Navigation – Plan du site

AccueilNuméros20-1The diffusion of French à travers...

The diffusion of French à travers from the 18th century onwards.

Some steps toward an explanation1
Thomas Hoelbeek

Résumés

Cet article examine des explications possibles de la diffusion soudaine de l'expression française à travers à partir du milieu du XVIIIe siècle. À partir de ce moment, à travers est devenu remarquablement plus utilisée que l'expression similaire au travers (de), et aussi relativement plus fréquente par rapport à par, une préposition avec laquelle elle est en concurrence dans certains contextes. Une première hypothèse suppose une concurrence avec par. Une deuxième hypothèse est liée à la fin de la liberté de à travers et au travers par rapport à la combinaison avec la préposition de. Les résultats indiquent que la croissance des fréquences de à travers (de) pourrait, en partie, découler d'une concurrence avec par pour la combinaison avec certains types de Sites (l'entité de référence d'une relation spatiale). L'augmentation du degré de figement - et donc de grammaticalisation - de à travers semble également avoir contribué à sa diffusion.

Haut de page

Texte intégral

1. Introduction

1.1. À travers (de), au travers (de) and par

  • 1 This research was supported by the research council (‘Onderzoeksraad’) of the Vrije Universiteit (...)

1À travers and au travers (de), both meaning ‘through’, belong to the limited group of the main markers of spatial dynamic relations with a medial polarity (any movement includes the initial phase of departure, a medial phase, and the final phase of arrival). In the past, they were often used and considered as synonyms, whereas this does not seem to be the case in Modern French (cf. Martin & Dominicy 2001; Dominicy & Martin 2005). Their respective semantics have progressively diverged, and the freedom in the use of the preposition de, ‘of/from’, which combined indifferently with au travers and à travers, disappeared; indeed, neither au travers (in its prepositional function) nor à travers de exist in Modern French anymore (see Hoelbeek 2015a; 2017).

2Another marker of spatial dynamic relations with a medial polarity is par, ‘through’. Both à/au travers (de) and par express a directional meaning that involves reference to a path (Borillo 1998: p. 49). However, in Modern French subtle differences exist between their specific usage patterns. For example, both à travers and par can introduce the medial Ground of a movement, but à travers, unlike par, does not take in consideration the relations of that entity with other elements of the environment and does not presuppose that the motion verb involves some kind of alteration of such relations (Aurnague & Stosic 2002: p. 137). The same holds true for au travers (de).

1.2. Research questions

3This paper investigates possible explanations for the sudden diffusion of à travers from the middle of the 18th century onwards. Indeed, when analyzing the evolution of the group of French complex dynamic expressions derived from the word travers (‘breadth’, from Latin tran(s)versu(m), ‘oblique, across’), we noticed that from that moment, à travers became remarkably more used than the similar expression au travers (de), and relatively more frequent in comparison with par, which always remained much more numerous, though.

4A first hypothesis (see Section 3.1) supposes a competition with the similar preposition par. À travers, which competes in certain contexts with par, may have taken over some semantic values of its rival. The frequencies of par in the Frantext database do in fact decrease for the period in question, which supports this idea. The combinations of à travers and par with specific types of Ground (Section 3.1.1) and with a particular set of verbs (Section 3.1.2) will be analysed, in order to put this idea to the test.

5A second assumption that will be investigated (see Section 3.3) is a possible causal connection that might exist between the end of the freedom of à travers and au travers with respect to the combination with the preposition de on the one hand, and the rise of à travers on the other. As said above, this preposition became unacceptable with à travers while the combination with the (prepositional) uses of au travers became obligatory. The combinatory freedom of de ceased to exist at the beginning of the 18th century for à travers, and at the beginning of the 17th century for au travers, and led to a semantic reorganization we described in depth in earlier work (Hoelbeek 2015a).

1.3. Theoretical framework

6In this paper, a diachronic perspective is adopted (on the diachronic evolution of prepositions and their grammaticalization, see, for example, Svorou (1994), Di Meola (2000), Hoffmann (2005); for a more general overview of grammaticalization and semantic change in French we refer the reader to reference works such as Marchello-Nizia (2006)), while the methodology is influenced by cognitive linguistics. This discipline has had a considerable influence on the theoretical and methodological development of semantic descriptions (see, for example, Ungerer & Schmid 1996); we believe, together with Blank & Koch (1999: p. 1), that, in turn, analyses of diachronic problems can help advancing cognitive theorizing, because historical semantics are the ideal terrain to put models and semantic theories to the test. More specifically, we apply a functional approach in the tradition of works by, for instance, Vandeloise (1986), Herskovits (1986), Aurnague (1991) and Vieu (1991).

2. Description of the corpus and methodology

2.1. The corpus

7The current study utilizes a corpus composed with data stemming from the Frantext database (for detailed discussions on the use of corpora the reader is referred to reference works such as, for example, McEnery & Wilson (2001)). The dataset used covers the period 1500-1899, and was designed to take a first step in the diachronic analysis of the group of French complex dynamic expressions containing French travers. It has been used by the author in several contributions analyzing the diachronic evolution of the expressions described above (Hoelbeek 2011; 2012a; 2012b; 2013; 2014; 2015a; 2015b; 2015c; 2017). Synchronic analyses remain comparatively few, and often deal only briefly with our subject (cf. Spang-Hanssen 1963; Langacker 1987; Somers 1988; Weinrich 1989; Schwarze 1989; Flageul 1997; Fong & Poulin 1998; Borillo 1998; Aurnague 2000; 2001; Martin & Dominicy 2001; Aurnague & Stosic 2002; Kwon-Pak 2002; Plungian 2002; Dominicy & Martin 2005; Sarda & Stosic 2007; Stosic 2002a; 2002b; 2005; 2007; 2009).

8Our corpus thus contains examples retrieved from the Frantext database, the referential data base for ‘standard’, written French (Marchello-Nizia 2005: p. 13, note 4). This online, diachronic database (continuously under development) contains 5 555 texts from the 9th century until the present (264 million words in January 20222), of which the large majority are literary texts, and a small portion scientific or technical works.

  • 3 The method used to automatically select an arbitrary sample is the following: after entering into a (...)

9In Table 1, we give an overview of the tokens containing à travers (de) and au travers (de) in our corpus, as well as the number of occurrences for each century. For the sake of feasibility, a representative random sample of 200 tokens was selected for periods where the amount of tokens provided by Frantext exceeds this number3:

  • 4 Data provided by Frantext, as consulted in June 2011.

Table Number of tokens for each expression analyzed4

10For all intervals, more tokens were found for à travers (de) than for au travers (de). The only exception is the period 1650–1699. It has been suggested by Vaugelas & Marzys (2009: p. 480) that the inverse proportion encountered here would be a result of Vaugelas’ remark (1647), stating that “[t]ous deux [au travers et à travers] sont bons, mais au travers, est beaucoup meilleur, et plus usité” (‘both [au travers and à travers] are correct, but au travers is much better and more frequently used’) (2009: p. 480). While that may be true, the data provided by Frantext suggest how Vaugelas’ observation was unfounded (at least in written language), and that its effect quickly faded away (for a more detailed discussion, see Hoelbeek 2017: p. 38).

