- 1 The recurrent lexico-syntactic constructions (RLCs) are conceived here from a contextualist perspec (...)
1In literary studies, subgenres of the novel tend to be defined first and foremost in terms of their subject matter and recurring structural features on the macro-level (plot patterns, setting, character constellations). Language usually plays a minor role in genre definitions and is often seen primarily in the context of establishing the setting (e.g. neologisms as world-building strategies in science fiction) or supporting characterisation (e.g. sociolects hinting at characters’ social background). In the following, we propose an alternative, corpus-driven approach to the definition and contrastive analysis of two subgenres, fantasy (FY) and the historical novel (HIST), based on the concept of the phraseological motif. This concept has proved useful in several case studies (cf. Jacquot et al. 2020, Sorba et al. 2020, Vidotto & Goossens 2020) since it facilitates linking literary and linguistic interests systematically. Our approach helps to understand the role played by recurrent lexico-syntactic constructions (RLCs)1 as semantic and structural genre-specific building blocks of literary texts.
- 2 See, for instance, Renauld’s (2016) review of the French translation of publications by the Stanfor (...)
2The focus on genres distinguishes our approach from corpus-stylistic studies that examine the style of a single author (cf. Fischer-Starcke 2010, Mahlberg 2013) or even that of a single literary text (e.g. Stubbs 2005, Ho 2012) as well as from the ‘distant reading’ approach developed by Moretti and other scholars affiliated with the Stanford Research Lab (cf. Allison et al. 2011), which have met with some scepticism, especially on the part of literary scholars.2 Our corpus-driven approach positions itself in between the two extremes sketched above as far as its scope is concerned. Instead of focusing on individual authors/texts or on large unstructured literary corpora, which are required for Moretti’s distant reading, this approach draws upon an automatic extraction of RLCs which serve as a starting point for identifying and characterizing the phraseological motifs that are statistically specific to FY and HIST.
3We present our corpora and our methodology in section 2. Section 3 proposes two case studies on phraseological motifs in English and French fantasy novels (FY) and historical novels (HIST).
4The case studies in section 3 are based on data extracted from the English and French PhraseoRom corpora.3 The criteria for compiling the literary corpora (consisting of approximately 2,000 novels) and for partitioning them into sub-corpora were provided by genre categories used in literary studies and by labels employed by publishing houses and library catalogues. Tables 1 and 2 below detail the composition of the six English and French sub-corpora: general fiction/littérature blanche (GEN), fantasy (FY), historical novels (HIST), crime fiction (CRIM), romance novels (ROM), and science fiction (SF):
Table 1. The number of authors, novels and tokens in the English PhraseoRom corpora
Subgenre
|
Authors
|
Novels
|
Tokens
|
GEN
|
104
|
169
|
19,461,070
|
FY
|
57
|
153
|
22,531,553
|
HIST
|
52
|
128
|
19,339,040
|
CRIM
|
71
|
163
|
19,223,169
|
ROM
|
39
|
103
|
14,440,646
|
SF
|
53
|
133
|
16,705,651
|
Total
|
376
|
849
|
111,701,129
|
Table 2. The number of authors, novels and tokens in the French PhraseoRom corpora
Subgenre
|
Authors
|
Novels
|
Tokens
|
GEN
|
172
|
442
|
34,593,554
|
FY
|
43
|
109
|
13,966,566
|
HIST
|
40
|
114
|
14,193,642
|
CRIM
|
85
|
196
|
18,010,790
|
ROM
|
38
|
110
|
9,516,633
|
SF
|
40
|
150
|
13,528,393
|
Total
|
418
|
1,121
|
103,809,578
|
- 4 The French and English PhraseoRom corpora have been annotated syntactically. Lexico-syntactic trees (...)
- 5 For a detailed description of this process, cf. Kraif (2019).
5Automatically extracted recurrent lexico-syntactic trees (RLTs)4 (Kraif 2019) include information on syntactic relations of dependency between lexical units that are specific to particular sub-corpora compared to all PhraseoRom corpora.5 Figure 1 and Figure 2 illustrate the RLTs which are specific to FY and HIST in both languages.
