Navigation – Plan du site

Locative Inversion In Discourse: a strategy of non-commitment

Christine Copy et Lucie Gournay


Dans cet article, l'inversion locative (IL) en anglais est étudiée en tant que réalisation syntaxique d'une prédication dénuée de toute prise en charge identifiée à un sujet-parlant.
Dans une première partie, on s'attache à montrer que les contraintes syntaxiques et les restrictions modales, connues dans la littérature et qui caractérisent la forme, signalent que l'inversion locative est la trace d’une prédication objectivée. Les contraintes observées sont par exemple l’incompatibilité avec les formes de remise en cause de l’assertion (négation ou modalisation épistémique), les fortes restrictions des déterminations aspectuelles (notamment BE+ING), incompatibilité avec l'interprétation générique et hors situation etc.
Dans une deuxième partie, la prédication objectivée marquée par l'IL est mise en relation avec son utilisation en discours dans deux contextes: i) les débuts de contes où son fonctionnement discursif et énonciatif est comparé à celui des constructions en there qui renvoient à une prise en charge par un énonciateur particulier; ii) les passages descriptifs, dont ceux des "crime scenes", où les occurrences d’IL apparaissent typiquement alors que la narration bascule d'un mode de focalisation interne à la représentation immédiate de la situation de référence dans le discours.
Dans les deux cas, on voit que la prédication objectivée intervient quand l’énonciateur est identifié pragmatiquement, et qu’il y a de fortes motivations pour que la prise en charge prédicative soit rendue trans-individuelle.

Haut de page

Texte intégral

0. Presentation of the two major issues concerning Locative Inversion

1Locative Inversion (hence LI) is a main clause construction that appears in English with the status of a non-canonical structure.

[1] Each floor of the Magnolia Street building had a short hallway with two apartments on either side. At the far end was a large window that let in the morning sun. (W. Mosley, A Red Death, 1993)

2As illustrated in [1], LI follows a specific pattern, composed of a preverbal spatial element (usually a PP or an AdvP), followed by a VP mainly restricted to spatial configuration verbs, and a postverbal subject NP, the realization of which can only be lexical, and not pronominal.

3LI is also associated to other syntactic restrictions than those just mentioned, and those restrictions have been analyzed either as evidence of a specific underlying structure (Bresnan, Kanerva, 1989; Coopmans, 1989; Hoekstra, Mulder, 1990; Culicover, Levine, 2001; among others) or as evidence of a specific ordered informational pattern (Prince, 1986; Rochemont, Culicover, 1990; Birner, 1992; Bresnan, 1994; among others).

4In recent work, those two trends of analyses have been heavily criticized (Chen, 2003) on the basis that neither accounted for all the syntactic properties of LI. Indeed, a relevant analysis of LI should provide an explanation for all the distributional fixed properties that have come to light, whether they concern the preverbal locative element, the VP, the postverbal NP or the sentence as a whole. Yet, some properties can be clearly sorted out as being evidence of the particular predication achieved by LI (see section 2).

5In the linguistic literature, yet another problem arises, pertaining not to the internal properties of the construction at issue, but to its particular discursive effects.

6First it should be noted that LIs never occur in certain types of contexts or, on the contrary, are typical of some types of discourse (as shown in Dorgeloh, 1997; Kreyer, 2006a; among others).

7But the real paradox is that when LI occurs, it is described as expressing either some sort of perceptual objectivity or some sort of perceptual subjectivity.

  • 1  Bolinger (1977) refers to Atkinson (1973: 15): "the reader has the impression not of being informe (...)
  • 2  The effect of the inversion is that the speaker "has moved his camera closer, due to his having ad (...)

8To illustrate these antinomic interpretations, one can compare Bolinger (1977) – when he speaks of a "stage effect"1 which erases the speaker’s point of view – to Dorgeloh (1997), when she claims that despite an empathy effect, the speaker’s subjectivity is still at work2. There are other references that might also be mentioned (Drubig, 1988; Chen, 2003; among others). One way of solving this opposition would lie in adopting an approach that does not rely on interpretation but rests on factual observations.

9In this paper, an attempt is made to rely on fixed theoretical concepts in order to account for both some internal properties of LI and its discursive specificities. We believe that these two aspects of LI are intertwined and thus should not be treated separately.

10This epistemological postulate explains why all through this paper, our proposals are based on naturally occurring data. Some of the evidence used in our line of argumentation has already been put forward in previous studies that were also based on real data (Birner, 1992; Dorgeloh, 1997; Chen, 2003); other evidence has been directly observed from our corpora (see section 2 and 3 for details).

11In order to justify the form and use of LI, we will resort to a specific definition of non-commitment. In the first section below, the theoretical concept of speaker-commitment will be defined in reference to Culioli’s framework.

  • 3  In this paper, the term "grammaticalization" refers to the gradual process by which a lexical item (...)

12Then, on the basis of some of LI’s complex internal properties, the inverted sentence will be presented as a grammaticalized evidential marker of a non-speaker-based predication3.

13In the last section, a confirmation of our claim will be found when discussing two actual contexts in which LI is (or is not) a typical structure.

1. The predication marked by LI

14It is necessary to define what will be called a non-speaker-based predication in this paper. This concept will be contrasted with a particular type of predication, referred to as non-commitment. In order to clarify what we mean by that term, we’ll start by explaining our use and definition of the term speaker.

1.1.Speaker and enunciator

  • 4  See Culioli (1990, 1999) and especially Culioli (1999, t.2: 49-52, 95-106, 133-137).

15In the Culiolian theory4, the notion of enunciator (for the French énonciateur) refers to a source which functions as a deictic origin and which is traceable in the sentence (as explained for instance in Wyld, 2007). More precisely, the enunciator is the abstract origin, from which, for instance, the personal pronoun references, the time references and the modality choices are calculated.

16In other words, there are grammatical markers, in the sentence, which refer by contrast or identification to a constructed deictic origin. This can be illustrated with [2] and [3]:

[2] In the front row, there might have been about ten people.

[3] In the front row, there were ten people.

  • 5  The use of the past tense (cf. might have been / were) is also most important to define the proper (...)

17In [2], the presence of a modal auxiliary marks a direct reference to a specific enunciative origin. This origin is constructed as calculating or evaluating the uncertainty value of the predicative relation <ten people be in the front row>. In [3], the choice for a non-modalized past tense is also the tangible trace of an evaluation, this time providing information on the degree of certainty concerning the predicative relation. In both cases, the verbal markers deictically refer to a subjective origin, which evaluates the epistemic values of the proposition <ten people be in the front row>5 and this subjective source corresponds with our intuitive representation of "the person who actually talks" (a locutor). Of course, in some cases, it is more complicated and it becomes necessary to distinguish the enunciator and the locutor.

[4] According to Mary, there were about ten people in the front row.

18In [4], a distinction must be drawn between the origin of the epistemic value of <ten people be in the front row> and the origin of the whole statement. Both origins can be inferred from the phrase "according to Mary", and characterized as referring to potential individuals.

19So, in the Culiolian theory, a distinction is usually made between an enunciator and a locutor, the former being defined as the ultimate subjective reference, and the latter referring to the subjective reference quoted.

20These two subjective origins however share an important property: indeed, they refer to an individual subjectivity. So, although the enunciator is supposed to be an abstract origin, it is usually conceptualized as a potential locutor, who is endowed with the faculty of evaluation. That’s why in many Culiolian studies, the actual ultimate enunciative origin is called "sujet énonciateur / enunciator as a speaking subject" (and not simply énonciateur / enunciator).