2.2. Methodological remarks

11Two entities are distinguished when describing space in the theoretical framework of cognitive linguistics: the reference entity (the object of the spatial relation) and the located entity (the subject of the spatial relation). We adopt Talmy’s (1983) terminology in this paper, which means that the located entity is named the ‘Figure’ and the reference entity the ‘Ground’.

12We attribute a crucial role to functional parameters, and thus follow a mainstream approach in cognitive linguistics, illustrated by scholars like Herskovits (1986), Vandeloise (1986; 1991) and Stosic (2002a; 2002b; 2005; 2007; 2009; 2015). Space is described in this approach by means of functional concepts inspired in naive physics and human bodily experience, like access to perception, potential meeting, and general vs lateral orientation, concepts that “are tied to the extralinguistic knowledge of space shared by the speakers of one language” (Vandeloise 1991: p. 13). In doing so, we are able to reformulate in fine-tuned terms the more conventional categorization of spatial, perceptual and metaphorical uses.

3. À travers (de) vs au travers (de): divergent frequencies

13As was shown in Table 1, for the period covering the last 150 years of our corpus (1750-1899), tokens of à travers (de) increase tremendously. This is partly explained by the composition of Frantext, which contains a higher number of texts for more recent periods. However, we see an important difference with au travers (de), which becomes more frequent as well, but much more gradually. The question arises what could explain this difference. Why did à travers (de) became significantly more frequent than au travers de from the 18th century on?

14In what follows, an explication is sought for this evolution. Various possible factors are analyzed. A first hypothesis posits a competition with a similar preposition, par, ‘through’; a second one assumes that à travers and au travers de were grammaticalized to unequal extents.

3.1. Hypothesis 1: Competition with par

15A possible explanation for the rising frequencies of à travers (de) could be that this expression, which competes in certain contexts with the preposition par, ‘through’ – though it would be wrongful to consider them synonyms, at least in Modern French (cf. Stosic 2002b) – took over some semantic values of its rival. Indeed, the frequencies of par in Frantext decrease for the period in question, which supports the hypothesis at hand.

16In the past at least, both à travers (de) and par were able to combine with medial Grounds where a movement takes place. Yet, in contemporary French these two expressions establish different relations with the elements of the environment: the Grounds that combine with par link two other entities (Stosic 2002b: p. 132); by contrast, the Grounds selected by à travers restrict the freedom of movement of the Figure, and ‘guide’ it. Elsewhere, Stosic states that unlike par, which introduces noun phrases referring to passages designed as such, the entities denoted by the noun phrases associated with à travers often are not functionally designed for passage (2002b: p. 129).

17Two examples (borrowed from Stosic 2002b) show that par and à travers focus on mutually distinct properties of the Ground:

(1) Il s’est levé et s’est mis à regarder par la fenêtre.
‘He stood up and began to look out of the window.’
(André Gide, Les Faux-monnayeurs, 1925)

(2) Il leva la tête et regarda à travers la fenêtre le ciel froid.
‘He raised his head and looked through the window at the cold sky.’
(Simone de Beauvoir, Les Mandarins, 1954)

18In sentence (1), with par, the window is not seen as an isolated entity but as a ‘point of passage’ (Kwon-Pak 1997): thus par establishes a spatial dynamic relation that belongs to the path-phase, where the Ground constitutes a communication zone between two other spatial entities (Stosic 2002b: p. 70). By contrast, Stosic claims that, in sentence (2), the capacity for the window to link the surrounding entities is not taken into account. The focus would rather be shifted to the guiding of the fictive path that the ‘probe’ follows (vocabulary borrowed from Talmy’s descriptions of Fictive Motion (2000a: pp. 99–175)), due to the fact that the window restricts the extension of the visual field (Stosic 2002b: p. 132).

3.1.1. Sub-Hypothesis 1.1: Type of Ground

19Though differences exist between par and à travers, both were able to select Grounds that were either ‘Not Functionally Designed for Passage’ or ‘Functionally Designed for Passage’ (Stosic 2002b). Before presenting the results of an analysis of the Grounds à travers combines with, let us briefly characterize the two types of Grounds in question.

a) Grounds Not Functionally Designed for Passage

20A Ground that is Not Functionally Designed for Passage (henceforward NFDP) may include entities; however, they do not succeed in blocking the movement of the Figure. In other words, the Figure manages to create itself a passage through the Ground by choosing a path more or less freely and by using the space available between the entities that the Ground may happen to contain. Most often, these Grounds belong to the group of entities called ‘places’, but they can be as well ‘portions of space’ or ‘material entities’ (for more detailed descriptions of the entities distinguished in the description of space, see, among others, Vandeloise 1988; 1995; Aurnague, Vieu & Borillo 1997; Aurnague 1998; 2000; Talmy 2000a; Stosic 2002b: pp. 74–78; Hoelbeek 2017: pp. 71–73).

21Grounds that are NFDP offer some (passive) force of resistance, and even if the Figure is always able to move from one position to another, i.e. to reach a point located on the opposite side, they are conceptualized as ‘obstacles’ in the speakers’ shared knowledge (Stosic 2002a: p. 150). This is due, sometimes, to the properties of the Ground, and sometimes to the properties of the Figure.

22The two phrases in (3)and (4)demonstrate how both par and à travers can be combined with a Ground that is NFDP. Notice that example (4)is preferred in Modern French.

(3) Se promener par les champs
‘Walking through/across the fields’

(4) Se promener à travers les champs
‘Walking through/across the fields’

b) Grounds Functionally Designed for Passage

23A Ground defined as Functionally Designed for Passage (henceforth FDP) usually lets the Figure pass through a portion of space it defines. Examples are ‘Conduits’, ‘Objects with interstices’ and ‘Communication Tracks’.

24In (5)à travers selects a Ground belonging to the group of Conduits (a pipe); in (6)par combines with the same entity (both examples were already given by Stosic 2002b):

(5) Glouglou glouglou fait la pipe à eau avec un pffeuhpffeuh en bruit de fond de la fumée qui monte à travers le tuyau transparent.
‘Glouglou glouglou does the water pipe with a pffeuhpffeuh in the background of smoke rising through the transparent pipe.’
(Frederic Lasaygues, Vache noire, hannetons, 1985)

(6) En se régalant de tabac français sentant à peu près aussi bon que ce qui sortait par le tuyau du derrière de sa voiture à caprices.
‘Feasting on French tobacco smelling nearly as good as what came out through the pipe at the back of his quirky car.’
(Remo Forlani, Gouttière, 1989)

c) Testing Sub-Hypothesis 1.1

  • 5 Stosic (2002b) analyzed only concrete occurrences of Modern French à travers. Since comparisons are (...)
  • 6 As is explained hereafter, Frantext was interrogated not only for par, but also for à travers. Cons (...)

