonnée la rend inapte à jouer un rôle de repère. Sa stabilisation passe par la recherche d’un site, qui se trouve dans la principale. L’effet corrélatif est donc induit par un conflit entre les contraintes de la syntaxe et celle de rénonciation. La réanalyse "en cours d’énonciation” ainsi opérée pourrait être vue comme le reflet iconique de l’effet de péripétie, décrit dans le Petit Robert comme un "changement subit de situation dans une action dramatique, un récit”.

Top of page

Full text

0. Introduction

1In the following pages, we shall be studying a rather particular, but not unusual, use of the subordinating conjunction when, such as can be exemplified in utterances like the following:

1. I was just beginning to write my article, when the telephone rang.

2Such a use of when, innocent as it may appear, poses in fact a number of serious theoretical problems to the description of this marker, and indeed to an enunciative characterisation of adverbial subordination in general.

3In what follows, we shall begin by first elucidating the problem. We shall see how the adverbial subordinating conjunction when is normally described, in both traditional and enunciative models of grammar, before demonstrating that both types of description are inadequate to account for the use made of when in utterances like (1).

4We shall then leave these more general theoretical considerations to one side to describe in greater precision utterances like (1), basing our description on a limited corpus of such utterances. The utterances can be divided into two main types, according to the relation established between the processes in the two clauses. We will also be excluding from our analysis a superficially similar type of utterance.

5This investigation will, finally, lead us to suggest ways in which the description of when mooted in the first section might be modified and refined to account for cases which could otherwise only be labelled “exceptions to the rule”. We shall consider three explicative hypotheses, each formulated within the framework of the theory of enunciative operations.

1. The problem

6By “peripeteic when constructions”, we understand constructions of the general form “PROPOSITION 1 when PROPOSITION 2”, in which the first proposition is, in syntactic terms, the main clause, and the second the subordinate clause. The Greek peripeteia from which we have derived “peripeteic” signifies an unforeseen event, and the word is used technically to refer to a “sudden change of fortune or reverse of circumstances” (SOED 1556). Such a description of the narrative relationship between the two clauses seems to correspond to an intuitive appreciation of the following examples:

2. [...] I made good my retreat to the narrow tunnel But I had scarce entered this when my light was blown out and in the blackness I could hear the Morlocks rustling like wind among leaves, and pattering like the rain, as they hurried after me. (Wells)
3.
1 had not gone more than 150 yards, however, when I heard a hideous outcry behind me, which caused me to run back again. (Conan-Doyle)
4.
I was just fitting my key into the door when I noticed a man at my elbow (Buchan)
5. I had come to the conclusion that he had dropped asleep, and indeed was nodding myself, when he suddenly sprang out of his chair with the gesture of a man who has made up his mind and put his pipe down upon the mantelpiece. (Conan-Doyle)
6.
I slept at Baker Street that night, and we were engaged upon our toast and coffee in the morning when the King of Bohemia rushed into the room. (Conan-Doyle)
7.
With an apology for my intrusion, I was about to withdraw when Holmes pulled me abruptly into the room and closed the door behind me. (Conan-Doyle)

7Despite their relative frequency, such utterances pose problems for the formalisation of relationships of adverbial subordination. In this section we shall be looking, firstly, at traditional approaches to adverbial subordination (1.1), and then,in more detail, at enunciative approaches to the question (1.2). This presentation will enable us more properly to appreciate the problems created by utterances like those above (1.3) and to show how the few studies in which such utterances are mentioned fail really to go beyond the level of a superficial description (1.4).

1.1. Traditional approaches to adverbial subordination

8Arrivé, Gadet and Galmiche base their discussion of circumstantial clauses on the following general definition of the category:

9“Type de subordonnée [...] la proposition circonstancielle est traditionnellement définie par sa forme (elle commence par une conjonction de subordination) et par son sens (elle exprime l’une des circonstances dans lesquelles se déroule l’action de la principale)”. (1986 : 104)

10Dubois-Charlier and Vautherin adopt a similar, semantic, definition when they define temporal clauses as follows:

11“Les circonstancielles de temps répondent à la question : ‘l’action s’est passée quand ?’ ; elles indiquent que l’événement décrit par GN + GV s’est passé avant P’, après P’, pendant P’, etc.” (1997 : 259)

12And so, if we take an utterance like the following:

8. When I realized this, I hurriedly slipped off my clothes, and, wading in at a point lower down, I caught the poor mite and drew her safe to land (Wells)

13the “action” would be, roughly speaking, I hurriedly slipped off my clothes, and, in response to the question “When did the action take place”, the answer would correspond to the subordinate clause: When I realized this.

14We might say that, in this, fairly traditional, approach to the problem, when adverbial clauses could be described as clauses which inform us as to the temporal circumstances of the event or situation referred to in the main clause.

1.2. Enunciative approaches to adverbial subordination

15A similar idea is expressed in linguistically more rigorous terms by Lapaire and Rotgé:

16“Dans la [...] structure WHEN Prop1, Prop2 WHEN sert à poser un point de repère temporel par rapport auquel la deuxième proposition sera effective. On a alors le fonctionnement suivant : ‘je-énonciateur pose un repère temporel (Prop1) tel que quand Prop1 est vrai, alors Prop2 l’est aussi [...]’. La deuxième proposition dépend ainsi de la réalisation de la première, indicateur d’ordre temporel.” (1991 : 685)

17The notion of repérage, or the localisation of one proposition with regard to another, is a recurrent feature in other formal linguistic descriptions of when:

18“En tant que conjonction de subordination, WHEN introduit une proposition subordonnée qui constitue le repère de la proposition principale ou proposition repérée (Bouscaren, Persec et al 1998 : 247).