21In order to account for the deictic origin that is traceable in LI, Gournay (2005) pointed out that the notion of enunciator as a speaking subject was too restricted. Sometimes the ultimate origin of the sentence is not identified to an individual subjectivity. Rather, it refers to a universal source that reports but does not evaluate.

22The universal source can be found in proverbs, for instance. In this type of sentence, there are no traces of subjective evaluation since all the grammatical markers are normally fixed, which means there are no potential paradigmatic variations. For example, in A friend in need is a friend indeed, the use of the tense and polarity is fixed. This absence of choice shows that the validity of the statement is grammatically encoded as encompassing any subjective evaluation pertaining to a speaking subject.

23Hence the need for a new distinction: enunciator vs. speaker. The concept of speaker is used here to make a distinction between two types of deictic ultimate origins: speaker-based enunciators vs non-speaker-based enunciators (see schema below).

Figure 1: The properties of the Enunciator

24To sum up, in the Culiolian theory, a well-established distinction is made between enunciator and locutor, which is only activated in cases where two subjectivities are at work. In this paper, another distinction is used, introducing two types of enunciator:

  • most ultimate origins traceable in naturally occurring examples refer to the representation of a potential speaker, endowed with cognitive capacities, including that of evaluation (speaker-based origin or speaker);

  • in some cases, especially in LI as will be demonstrated in section 2, the ultimate origin traceable via markers and their lack of paradigmatic potentials, is a universal indefinite source of knowledge, that does not calculate.

1.2. Non commitment vs non-speaker-based predication

25In English, one classic example of what is frequently called "commitment" (i.e. the fact that the speaker assumes responsibility for the expressed statement) is the use of modal auxiliaries compared with the use of more periphrastic markers.

[5] I must go

[6] I have to go

26In the Culiolian framework, it is argued that in [6], contrarily to [5], the choice for a periphrastic deontic marker explicitly states the speaker’s non commitment to the necessity expressed. Yet, in [6], some degree of commitment is still expressed thanks to the positive valuation. Indeed, the polarity in [6] can be considered as resulting from a series of possibilities, in which the well-formed sentence I don’t have to gowas included. In other words, the presence of a positive polarity, when the negation is allowed in the paradigm available, reveals a minimal degree of speaker-commitment.

27To account for certain types of sentence where neither modal nor polar variations are allowed in the VP, it is crucial to introduce the concept of "non-speaker-based predication".

[7] Off they went.

28The sentence in [7] can be seen as echoing a paradigm devoid of a negative or a modalized counterpart:

[7’] *Off they didn’t go / *Off they could go

29The claim made in the next section is that LI reveals this particular type of non-speaker-based predication (i.e. a complete form of non-commitment) that indicates that the statement expressed is not packaged as being the result of a specific speaker’s evaluation.

2. Constraints in LI as evidence of a non-speaker based predication

  • 6  The corpus used here is composed of 350 naturally-occurring locative inversions, taken from a vari (...)

30In English, LI is constrained by a series of syntactic properties. A vast majority of them have already been put forward in the literature on the subject. They are confirmed in the corpus of naturally occurring examples of contemporary English used for this section6.

  • 7  In this paper, we only discuss the properties that are evidence of the particular informational fu (...)

31In this section, important distributional fixed properties will be illustrated7 and we will show that they form a coherent set of constraints: indeed, it appears that in LI, there is no syntactic slot left for grammatical markers implying some degree of subjective calculation.

2.1. Negation in LI

32First, there is a strong constraint on negation in LI: this constraint, already noted in the literature (Dorgeloh, 1997; Chen, 2003; among others), is clearly confirmed in the corpus used for this study, so that the acceptability judgement in [8] is quite realistic:

[8] *At the end of the corridor wasn’t any large window

33In other words, it can be observed that, in LI, negation is not felicitous on the predicate. Yet, a negative adverb can appear in locative there-sentences like [9].

[9] The carpet in the living room was maroonish, and in the hallway there wasn't any carpet (

34Example [9] shows that there is an important difference between LI and its canonical counterpart containing there. In the first type of sentence only, the existential verb cannot be negated. This observation can be rephrased as follows: there is no paradigm for polar variation in LI.

35Now, the same constraint actually applies on the postverbal NP since negative items cannot be found in the postverbal NP. However, there is one counter-example in our corpus:

[10] The sign on that wall doesn’t read BEYOND THIS POINT LIE MONSTERS, like the legend on ancient maps. It reads BEYOND THIS POINT LIES NOTHING. Nothing. Does such a thing exist? I’ve heard children ask this question: Isn’t even "nothing" something? (G; Iles, Blood Memory, 2005)

36In [10], a negation marker is attested in the postverbal NP. Yet, in this context, its polar value becomes positive and nothing clearly refers to something, as can be inferred from the question "isn’t even "nothing” something?".

  • 8  The same research on the BNC gave a result that is not conclusive: 0 occurrences for sentences lik (...)

37Again, the constraint discussed here does not apply with locative there-sentences as can be checked on the Internet (see [11 a-b])8.

[11a] Behind me was no one. (0 occurrence on Google, 02/01/2007)

[11b] Behind me there was no one. (54 checked occurrences on  Google, 02/01/2007)

38In short, although there-sentences allow polar variations in the verbal and postverbal paradigmatic slots (at least to some extent), such a variation is not possible in LI. Presumed grammatical counter-examples (cf. 10) appear to confirm this constraint at the referential level.

2.2. Modality and LI

  • 9  Chen notices that sentences composed of a modal+past infinitive (like *Over the mountain could hav (...)

39Apart from the general constraint on negation, some authors have noticed an interesting constraint on modality (Coopmans, 1989; Chen, 2003; among others). Although modal markers are attested in LI, their uses are restricted to only a few patterns. For instance, they are never found followed by a past infinitives as noted in Chen (2003: 178)9, and they only appear in contexts where they do not have epistemic or deontic values. These observations lead us to the acceptability judgement in [12]:

[12] ?* At the end of the corridor may/must/should be a large window

40Now, one recurrent pattern in which a modal can be found in LI is [COULD+ passive infinitive of perception verbs], as illustrated in [13]:

[13] Farther to the west could be seen a strip of blue where the ocean met Mission Bay, then just a lot of mud flats. ( journal/83fall/gem.htm)

41In examples like [13], the modal auxiliary COULD is used assertively: the perception expressed is effective, and the existence of a strip of blue is not questionable. Thus, the modality marked by COULD only qualifies the representation of the perception, by referring to a potential group of viewers in the situation described. In passing, let’s note that the reference to the potential viewers remains totally implicit, via the use of the agentless passive (the only type of passive found in LI).

2.3. Genericity and atelicity in LI

  • 10  As shall be discussed in the final section.

42Two relevant examples of the constraints at play in LI are the blatant absence (mentioned in Bolinger, 1977)10 of mere indefinite markers like some (see [14]) and the impossibility of a generic interpretation of sentences like [15]:

[14] *?At the end of the corridor was someone / something strange.

[15] Outside my window is a tree.

Outside my window is a tree.

There only for me.

And it stands in the gray of the city,

No time for pity for the tree or me.

There is a world of pain.