25An analysis was conducted for the whole period covered by our corpus (from 1500 to 1899) so as to determine whether an evolution could be observed concerning the type of Grounds à travers (de) combined with.5 We aimed at verifying whether the two types of Grounds described above (or one of them), which were originally compatible with both par and à travers (de), became significantly more frequent in combination with à travers (de) from the 18th century onward. If such an evolution was found, specific research in the Frantext database could be done for combinations of the form [par + Type X of entity frequently constituting Type Y of Ground] in order to corroborate that the opposite tendency could be observed for combinations with par6.

26The results (shown in Tables 2 and 3) indicate that at least since the 16th century à travers prefers the combination with Grounds that are NFDP (never less than 49,59%) over Grounds that are FDP (never more than 10,90%). According to the descriptions of Stosic (2002b: p. 129), this is still the case in Modern French.

Table Evolution of the NFDP-Grounds and FDP-Grounds for à travers (de) – 1500-1699

Table Evolution of the NFDP-Grounds and FDP-Grounds for à travers (de) – 1700-1899

27As for the specific evolution of each type of Ground, we see that à travers gradually combined more easily with FDP-Grounds. However, the second half of the 18th century and the period 1850-1899 present a (small) decline. Generally speaking, FDP-Grounds become slightly more frequent from the 18th century on, but an exponential rise in the combination with à travers does not seem to have taken place.

28When looking at the NFDP-Grounds, we see a gradual increase during the 16th and 17th Centuries (from 50,00% to 64,84%). The beginning of the 18th century marks a strong decline (to 49,59%), followed by a gradual growth up to the 19th century, when the number of NFDP-Grounds reaches the range of 75,64% to 77,48%.

29It is especially the latter evolution that could have played a role in the diffusion of à travers at the expense of par. Indeed, as shown in examples (3)and (4)(on page 5), par is still able to combine with NFDP-Grounds today, but à travers is preferred. The data in Tables 2 and 3, which suggest an increasing frequency of the combination [à travers + NFDP-Ground], could have been caused by a decline of the construction [par + NFDP-Ground].

30In order to test this hypothesis, an additional test was led. First, on the basis of the data available in the corpus used, a list was drawn up of entities that often constitute NFDP-Grounds (a minimum of 10 occurrences over the whole data set was used as a criteria) in combination with à travers. These entities are listed in Table 4.

Table Most frequent (concrete) entities constituting NFDP-Grounds for the period 1500-1899

31Then, Frantext was systematically interrogated for the co-occurrences of these entities with both à travers and par, with a maximum distance of 3 interspaces for par (as to include occurrences like par_ tout_ le_ pays, ‘across the whole land’) and 4 interspaces for à travers (in order to include instances like à_ travers_ d’_un_ linge, ‘through a sheet’).

  • 7 Relative frequencies in function of the number of words for each period are given here.

32It is important to remain prudent when interpreting these rough and unanalyzed data. Moreover, some results are problematic, for example because of the low tokens that were found for some combinations. Nevertheless, three main observations can be made7:

1) For a relatively large group of combinations, it appears that, during all periods, the combination with à travers is less frequent than the combination with par, if we leave aside some small exceptional time intervals. An illustration is the case of pays (Figure 1), where we see, in the 19th century, a decrease of the use with par, while the use of à travers rises, which could indicate some kind of correlation:

Fig. Co-occurrences of à travers and par with pays – 1500-1899

2) In other cases, the combination with à travers becomes clearly more frequent than the combination containing par (champ(s), bois/boys, arbres, forest(s)/forêt(s)), but only from 1750 onwards. Before that moment, the two types of combinations are characterized by parallel evolutions, or by a constant decline of the uses with par. The evolution with arbres (Figure 2) illustrates the former kind of evolution, the one with champ(s) (Figure 3) the latter. In both cases, the decline of the combinations with par, happening just before the rise of the uses of à travers, seems to be a condition to trigger the diffusion of the latter:

Fig. Co-occurrences of à travers and par with arbres – 1500-1899

Fig. Co-occurrences of à travers and par with champ(s) – 1500-1899

3) Finally, a third group of combinations (feuillage/feuilles and terres) shows analogous evolutions for both combinations, until the second half of the 19th century. From that moment on, an inversely proportional evolution can be observed, as is shown in Figure 4:

Fig. Co-occurrences of à travers and par with feuillage/feuilles– 1500-1899

33From these results it can be confirmed that various combinations containing à travers became more recurrent, sometimes even more so than the rival construction with par. In some cases, the findings suggest that the combinations with à travers might have become more frequent at the expense of the constructions with par. However, this phenomenon was not detected for all combinations with à travers. Most importantly, the rise of uses with à travers does not always seem to be conditioned by the evolution of uses including par, since the former become more frequent even in combination with entities for which the frequencies of the latter did not drop.

3.1.2. Sub-Hypothesis 1.2: Type of motion expressed

a) A typology of motions

34First of all, we should give an overview of the different types of motions distinguished in this paper. The following basic distinction is used (see a.o. Sablayrolles 1995; Talmy 2000a; 2000b; Aurnague & Stosic 2002; Hickmann 2003; Hoelbeek 2017: pp. 63–65):

- ‘Change of Location’ (henceforward CoL): Aurnague and Stosic (2002: pp. 118–123) explain that this kind of motion implies a change of position and a change of the relation between the Figure and the Ground. Ex.: He is running into the kitchen.

- ‘Change of Position’ (henceforward CoP): Motion within a general location, implying a change of position without a change of relation between the Figure and the Ground. Ex.: He is running in the kitchen.

  • 8 These examples are borrowed from Hickmann (2003: p. 81).

- ‘Static position’: No change of position, nor any change in the relation between Figure and Ground. Ex.: He is in the kitchen.8

35The starting point for Sub-Hypothesis 1.2 is the observation by Aurnague and Stosic (2002: p. 137) that Modern French par favors situations where a connection or relation between locations is described; indeed, par often combines with Grounds that function as Conduits. These medial Grounds define portions of space that allow the Figure to pass through, and they create (in a direct or indirect way) a change between the initial and final state of the movement. This characteristic brings about that in Modern French par is found almost exclusively in ‘motion complexes’ (a (motion) verb followed by a preposition, cf. Sablayrolles 1995) where it implies (in combination with the verb) a CoL, which means that not only a change of position of the Figure is expressed, but also a change in the relation between the Figure and the Ground (Aurnague and Stosic 2002: pp. 136–137).

b) Testing Sub-Hypothesis 1.2

36Our aim was to verify whether à travers (partly) replaced par in motion complexes where a motion within a general location is expressed, i.e. motion complexes implying a CoP, without any change in the relation between the Figure and the Ground. Indeed, even if par is still able to be used in motion complexes expressing motion within a general location today (see example (3), page 5), à travers is preferred in this kind of situations (see example (4), page 5).