19“Les conjonctions de subordination indiquent le repérage entre une proposition principale (le repéré) et une proposition subordonnée (le repère)” (Groussier & Rivière 1996:45)

20And so, if we return to our example above:

8. When I realized this, I hurriedly slipped off my clothes [...] (Wells)

21here we can consider that the subordinate proposition I realized this serves as the temporal locator (in French: repère) for the main proposition I hurriedly slipped off my clothes, which is the locatum (repéré).

22So far, then, we have seen that, intuitively and linguistically, the idea of considering a when subordinate proposition as a sort of locator, serving to situate temporally the predication in the main proposition, is a fairly appealing one. This idea is, however, seriously challenged by the existence of one particular category of when subordinate propositions which we shall be turning to now.

1.3. The problem posed by peripeteic when clauses

23The problem is that if we take an utterance like the fabricated example presented in the introduction:

1. I was just beginning to write my article, when the telephone rang.

  • 1 Things are a little more complicated than this, but for the time being, we are dealing with “first (...)

24it is very difficult to consider that the when proposition informs us of “when the action takes place”, to paraphrase Dubois-Charlier and Vautherin above. In fact, one almost has the impression that the “action” (or rather, the event) in the utterance in question is the telephone rang, and that the main proposition (syntactically speaking), i.e. I was just beginning to write my article, informs us as to when the action in the subordinate clause takes place!1 In other words the relationship usually described between a locating subordinate proposition, and a located main proposition appears here to be inverted.

  • 2 There is some change, of course, as the following discussion will make clear, but the relationship (...)

25One argument in favour of this is that, when one changes the position of the conjunction, leaving the proposition where it is, the meaning does not seem to change significantly2:

1. When I was (just) beginning to write my article, the telephone rang.

26There are, however, other arguments, related to certain syntactic properties of the utterances in question. Firstly, the construction allows subject-verb inversion, while, to the best of my knowledge, the same type of inversion is impossible in normal when subordinate clauses:

9. Away they went, and I was just wondering whether I should not do well to follow them when up the lane came a neat little landau, the coachman with his coat only half-buttoned, and his tie under his ear, while all the tags of his harness were sticking out of the buckles. (Conan-Doyle)

27The above example is authentic, but it is easy to fabricate, from authentic utterances, other perfectly acceptable similar examples:

10. There was no one in the little street, so I dropped the milk-cans inside the hoarding and sent the cap and overall after them. I had only just put on my cloth cap when a postman came round the corner. I gave him good morning and he answered me unsuspiciously (Buchan)
10’.
I had only just put on my cloth cap when round the corner came a postman.

28Now, it would obviously be impossible to say:

11”. *When round the corner came a postman, I had only just put on my cloth cap

29This remark brings us to a second property of this sort of when proposition: in such utterances, the when proposition is always in second position. Even when, as in (12), there is no inversion, changing the order of the two propositions produces a very unnatural result:

11’”. ?? When a postman came round the corner, I had only just put on my cloth cap.

30Another remarkable property of this sort of utterance is the fact that one can find interrogatives in the subordinate proposition. We have found no examples in a fairly restricted corpus we were working with, but the following utterances (12) and (6) may easily be modified to give (12’) and (6’):

12. One night - it was on the twentieth of March, 1888 - I was returning from a journey to a patient (for I had now returned to civil practice), when my way led me through Baker Street. (Conan-Doyle)
12’.
One night [...] I was returning from a journey to a patient [...], when where should I find myself but in Baker Street?
6.1 slept at Baker Street that night, and we were engaged upon our toast and coffee in the morning
when the King of Bohemia rushed into the room. (Conan-Doyle)
6’.
[...] we were engaged upon our toast and coffee in the morning when who should rush into the room but the King of Bohemia?

  • 3 A study of this very unusual interrogative form would merit a separate article!

31Admittedly, the type of question involved is always of the same form (i.e. wh- should... but...?), and this is, remarkably, the only syntactic environment in which such questions occur3, but the very fact of finding an interrogative — which is normally an independent speech act — in a subordinate proposition — which is syntactically dependent — is noteworthy.

  • 4 That is, something which, if it is not necessarily already known to the co-speaker, is at least pre (...)

32It should by now be obvious that such when propositions cannot be considered to locate the main proposition in the same way as subordinate when propositions normally do. After all, both subject-verb inversions of the type mentioned and interrogatives normally function as independent utterances. The fundamental problem is that it is hard to consider that the when clause locates the main clause, given that the when clause is here used to vehicle new information, while a locating clause ought normally to constitute a stable reference point4.

1.4. Linguistic approaches to peripeteic when clauses

33There has not, to the best of my knowledge, been any modem enunciative linguistic study of peripeteic when constructions, but the peculiarities of such utterances have been remarked upon.

34We owe our use of the name “peripeteic” to Adamczewski and Delmas, who write:

35“[...] il existe d’autres constructions intéressantes du type S (,) when S2 (avec ou sans la virgule) :

  • 5 The numbering of examples follows that of the present article and not that of the original.

36[13]5 I was weeding the dahlias, when my telephone began to ring.

37(When introduit bien un énoncé rhématique ici ; d’ailleurs cette espèce particulière de when a été appelée parfois le “when de péripétie”.)” (1982:347)

38The authors do not, unfortunately, mention the source of this appellation.

39Lapaire and Rotgé also signal the exist of such utterances:

40“WHEN sert le plus souvent à ancrer un procès temporellement par rapport à un autre, comme en [14] :

[14] I said that I was not averse to talking, that I had just been rather immersed in something WHEN he arrived, and I begged him...