In the falling rain

Around me. (Song by Cream)

43The two constraints exemplified in [14] and [15] illustrate the fact that in LI "abstractions make poor actors" (Bolinger, 1977: 96).

44The infelicity of [14] is due to the use of an indefinite pronoun in the postverbal NP. Indefinite markers implied a lack of referential stability that is incompatible with LI. This incompatibility is due to the fact LI erases any representation of subjective, on-going evaluation. On the other hand, in there-sentences indefinite markers are quite acceptable: At the end of the corridor, there was someone / something strange. So, the constraint on indefinite markers cannot be generalized to existential sentences: it only applies with the LI predication.

  • 11  Of course, actual perception may not take place: if I am blind, I can still describe the scene in (...)
  • 12  LI in French has been widely studied by many authors since Le Bidois (1950). For a more recent acc (...)

45In [15], the first line of the song refers to a perception event located in a specific situation11. By contrast (as shown for the French by Willems, 198912), if the first line of the song was "Outside my window there is a tree", a potential ambiguity would arise: the existential sentence would either refer to a property (the view from my bedroom window) or an actual perception event (what I see, now, standing at my window).

46Thus, in LI, as opposed to what can be observed for there-sentences, there is no room for abstraction markers i.e. markers that refer to an entity or event that is not fully identifiable or time-and-space located. In the preverbal slot, the locative element must be endowed with a definite and specific value that is clearly recoverable from the context.

47One last interesting constraint is the aspectual restriction that applies in the case of inverted sentences. Indeed, it is established that the BE+ING form cannot be found in naturally occurring examples (as noted in Quirk et al., 1972: 949; Chen, 2003:181; among others).

[16] ?* At the end of the corridor was sitting John

48This constraint comes out clearly in examples like [17] where two sentence patterns can be compared:

[17] Jeannine was sitting in one of the wing chairs at the windows overlooking Lake Michigan. Her face was carefully made up and it was hard to tell how she felt about her husband's death. Across the room, feet tucked up under her on an armchair, sat Paige Carrington. (S. Paretsky, Deadlock, 1987)

  • 13  For an analysis of BE+ING, see Dufaye (2009: 65).

49In the first sentence, the progressive form appears in a canonical sentence. In the last sentence, the change of word order goes with the appearance of a simple form. This constraint can be attributed to the function of BE+ING. Indeed, this aspectual form deictically refers to a speaker’s evaluation of an event (whether as a process, an on-going commitment or refusal, etc.)13. Furthermore, the constraint on BE+ING echoes other verbal constraints that demonstrate that the predication expressed is grounded on trans-individual knowledge and perception.

50In Levin & Rappaport Hovav (1996), although the absence of BE+ING in LI is not discussed, it is claimed that the interpretation of the verb in LI is motivated by a constraint on telicity and, for instance, verbs of change of state like sit can only appear in LI when a statal interpretation (ibid, 241) can be triggered. In other words, the fact expressed in LI is represented as instantaneously captured but not described. This particular type of representation also explains why there are strong verbal lexical constraints (Bolinger, 1977: 97; Levin, Rappaport Hovav, 1996: 253) which can be formulated as follows: the verb has to denote the typical or expected action of the referent of the subject.

51In this section, it has been shown that the set syntactic pattern of LI does not leave any paradigmatic slot for commitment markers: the predication expressed by LI is not syntactically compatible with epistemic calculation (true, false, probable) and it does not accept any marker that directly refers to a speaker’s assessment (indefinite, generality etc.).

52The claim defended here is that LI is a grammaticalized evidential marker that encodes a source of information that is non-subjective but rather universal, or trans-individual (since it encompasses any individuality). This particular predication occurs in narrative discourse, according to a discursive strategy, when a clear erasing of the speaker’s subjective view is wanted or expected. This is what will be put forward in the next section, with a focus on LI at the beginning of tales as well as in the description of crime scenes and other descriptive contexts.

3. The notion of speaker’s commitment in narrative contexts

53In section 1 above, we have defined three types of enunciative origins, speaker-based enunciator, non-speaker-based enunciator and locutor, as construed relatively to the degree of commitment, or of non-commitment expressed in a predication. Most utterances, however, rely on a combination of enunciative sources. Such is the case in narratives where the referential system of each utterance is the result of a more or less complex association of points of view. This has been discussed in narratological studies (Booth, 1961; Chatman, 1978; Cohn, 1978; Rimmon-Kenan 2002; among others) where a text is typically described as resulting from a multiplicity of points of view that can be reconstructed through the identification of the sources of narration and focalization.

54In order to describe the complex intertwining of sources involved in the production of a text in the perspective developed in this article, it is necessary to consider how the above-mentioned concepts apply to narration.

  • 14  For a definition of the implied author as used in this article, see Booth (1961: 70-71) : “As he w (...)
  • 15  See Wyld (2007) for definitions more detailed.

55In a Culiolian enunciative perspective, narratives have been described as relying on a complex multi-faceted entity constituting each utterance’s referential system (Danon-Boileau, 1982; Rivara, 2000; Wyld, 2001, 2007). In Wyld (2007), this referential system has been formalized as minimally comprising two mutually located enunciative origins. The first one is absolute (cf.the implied author14), and the second one is derived (cf. the narrator). They are both construed as sharing responsibilities in the genesis of the text, though at different levels. The absolute origin is associated with operations pertaining to the predicative level (choice of the notions and of the diathesis in the predicative relation) but also to the pragmatic level (mode of organization of the text, use of tropes, etc.). The derived origin, on the other hand, is associated with operations pertaining to the enunciative level, i.e. the aspectual and modal determination. In a Culiolian approach, this derived origin is seen as the main source of subjectivity in a text and as the deictic centre of the narration15.

56Furthermore, the narratorial level, i.e. the derived origin in our model, may also be multi-faceted and composed of more than one origin, thus representing the enunciative heterogeneity of an utterance. In that case, the deictic center of the narration is complex.

  • 16  According to Maingueneau, the hyperénonciateur is involved in the interpretation of particitation (...)

57However, narratorial agencies differ in their characteristics according to the degree of commitment they are associated with. Hence, they are traceable via linguistic markers and construed in our model as speaker-based enunciators. Yet, in some cases, such as in proverbs or sayings, there is no speaker’s commitment at work. The derived enunciative source is therefore associated with a proto-narratorial, non-subjective source similar to what Maingueneau (2004), in a polyphonic approach, calls the hyperénonciateur16. In the approach developed in this article, the proto-narrator is construed as a non-speaker-based enunciator and is traceable in the linguistic markers.

58In that perspective, the use of prototypical markers in certain types of texts, at precise moments, so to speak, of the narrative is related to the narratorial strategy at work. Following Delmas (2004: 33), we consider that these constructions play a major role in the constitution of what he calls the co-enunciative protocol resulting from the relations established between the actors of the on-going enunciation.

59As far as LI is concerned, data show that it scarcely occurs at the beginning of transcribed folktales compared with there-sentences. The aim in sub-section 3.1. is to explain why it is rare in that context where the narrative apparatus focuses on the teller’s commitment in the narration. In sub-section 3.2, LI will be analyzed in other narrative contexts, among which descriptions of crime scenes in detective stories. In that context, the shift from subjective perception to disembodied perception, resulting in the use of LI, will be accounted for.

3.1. Narrative strategy at the beginning of folktales

  • 17  For more details on folk tales and their specificities as compared to literary fairy tales, see Br (...)