Figure 1. Structure of the RLTs draw sword and dégainer épée (automatically extracted from the English and French corpora)
Figure 2. Structure of the RLTs raise his voice and dire d’une voix (automatically extracted from the English and French corpora)
6In accordance with our methodology (cf. Diwersy et al. 2021), RLTs have to fulfil a number of criteria. The minimal morpho-syntactic criterion is containing a verb. The specificity of an RLT to a (sub)genre is determined by its log-likelihood ratio (LLR, cf. Dunning 1993). The LLR index should be higher than 10.83 (= threshold of statistical specificity), using the entire PhraseoRom corpus as backdrop. Moreover, only those RLTs were examined whose absolute frequency amounts to 10 or more occurrences in a sub-corpus and which occur in 50% or more of the authors in a sub-corpus, which ensures a sufficient dispersion.
- 6 For more details, cf. Novakova & Siepmann (2020: 291-293).
7RLTs like <draw sword> and <dégainer l’épée>, for example, give rise to statistically relevant phraseological motifs, i.e., recurrent “multidimensional units” (Legallois 2012: §65) that often fulfil specific discursive functions (DFs) in the literary text. A phraseological motif constitutes a “collocational frame” consisting of stable as well as variable lexical and syntactic components (cf. Longrée & Mellet 2013: 66). In other words, one and the same phraseological motif typically encompasses syntagmatic extensions of its nucleus as well as paradigmatic variation on its components, as the case studies in section 3 will illustrate. Moreover, through their specific discursive functions (DFs), the phraseological motifs play an important role for structuring the literary texts. Following Adam (2011: 267), a basic distinction can be made between the narrative DF (referring to actions that drive the plot) and the descriptive DF, associated with descriptions of places, characters, etc. in the text. Research within the PhraseoRom project has identified further typical DFs,6 for instance the affective DF (depicting emotions), the cognitive DF (representing characters’ thoughts), and the pragmatic DF (depicting characters’ speech). There is neither a one-to-one correspondence between phraseological motifs and DFs, nor are DFs necessarily mutually exclusive.
- 7 Lexicoscope makes it possible to extract 5,000 occurrences of collocations and colligations that ar (...)
- 8 For more details on “unexpected” statistically specific RLTs in English and French FY, cf. Gossens (...)
8The data has allowed us to identify statistically significant expressions,7 some of which may not be very surprising from the point of view of genre criticism since they quite obviously contribute to the world building in the two subgenres in the two languages. Thus, in English FY collocations including the following ones, which are in accordance with the criteria presented in section 2, are highly specific: <life save> (98,62), <draw breath> (96,99), <crackle fire> (93,18), <slash sword> (67,18), <wrinkle brow> (51,02). For French FY the following constructions are statistically relevant: <utiliser magie> (233,41) ‘use magic’, <pousser cri> (220,68) ‘shout a cry’, <sourcil froncer> (222,41) ‘furrow his brow’, <sang couler> (206,16), ‘blood flow’, <échanger regard> (196,71), ‘exchange glance’. As the results show, our corpus driven approach has also revealed less likely results, for example <draw breath> or <wrinkle brow> in English or <sourcil froncer> ‘furrow his brow’ or <échanger regard> ‘exchange glance’ in French8.
- 9 For more details, see Tables 3 and 4.
9For the English HIST sub-corpus, expressions like the following ones stand out: <anchor drop> (199,32), <cathedral build> (195,94), <telescope raise> (181,84), <harbour enter> (159,22), <trumpet raise> (41,16). In French HIST the following expressions are statistically significant: <pape élire> ‘elect pope’ (115,04), <confier commandement> (62,05) ‘give command’, <célebrer fête> (42,28) ‘celebrate feast’, <trône hériter> (27,31) ‘inherit the throne’, <psaume chanter> (53,06) ‘sing a psalm’. Again, there are also constructions that seem to be less likely candidates as basic building blocks of the genre, in particular expressions that refer to verbal communication in the two languages, which we will analyze in section 3.2: <raise his voice>, <give order>, <dire d’une voix> [say in a voice], <donner l’ordre> [give the order].9
10For the following contrastive analysis of RLTs in English and French FY and HIST, four specific RLTs have been selected: <draw sword> and <dégainer l’épée> (FY) and <raise voice> and <dire d’une voix> (HIST).10 These RLTs, which correspond to binary collocations, generate different phraseological motifs that characterize the subgenre thematically, structurally and stylistically. Moreover, the contrastive approach serves to highlight stylistic differences between literary traditions that are characteristic of English and French fiction.