41Il est des cas néanmoins où cette conjonction est utilisée pour relancer la narration, c’est-à-dire pour signaler une nouvelle étape dans la narration (WHEN = AND THEN dans ce cas de figure) :

[15] We had been talking together for hours on end WHEN suddenly he got up and rushed to the door.

42“Il serait alors légitime de considérer WHEN he got up... en [96] comme une sorte de coordination.” (1991 : 586)

43Again, the idea that these utterances are used to signal a new step in the narrative is interesting, but the authors do not take this idea further. They legitimately paraphrase when with and then, and describe it in this case as a “sort of coordination”, but do not linger on how unusual it is for a construction which, in syntactic terms, has nothing to distinguish it from a case of adverbial subordination, to function, here, as a coordinating construction.

44Among attempts to explain this use of when, we might mention Arrivé, Gadet and Galmiche, who describe the phenomenon (in the case of quand and lorsque) as a sort of “inverse subordination” (i.e. the clause marked as a subordinate clause in fact functions as the main clause and vice versa):

45“La conjonction introductrice [de circonstancielles de temps] la plus fréquente est quand, souvent en tête de phrase. Quand exprime la simultanéité quand le temps des deux proposition est le même (quand il partait, je souriais), et l’antériorité quand un temps simple s’oppose à un temps composé (quand il fut parti, je souris). Quand et lorsque se prêtent à la subordination inverse, où la subordonnée, placée en fin de phrase, constitue le propos (on était au fromage quand un orage éclata).” (1986: 110)

46Now, although we would entirely agree that the when clause has to be postposed for a peripeteic interpretation to be possible, this is not enough. More often than not, in fact, postposed when clauses fonction as standard temporal adverbials, even if they do also constitute the comment (or “propos”, according to one definitionof this term). Even the type of example cited above need not necessarily be interpreted as peripeteic. Let us take the two examples below to illustrate our point:

16. We had just started on the cheese and biscuits when the storm broke.
17. A : What were you doing when the weather turned?
B : (Let me think, oh yes) We had just started on the cheese and biscuits when the storm broke.

47(16) would, of course, most naturally be interpreted as a peripeteic use of when, while the topicalisation of the storm broke, created (rather artificially, it must be admitted) in (17) by the question-answer sequence prevents a peripeteic reading.

48In the authentic examples below, the when clauses are also postposed, and could perhaps be described as the "comment", but still do not construct a peripeteic meaning:

18. But I made a sudden motion to warn them when I saw their little pink hands feeling at the Time Machine. (Wells)

19. I used to thank God for such mornings way back in the Blue-Grass country, and I guess III thank Him when I wake up on the other side of Jordan (Buchan)

49Some utterances, finally, remain ambiguous in the absence of sufficient context:

20. I married, too, and though my wife died young she left me my dear little Alice. Even when she was just a baby her wee hand seemed to lead me down the right path as nothing else had ever done. In a word, I turned over a new leaf and did my best to make up for the past. All was going well when McCarthy laid his grip upon me (Conan-Doyle)

50Either: i. we already know that McCarthy is going to lay his grip on the speaker, in which case the utterance All was going well when McCarthy laid his grip upon me might be considered as the answer to the (hypothetical) question : How were things when McCarthy laid his grip upon you? and, in this case, the subordinate clause will be largely unstressed, or: ii. the arrival of McCarthy unexpectedly spoils an otherwise idyllic situation, in which case both clauses will probably be stressed as if they were independent of each other.

51In the light of the above remarks, it seems impossible to consider that the peripeteic interpretation is created uniquely by the order of the propositions. The postposition of the when clause is a necessary, but not a sufficient condition.

52We have just mentioned the fact that a difference in readings provokes a prosodic difference. Some authors have indeed suggested that the peripeteic interpretation is a function of the intonation of the utterance. Bally, for example, writes:

  • 6 An AZ reading corresponds to a rising first segment, and a second segment carrying "l'intonation mo (...)

53“Le jeu de la mélodie suffit aussi à transformer une subordonnée en principale, et inversement. “ Nous étions au jardin lorsque l’orage éclata ” est une phrase liée contenant une principale et une subordonnée ; mais prononcez-la en AZ, elle équivaudra à “Alors que nous étions au jardin (A), un orage éclata (Z).” De même : “Nous étions à peine rentrés, que l’orage éclata” signifie en réalité ‘"Un peu après que nous étions rentrés (A), l’orage éclata (Z)”; là aussi, la succession des proposition est l’inverse de ce que fait prévoir la forme matérielle, et seule la mélodie décide de l’interprétation [...]”6 (1950 : 64)

  • 7 "Des phénomènes de prosodie entrent alors en jeu. Nous ne les examinons pas ici" (1993:87). In fact (...)

54Similarly, in her article on when, Zeitoun (1993) mentions in a note the fact that “la principale peut parfois être le repère et la subordonnée le repéré” (1993:87), considering this to depend upon the prosody7.

  • 8 Note that our examples have, for the most part, been drawn from written texts.

55This position is not easy, however, to defend: if it is indeed true that one does not pronounce such utterances in the same way as standard uses of the subordinating conjunction when, it is nonetheless undeniable that even the meagre context given above in (l)-(7)8, for example, and elsewhere, is enough for us to understand, without having heard the utterance spoken, that the when in question is not like the standard when. The only time when the prosody might play a significant role is in the case of utterances which are otherwise ambiguous, or where we do not have sufficient context.