60The focus here will be on the particular case of transcribed folktales in English as found in books of collected folk and fairy tales. Collected folktales are tales that were first orally told then transcribed and edited17. Though of course modified to be written down and published in a more standard English, they display textual characteristics which are evidence of their oral origin as it is reconstructed in the transcription. This shows through the occurrences of dialectal forms remaining in the text, either lexical as in [18] or syntactical as in [19]:

[18] You see, sir, there was a colonel wanst, in times back, that owned a power of land about here. (Fairy and Folk Tales of Ireland)

[19]You see, there was a waiver lived, wanst upon a time, in Duleek here, hard by the gate and a very honest, industherous man he was, by all accounts. (Fairy & Folk Tales of Ireland)

61In both examples, once is transcribed in its dialectical form wanst. In [19] the introductory sentence contains a relative clause considered as non canonical (there was a waiver lived in Duleek) as well as non standard spelling which aims at reproducing a dialectal pronunciation (industherous).

62This oral origin is also reconstructed through the occurrence of words or phrases that typically initiate oral communication such as you see as shown in the two above-mentioned examples.

63Hence, the opening of such tales is typically composed of two types of utterances representing two intertwined fictitious dimensions as exemplified in [20] and [21]:

  • 18 Folktales of Ireland, edited and translated by Sean O’Sullivan, The University of Chicago Press, 19 (...)

[20] A long, long time ago — if I were there then, I wouldn't be there now; if I were there now and at that time, I would have a new story or an old story, or I might have no story at all — there was a king and a queen in Ireland, and they were married. (Folktales of Ireland18)

  • 19 A Dictionary of British Folk-Tales in the English Language, ed. by Briggs, K. M. Part A: "Folk Narr (...)

[21] Once upon a time, just before the monkey tribe gave up the nauseous custom of chewing tobacco, there lived an old hag who had conceived an inordinate desire to eat an elf: a circumstance, by the way, which indubitably establishes that elves were of masticable solidity, and not, as someone has it, mere /"Shadowy dancers by the summer streams"/ so the old lady went to the place where the fairies dwelt, and knocked at the hill-top. (A Dictionary of British Folk-Tales, Part A: Fairy Tales19)

64These two fictitious dimensions have been described by Vuillaume (1990), as the primary fiction (fiction primaire), i.e. the story proper, and the secondary fiction (fiction secondaire, in bold in the examples above), composed of commentaries on the act of telling as in [20], or on facts of the story as in [21].

65The intertwining of primary and secondary fiction is not specific to folktales and it is often found in 18th century novels for instance.

66However, in folktales, the narrator of the primary fiction has characteristics typical of the genre: it is not posited as specific, but as non-subjective, i.e. not referring to a particular speaker, conversely to what happens in novels where the narrator is construed as unique, bearing specific characteristics displayed for one narration only. In folktales, it is construed as a proto-narrator which represents the trans-individual entity responsible for the perpetuation of the narrative. This type of proto-narrator is common to utterances that are not related to any specific subjective entity where the reference values are not calculated relative to a specific speaker. In this paper, those utterances have been tagged "non-speaker-based predication".

67Yet, the actual narration of a folktale proceeds from the actualization of the narrative by a specific teller, endowed with the qualities of a speaker, who thus perpetrates, in his name, so to say, the transmission of the tale.

68The discursive specificities of beginnings of folktales come from the tension created by the association of these two types of origins in one, complex, multi-layered enunciative entity: a speaker-based enunciator identified with the teller and a non-speaker-based enunciator identified with the proto-narrator:

Figure 2: The complex system of origins at play in transcribed folktales

69The recurrent surfacing of the speaker-based enunciator, i.e. the teller, at the beginning of the telling, is specific to the genre, as is the occasional intertwining, along the narration, of this entity with the proto-narrator. In the present analysis, it is directly linked to the types of linguistic constructions prevailing at the beginning of transcribed folktales and pertaining to the functioning of fiction in such texts.

3.1.1. LI and the erasing of the speaker’s point of view

  • 20  See note 1.

70Bolinger (1977) associates the use of LI with the erasing of the speaker’s point of view. In his analysis, the use of LI is related to the concept of vividness, i.e. the state of affairs expressed is constructed as "on stage20 (…) as though directly perceptible to a reader, in the course of a narrative" (ibid, 94). On the contrary, the use of there-sentences is associated with the expression of a subjective point of view.

71This is quite coherent with our corpus results where it is clear that there-sentences are recurrently preceded by segments belonging to the secondary fiction which are the expression of a subjective point of view, or which emphasize the act of telling, as [18] and [19] show, while such configuration is not found with LI in our corpus.

  • 21 Fairy and Folk Tales of Ireland, with a foreword by Benedict Kiely, ed. by W.B. Yeats, Touchstone, (...)

[18] You see, sir, there was a colonel wanst, in times back, that owned a power of land about here. (Fairy and Folk Tales of Ireland21)

[19] You see, there was a waiver lived, wanst upon a time, in Duleek here, hard by the gate and a very honest, industherous man he was, by all accounts. (Fairy & Folk Tales of Ireland)

72Consistently, in our corpus, LI is not found to occur at the beginning of tales following utterances belonging to the secondary fiction contrary to there-sentences. Nevertheless, LI can be followed by subjective commentaries as exemplified in [24]:

[22] On the shore of Smerwick harbour, one fine summer's morning, just at daybreak, stood Dick Fitzgerald "shoghing the dudeen", which may be translated, smoking his pipe. The sun was gradually rising behind the lofty Brandon, the dark sea was getting green in the light, and the mists clearing away out of the valleys went rolling and curling like the smoke from the corner of Dick's mouth. (Fairy & Folk Tales of Ireland)

73Furthermore, according to Bolinger, the type of vivid representation expressed by LI functions as a cohesive discourse marker. For him, this explains why it is ruled out of beginnings of narratives where there-sentences are felicitous– see Bolinger’s examples in [23] and [24]:

[23] *In Xanadu lived a prince of the blood. (Bolinger, 1977: 110)

[24] In Xanadu there lived a prince of the blood. Near him lived a beautiful princess whose name was Divinapreciosa. (ibid)

  • 22  These figures come from a corpus of 508 tales.

74As predicted by Bolinger, there-sentences are very common in such contexts (about 60% of the corpus22). However, contrary to his prediction, the corpus shows that, albeit rare (only 2,6% of the corpus), LI can be found at the beginning of folktales:

  • 23 A Dictionary of British Folk-Tales in the English Language, ed. by Briggs, K. M. Part B: "Legends", (...)

[25] Near the town of Aberdeen, in Scotland, lived James Campbell, who had one daughter, named Mary, who was married to John Nelson, a young man of that neighbourhood. (A Dictionary of British Folk-Tales, Part B: Legends23)

75The small number of LI in that context implies that the level of constraint is high. Indeed, the examples found in the corpus of folk tales show that the level of determinacy of locative item is essential in the acceptability of the construction. Those determinacy constraints are corollary to the erasing of the speaker's point of view in LI and correspond to the necessity of creating a non-subjective stabilized reference, specific to the on-going enunciation and construed as the origin of the predication. Actually, LI at the beginning of folktale allows a reference to a broader situation than what we saw in most of the examples seen in 2. The locative PP refers to a city as in [25], a region, an unnamed country etc. Yet, despite this change that bars any representation of perception, all the constraints seen in 2 are valid. Here again, LI syntactically embodies a non-speaker-based predication.