- 11 This happens to be true even for children’s FY, as novels including Tolkien’s The Hobbit, Lewis’s T (...)
- 12 Even the Harry Potter series, where wands are the predominant weapon, evokes the idea that swords a (...)
11A comparison of the data extracted from the English and French FY corpora reveals a striking over-representation of RLTs referring to physically violent or threatening actions. The RLTs <draw sword> (LLR 546.30) and <dégainer l’épée> (LLR 454.01) stand out, being among the most specific RLTs for FY in the two languages. These quantitative findings correlate with the fact that physical combat tends to be a central component of the plot in FY.11 Moreover, these findings suggest that the sword is the prototypical FY weapon, which also correlates with the appearance of many named swords in the genre – a strategy used to stress the significance of a particular sword that can be traced back to weapons like the famous Excalibur in medieval literature.12
12As the data show, there is considerable paradigmatic variation with regard to the verbs combined with the nouns épée | sword. Tables (1) and (2) illustrate the ten most frequent verbal collocates of épée | sword specific to FY in the corpora:
Table 1. The ten most frequent verbal collocates of <sword> in the English FY sub-corpus, in descending order based on the LLR index
FY EN
|
LLR
|
<draw sword>
|
546.30
|
<raise sword>
|
82.51
|
<sheathe sword>
|
81.24
|
<swing sword>
|
75.26
|
<sell sword>
|
63.45
|
<drop sword>
|
62.36
|
<wield sword>
|
49.25
|
<unsheathe sword>
|
40.81
|
<brandish sword>
|
24.88
|
<thrust sword>
|
24.44
|
Table 2. The ten most frequent verbal collocates of <épée> in the French FY sub-corpus, in descending order based on the LLR index
FY FR
|
LLR
|
<dégainer épée> [draw sword]
|
454.01
|
<tirer épée> [draw sword]
|
276.62
|
<lever épée> [raise sword]
|
206.95
|
<lâcher épée> [drop sword]
|
157.31
|
<manier épée> [wield sword]
|
145.88
|
<brandir épée> [brandish sword]
|
141.23
|
<porter épée> [carry sword]
|
130.86
|
<rengainer épée> [sheathe sword]
|
124.65
|
<sortir épée> [draw/take out sword]
|
112.04
|
<planter épée> [plant sword]
|
93.91
|
- 13 Cf., for example, on the phraseological motifs generated by the noun dragon in French FY Sorba & No (...)
13The remarkable variation of verbal collocates of épée | sword fits perfectly into the (pseudo-)medieval settings characteristic of many English and French FY novels. The different verbs combined with sword | épée refer to distinct stages of a sword fight: draw, unsheathe, dégainer, tirer, sortir (initial stage/preparation for the fight), wield, swing, brandish, thrust, lever, brandir, manier (middle stage), sheathe, drop, rengainer, lâcher (final stage/end of the sword fight). In medieval epics, from which FY continues to draw inspiration, this structure of different stages of fighting gave rise to the trope of single combat (“combat singulier”) (Rychner 1955), which constitutes a “recurrent narrative configuration” in French pre-modern literature (Weil 2016 [1990]) and English medieval epics alike. Similar to medieval epic literature ranging from the Old English Beowulf to Chrétien de Troyes’s Arthurian romances, FY still tends to feature plots in which single combat is perceived as crucial for the characters’ development as well as for the fate of the fictional world. Epic conflicts between different factions, who are often stylised as forces of good and evil, is characteristic of fantasy. To stress that the stakes are high in a sword fight, supernatural and monstrous beings are among the enemies,13 which are often based on folklore, fairy tales, and/or mythology (cf. Baudou 2005: 46).