  • 9 As for sub-hypothesis 1.1, Frantext was interrogated not only for par, but also for à travers, for (...)

37A first step consisted in analyzing, within our corpus, the kind of motion expressed by motion complexes containing à travers. We expected to find a significant change (specifically, more motion complexes expressing a CoP) around the moment when à travers became more frequent. If such a change could be found, the next step would consist in verifying whether an opposite evolution had taken place for par, which could indicate an inversely proportional connection9.

  • 10 Again, abstract configurations are filtered out of this analysis.

38The results of the analysis are displayed in Tables 5 and 6. Static positions are not relevant here, but are included for the sake of completeness:10

Table Evolution of the type of motion complex containing à travers (de) – 1500-1699

Table Evolution of the type of motion complex containing à travers (de) – 1700-1899

39The results show that motion complexes containing à travers express a CoL in between 62,18% and 80,49% of the cases. The evolution is characterized by up and downs, but from the second half of the 18th century onwards the numbers are slightly lower, so a takeover at the expense of par seems excluded for this type of motion.

40Yet, from the second half of the 17th century motion complexes expressing a CoP become more frequent and their numbers stabilize around 35% (with exception for the period 1700-1749; no clear reason could be discovered for the divergent result in this period). This could have been a ‘danger’ for par, in that this preposition typically occurs in motion complexes expressing this very semantic value.

41In order to put this idea to the test, Frantext was again consulted. The most frequent verbs forming motion complexes containing à travers in our corpus (at least 5 occurrences) that express a CoP are listed up in Table 7:

Table Most frequent (concrete) verbs found in motion complexes expressing a CoP for the period 1500-1899

42The co-occurrences of these verbs with par and à travers were then tested in Frantext for the period 1500-1899. The following parameters where used:

- Element 1 (the verb) of the search before element 2 (the preposition);

- A distance of 3 interspaces between the elements, as to include occurrences like [Des vaches] courant_sans_queue_à travers champs, ‘[Cows] running without tail across the fields’. Restricting the distance to 3 interspaces unfortunately excludes some relevant configurations, but it is necessary in order to exclude irrelevant configurations where another verb occurs between element 1 and element 2;

- For the analysis of the verbal clause ‘prendre (la fuite/sa route)’, the distance between element 1 and 2 was set to 3 interspaces, 6 between element 1 and 3 and finally 3 between element 2 and 3. Instances like (7)where included this way.

(7) Monsieur, j’avois tousjours esperé que Mr Du Thou prendroit_sa_route_d’_Italie_par icy et que je luy pourrois moy mesmes bailler mes lettres, […]
‘Sir, I had always hoped that Mister Du Thou would take his road to Italy past here and that I could give him my letters myself, […]’
(Nicolas de Peiresc, Lettres, 1627)

43Some limitations have to be mentioned. Frantext allows to automatically retrieve co-occurrences with whole paradigms of verbs, but it should be mentioned that homophonous nominal forms are inevitably included in the results; this is the first reason why the figures should be interpreted with some reservations. Further, some ancient spellings of verbs (ex. courrir, instead of modern courir) are not registered as lemmas in Frantext and, with the verb aller, auxiliary uses cannot be excluded automatically (va regarder par […], ‘go looking through […]’). In addition, it is probable that some other unwanted results were obtained, that could only be filtered out manually: abstract configurations; cases where par does not indicate the movement of an entity through a Ground, but rather the manner in which something is done, the reason behind an event, or the amount of something per some unit of measurement; etc. The same holds for à travers, but an in-depth analysis of the data provided by Frantext falls outside the scope of this paper.

  • 11 Again, relative frequencies in function of the number of words for each period are given here.

44For all the combinations analyzed, we see a rise of the configurations with à travers from (the beginning or end of) the 18th century onward, more or less significantly according to the collocation at hand. As for the relation with combinations containing par, we can see several different patterns11:

1) Four of the combinations are characterized by a declining number of uses with par for the first part of the period under study, and a (small) revival from the 18th century on (the same period where uses with à travers increase lightly). This kind of evolution concerns the motion complexes including the verbs cour(r)ir, aller, marcher and (re)venir, and is illustrated by the evolutions with courir in Figure 5:

Fig. Evolutions of [courir+ par] and [courir+ à travers]

2) Other combinations (with prendre (la fuite/sa route) and guider) show a peak of uses with par in the 18th century, and a sharp decline afterwards. As stated above, uses with à travers rise from the 18th century onwards, as shown in Figure 6:

Fig. Evolutions of [guider + par] and [guider + à travers]

3) A third type of evolution (motion complexes including the verbs conduire and chercher) is characterized by an important rise of the configurations with par until the 18th century, after which a sharp decline takes place (see Figure 7). One could interpret the decline of the uses with par as a consequence of the increase of those with à travers, even if this growth is much less important:

Fig. Evolutions of [conduire + par] and [conduire + à travers]

4) Regarding the motion complexes with the verbs emporter, traisner/traîner/trainer and (se) sauver, a strong rise of the uses with par can be observed from the 18th century onward (see Figure 8). These combinations thus do not seem to be influenced by the putative rise of uses with à travers for that same period:

Fig. Evolutions of [traisner/traîner/trainer + par] and [traisner/traîner/trainer + à travers]

5) Finally, the combinations with suivre show how some constructions with par remain more or less stable during the whole period analyzed, and do not seem to be influenced at all by the (modest) increase of the uses with à travers from the 18th century on (Figure 9):

Fig. Evolutions of [suivre + par] and [suivre + à travers]

45Even if some motion complexes containing à travers seem to benefit from the decline of the constructions with par from the 18th century onwards, this does not seem to be the case for all constructions analyzed. But for certain combinations with par it do holds that they become less frequent in the 18th century, and that the amount of corresponding combinations with à travers rises. Some kind of connection thus appears, even if the climb of uses with à travers proves to be less substantial than the decline of uses including par. Moreover, it is only in the 18th century that the evolution of uses with par has some kind of impact on the development of those with à travers.

46Keeping in mind the constraints of the test performed, indications are provided that there might be a connection between the developments of both types of constructions for the combinations with some specific verbs, especially in more recent times. However, on the basis of the results obtained it seems unjustified to claim that the diffusion of uses with à travers is the only factor that caused the decline of those with par, all the more so since combinations with some verbs do not seem affected in any way by the combinations with à travers. In addition, the increase of à travers appears to be less significant than initially assumed.

3.2. Re-evaluation of the data

47A closer examining of the tokens of both par and à travers in Frantext (see Table 8 and Figure 10) enables to objectify the figures, and make the following observations:

481) The percentages of par are characterised by a constant decline from 1500 until 1899. Since the use of à travers increased only around 1750, it is very likely that other factors influenced the evolution of par besides a possible influence from à travers.

492) À travers has always been much rarer than par. Therefore, an influence of à travers, though not to be excluded completely, does not seem to have been the only reason why par continued to decline.