1.5. Recapitulation

56In this section we have endeavoured to explain the nature of the problem posed by peripeteic when clauses. Both traditional and linguistic models of adverbial subordination consider the subordinate clause to be a locator, and thus to constitute a stable reference point. Peripeteic when, however, introduces new, or unexpected elements into a narrative, and does not correspond to the general model. It is a rarely noted phenomenon, which can, we have seen, be explained satisfactorily neither by the order of the propositions nor by the intonation. Both factors are certainly important in such utterances, but are not sufficient to explain the peripeteic interpretation.

2. A description of peripeteic when clauses

57In this section, we shall be attempting to establish a basic typology of peripeteic when clauses. Before we go any further, it is perhaps worth mentioning the fact that, stylistically, such clauses typically belong to narrative texts, generally written in the preterite tense, and to the adventure story genre. It is not, however, this feature upon which we will base our typology but the relation of consecution obtaining between the process in the first clause and the one in the second.

58Let us call the process featured in the first clause PI, and the process in the second, when, clause, P2. Let us represent the processes in question as intervals on the class of instants.

59In the vast majority of examples of peripeteic when in our corpus, PI is either in the be+ing form or in the have -en form (both coupled with the preterit tense). In both cases the process might be said to function continuously — i.e. both refer to a stable situation in which one instant validating P is potentially identifiable to the surrounding instants.

60P2, on the other hand, is always in a simple form (the preterit), and functions punctually — i.e. we can consider that the distance separating the boundaries representing the beginning and the completion of the process is, linguistically — if not realistically —, zero. Another, more subtle point concerns the nature of the processes in question which is not merely “punctual”, but which introduces a discontinuity into a stable referential frame which is, after P2 has occurred, irremediably altered.

61To take a banal example, the following utterance features a PI in the have - en form, and a simple form of a punctual process in P2.

21. We had been sitting in silence for some minutes when Joe coughed.

62The result is rather strange. One would expect (21) to be followed by some extra information, making it clear how the fact that Joe coughed changed the situation. If we modify (21) to give (21 ’):

21’. We had been sitting in silence for some minutes when suddenly Joe coughed so hard that...

63then the addition of suddenly and the adverbial qualification seem to render the utterance more natural.

64Normally, however, P2 is a process which functions, either as the predication of existence of some event {my light was blown out, in (2), he suddenly sprang out of his chair, in (5), the king of Bohemia rushed into the room, in (6), Holmes pulled me abruptly into the room (7)) or as the perception of something new (I heard a hideous outcry (3), I noticed a man at my elbow (4)).

65We might then represent the situation figured in such utterances as follows:

66Thus P2 coincides with the end of PI. In fact, "coincides with” is a little misleading, because P2 is really contiguous with the end-point of PI. Either contiguous and to the right of the end-point of PI (2.1) or contiguous and to the left of the end-point of PI (2.2).

2.1. Immediate succession

67The first type of peripeteic when clause, then, is that in which P2 is contiguous to and to the right of the end-point of PI. This is what we find in utterances like the following:

2. [...] I made good my retreat to the narrow tunnel But I had scarce entered this when my light was blown out and in the blackness I could hear the Morlocks rustling like wind among leaves, and pattering like the rain, as they hurried after me. (Wells)
3.
I had not gone more than 150 yards, however, when I heard a hideous outcry behind me, which caused me to run back again. (Conan-Doyle)
22.
I had only just put on my cloth cap when a postman came round the corner. I gave him good morning and he answered me unsuspiciously (Buchan)
23.
‘It was not [Lord Alloa], ’ I cried; ‘it was his living image, but it was not Lord Alloa. It was someone who recognized me, someone I have seen in the last month. He had scarcely left the doorstep when I rang up Lord Alloa’s house and was told he had come in half an hour before and had gone to bed.’ (Buchan)
24.
She had hardly said the words when young Mr. McCarthy came running up to the lodge to say that he had found his father dead in the wood, and to ask for the help of the lodge-keeper. (Conan-Doyle)

68PI, in the had -en form, creates the representation of a process which is complete from a given reference point. The reference point is provided by P2. It is remarkable how often such utterances include aspectuo-modal adverbial expressions in the first clause; we find, respectively, scarce, not... more than, only just, scarcely, hardly. Such expressions are used, roughly speaking, to more or less the same end. They express the fact that the gap separating the end-point of PI from P2 tends towards zero. This is achieved by using negatively charged expression like scarce, scarcely or hardly, by denying (not) any increment (more than) between the completion of PI and P2, as in (3), or by asserting a "perfect fit”, or "adjustment” (just) between the end-point of PI and P2, as in (4).

69The effect obtained from the minimalising of the distance between the completion of PI and P2 in all the above examples is that of an immediate succession of events. We might represent this schematically as follows:

70----------------P1----------------------------][P2]

2.2. Interruption

71The second type of peripeteia when clause is that in which P2 is contiguous to and to the left of the end-point of PI. The following examples illustrate what we mean by this description:

4. I was just fitting my key into the door when I noticed a man at my elbow (Buchan)
5.
I had come to the conclusion that he had dropped asleep, and indeed was nodding myself, when he suddenly sprang out of his chair with the gesture of a man who has made up his mind and put his pipe down upon the mantelpiece. (Conan-Doyle)
6.
I slept at Baker Street that night, and we were engaged upon our toast and coffee in the morning when the King of Bohemia rushed into the room. (Conan-Doyle)
7.
With an apology for my intrusion, I was about to withdraw when Holmes pulled me abruptly into the room and closed the door behind me. (Conan-Doyle)
25.
So determined was their denial that the inspector was staggered, and had almost come to believe that Mrs. St. Clair had been deluded when, with a cry, she sprang at a small deal box which lay upon the table and tore the lid from it. (Conan-Doyle)
26
. So he sat as I dropped off to sleep, and so he sat when a sudden ejaculation caused me to wake up, and I found the summer sun shining into the apartment. (Conan-Doyle)
27
. We had reached Baker Street and had stopped at the door. He was searching his pockets for the key when someone passing said:
“(Good-night, Mister Sherlock Holmes.”
(Conan-Doyle)