14Being part of the literary scenario of single combat, the phraseological motifs <draw sword> and <dégainer l’épée> typically fulfil a narrative DF; they tend to correlate with the preparation for or the onset of a sword fight. The motifs function as building blocks of the story that often coincide with turning points, with stages in the quest plot that is characteristic of many FY novels (cf. Baudou 2005: 45), or even with the climax of the entire plot (killing the enemy) and may contribute to a character’s development (gaining experience in combat, acting fearlessly, heroically, cruelly). The examples below illustrate these possibilities:
(1) Il dégaina son épée à la lame coupante et, sans pitié, l’enfonça dans le coeur d’Opale. (Bujor F., La prophétie des pierres, 2002)
- 14 Translations (1a), (2a), (6a) are based on the parallel corpus of literary translations that is par (...)
(1a) He drew his keen sword and with a brutal smile, plunged it pitilessly into Opal’s heart.14
(2) ll maniait avec dextérité une épée apparemment enchantée et, surtout, il ne redoutait pas la mort (Bujor F., La prophétie des pierres, 2002)
(2a) He wielded an apparently enchanted sword with uncommon skill and, above all, he did not fear death.
(3) As they came out into the light at the other end of the alley Cohen flung himself against the wall, drew his sword, stood with his head on one side as he judged the approaching footsteps, and then brought the blade around in a dead flat sweep at stomach height. (Pratchett T., The Light Fantastic, 1986)
(4) He then raised his sword and began to scrutinise the blade, apparently taken aback to discover it streaked with blood. (Ishiguro K., The Buried Giant, 2016)
15As the examples above already suggest, there tend to be more syntagmatic extensions of the RLTs in French than in English, which renders the motif linguistically more elaborate in French. Examples (1) and (2) as well as (5) and (6) below exemplify an overall tendency of French FY authors to provide more details when referring to the noun sword. In the collocation dégainer l’épée, which forms the core of the French motif, the nucleus épée is often accompanied by an adjectival modifier (énorme [huge], large [broad], fine [fine], courte [short] (ex. 5), tranchante [sharp] (ex. 5), or enchantée [enchanted], as in (2)), by a prepositional phrase (à la lame coupante in (1), à la lame étincelante in (6)), or by a participle extension (maculée de son propre sang in (5)):
(5) Tout contre lui surgit le visage grimaçant d’un homme d’armes coiffé d’un bassinet de fer, brandissant une épée courte et tranchante, maculée de son propre sang. (Fetjaine J.-L., La trilogie des elfes 2 La nuit des elfes, 1999) [The grimacing face of a man armed with an iron pommel, brandishing a short and sharp sword, stained with his own blood, appeared opposite him.]
(6) De sa main valide, il dégaina rapidement son épée à la lame étincelante. (Bujor F., La prophétie des pierres, 2002)
(6a) Suddenly he heard a sound, and with his good hand, swiftly drew his gleaming sword.
(7) Falk dégaina son épée d’un geste qui en disait long sur son expertise des armes. (R. Michel, L’agent des ombres 8 Ange et Loki, 2013) [Falk drew his sword with a gesture that spoke volumes about his expertise with weapons.]
16Adverbial modifiers accompanying the verb in the French collocation convey an impression of the fighting style, inviting readers to visualise an action sequence: dégainer lentement / vivement / rapidement (ex. 6); rengainer rageusement / tranquillement son épée. In some contexts, these modifiers can add an affective FD to the narrative FD. Moreover, as example (7) demonstrates, the motif encompasses complex syntagmatic extensions of the collocation frame <dégainer l’épée>: a prepositional phrase (d’un geste [with a gesture]) enriched by a relative clause (qui en disait long sur son expertise des armes [that spoke volumes about his expertise with weapons]). These syntagmatic extensions add a descriptive dimension to the narrative DF.
17In contrast to French, where many multi-word syntagmatic extensions occur within the extracted RLTs, the corresponding expressions in English turn out to be less embellished and the motifs are also less variable on the lexical level, as examples (3)-(4) above illustrate. Although adjectival modifiers like polished (8), silver, flamed occur, they are far less frequent than similar syntagmatic extensions in the French examples:
(8) Urtha drew his polished sword, rested it in his hand. (Holdstock R., Celtika, 2001)
(9) He drew his sword again and beat on the stone with the hilt, but it only gave out a dull sound. (Tolkien J.R.R., The Two Towers, 1954)
18As the results show, motifs generated by <draw sword> in English FY fulfil only a narrative DF. The scarcity of syntagmatic extensions of the motif in the English examples arguably accelerates the pace of the action, while making it perhaps more difficult for readers to visualise the fight scene in detail. Moreover, among the 437 occurrences of <draw sword> and 88 occurrences of <raise sword> in English FY, there are only three instances of adverbial modifiers (instinctively, reflexively, again (ex. 8)).