  • 12 Another interesting observation that can be made is that the numbers of à travers are caracterised (...)

503) Even if, at first sight, à travers seems to have become dramatically more frequent around 1750, the relative ascent (in function of the number of words of Frantext) shows a much less impressive evolution (see Table 8): we see an evolution from 0,0022% for the period 1700-1749, to 0,0150% between 1850 and 1899. Though still remarkable, the relative augmentation is less striking than the rough numbers: from 270 to 5653 tokens. What has happened should thus rather be interpreted as a combination of a decline of au travers de and a rise of à travers, rather than a sudden diffusion of the latter exclusively12.

514) The percentages of par was relatively stable from 1750 to 1849, even if à travers went from 0,0044% to 0,0111%. Thus it seems that à travers did not have a (strong) influence on par in this particular period:

Table Frequencies of tokens of à travers (de) and par in function of the number of words in Frantext

Fig. Frequencies of tokens of à travers (de) and par in function of the number of words in Frantext

3.3. Hypothesis 2: Increased bondedness

52The second hypotheses we will investigate is linked to the combinatory freedom of à travers and au travers with respect to the preposition de. Indeed, à travers became more frequent from the moment the combination with this preposition ceased to be possible. Possibly part of the explanation for the ascent of à travers (de) must thus be searched for in the competition with au travers (de). In fact, even if today the semantics they express are not exactly identical (cf. Martin & Dominicy 2001; Dominicy & Martin 2005), they prove equivalent in numerous contexts and they were considered synonyms in the past.

53Our corpus shows, that the combinatory freedom ceased to exist at the end of the 18th century for à travers, and at the beginning of the 17th century for au travers (see also Hoelbeek 2015). Figure 11 illustrates graphically the chronology of this evolution:

Fig. Visualization of the combinatorial evolution of the preposition de with à travers and au travers

54According to Fagard and De Mulder (2007: p. 18), an evolution of a complex preposition from [Prep1 + article + Base + Prep2] to [Prep1 + Base + Prep2] is a sign of increasing bondedness. This kind of rule is not universal, and Adler (2001) shows its restricted relevance; but as will be expounded below, it could provide an appropriate explanation in our case. The concept of bondedness is described by Lehmann as “the intimacy with which it [a sign] is connected with another sign to which it bears a syntagmatic relation”, and is understood as an indication of the degree of grammaticalization. Indeed, “[t]he degree of bondedness of a sign varies from juxtaposition to merger, in proportion to its degree of grammaticality” (2002: p. 131).

55In French, both combinations (à travers (de) and au [=à+le] travers (de)) have coexisted for a very long time. However, it is striking that, from the moment the vagueness as to the combination with the Prep2 (de) disappeared completely (possibly also triggered by an extralinguistic factor, namely Vaugelas’ remark (1647) that au travers should be used with de, and à travers without (Vaugelas & Marzys 2009: p. 480)), the more ‘evolved’, or more grammaticalized preposition (à travers) became much more used (although, as stated in section 2.1, Vaugelas prescribed the use of au travers, and the frequency of à travers first descended before starting its rise).

56If we consider à travers to be more bonded, and thus more grammaticalized than au travers de, it is likely that this state of affairs had an influence. Tables 9 and 10, and Figure 12 demonstrate (using relative frequencies) that à travers (de) has always been more frequently used, except for the period 1650-1699 where it reaches a historically low level (cf. Vaugelas’ remark in section 2.1). From that point on, à travers (de) starts a continuous ascent and becomes significantly more frequent than au travers (de) from 1750 on. Between 1750 and 1799 au travers (de) drops to 0,0007527%, and then slowly increases again to 0,0013492% in the period 1850-1899.

Table Frequencies of tokens of à travers (de) and au travers (de) in function of the number of words in Frantext

Table Frequencies of tokens of à travers (de) and au travers (de) in function of the number of words in Frantext

Fig. Frequencies of tokens of à travers (de) and au travers (de) in function of the number of words in Frantext

57If we compare the evolution of à travers (de) with that of au travers (de), we observe thus again a sudden growth for à travers (de) from 1750 on, and a more advanced degree of grammaticalization seems to be an explanatory factor. Elsewhere (Hoelbeek 2017: pp. 91–92), we discussed some other factors that point to an increasing specialization of à travers (de) (viz its adverbial use that diminisches as from 1700, and a semantic stability in conveying mostly dynamic semantics, at least from 1550 onwards). The evolution we mentioned in section 3.1 c), i.e., Grounds that are NFDP making up more than 75.00% of the entities à travers combines with from 1800 onwards, is an indication of an increasing stability in the semantics conveyed as well. In addition to bondedness, several other elements thus seem to indicate that à travers became ever more a prototypical lexical preposition (see also Fagard, de Mulder & Hoelbeek 2020). The expression au travers (de), on the other hand, continued to present behaviour of a prepositional expression, characterized by a lower degree of bondedness (for a more detailed description, see Hoelbeek 2017: pp. 95–97). Baring in mind that frequency is assumed to be an indication of grammaticality (Hopper & Traugott 2003: p. 6), à travers (de) becoming ever more grammaticalized could be a plausible explanation for its rise.

58To conclude, let us remark that au travers (de) does not seem to head for extinction. On the contrary, its frequencies increase slowly after the regression at the end of the 17th century. This is not an evolution one would expect for the less grammaticalized competitor of à travers. Possibly its relative diffusion could, as seems to be the case for à travers (de), be explained partly by some takeover process at the expense of par. Another explanation for the survival of au travers de could be searched in a mechanism described by Hoffmann (2005: p. 153), i.e. the grammaticalization of low-frequency constructions (au travers de) by analogy towards (formally) parallel combinations that are more frequent (à travers). Yet, interesting as the relative frequency of au travers de may be, it does not form the subject of the present contribution.

4. Discussion

59The growth of combinations with NFDP-Grounds, which applies to à travers from the second half of the 18th century on, could in some cases be explained by a decline of combinations containing the same NFDP-entities and par, as showed by the additional analysis realized by consulting Frantext. However, the results are not systematic, and therefore cannot be considered to be conclusive proof of a semantic competition that would have led to an increase of combinations with à travers and a decrease of those with par.

60As for the type of motion expressed by the motion complexes containing à travers and par, it turns out that a competition in the expression of a CoL seems unlikely. However, the complexes containing à travers that express a CoP do seem to have become more frequent, and may thus have become preferable in some contexts. Frantext was consulted for a more detailed analysis; it was shown that most combinations with par become less recurrent in the 19th century, when those with à travers climb slightly. Some kind of mutual impact seems probable, even if the diffusion of uses with à travers is unlikely to be the only factor that caused the decline of those with par.