72Here PI, typically, though not always, in the be+ing form, creates the representation of a process which has not yet reached completion from a given reference point. This reference point is again provided by P2. Aspectuo-modal expressions are less frequent in these utterances, though we might mention just, in (4). Examples (4) and (5) feature the be+ing form, while in (6) and (7) another form is used, but with similar meaning. We might indeed modify (6) and (7) to include be+ing forms without significantly altering the aspectual representation of PI :

6’. ... we were eating our toast and drinking our coffee...
7’. ... I was going to withdraw...

73Interestingly, (25) includes the have -en form, but nonetheless constructs a representation of an uncompleted process, as almost in combination with have -en in fact asserts that the end-point of the process is very close, but has not yet been crossed.

74(26) does not feature be+ing either, but contains the anaphorically reinforced repetition of the verb sit thus allowing us to construct a similar aspectual form to that of the other utterances.

75(27), finally, is rather amusing, for the event in P2 remains unexplained until the last lines of the story, where we learn that the "someone" referred to was in fact the lady Holmes had been following, but in disguise and unknown to him. The peripeteic aspect of the clause — the sudden or unforeseen event — is thus not immediately exploited, but contributes to creating a problem for the reader (why mention something so banal?) which is only resolved at the end of the short story.

76The effect obtained here is of the interruption of a stable situation represented continuously in P1 by P2. In schematic terms:

77-----------------P1------------------------][P2]

2.3. A similar case excluded

78Before going further, it is perhaps worth differentiating between the two above cases, and a related case which, although it does share common features with peripeteic when clauses, does not function in quite the same way.

28. And it was already long past sunset when I came in sight of the palace, silhouetted black against the pale yellow of the sky. (Wells)
29.
It was a quarter past six when we left Baker Street, and it still wanted ten minutes to the hour when we found ourselves in Serpentine Avenue. (Conan-Doyle)
30.
It was nearly four o 'clock when we at last, after passing through the beautiful Stroud Valley, and over the broad gleaming Severn, found ourselves at the pretty little country-town of Ross. (Conan-Doyle)

79In the above examples, the first clause features a definite time expression, while the when clause contains the predicative relation located by the time expression. Superficially, then, the case is similar, as the subordinate clause seems to vehicle most of the relevant information. However, there are differences. The syntactic peculiarities of peripeteic when mentioned above (interrogatives and inversions) are not possible in such utterances:

31. *It was five o ’clock when who should phone me up but Bob?
32.
??It was five o ’clock when along came Bob.

  • 9 That is to say that the first clause is pronounced as an independent clause while the second is una (...)

80In fact, such constructions are best analysed as cleft constructions, in which the information in the when clause has been preconstructed (cf. already in (28) and {at last in (30)), and in which the relevant information is really the temporal local-isation of the predicative relation in the second clause. By way of a confirmation, such constructions behave prosodically as clefts9, and not as peripeteic when clauses.

2.4. Recapitulation

81To recapitulate briefly, in this section, we have tried to establish a typology of peripeteic when clauses. Such utterances may be classed into two basic types. In the first type, P2 intervenes immediately after completion of P1, constructing an immediate succession of events. In the second type, P2 intervenes immediately before completion of P1, in fact interrupting P1. These cases may be represented schematically as situations of contiguity between the end-point of P1 and P2.

3. Hypotheses to account for when

82The typology we have established above is descriptive, and takes into account the relationship between the two processes. We have not yet looked closely at the way in which when, the element conjoining the two clauses, functions. If we consider that when serves to mark a single operation, then the problem is to explain how this operation can unambiguously generate both standard and peripeteic uses of when. In this section, we shall be looking at three explicative hypotheses.

3.1. A non-oriented localisation

83We saw earlier that enunciative studies of when describe the marker as establishing a relation of localisation between two propositions. Zeitoun, for example, writes:

84“le procès contenu dans la subordonnée (terme repère) localise dans le temps celui auquel renvoie la principale (terme repéré)” (1993 : 83)

  • 10 This form of symbolic representation here, and more particularly later on, is based on the formal m (...)

85If we represent the temporal coordinates of the first clause and the second clause as t1 and t2 respectively, then the description proposed by Zeitoun yields the following symbolic representation10:

86PROPOSITION 1 when PROPOSITION 2 <t1 t2>

87when PROPOSITION 2, PROPOSITION 1 <t2 t1 >

88It is clear from this that the formal description generally proposed for when is oriented. In other words, when always indicates the locating lexis in an interpropositional relationship. Now, we have shown that such a description is inadequate to account satisfactorily for peripeteic uses. Our first hypothesis to account for peripeteic uses is that when still posits an interpropositional localisation, but does not indicate the orientation of the relation. According to this hypothesis, the same utterance, of the form PROPOSITION 1 when PROPOSITION 2 may potentially generate two inversely oriented representations.

89Either : <t1 t2> Or : <t1 t2>

90One first objection might be to say that such a model is too productive. If PROPOSITION 1 when PROPOSITION 2 may equally express a situation in which t1 is localised by t2 and one in which t2 is localised by tl, then how is one to choose the correct interpretation of any given utterance?