3.1.2. Constraints on the locative in LI at the beginning of folktales

  • 24  We underline.

76In some analyses, as opposed to Bolinger (1977), the occurrence of LI at the beginning of folk tales is presented as self-evident. For instance, Birner and Ward (2002), consider that "an inversion commonly24 performs a scene-setting function at the outset of a narrative". As we have seen, though, this is hardly the case in our corpus.

77Furthermore, according to these authors, "the discourse is far more coherent when the topic of the following clause or clauses is the entity represented by the postposed NP rather than that represented by the NP in the preposed phrase" (ibid, 1387). Example [24], however, shows that the postposed NP does not have to be the topic of the clause that follows. Therefore, the scene-setting function as described by Birner and Ward is not sufficient to predict the occurrence of LI in discourse.

78Yet, a key to the use of LI lies in the notion of "minimally informative setting relative to which the postposing can be interpreted" (ibid) provided, according to Birner and Ward, by the preposed PP. This notion can be rendered more precisely by spelling out the formal properties of the preposed PP(s), essential to the acceptability of LI. Those properties are corollaries of what is called in this paper a non-speaker based predication and of the determinacy constraints attached to the item that fills the preverbal slot.

79Indeed, when studying the preverbal PPs in the context of folktales, it becomes obvious that the level of determinacy required is constrained and that most of the time this preverbal PP contains names of existing places (or presented as such). To illustrate this, one can compare the difference of acceptability between [25], a naturally occurring utterance, and [26], a made-up sentence not found in the corpus:

[25] Near the town of Aberdeen, in Scotland, lived James Campbell, who had one daughter, named Mary, who was married to John Nelson, a young man of that neighbourhood. (A Dictionary of British Folk-Tales, Part B: Legends)

[26] ?Far away lived a priest connected with the old priory church of this parish.

80When the locative refers to a place that is under-specified qualitatively speaking, LI is not considered as good as there-sentences:

[27] ?In London lived a man who had three sons.

[28] In London there lived a man who had three sons.

81In order for LI to be felicitous, a precise mapping of the locative reference in the preposed PP is required.

82We have seen in 2.3. above that the postverbal NP cannot be indefinite. Similarly, examples with somewhere or any indefinite pronoun functioning as a locating item is not found in the corpus:

[29] ??Somewhere lived a man who had three sons.

83However, the following example found in the corpus shows that it is possible:

[30] Somewhere very far away lived a quarryman. He was old, and his wife had never borne him any children. (A Dictionary of British Folk-Tales, Part B: Fairy Tales)

84Yet, this example is unique in the corpus and the frequency rate implies here again that the level of constraint is high. Accordingly, the felicity of [30] can be compared with the acceptability judgments in [30’] and [30’’]:

[30’] ?Somewhere lived a quarryman. He was old, and his wife had never borne him any children.

[30’’] ?Very far away lived a quarryman. He was old, and his wife had never borne him any children.

85In [30] the indefinite somewhere is specified by very far away. This specification is crucial and contributes to the building up of a locating item that is both specific and recoverable from the context. Indeed, this minimal specification allows the construction of a qualitative otherness between somewhere very far away and somewhere not very far away or even very very far away and therefore restrains the field of vision, so to speak. This accounts for the acceptability judgment in [27] where no such restriction is operated and in London is not contrastable in context with any other locating item. In other words, somewhere very far away is sufficiently determined in [30] to achieve a deictic function, which rules out any genericity or ambiguity of reference. Any lower level of determinacy is incompatible with the necessary deictic reference of the preposed NP.

  • 25  It is probably this distancing effect that native speakers have in mind when they assume that LI i (...)

86We have seen in section 2 that the constraints on LI rule out any form of valuation by a speaker. At a discourse level, the use of LI hinders the "play upon voices" and the commitment of the teller which are so characteristic at the very beginning of folktales. This explains why there are so few occurrences of inversion in that context: in transcribed tales, the low frequency rate of LI in the initial sentence is coherent with the tendency to construct a direct reference to the teller. The type of non-commitment underlying any occurrence of LI is then clearly incompatible with the narrative strategy operative in transcribed folktales. Therefore, when LI nonetheless appears in that context, the discursive outcome is that of a tale not told by someone in particular but that of a situation for everyone to witness, just as if the proto-narrator identified with a trans-individual point of view remained the only entity involved. LI has thus a clear distancing effect, marking a deictic centre which cannot be associated with a speaker25.

  • 26 This is also antithetical with the relative referential vagueness characteristic of folk and fairy (...)

87The locative item in LI is built up to achieve reference to a unique spot, perfectly defined and accessible to any speaker26. In context, it often represents strong evidence of the shift from a subjective predication to a trans-individual predication. Such a textual strategy becomes even more obvious when it comes to descriptions of crime scenes in detective stories.

3.2. Evidence…from crime scenes and other descriptive contexts

88As already mentioned, several authors have put forward the link between LI and some narrative strategies. Dorgeloh (1997: 3) notices that LI can be totally absent from one book while appearing in clusters in others, mainly in descriptive passages.

89Moreover, LI has been associated with visual description in narratives. Hence Kreyer (2006b: 16) who, following a cognitive approach, describes LI as creating "the illusion of immediate-perception by stimulating ‘natural perception’".

  • 27  For instance, not a single locative inversion was found in 8 contemporary romance novels.

90The consequences of these two facts, which are confirmed by our corpus27, is that LI does not appear in just any descriptive passage. Once again, the on-going narrative strategies must be taken into account. Our goal in this last part is to try and put forward tracks of research that would lead us to understand the motivations behind the occurrences of LI in descriptive, narrative contexts, where existential THERE seems to be the "unmarked" existential marker.

91To illustrate this, here is the beginning of Kew Gardens by Virginia Woolf. The text, which is almost entirely descriptive, starts with two existential statements. The second one is an instance of LI. It is the only instance in the whole text.

[31] From the oval-shaped flower-bed there rose perhaps a hundred stalks spreading into heart-shaped or tongue-shaped leaves half way up and unfurling at the tip red or blue or yellow petals marked with spots of colour raised upon the surface; and from the red, blue or yellow gloom of the throat emerged a straight bar, rough with gold dust and slightly clubbed at the end. (V.Woolf, "Kew Gardens", in Monday or Tuesday, 1921)

92Several questions can be addressed: does [31] illustrate a recurrent pattern, where LI can only appear once the setting described has already been posited, contrary to what happens with there-sentences? At another level, what is the function of LI at this point in the beginning of the short story and why does it seem incompatible with the strategy at play in the rest of the text? In our perspective, those two questions would be treated with regard to the non-speaker-based predication LI encodes.

  • 28  For this section, 40 examples taken from A Series of Unfortunate Events were added to the literary (...)

93What our literary corpus28 brought to light is that there exist sub-types of descriptive contexts in which LI can be expected to occur. But more important, even in these rather precise sub-types, the alternation LIs/there-sentences remains relevant and worth analysing.

94One typical context in which LI is very frequent is the description of crime scenes in detective novels. In that context, LI appears inside passages of internal focalization (usually of first-person narration but not only as we shall see) where the source of perception has been clearly identified with a subjective point of view. Thus, the on-going enunciation relies on a speaker’s evaluation. The description expressed bears on a concrete element which serves as a location (room, table, wall, piece of paper, etc.) for an object or a character. At some point in the description, an occurrence of LI stands out.