61Even if the analysis seems to points to (combinations with) par as a factor in the relative rise of à travers, this is probably not the only aspect that played a role. The increased level of bondedness – and thus of grammaticality – of à travers in comparison with au travers de must have exerted some influence too. Indeed, the end of the combinatory freedom that existed for à travers and au travers with respect to the use of the preposition de seems to have triggered the diffusion of the more ‘evolved’, the more grammatical preposition.

5. Conclusion

62In this contribution, we have dealt with the question as to know why à travers became significantly more frequent than au travers de from the 18th century on. We established that the rise of the frequencies of à travers (de) for the period 1750-1899 could, in part, follow from a competition with par for the combination with NFDP-Grounds on the one hand, and the expression of motions involving a CoP on the other. However, on the basis of the obtained results it does not seem possible to indicate a clear causal relation, in one direction or the other. As we indicated, though, the increased level of bondedness – and thus of grammaticalization – of à travers is likely to have contributed to its diffusion, at the expense of the less grammaticalized au travers (de).

63Other elements could have played a role as well, and therefore remain to be explored. For example, the evolution of the Figurative, epistemic uses of à travers; the absolute frequencies of certain verbs in the Frantext database, independently of their combining with à travers or par; and finally, the evolutions of verbs like (faire) passer, ‘to pass’.

Haut de page

Bibliographie

Adler, S. 2001. Les locutions prépositives : questions de méthodologie et de définition. Travaux de linguistique 42–43(1). 157–170.

Aurnague, M. 1991. Contribution à l’étude de la sémantique formelle de l’espace et du raisonnement spatial : la localisation interne en français, sémantique et structures inférentielles (PhD Dissertation). Toulouse: Université Paul Sabatier.

Aurnague, M. 1998. Basque genitives and part-whole relations: typical configurations and dependences. Carnets de Grammaire 1. 1–60.

Aurnague, M. 2000. ’Entrer par la petite porte, passer par des chemins de traverse’ : à propos de la préposition “par” et de la notion de “trajet.” Carnets de Grammaire 7. 1–65.

Aurnague, M. 2001. Entités et relations dans les descriptions spatiales : l’espace et son expression en basque et en français. Habilitation à Diriger les Recherches en Linguistique, Université de Toulouse-Le Mirail, Équipe de Recherche en Syntaxe et Sémantique.

Aurnague, M. & D. Stosic. 2002. La préposition “par” et l’expression du déplacement. Vers une caractérisation sémantique et cognitive de la notion de “trajet.” Cahiers de Lexicologie 81. 113–139.

Aurnague, M., L. Vieu & A. Borillo. 1997. Représentation formelle des concepts spatiaux dans la langue. Langage et cognition spatiale 69–102.

Blank, A. & P. Koch. 1999. Introduction: historical semantics and cognition. In Historical Semantics and Cognition (Cognitive Linguistic Research), vol. 13, 1–14. Berlin/New York: Walter de Gruyter.

Borillo, A. 1998. L’espace et son expression en français. Gap/Paris: Éditions Ophrys.

DiMeola, C. 2000. Die Grammatikalisierung deutscher Präpositionen. Tübingen: Stauffenburg.

Dominicy, M. & F. Martin. 2005. “À travers” et “au travers de”. Des emplois perceptuels aux emplois spatiaux. (Ed.) P. Dendale. Recherches Linguistiques, Metz : Université de Metz (Le Mouvement Dans La Langue et La Métalangue) 27. 151–190.

Fagard, B. & W. De Mulder. 2007. La formation des prépositions complexes: grammaticalisation ou lexicalisation ? Langue française 156(4). 9–29.

Fagard, Benjamin, Walter de Mulder & Thomas Hoelbeek. 2020. Sémantique des prépositions. In Grande Grammaire Historique du Français. De Gruyter Mouton. https://doi.org/10.1515/9783110348194.

Flageul, V. 1997. Description sémantico-cognitive des prépositions spatiales en français (PhD Dissertation). Paris: Université de Paris IV – Sorbonne.

Fong, V. & C. Poulin. 1998. Locating Linguistic Variation in Semantic Templates. (Ed.) J.P. Koenig. Discourse and Cognition, Stanford, CA, CSLI Publications. 29–39.

Herskovits, A. 1986. Language and spatial cognition: An interdisciplinary study of the prepositions in English. Cambridge: Cambridge University Press.

Hickmann, M. 2003. Children’s Discourse: Person, Space and Time Across Languages. Cambridge University Press.

Hoelbeek, T. 2011. La locution spatiale “à travers (de)” et le concept de ‘guidage’. Présentation d’un cadre interprétatif illustré par des attestations du XVIe siècle. In A. De Wit, Ph. De Brabanter, L.-A. Cougnon, T. Colleman, P. Dendale & E. Coussé (eds.), Studies van de BKL / Travaux du CBL / Papers of the Linguistic Society of Belgium [online], 1–14. http://webh01.ua.ac.be/linguist/SBKL/sbkl2011/hoe2011.pdf.

Hoelbeek, T. 2012a. The use of French ‘à travers (de)’ and ‘au travers de’ for describing a movement along a curved surface: peripheral membership or anomaly? In M. van Hattum, J. Morris & D. Hoffman (eds.), Salford Working Papers in Linguistics and Applied Linguistics [online], 2–14. http://www.salford.ac.uk/__data/assets/pdf_file/0017/121607/Hoelbeek,-2-14.pdf.

Hoelbeek, T. 2012b. ‘La contrainte du “parcours minimal” et les usages des expressions “à travers (de)” et “au travers (de)”. La situation aux XVIe et XVIIe siècles.’ In S. Polis, Ph. De Brabanter, L.-A. Cougnon, T. Colleman, P. Dendale & E. Coussé (eds.), Studies van de BKL / Travaux du CBL / Papers of the Linguistic Society of Belgium [online], 1–15. http://webh01.ua.ac.be/linguist/SBKL/sbkl2012/hoe2012.pdf.

Hoelbeek, T. 2013. Les expressions spatiales contenant le mot ‘travers’: analyse sémantique de l’usage aux XVIe siècle. CogniTextes [online] 9. 1–24.

Hoelbeek, T. 2014. The spatial expressions containing French “travers” and Italian “traverso”. A functional semantic description from a diachronic perspective. (PhD Dissertation). Brussels: Vrije Universiteit Brussel/Université Libre de Bruxelles.

Hoelbeek, T. 2015a. À travers de? A diachronic analysis of French à/au travers in combination with the preposition de. Neuphilologische Mitteilungen. Modern Language Society 116(1). 55–84.

Hoelbeek, T. 2015b. La contrainte du ‘parcours minimal’ pour la description des usages des expressions ‘à travers (de)’ et ‘au travers (de)’’ : mise à l’épreuve d’un point de vue diachronique. Sciences pour la communication 31.

Hoelbeek, T. 2015c. “Par à travers”: Disparition et réapparition d’une combinaison de deux prépositions rivales. Disparitions. Contributions à l’étude du changement linguistique (Linguistique Historique) 251–266.