91The answer would be to say that the orientation of the relation of localisation depends upon other factors in the utterance. So, for example, with an utterance of the form when PROPOSITION 2, PROPOSITION 1, i.e. featuring when in initial position, it is always t2 which localises tl, because the topicalisation of the when clause implies its preconstruction and this, in turn, is a typical property of a locating lexis.

92In the case of peripeteic when clauses, we might say that the inversion in the orientation of the relation is constructed by a combination of factors including, for example, the aspectual form of the process and the presence of aspectuo-modal markers in the first clause, the simple form and the aktionsart of the process in the second clause.

93Thus, to take an example, in (2):

2. [...] I made good my retreat to the narrow tunnel. But I had scarce entered this when my light was blown out and in the blackness I could hear the Morlocks rustling like wind among leaves, and pattering like the rain, as they hurried after me. (Wells)

94the presence of had -en, and scarce in the first clause and the "punctual" process blow out in the second form mean that we interpret the second clause as localising the first.

95This hypothesis is at first sight appealing, but presents some problems which will prevent us from retaining it.

96The first objection is an interpretative one. If one considers that in such utterances one is indeed faced with a case of what Arrivé, Gadet and Galmiche refer to as “inverse subordination”, then things would be straightforward, but, as we saw in our second section, things are a little more complicated. In fact, if it is true that the process in the first clause provides the “backdrop” to the event represented in thesecond, and thus localises it, the event in the second clause serves at the same time to localise the end-point of the process in the first. Put differently, each clause determines the other, in a correlative pattern analogous with that found in comparative constructions.

97Indeed the example above (and many other examples of peripeteic when) can be manipulated to give recognised correlative expressions:

2’. ... scarcely had I entered this when...
2
”. ... no sooner had I enterd this when...

98And so it is not enough to say that in peripeteic when clauses t1 localises t2, for, at the same time, t2 localises t1. Our first hypothesis thus appears to be a little too constraining.

3.2. A doubly oriented localisation

99Our second hypothesis originates in an attempt to avoid the problems of the first, and to account for the correlative aspect of peripeteic when clauses. Instead of claiming that the operation marked by when is “either t2 localises t1 or t1 localises t2”, we might suggest that when always marks a double localisation, i.e. “t2 localises t1 and t1 localises t2”, but, according to the context, one or other of the two possible orientations is often “filtered out”.

100In other words, according to this hypothesis, when is always correlative, i.e. it is always used for relations of mutual determination, but this doubly oriented relation is polarised, so to speak, by other factors in the utterance. Thus the preconstruction by topicalisation of one of the clauses would lead to its being considered the locator, whereas second position would be privileged for the locatum. Aspectuo-temporal forms, adjectives and the previous context might all contribute to the polarisation of a potentially double localisation. We can now modify the formalisation sketched above to give:

101(Potentially) Both: <t1 t2> And: <t1 t2>

102Intellectually, such a radical solution is tempting. Correlative uses of when, such as are to be found in peripeteic utterances and in hardly constructions are, according to this hypothesis, in fact, indicators of the double localisation which when always marks potentially. Normally, however, other features in the utterance “disactivate” one of the orientations.

103Diachronically too, the explanation seems to hold water. In Old English the word hwonne was used interrogatively, and other uses of Modem English when were expressed with a number of correlative devices, notably, tha... tha, thonne... thonne. Interestingly for us, Mitchell and Robinson note:

104“Much of the difficulty with correlative pairs arises from the fact that [...] the conjunction and the adverb have the same form, e.g. tha can mean both ‘when’ and ‘then’. [...] But the word order is [a] useful and reliable guide, for it may be taken as a pretty safe rule for prose that, when one of two correlative tha clauses has the word-order V. S., it must be the principal clause and tha must mean ‘then’.” (1992:68)

105When replaces the conjunction tha under the Norman influence in Middle English (cf. French quand), while the “adverb” tha (i.e. the correlative element in the main clause) disappears. We have seen that peripeteic constructions often contain correlative aspectuo-modal adverbs in the first clause, which bear on the gap between the completion of PI and P2. And so one might want to establish a rough parallel between, in Old English, tha... tha and, in Modem English peripeteic and correlative constructions, sequences like hardly/scarcely/just... when.

106Interesting as the solution is, it does pose a certain number of theoretical problems. If, as the hypothesis suggests, when is always potentially correlative, the fact remains that this potential is only realised in a very limited number of constructions in Modem English. To account for this fact we would need to have at our disposal a formidable set of “rules” to show how the surrounding context reduces the interpretative possibilities of an utterance. Worse still, the set of “rules” would in fact all tend towards the same conclusion, because when when is not correlative, it always indicates the locating lexis to the right. In other words, correlative uses of when seem to represent the only instance of when being used to mark a locator to its left.

107A last, related problem, if when is always correlative, is that of how to differentiate between PROPOSITION 1 when PROPOSITION 2, and when PROPOSITION 2, PROPOSITION 1. There can be no point in putting when at the beginning of the utterance, or between the two propositions, if when always indicates a double localisation.

3.3. The reanalysis of a right-oriented localisation

108With the last hypothesis we shall endeavour to reconcile, on the one hand, the infinitely more frequent use of when to indicate a locating lexis to its right, with the correlative function it can assume in certain types of utterance.

109Let us begin by positing that, in an utterance of the form PROPOSITION 1 when PROPOSITION 2, standard when normally establishes a relationship of the general form:

110< λ1 ( )t1 (t2) (λ2) >

  • 11 A similar formula might be given for utterances in which the when clause cornes first: <<( )t1 .. (...)