95The two extracts below exemplify the sudden occurrence of LI in the context studied:

  • 29  We underline.
  • 30  We underline.

[32] As I sit, I glance quickly29 around the office, but the place is so sparsely decorated that I only register a few details30. Soft white walls, teak shelves, a couple of long, vertical paintings that look Chinese. To my left hangs a samurai sword, its truncated blade gleaming with threatening purpose. To my right, on a sideboard,sits a stone Buddha that looks authentic enough to have been stolen from an Asian jungle somewhere. (G. Iles, Blood Memory, 2005)

[33] Chaim was hunched over a chair. Most of the blood was right under him. But there was also blood on the dresser and in the bathroom. Blood on the phone, in the dial. There were bloody handprints on the wall. He'd gone all the way around the room, propping himself up with his bloody hand. Next to his body was a light green cushion, splattered and clotted with blood. He'd pressed the cushion to his chest, trying to staunch the bleeding, but he must have known that it wasn't going to work. (W. Mosley, A Red Death, 1993)

96In these examples, which are first-person narration, the situation of reference is first described in details through the narrator’s eyes: it is explicit in [32] with the perception verbs in I glance quickly, I only register… It is implicit in [33] but the italics used in the third sentence, among other clues, give an orality to the passage, indicating that the speaker represented here can be identified to the protagonist. The perception events are represented in an iconic way, so that the reader gets the impression of a perfect match between the representation and the reality of the situation.

97From what can be observed from our corpus, LI only appears once the reference of the preverbal location is given in the left-context and considered as shared knowledge (in [32], To my left/As I sit…, in [33], his body/Chaim was hunched over a chair). The preverbal PP then functions as a deictic marker, its full reference depending on the situation represented (as is clear in [32]). In other words, the "stage effect" first described for LI in English by Bolinger (1977) is only possible if certain referential determinations are given in the left-context.

98Yet, this necessary condition does not explain why LI appears in contexts where existential there-sentences or verbless sentences (as is underlined in [32]) can also be found. What then is the function of LI? How can we account for the fact that the context in which it typically appears is highly subjective?

99From the examples gathered, LI is used as a discursive close-up marker. The entity referred to in the postverbal NP bears a dramatic function in the narrative. The sudden change of enunciative origin is used to underline the relevance of the entity presented in the narrative, whose existence in the situation is represented as being universally blatant. So that the distribution of LIs versus there-sentences is not meaningless: the non-speaker-based predication marker is used to posit the existence of an entity that distinguishes itself in the setting. This entity is usually associated to a certain amount of climax in the succession of perceptions expressed, as can be shown in the following extract:

[34] [A man is intrigued by a strong smell coming from the apartment of his neighbour]

The door to Apartment J was ajar. That was where the smell came from.(…)

The living room was a mess. The shades were drawn and the curtains pulled, so it was twilight in the musty rooms. Ghostly white cartons of Chinese food were open and moldering on the table, trash everywhere. I flicked the light switch, but the bulb had burned out. (a) Against a far wall there sat an altarshe had made from a small alcove. Inside she had glued a picture of Jesus. It was painted like a mosaic. He had a halo and held two fingers and a thumb above three saints who were bowing to receive his blessing. (b) All around the painting there were old flowerswired to the walls. They were unidentifiable brown things that she'd probably brought home from mass or after funeral.

(c) At the foot of the painting was the bronze dish that she also used to burn the incense. The smell was much stronger there. Little ashes, like white maggots, were littered around the brimming dish. (d) And there was a black, gummy substance on the ledge and down the wall to the floor. The bathroom was disgusting. All kinds of cosmetic bottles open and dried until the liquids had caked and cracked. (e) Mildewed towels on the floor. A spider spun its web over the bathtub faucet. The worst smells came from the bedroom, and I hesitated to go in there. It's a funny thing how smell is such an animal instinct. The first thing a dog will do is sniff. And if it doesn't smell right there's a natural reluctance to get any closer.

Maybe I should have been a dog.

Poinsettia was hanging from the light fixture in the middle of the ceiling. She was naked and her skin sagged so that it seemed as if it would come right off the bone any second. (f) Directly under her was the cause of the worst smells. Even as I watched a thick drop of blood and excrement fell from her toe. (W. Mosley, A Red Death, 1993)

100In [34], there is a succession of spatial NPs (the living room, a far wall, altar, small alcove, picture of Jesus) leading to a climatic though momentary ending of the description with a LI (c) At the foot of the painting (...) referring to the portrait of Jesus. The same long succession can be found before the occurrence of LI in (f) with bedroom, Poinsettia, ceiling, directly under her. Along with the two instances of LI are utterances which often occur after a preposed locative item as well and are typically related to locative inversion in contexts where there is a precise topological description: there-sentences (a), (b), (d) and a verbless sentence (e).

101A close look at the narrative organisation of the text provides hints as to the distribution and the discursive function of each constructions. As underlined before, the description is here closely linked to a subjective perception. As the description of the setting unfolds from the focalizer’s point of view, a there-sentence occurs each time there is a change of focus. In [34], this is the case in (a) an altar, (b) old flowers and (d) a black gummy substance which are posited as new landmarks in the setting.

102This discursive pattern is recurrent in our corpus as shown by the following example, yet another crime scene:

[35] [as he is watching some game he has just missed with his binoculars, a man makes out what appears to be a crime scene in the distance]

  • 31  We underline.

He lowered the binoculars and looked over the country at large. Then he raised them again. (a) There looked to be men lying on the ground. He jacked his boots into the rocks and adjusted the focus. The vehicles were four wheel drive trucks or Broncos with big all-terrain tires and winches and racks of roof lights. The men appeared to be dead31. (…).

When he approached the trucks he had the rifle unslung and cradled at his waist with the safety off. (…). He stood there. Listening.

  • 32  We underline.

In the first vehicle (b) there was a man slumped dead over the wheel. (c) Beyond were two more bodies lying in the gaunt yellow grass. (d) Dried blood black on the ground32. He stopped and listened. Nothing. The drone of flies. He walked around the end of the truck. (e) There was a large dead dog there of the kind he'd seen crossing the floodplain. The dog was gutshot. (f) Beyond that was a third body lying face down. (C. McCarthy, No Country for Old Men, 2008)

103This example is third-person narration and the different sources of perception in the text are indicated through focalization. This passage starts with external focalization as the character’s situation is described by an omniscient narrator (He lowered the binoculars and looked over the country at large. Then he raised them again). Then the narration shifts to internal focalization when the main character (now the source of a subjective perception) watches the landscape with his binoculars. In the text, this change of focalization is marked by a there-sentence (There looked to be men lying on the ground). Here again, this narrative device is used each time the character focuses on a new spot as in (b) a man slumped dead and (e) a large dead dog. Then eventually, the description ends up with a LI as in (c) and (f).

104This however does not account for the distribution of verbless sentences and LI. Why in [35], is existential sentence (d) Dried blood black on the ground verbless and not inverted (vs. On the ground was dried black blood)? The same question arises in [34] with (e) Mildewed towels on the floor / On the floor were mildewed towels. According to the data of our corpus, this distribution appears to be linked to a cohesive function of LI and to its role as regard to the unwinding of the main theme in the passage.