Hoelbeek, T. 2017. The Evolution of Complex Spatial Expressions within the Romance Family: A Corpus-Based Study of French and Italian. Brill. https://brill.com/view/title/32955.

Hoffmann, S. 2005. Grammaticalization and English complex prepositions: A corpus-based study. London/New York: Routledge.

Hopper, P. J. & E. C. Traugott. 2003. Grammaticalization. Second Edition. (Cambridge Textbooks in Linguistics). 2nd edn. Cambridge: Cambridge Univerity Press.

Kwon-Pak, S.N. 1997. Les prépositions spatiales: sur quelques emplois de la préposition “par” (PhD Dissertation). Université de Strasbourg 2. Strasbourg.

Kwon-Pak, S.N. 2002. Par–spatial : l’espace temporalisé ; par–temporel : le temps spatialisé ? (Ed.) L. Kupferman. Scolia. (La Préposition Française Dans Tous Ses États - 4, Actes Du Colloque PREP AN 2000) 15. 43–56.

Langacker, R. W. 1987. Mouvement abstrait. Langue française (76). 59–76.

Lehmann, C. 2002. Thoughts on grammaticalization – second, revised edition. Erfurt: Universität Erfurt, Philosophische Facultät.

Marchello-Nizia, C. 2005. A NLP-driven approach to historical linguistics. (Ed.) D. D. Pusch, J. Kabatek & W. Raible. Romanistische Korpuslinguistik II: Korpora und diachrone Sprachwissenschaft (Romance Corpus Linguistics II: Corpora and Diachronic Linguistics). Tübingen: Gunter Narr Verlag 11–30.

Marchello-Nizia, C. 2006. Grammaticalisation et changement linguistique. Bruxelles: De Boeck.

Martin, F. & M. Dominicy. 2001. “À travers”, “au travers (de)” et le point de vue. (Ed.) L. Kupferman, E. Katz & M. Asnès. Travaux de Linguistique (La Préposition) 42–43(1). 211–227.

McEnery, T. & A. Wilson. 2001. Corpus linguistics. 2nd edn. Vol. 6. Edinburgh University Press Edinburgh.

Plungian, V. A. 2002. “À travers” et les autres mots de sa famille : grammaticalisation et polysémie. (Ed.) L. Kupferman. Scolia. (La Préposition Française Dans Tous Ses États - 4, Actes Du Colloque PREP AN 2000) 15. 123–132.

Sablayrolles, P. 1995. The semantics of motion. In Proceedings of the seventh conference on European chapter of the Association for Computational Linguistics (EACL’ 95), 281–283. San Francisco, CA, USA: Morgan Kaufmann Publishers Inc. https://doi.org/10.3115/976973.977015.

Sarda, L. & D. Stosic. 2007. Les compléments spatiaux dynamiques détachés en tête : analyse des compléments en “par” et “à travers” dans la perspective de l’encadrement du discours. (Ed.) D. Stosic & N. Flaux. Les constructions détachées : entre langue et discours, Arras/ Artois : Presses Université 41–56.

Schwarze, C. 1989. Polysemie als Prozedur, am Beispiel von Frz. “à travers” und “chez.” (Ed.) C. Habel, M. Herweg & K. Rehkämper. Raumkonzepte in Verstehensprozessen. Interdisciplinäre Beiträge zu Sprache und Raum, Tübingen: Niemeyer 310–338.

Somers, A. 1988. “En travers”, une analyse sémantique (Master Thesis). Brussels: Vrije Universiteit Brussel.

Spang-Hanssen, E. 1963. Les prépositions incolores du français moderne. Copenhagen: Gads Forlag.

Stosic, D. 2002a. Les prépositions spatiales “par” et “à travers” – convergences et divergences. (Ed.) L. Kupferman. Scolia. (La Préposition Française Dans Tous Ses États - 4, Actes Du Colloque PREP AN 2000) 15. 143–156.

Stosic, D. 2002b. “Par” et “à travers” dans l’expression des relations spatiales : comparaison entre le français et le serbo-croate (PhD Dissertation). Toulouse-Le Mirail: Équipe de Recherche en Syntaxe et Sémantique, Université de Toulouse-Le Mirail.

Stosic, D. 2005. “Prendre par le sentier à travers le bois” ou comment ‘à travers’ (se) fraie un chemin. (Ed.) M. Tenchea & A. Tihu. Actes du Colloque franco-roumain de linguistique (Timisoara, 29-31 mai 2001), Timisoara, Editura Excelsior (Prépositions et Conjonctions de Subordination. Syntaxe et Sémantique) 207–218.

Stosic, D. 2007. The Prepositions “par” and “à travers” and the Categorization of Spatial Entities in French. (Ed.) M. Aurnague, M. Hickmann & L. Vieu. Human Cognitive Processing, Amsterdam/Philadelphia: John Benjamins (The Categorization of Spatial Entities in Language and Cognition 20) 71–92.

Stosic, D. 2009. Comparaison du sens spatial des prépositions “à travers” en français et “kroz” en serbe. Langages 173. 15–33.

Stosic, D., B. Fagard, L. Sarda & C. Colin. 2015. Does the road go up the mountain? Fictive motion between linguistic conventions and cognitive motivations. Cognitive Processing 16(1). 221–225. https://doi.org/10.1007/s10339-015-0723-8.

Svorou, S. 1994. The grammar of space (Typological Studies in Language). Vol. 25. Amsterdam/Philadelphia: John Benjamins.

Talmy, L. 1983. How language structures space. (Ed.) H.L. Pick & L.P. Acredolo. Spatial orientation: Theory, research, and application. New York: Plenum Press. 225–282.

Talmy, L. 2000a. Toward a Cognitive Semantics, Volume 1: Concept Structuring Systems. Cambridge (Mass.)/London: The MIT Press.

Talmy, L. 2000b. Toward a Cognitive Semantics. Volume 2: Typology and Process in Concept Structuring. Cambridge (Mass.)/London: The MIT Press.

Ungerer, F. & H.J. Schmid. 1996. An introduction to cognitive linguistics. London/New York: Longman.

Vandeloise, C. 1986. L’espace en français. Paris: Éditions du Seuil.

Vandeloise, C. 1988. Les usages spatiaux statiques de la préposition “à.” Cahiers de Lexicologie 53(2). 119–148.

Vandeloise, C. 1991. Spatial prepositions: a case study from French. University of Chicago Press.

Vandeloise, C. 1995. De la matière à l’espace: la préposition “dans.” Cahiers de Grammaire 20. 123–145.

Vaugelas, C. F. de & Z. Marzys. 2009. Remarques sur la langue françoise. Genève: Librairie Droz.

Vieu, L. 1991. Sémantique des relations spatiales et inférences spatio-temporelles : une contribution à l’étude des structures formelles de l’espace en langage naturel (PhD Dissertation). Toulouse: Université Paul Sabatier.

Weinrich, H. 1989. Grammaire textuelle du français. (Trans.) G. Dalgalian & D. Malbert. Paris: Didier/Hatier.