111i.e. when indicates that a first proposition XI is located with regard to an empty form ( )t1 representing a temporal coordinate t1. This is located with regard to a temporal coordinate t2 which in turn localises a predicative relation λ2.11

112When, in the above formula, still indicates the locating lexis to the right. In peripeteic uses of when, however, the locating lexis does not possess the referential stability required in order to serve as a locator. It predicates the existence of a new event, and requires, in turn, a locator to stabilise it. This locator is to be found in the first clause, where it is often announced by an aspectuo-modal adverb.

113To return to the example (2),

2. [...] I made good my retreat to the narrow tunnel. But I had scarce entered this when my light was blown out and in the blackness I could hear the Morlocks rustling like wind among leaves, and pattering like the rain, as they hurried after me. (Wells)

114Here the second clause, my light was blown out is marked by when as the locator for the first clause, I had scarce entered this. However, the process in the second clause functions punctually, introducing a new event — a discontinuity — into the referential frame. It is the second clause which, in turn, requires localisation, and this is obtained in the reference point implied by the use of the had -en form and scarce in the first clause. It is as if we are being told at the same time “I had entered this just before my light was blown out” and “my light was blown out just after I entered this”.

115And so the enunciative properties of λ2 (the fact in particular that in contains the predication of the existence of a new event) lead us to establish a second relationship whereby λ2 gains in determination form λI. In terms of the representation we obtain:

116a) < λ1 ( )t1 (t2) (λ2) > but λ2 is quantitatively unstable and requires locating, giving us a second representation:

117b) < λI (t1) ( )t2 (λ2) > in which the roles of t1 and t2 are inverted.

118In other words, peripeteic when does function correlatively, but this correlation is not there to begin with, so to speak, but is constructed when there is a conflict between when’s role of indicating a locator to the right, and the actual enunciative status of what is to the right.

119Such a hypothesis is perhaps less appealing intellectually than the more radical approaches in 3.1 and 3.2 but corresponds more precisely, we feel, with the way in which peripeteic when actually works. One has the impression, in the following (already quoted) phrases, that the standard when is somehow subverted, and that one is obliged to readjust one’s expectations and to make a reanalysis of when in mid-phrase.

3. I had not gone more than 150 yards, however, when I heard a hideous outcry behind me, which caused me to run back again. (Conan-Doyle)
4.
I was just fitting my key into the door when I noticed a man at my elbow (Buchan)

120The effect of reanalysis with peripeteic when is what distinguishes (3) and (4) from (3’) and (4’):

3’. When I had not gone more than 150 yards, 1 heard a hideous outcry behind me, which caused me to run back again. (Conan-Doyle)
4’.
When I was just fitting my key into the door, I noticed a man at my elbow (Buchan)

121(3’) and (4’) give us surprising events, but (3) and (4) iconically mirror the surprising events with a syntactic surprise in mid-phrase too! Put another way, it is not merely the relationship between events in peripeteic when clauses which creates the effect of peripeteia but also the very syntactic relationship between lexes, the representation of which has to be modified in mid-reception, so to speak. This imposed reanalysis is the direct consequence of a conflict between the syntactic markers and the enunciative status of the second clause.

4. Conclusions

122To sum up, in the preceding lines, we tried firstly to show that certain, peripeteic, when constructions cannot be analysed as standard uses of the adverbial subordinating conjunction when. We then, in a brief description of peripeteic uses of when, distinguished two main sub-types. In both subtypes there is a relation of contiguity between the moment of the process in the second clause and the end-point of the process in the first. When the end-point comes first, and the second process is contiguous to the right, then the interpretation is one of "immediate succession”. When the second process is contiguous and to the left of the end-point, then the interpretation is one of “interruption”. The distance between the two points is often assessed by means of aspectuo-modal adverbs in the first clause. We finally endeavoured to account for such peripeteic uses of when by testing a number of hypotheses. We preferred to adopt the hypothesis according to which peripeteic uses of when are the result of a conflict between the enunciative determinations of the second clause, and the syntax, thus provoking a reanalysis of the use made of when.

123The phenomenon of reanalysis provides a demonstration of the plasticity and complexity of linguistic information, such that a single utterance may receive different, coexisting and yet conflictual interpretations, the "signification" obtained resulting, in the end, from the conflicts and resolutions such a reanalysis implies. The point is probably that it is illusory to speak of the "final” interpretation of a given utterance. The interpretation, in such cases as these, is more a process than a result, and so to talk of finality would be irrelevant.

124A final remark prompted by the above lines concerns the role of the analysis of "micro-problems” in linguistic investigation. Grammars often consider exceptional uses of linguistic markers as fundamentally uninteresting curiosities. An alternative is to ask oneself how a marker which normally functions in one way can, in certain circumstances, often come to function in another. This in turn might lead us to modify the description for a marker in such a way as to recognise the plasticity and complexity of the construction of linguistic meaning. We might also be led to reconsider linguistic descriptions which are unproblematical, all the while one works with the most frequently occurring phenomena, but which reveal their limitations when tested on marginal cases.

Top of page

Bibliography

CORPUS

The corpus is based on the following electronic texts from the Gutenberg project :

BUCHAN, John, (1915) The Thirty-Nine Steps.

CONAN-DOYLE, Arthur, (1892) Selected stories from The Adventures of Sherlock Holmes (“A Scandal in Bohemia” “The Red-headed League” “A Case of Identity” “The Boscombe Valley Mystery” “The Five Orange Pips” “The Man with the Twisted Lip”).