19To sum up, the linguistic complexity of the phraseological motifs identified in French FY has an impact on their discursive functions, which are often mixed in French, i.e. narrative and descriptive, narrative and affective. A more pronounced tendency towards syntagmatic extension and paradigmatic variation in the French sub-corpus can also be observed with other motifs (cf. Novakova & Gymnich 2021), which seems to reflect the strong emphasis on variation as a stylistic ideal of French literary language in general (cf. Philippe 2016).
20While there are clear differences between the phraseological motifs <dégainer épée> and <draw sword> with respect to their syntagmatic extensions, the syntactic distribution of the two motifs is almost identical. Both occur primarily in paratactic constructions or in coordination. The following examples illustrate these similarities:
(10) Orville sortit l’épée à lame large qui lui avait servi de pelle, la jeta dans la ravine et mit à sa place l’arme qu’il s’était choisie. (Goddyn R., 1 Le sang des 7 rois 1 Livre premier, 2013) [Orville took out the sword with the broad blade which had served him as a shovel, threw it into the gully and put into its place the weapon he had chosen.]
(11) Then, as one man or troll, or demon they drew their swords, turned around, and headed towards Lord Hong’s army. (Pratchett T., DW 17 – Interesting Times, 1994)
21The syntactic distribution in examples (10) and (11) shows clearly that the RLTs typically form part of a sequence of actions. This seems apt in a genre like FY, which tends to be action-driven. More specifically, this syntactic distribution of the RLTs coincides with their function of marking stages in sword fights and, more generally, with their narrative DF.
22Moreover, the motifs <dégainer épée> and <draw sword> often appear in gerund constructions in both languages. In this case, the motif seems to fulfil an infra-narrative DF, since the focus is on other actions (e.g. movements), while drawing a sword is something that happens simultaneously and somewhat in the background, as in (12) and (13):
(12) Côme en profita pour faire balancer son adversaire sur le flanc et se relever tout en dégainant son épée. (Gaborit M., Le Roi des Cendres, 2002) [Côme took advantage of it to swing his opponent to the side and stand up while drawing his sword.]
- 15 The motif appears in gerund constructions in 10% of all occurrences of <raise sword>.
(13) With a great effort he turned in his saddle and, lips curled back against the pain, stood in his stirrups, raising his sword. (Abercrombie J., The Heroes, 2011)15
- 16 McIntyre and Walker (2019, 16) define “text style” as the result “of how the linguistic choices evi (...)
23In sum, the specificity of the motifs <dégainer épée> and <draw sword> to French and English FY is evidence of thematic and structural similarities as far as the depiction of single combat is concerned. While sword fights are typical of FY and are associated with narrative or infra-narrative DFs (depending on the syntactic distribution of the motifs) in both languages, an overall tendency to feature a higher degree of paradigmatic variation as well as more syntagmatic extensions gives rise to a more elaborate style in French FY.16
- 17 The PhraseroRom semantic grid presented in Diwersy et al. (2021) is based on the tradition of WordN (...)