Haut de page

Notes

1 This research was supported by the research council (‘Onderzoeksraad’) of the Vrije Universiteit Brussel, Belgium. The author is very grateful to Marc Dominicy for his comments on an earlier version of this paper. The author also thanks the anonymous referees for their valuable comments. Of course, any errors remain his own responsiblity.

2 See www.frantext.fr

3 The method used to automatically select an arbitrary sample is the following: after entering into a spreadsheet all the relevant data for a particular expression (e.g. 3677 hits of à travers (de) for the period 1800-1849), an arbitrary number is automatically assigned to each fragment. Then the data are sorted in ascending order, and finally the representative random sample is obtained by conserving the first 200 sequential numbers only.

4 Data provided by Frantext, as consulted in June 2011.

5 Stosic (2002b) analyzed only concrete occurrences of Modern French à travers. Since comparisons are made with his observations, the analysis is conducted here under the same conditions, filtering out abstract entities functioning as Figure or Ground.

6 As is explained hereafter, Frantext was interrogated not only for par, but also for à travers. Consulting Frantext at the same moment for the two prepositions was done in order to obtain comparable results within the scope of this paper (Frantext being continuously under development, it contained less data when our corpus was created, in June 2011).

7 Relative frequencies in function of the number of words for each period are given here.

8 These examples are borrowed from Hickmann (2003: p. 81).

9 As for sub-hypothesis 1.1, Frantext was interrogated not only for par, but also for à travers, for the same reasons explained in note 6. See also below.

10 Again, abstract configurations are filtered out of this analysis.

11 Again, relative frequencies in function of the number of words for each period are given here.

12 Another interesting observation that can be made is that the numbers of à travers are caracterised by a descending tendency during the first two centuries. The period 1650-1699 presents the lowest percentage: the relative numbers thus also affirm the idea that Vaugelas’ remark (see Section 2.1) had a restraining effect on the use of à travers, but it only reinforced an existing evolution. Strangely enough, from 1700 onwards the percentages rise again and continue doing so, until reaching unseen numbers as from the middle of the 18th century.

Haut de page

Table des illustrations

Légende Table Number of tokens for each expression analyzed4
URL http://journals.openedition.org/corela/docannexe/image/14959/img-1.jpg
Fichier image/jpeg, 145k
Légende Table Evolution of the NFDP-Grounds and FDP-Grounds for à travers (de) – 1500-1699
URL http://journals.openedition.org/corela/docannexe/image/14959/img-2.jpg
Fichier image/jpeg, 125k
Légende Table Evolution of the NFDP-Grounds and FDP-Grounds for à travers (de) – 1700-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-3.jpg
Fichier image/jpeg, 127k
Légende Table Most frequent (concrete) entities constituting NFDP-Grounds for the period 1500-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-4.jpg
Fichier image/jpeg, 196k
Légende Fig. Co-occurrences of à travers and par with pays – 1500-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-5.jpg
Fichier image/jpeg, 120k
Légende Fig. Co-occurrences of à travers and par with arbres – 1500-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-6.jpg
Fichier image/jpeg, 125k
Légende Fig. Co-occurrences of à travers and par with champ(s) – 1500-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-7.jpg
Fichier image/jpeg, 132k
Légende Fig. Co-occurrences of à travers and par with feuillage/feuilles– 1500-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-8.jpg
Fichier image/jpeg, 143k
Légende Table Evolution of the type of motion complex containing à travers (de) – 1500-1699
URL http://journals.openedition.org/corela/docannexe/image/14959/img-9.jpg
Fichier image/jpeg, 151k
Légende Table Evolution of the type of motion complex containing à travers (de) – 1700-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-10.jpg
Fichier image/jpeg, 144k
Légende Table Most frequent (concrete) verbs found in motion complexes expressing a CoP for the period 1500-1899
URL http://journals.openedition.org/corela/docannexe/image/14959/img-11.jpg
Fichier image/jpeg, 201k
Légende Fig. Evolutions of [courir+ par] and [courir+ à travers]
URL http://journals.openedition.org/corela/docannexe/image/14959/img-12.jpg
Fichier image/jpeg, 116k
Légende Fig. Evolutions of [guider + par] and [guider + à travers]
URL http://journals.openedition.org/corela/docannexe/image/14959/img-13.jpg
Fichier image/jpeg, 142k
Légende Fig. Evolutions of [conduire + par] and [conduire + à travers]
URL http://journals.openedition.org/corela/docannexe/image/14959/img-14.jpg
Fichier image/jpeg, 113k
Légende Fig. Evolutions of [traisner/traîner/trainer + par] and [traisner/traîner/trainer + à travers]
URL http://journals.openedition.org/corela/docannexe/image/14959/img-15.jpg
Fichier image/jpeg, 133k
Légende Fig. Evolutions of [suivre + par] and [suivre + à travers]
URL http://journals.openedition.org/corela/docannexe/image/14959/img-16.jpg
Fichier image/jpeg, 107k
Légende Table Frequencies of tokens of à travers (de) and par in function of the number of words in Frantext
URL http://journals.openedition.org/corela/docannexe/image/14959/img-17.jpg
Fichier image/jpeg, 195k
Légende Fig. Frequencies of tokens of à travers (de) and par in function of the number of words in Frantext
URL http://journals.openedition.org/corela/docannexe/image/14959/img-18.jpg
Fichier image/jpeg, 135k
Légende Fig. Visualization of the combinatorial evolution of the preposition de with à travers and au travers
URL http://journals.openedition.org/corela/docannexe/image/14959/img-19.jpg
Fichier image/jpeg, 149k
Légende Table Frequencies of tokens of à travers (de) and au travers (de) in function of the number of words in Frantext
URL http://journals.openedition.org/corela/docannexe/image/14959/img-20.jpg
Fichier image/jpeg, 137k
Légende Table Frequencies of tokens of à travers (de) and au travers (de) in function of the number of words in Frantext
URL http://journals.openedition.org/corela/docannexe/image/14959/img-21.jpg
Fichier image/jpeg, 144k
Légende Fig. Frequencies of tokens of à travers (de) and au travers (de) in function of the number of words in Frantext
URL http://journals.openedition.org/corela/docannexe/image/14959/img-22.jpg
Fichier image/jpeg, 140k
Haut de page

Pour citer cet article

Référence électronique

Thomas Hoelbeek, « The diffusion of French à travers from the 18th century onwards. »Corela [En ligne], 20-1 | 2022, mis en ligne le 28 juin 2022, consulté le 28 mars 2024. URL : http://journals.openedition.org/corela/14959 ; DOI : https://doi.org/10.4000/corela.14959

Haut de page

Auteur

Thomas Hoelbeek

Linguistics and Literary Studies (LIST), Vrije Universiteit Brussel

Haut de page

Droits d’auteur

CC-BY-NC-SA-4.0

Le texte seul est utilisable sous licence CC BY-NC-SA 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search