WELLS, H. G., (1895) The Time Machine.

BIBLIOGRAPHY

The Shorter Oxford English Dictionary (1973) (ed. C. T. Onions), Oxford: Oxford University Press.

ADAMCZEWSKI, Henri & DELMAS, Claude, (1982) Grammaire linguistique de l’anglais, Paris : Armand Colin.

ARRIVÉ, Michel, GADET, François & GALMICHE, Michel, (1986) La grammaire d’aujourdhui, Paris : Flammarion.

BALLY, Charles, (1950) Linguistique générale et linguistique française, Berne : A. Francke.

BOUSCAREN, Janine & CHUQUET, Jean, (1987). Grammaire et textes anglais. Guide pour l’analyse linguistique, Gap : Ophrys.

BOUSCAREN, Janine et PERSEC, Sylvie, (1998) Analyse grammaticale dans les textes, Gap Ophrys.

CULIOLI, Antoine, (1999 [1982]) “Rôle des représentations métalinguistiques en syntaxe”, in Pour une linguistique de lénonciation t. 2, Gap : Ophrys.

DUBOIS-CHARLIER, Françoise & VAUTHERIN, Béatrice, (1997). Syntaxe anglaise, Paris : Vuibert.

GROUSSIER, Marie-Line et RIVIERE, Claude, (1996) Les mots de la linguistique. Lexique de linguistique énonciative, Gap : Ophrys.

LAPAIRE, Jean-Rémi & ROTGÉ, Wilfrid, (1991) Linguistique et grammaire de l’anglais, Toulouse : Presses Universitaire du Mirail.

MITCHELL, Bruce & ROBINSON, Fred, (1992) A Guide to Old English, Oxford: Blackwell.

ZEITOUN, Elizabeth, (1993) “When et la temporalité”, in Cahiers de recherche t. 6, Types de procès et repères temporels, Gap : Ophrys.

Top of page

Annex

I would like briefly to come back to the formula postulated for when in 3.3. We said that, in an utterance PROPOSITION 1 when PROPOSITION 2, when establishes a relationship of the general form:

< λ1 ()t1 (t2) (λ2)>

In fact we would posit that, whatever its function, when can be represented as follows:

<()t1 (t2) (λ2)>

Where t1 is the temporal coordinate of another lexis, we obtain the formula already given (i.e. the formula for adverbial subordination).

Where t1 is a temporal coordinate of a (temporal) notion, we obtain a relative:

33. Our chairs, being his patents, embraced and caressed us rather than submitted to be sat upon, and there was that luxurious after-dinner atmosphere when thought roams gracefully free of the trammels of precision. (Wells 14)

Where t1 stands alone, we obtain a nominal relative:

34. The bit I liked best was when Orson Welles and Joseph Cotton went up into the big wheel together.

Where t1 is combined with a scanning operation, we obtain an interrogative. The scanning operation being the result either of a direct question (34), or constructed with an “interrogative expression” (35):

35. When shall you be able to enter upon your new duties? 1217
36. Still, of course, I never dared to leave the room for an instant, for I was not sure when he might come, and the billet was such a good one, and suited me so well, that I would not risk the loss of it. 1282

Top of page

Notes

1 Things are a little more complicated than this, but for the time being, we are dealing with “first impressions”.

2 There is some change, of course, as the following discussion will make clear, but the relationship between locator and locatum seems to be the same, whatever the order.

3 A study of this very unusual interrogative form would merit a separate article!

4 That is, something which, if it is not necessarily already known to the co-speaker, is at least presented a unproblematical. Here the assertions vehicled in the when clauses are, on the contrary, completely unexpected in the context.

5 The numbering of examples follows that of the present article and not that of the original.

6 An AZ reading corresponds to a rising first segment, and a second segment carrying "l'intonation modale de toute phrase indépendante" (Bally 1950:62)

7 "Des phénomènes de prosodie entrent alors en jeu. Nous ne les examinons pas ici" (1993:87). In fact, Zeitoun does study a number of peripeteic uses of when, in which the relation between locatum and locator appears to be inverted, but she does not unfortunately draw attention to these particularities

8 Note that our examples have, for the most part, been drawn from written texts.

9 That is to say that the first clause is pronounced as an independent clause while the second is unaccentuated.

10 This form of symbolic representation here, and more particularly later on, is based on the formal model elaborated by Antoine Culioli (1982, for example) and presented for English in Bouscaren and Chuquet (1987).

11 A similar formula might be given for utterances in which the when clause cornes first: <<( )t1 .. (t2) (λ2)> λ1 >. The difference here is that the located lexis λ1 is in second position, while the locating expression is parenthesised (a reflection of the preconstruction necessary when the when clause comes first).

Top of page

List of illustrations

URL http://anglophonia.revues.org/docannexe/image/681/img-1.png
File image/png, 7.8k
Top of page

References

Bibliographical reference

Graham Ranger, « Notes on peripeteic when clauses », Anglophonia/Sigma, 10 (6) | 1999, 113-133.

Electronic reference

Graham Ranger, « Notes on peripeteic when clauses », Anglophonia/Sigma [Online], 10 (6) | 1999, Online since 15 June 2016, connection on 20 August 2017. URL : http://anglophonia.revues.org/681 ; DOI : 10.4000/anglophonia.681

Top of page

About the author

Graham Ranger

Université d’Avignon et des Pays de Vaucluse

By this author

Top of page

Copyright

Licence Creative Commons
Anglophonia – French Journal of English Linguistics est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Top of page
  • Logo Presses universitaires du Midi
  • Revues.org