24While the statistical relevance of phraseological motifs associated with single combat in FY might not come as a surprise due to the impact medieval epic poetry and courtly romances have had on the genre, our corpus-driven approach may also reveal RLTs that serve as building blocks in a genre without being clearly linked with themes which might be deemed characteristic of the genre. While it is difficult to specify themes that are typical of HIST in general beyond the genre-defining setting in the past, it seems safe to assume that a focus on large-scale violent historical events, such as wars, revolutions, uprisings, conquests, and persecution, is at least very common since it makes it easier to generate exciting plots. Yet, the data extracted from the English and French HIST sub-corpora show that RLTs belonging to the semantic category of verbal communication17 prove to be specific to both English and French HIST novels. The expressions of verbal communication are more specific than expressions of the semantic category actions that might be thought of as being linked to HIST due to their thematic focus (as leave the ship, restore order, faire la guerre or quitter la cour). Tables (3) and (4) summarise these results:
Table 3. RLTs specific to English HIST
Semantic value
|
RLTs
|
LLR
|
Verbal communication
|
<raise his voice>
<the captain said>
<give order>
<say in his voice>
|
109.34
82.57
67.12
26.54
|
Action
|
<leave the ship>
<broke the seal>
<leave the court>
<restore order>
|
92.89
56.58
31.94
20.64
|
Table 4. RLTs specific to French HIST
Semantic value
|
RLTs
|
LLR
|
Verbal communication
|
<dire d’une voix> [say in a voice]
<dire avec un sourire> [say with a smile]
<donner l’ordre> [give the order]
<dire à l’oreille> [say in the ear]
|
370.34
343.72
283.08
177.21
|
Action
|
<faire la guerre> [make war]
<mettre le siège> [lay siege]
<quitter la cour> [leave the court]
<rétablir l’ordre> [restore order]
|
119.60
118.28
107.10
67.84
|
- 18 A pragmatic DF is attributed to motifs that express speech acts involving the characters in a novel (...)
25These results allow the hypothesis that HIST can be defined in terms of interaction or verbal communication as much as in terms of actions which are considered typical of HIST. The specificity of RLTs such as <donner l’ordre> (LLR 283.08) or <rétablir l’ordre> (LLR 67,84) (see Table 4) or <give order> (LLR 67.1) and <restore order> (LLR 20,64) (see Table 3) could be explained by the emphasis on depicting social and/or military hierarchies in this genre (Gonon & Sorba 2020: 13). The RLT <give order> is typically combined with infinitive constructions: give order to march up to the citadel, to move out, to capture prisoners, etc. The actions referred to by the infinitive constructions may lead up to major actions on the plot level. Due to their (likely) impact on the plot, the motifs <give order> and <donner l’ordre> can be associated both with a narrative DF and a pragmatic DF18 in many contexts.
26An analysis of the more specific RLTs <dire d’une voix> (LLR 370,34) and <dire avec un sourire> (LLR 343,72) in the French HIST corpus (see Table 4) reveals that these RLTs appear to introduce or mark direct or reported speech, as the following examples illustrate:
(14) Le soir, en me versant du thé, Charlotte dit d’une voix distraite: – C’est drôle, la Koukouchka n’est pas encore passée... (Makine A., Le testament français, 1995)
(14a) That evening, pouring me a cup of tea, Charlotte said in a preoccupied voice, “It’s funny, the Kukushka hasn’t been past yet...
(15) Le duc de Berry, notre bon duc Jean comme disait mon père avec un douloureux sourire, avait été empêché d’entrer dans Paris, où il possédait une résidence. (Rufin J.-Ch., Le grand cœur, 2012)
(15a) The Duke of Berry, our good Duke John as my father used to call him, with a sorrowful smile, had been prevented from entering Paris, where he had a residence.
(16) – Pas de lui-même. Louis, poursuivit Monsieur de Bonneuil avec un discret sourire, ne lui a jamais pardonné les suggestions qu’il lui avait faites pour remédier au «délaissement» de la reine. (Merle R., Fortune de France 9 Les Roses de la vie, 1995) [‘Not from himself. Louis,’ Monsieur de Bonneuil continued with a discreet smile, ‘never forgave him for the suggestions he had made to remedy the Queen’s “neglect”.’]
(17) – Vous raisonnez fort justement, monsieur, approuva Richelieu avec un sourire sans joie. Que savez-vous d’autre? (d’Aillon J. Guerre des trois Henri 1 Les rapines du duc de Guise, 2008) [‘Your reasoning is correct, Monsieur,’ Richelieu agreed with a mirthless smile.’ ‘What else do you know?’]
- 19 Cf. Diwersy et al. (2020) on speech verbs in motifs in French and English.
- 20 On this, cf. also Gonon & Sorba (2020).