105In [34] for instance, the occurrences of LI can be associated to the main theme of the quotation: SMELLS. In this passage, LIs highlight the propositional contents referring to an identification of the smells: in (c), there is The smell was much stronger there ; in (f) the postverbal NP refers to "the cause of the worst smells" and comes as a climatic elucidation of what the protagonist has been looking for.

106The same thematic pattern can be unraveled in [35]: as he watches from a distance, the character makes out what he believes to be dead bodies. When he finally reaches the spot he had been looking at, what he sees proves him right. At this point in the narrative, the postverbal NP in LI refers to the identification of corpses.

107Another typical descriptive context in which LI can be found is when the postverbal NP refers to something that has the property of being readable as illustrated in [36], [37] and [38]:

[36] We came upon five men wearing identical black suits and white gloves. Above the left-hand breast pocket of each jacket was sewn a green flag that said First African in bright yellow letters. Each man carried a dark walnut tray with a green felt center. (W. Mosley, A Red Death, 1993)

[37] For at the bottom of the list of villages, below Paltryville and Tedia and Ophelia, was the most important thing they had read all morning. Printed in the flowery script, on the back page of the brochure Mr. Poe had given them, were the letters V.F.D. (L. Snicket, A Series of Unfortunate Events, 2000)

[38] There was a photograph, stapled to the page, which showed the Baudelaire parents, standing next to one man the Baudelaires had met briefly at the Village of Fowl Devotees, and one man the children did not recognize, and below the photograph was a sentence Klaus had read so many times that he did not need his glasses to read it again. "'Because of the evidence discussed on page nine, experts now suspect that there may in fact be one survivor of the fire, but the survivor's whereabouts are unknown,'" he recited. (L. Snicket, A Series of Unfortunate Events, 2000)

108In these examples, the use of LI highlights the fact that this marker is mostly found in contexts where the field of vision is particularly well-determined to the extent that it merges, so to say, with the object of the perception. Hence, the entity referred to in the postverbal NP (a green flag in [36], the letters in [37] and a sentence in [38]) is construed as the sole object of perception from all points of view.

109The change of sentence order thus displays a particular function in the discourse strategy. Indeed, when the perception is well-identified to an individual and once the "stage" described is outlined, the sudden change of predication goes with a dramatic focus on a particularized spot. The fusion of the subjective point of view with the trans-individual point of view erases all types of modal distance. The use of inversion thus appears as a narrative device to shift from the subjective point of view of an internal focalization (the-world-seen-through-my-eyes type of description) to a neutral presentation of facts (the-world-as-it-presents-itself) which suppresses other possible points of view or angles. This shift is possible when pragmatically the source of the perception represented is completely identified. And finally, this shift can be used as a stylistic device to posit the existence of an entity thus granted with special narrative value. It is also highly compatible with the representation of actual close-up acts of perceptions.

4. Conclusion

110The particular type of non-commitment put forward in this article is specific of the absence of any subjective evaluation in the predication. Our claim is that the "non-speaker-based predication" is categorized in English through constituent order in assertive sentences.

111LI has been found to exemplify this type of non-commitment particularly well, on the evidence of its linguistic properties. These constraints, specific to LI, as opposed to its canonical counterpart with there, are evidence of the absence of paradigmatic slots for cognitive calculation markers. It implies that the predication is not the result of a speaker’s evaluation. On the contrary, it seems to be presented as completely independent of a potential subjective source. This is why we consider that LI in English is an evidential marker that refers to a trans-individual (or universal) source of predication.

112This is confirmed at a functional level by the narrative use of LI, which occurs, as opposed to there-sentences, at stages where the erasing of a subjective point of view is needed to the benefit of a trans-individual, neutral point of view, hence the quasi-absence of LI at the beginning of folktales where the narrative strategy requires the endorsement of the narrative by a specific teller. On the other hand, LI appears, also strategically, at the climax of descriptions of crime scenes in detective stories, after long passages in internal focalization where the situation of reference has been built up step by step so as to become perfectly stabilized and presented as the direct reference (or deictic centre) of the utterance.

Haut de page


Atkinson, J. C. 1973. The Two Forms of Subject Inversion. The Hague, Paris: Mouton.

Belmont, N. 1999. Poétique du conte: Essai sur le conte de tradition orale. Paris: Gallimard.

Birner, B. J. 1992. The Discourse Function of Inversion in English, Doctoral dissertation, Evanston, I11, Northwestern University.

Birner, B. and Ward, G. 1993. There-Sentences and Inversion as Distinct Constructions: A Functional Account, B.S.L. 19: 27-39.

Bolinger, D. 1977. Meaning and Form. London: Longman.

Booth, W. 1961. The Rhetoric of Fiction. Chicago: University of Chicago Press.

Bresnan, J., Kanerva J. 1989. Locative Inversion in Chichewa: A Case Study of Factorization in Grammar. Linguistic Inquiry 20: 1-50.

Bresnan, J. 1994. Locative inversion and the architecture of the universal grammar. Language 70 (1): 72-131.

Bricout, B. 1992. Le savoir et la saveur. Henri Pourrat et "Le trésor des contes". Paris: Gallimard.

Chatman, S. 1978. Story and Discourse: Narrative Structure in Fiction and Film. London: Cornell University Press.

Chen, R. 2003. English Inversion: A Ground-before-Figure Construction. Berlin: Mouton de Gruyter.

Cohn, D. 1978. Transparent Minds. Princeton: Princeton University Press.

Coopmans, P. 1989. Where Stylistic and Syntactic Processes Meet: Locative Inversion in English. Language 65: 728-51.

Culioli, A. 1990. Pour une linguistique de l'énonciation, t1. Gap: Ophrys.

Culioli, A. 1999. Pour une linguistique de l'énonciation, t2. Gap: Ophrys.

Danon-Boileau, L. 1982. Produire le fictif. Paris: Klincksieck.

Delmas, C. 2004. Fragment d'un discours culinaire. In L. Gournay, J.-M. Merle (eds.), Contrastes, Mélanges offerts à J.Guillemin -Flescher. Paris: Ophrys: 25-34.

Dorgeloh, H. 1997. Inversion in Modern English, Form and Function. Studies in Discourse and Grammar, 6. Amsterdam/Philadelphia: John Benjamin.

Drubig, H. 1988. On the Discourse Function of Subject Verb Inversion. In J. Klegraf, D. Nels (eds.), Essays on the English Language and Applied Linguistics on the Occasion of Gerhard Nickel’s 60th Birthday. Heidelberg: Groos.

Dufaye, L. 2009. Théorie des opérations énonciatives et modélisation. Paris: Ophrys.

Gournay, L. 2005. Quand l’agencement est la trace d’une prédication ‘objectivée’. Analele Universitatii din Craiova, Annales de l'Université de Craiova, Série Langues et littératures romanes, numéro spécial, IX. Craiova: editura Universitaria: 65-74.

Gournay, L. 2006. Approche énonciative des catégories de marqueurs, Document de synthèse, Université Paris 7-Denis Diderot.

Hoekstra, T., Mulder, R. 1990. Unergatives as copular verbs; location and existential predication. The Linguistic Review 7: 1-79.

Kreyer, R. 2006a. Inversion in Modern Written English. Syntactic Complexity, Information Status and the Creative Writer. Tübingen: G. Narr.

Kreyer, R. 2006b. ‘Observer effect’, ‘eyewitness perspective’ and ‘imaginary guided tour’. What is so Natural about the Way Inversions Represent Spatial Relations?. In C. Copy, L. Gournay (eds.), Points de vue sur l'inversion. Paris: Ophrys: 115-132.

Lahousse, K. 2003. The distribution of postverbal nominal subjects in French. A syntactic, semantic and pragmatic analysis. Thèse nouveau régime, Université Paris 10, Université de Louvain.

Le Bidois, R. 1950. L'inversion du sujet dans la prose contemporaine. Paris: Ed. d'Artrey.

Lehmann, C. 1995/1982. Thoughts on Grammaticalization. Muenchen: Lincom Europa.

Levin, B., Rappaport Hovav, M. 1996. Unaccusativity: at the syntax-lexical semantics interface. Linguistic Inquiry Monograph, 26. London: The MIT Press.

Maingueneau, D. 2004. Hyperénonciateur et ‘particitation’. In A. Rabatel (ed.), Langages: Effacement énonciatif et discours rapportés 156. Paris: Larousse:111-126.

Prince, E.-F. 1986. On the syntactic marking of presupposed open propositions. In A. Farley, P. Fraley, K.-E. McCullough (eds.), Papers from the Parasession on Pragmatics and Grammatical Theory, 22nd Regional Meeting, Chicago Linguistic Society: 208-222.

Rimmon-Kenan, S. 2002, Narrative Fiction. London, New York: Routledge.

Rivara, R. 2000. La langue du récit: Introduction à la narratologie énonciative. Paris: L’Harmattan.

Rochemont, M., Culicover, P. 1990. English Focus Constructions and the Theory of Grammar. Cambridge: Cambridge University Press.

Vuillaume, M. 1983. Grammaire temporelle des récits de fiction. Semantikos 7: 62-76.

Vuillaume, M. 1990. Grammaire temporelle des récits. Paris: Minuit.

Ward, G., Birner, B. & Huddleston, R.. 2002. "Information structure". In G. Pullum, R. Huddleston, (eds.), The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press: 1363-1447.

Wyld, H. 2001. Subordination et énonciation. Paris: Ophrys.

Wyld, H. 2007. Some remarks on the nature, identity and locus of aspectual and modal sources in narrative texts: towards an enunciative treatment of point-of-view. Cahiers Charles V 42: 9-79.

Willems, D. 1989. Généricité, spécificité et constructions verbales: les structures à sujet inversé. Equivalence 17-18: 175-181.

Haut de page


1  Bolinger (1977) refers to Atkinson (1973: 15): "the reader has the impression not of being informed by the author of what is happening, but rather of being 'on stage' himself, receiving directly the impressions of the moment".

2  The effect of the inversion is that the speaker "has moved his camera closer, due to his having adopted a reference point from the discourse world established and in relation to which he then focuses a new discourse entity" (Dorgeloh, 1997: 105).

3  In this paper, the term "grammaticalization" refers to the gradual process by which a lexical item or a structure gains in grammaticality, i.e. serves a grammatical function (see Lehmann, 1995).

4  See Culioli (1990, 1999) and especially Culioli (1999, t.2: 49-52, 95-106, 133-137).

5  The use of the past tense (cf. might have been / were) is also most important to define the properties of the deictic source. In (1) and (2) a contrast is constructed between the time-reference of the situation expressed and the time-reference associated with the enunciative source.

6  The corpus used here is composed of 350 naturally-occurring locative inversions, taken from a variety of literary texts. Other studies have been based on larger and more diversified corpora (see Birner; 1992; Dorgeloh,1997; Chen, 2003), but with a different aim than the one developed in our research: clearly, our aim is not to discuss the discursive effects of LI but rather to account for the reasons why this type of sentence is syntactically constrained.

7  In this paper, we only discuss the properties that are evidence of the particular informational function of LI. For a more global account of LI (for instance, concerning the lexical postverbal NP), see Gournay (2006).

8  The same research on the BNC gave a result that is not conclusive: 0 occurrences for sentences like 11a, and one occurrence like 11b.

9  Chen notices that sentences composed of a modal+past infinitive (like *Over the mountain could have been seen a dark smoke) are not attested in natural occurring data. In English, modals + past infinitives refer to the field of irreality (past possibility, or past hypothesis). For an analysis of this constraint, (see Chen, 2003; Gournay, 2005).

10  As shall be discussed in the final section.

11  Of course, actual perception may not take place: if I am blind, I can still describe the scene in front of me (if for example, somebody has previously described it to me). The point is that the inverted sentence cannot refer to a generic property (that of the view from my window). LI constructs a time-and-space-located situation.

12  LI in French has been widely studied by many authors since Le Bidois (1950). For a more recent account of LI, see Lahousse (2003).

13  For an analysis of BE+ING, see Dufaye (2009: 65).

14  For a definition of the implied author as used in this article, see Booth (1961: 70-71) : “As he writes, [the real author] creates not simply an ideal, impersonal ‘man in general’ but an implied version of ‘himself’ that is different from the implied authors we meet in other men’s work (…) it is clear that the picture the reader gets of this presence is one of the author’s most important effects.”

15  See Wyld (2007) for definitions more detailed.

16  According to Maingueneau, the hyperénonciateur is involved in the interpretation of particitation utterances. Particitations belong to a fuzzy Thesaurus of utterances that circulates within a community.

17  For more details on folk tales and their specificities as compared to literary fairy tales, see Bricout (1992), Belmont (1999) as well as the introductions to the collections of tales used as corpus in this article, and particularly the introduction to A Dictionary of Folk Tales in the English Language by K.M. Briggs.

18 Folktales of Ireland, edited and translated by Sean O’Sullivan, The University of Chicago Press, 1966.

19 A Dictionary of British Folk-Tales in the English Language, ed. by Briggs, K. M. Part A: "Folk Narratives & Fairy Tales", London, Routledge, 1991.

20  See note 1.

21 Fairy and Folk Tales of Ireland, with a foreword by Benedict Kiely, ed. by W.B. Yeats, Touchstone, 1998.

22  These figures come from a corpus of 508 tales.

23 A Dictionary of British Folk-Tales in the English Language, ed. by Briggs, K. M. Part B: "Legends", London, Routledge, 1991.

24  We underline.

25  It is probably this distancing effect that native speakers have in mind when they assume that LI is a prototypical form at the beginning of folktales.

26 This is also antithetical with the relative referential vagueness characteristic of folk and fairy tales and perfectly exemplified by the prototypical expression once upon a time, there was.

27  For instance, not a single locative inversion was found in 8 contemporary romance novels.

28  For this section, 40 examples taken from A Series of Unfortunate Events were added to the literary corpus used in section 2, plus a dozen more we accidentally came upon in the course of our personal readings.

29  We underline.

30  We underline.

31  We underline.

32  We underline.

Haut de page

Table des illustrations

Légende Figure 1: The properties of the Enunciator
Fichier image/jpeg, 88k
Légende Figure 2: The complex system of origins at play in transcribed folktales
Fichier image/jpeg, 98k
Haut de page

Pour citer cet article

Référence électronique

Christine Copy et Lucie Gournay, « Locative Inversion In Discourse: a strategy of non-commitment », Discours [En ligne], 5 | 2009, mis en ligne le 04 novembre 2009, consulté le 27 juin 2017. URL : ; DOI : 10.4000/discours.7355

Haut de page


Christine Copy

CLILLAC, P7-Diderot

Lucie Gournay

IMAGER, P12-Val de Marne

Haut de page