27Since the motifs primarily embed characters’ speech into the text, the readers’ attention is directed towards what is said (i.e. the content of the utterance) rather than to the act of speaking, which implies that an infra-narrative DF should be attributed to the motifs. The examples also suggest that the phraseological motifs are very rich, featuring a high degree of paradigmatic variation in French FY with regard to the verb (dire [say]19 | ordonner [order] | répliquer [reply] | répondre [answer] | expliquer [explain] | lancer [launch] | prononcer [pronounce] | ajouter [add] | crier [shout] | reprendre [resume]), the noun (voix [voice] | ton [tone]), and adjectival extensions (voix forte [loud] | soured [sultry] | claire [clear] | étouffée [muffled] | ferme [firm] | calme [quiet] | grave [deep] | tremblante [shaky] | sèche [dry] | entrecoupée [interspersed]).20 The high degree of paradigmatic variation and the syntagmatic extensions occurring with the motifs are apt to contribute to the stylistic variation that tends to be deemed eminently desirable by French writers (cf. Phillippe 2016: 96). Moreover, adjectival modifiers in French contribute to depicting the state of mind or the emotions of characters (voix distraite ‘preoccupied voice’ (ex. 14), avec un douloureux sourire ‘with a sorrowful smile’ (ex. 15), avec un discret sourire ‘with a discreet smile’ (ex. 16), un sourire sans joie ‘with a mirthless smile (ex. 17)). In these cases, the DF of the motifs is twofold: infra-narrative and affective.
28In contrast to the French motifs associated with verbal communication in HIST, the motif <raise voice> (LLR 109.34) specific to English HIST displays few syntagmatic extensions. There are some adverbial modifiers – slightly, a little, angrily – and infinitive constructions that explain the characters’ reasons for raising their voice (to be heard):
(18) A murmur which seemed to contain both relief and triumph went around the table, but this time M. Dupont raised his voice slightly and pronounced over it. (Ishiguro K., The Remains of the Day, 1989)
(19) Jack raised his voice to be heard over their exclamations. (Follett K., The Pillars of the Earth, 1989)
29If speakers raise their voice to be heard, the motif corresponds to a minor action and is consequently associated with an infra-narrative DF. In some cases, however, raising one’s voice may be caused by excitement, by feelings of triumph, anger, even madness:
(20) His eyes took on their mad look and he raised his voice. ‘The Prophetess told me,’ he said in a tone of rising triumph, ‘that the twenty-third day of Mechir would bring glad tidings for all who fight against subjugation by light-eyed foreigners. This may be it.’ (Blake R., The Blood of Alexandria, 2010)
(21) ‘There’s a bloody great mine, sir!’ Sharpe raised his voice angrily. ‘Just waiting to kill our lads!’ (Cornwell B., Sharpe’s Tiger, 1997)
30In these examples, the motif contributes to the characterisation of the speaker and to expressing his emotions (= affective and pragmatic DFs), which is only made clear by the context, however. In other words, the motif <raise voice> appears to depend more on the semantic context for disambiguation and for identifying its DF than some other motifs.
- 21 On the perimeters of the functions of phraseological motifs and their granularity (motifs that are (...)
- 22 Philippe (2024) speaks of “tiraillements” [tug-of-war] and ‘tensions’ between the symbolist/impress (...)
31The contrastive analysis of phraseological motifs generated around the nouns sword | épée in FY or speech verbs in HIST, which proved to be specific in English and in French in the two genres, shows that our textometric approach based on the extraction of RLCs adds a new dimension to the analysis of ‘single combat’ in FY and verbal interaction in HIST, the latter being a particularly surprising result. Our findings allow us to conclude that phraseological motifs are reliable generic markers21 and resemble what Molinié (1996) felicitously refers to as “stylèmes”, i.e. distinctive features of style. This means we can go beyond existing stylistic analyses as well as literary analyses that rely on the study of recurrent themes or features of fictional universes (e.g. a fantastic world in FY, a setting in the past in HIST). Our approach provides empirical evidence of the existence of considerable similarities between genre conventions in English and French literature. Moreover, while the motifs fulfil almost identical DFs in the two languages (narrative, infra-narrative, affective and pragmatic DF), the main difference between English and French FY and HIST lies in how phraseological motifs are realised linguistically. The motifs extracted from the French sub-corpora invariably exhibit higher degrees of both paradigmatic and syntagmatic variation than the English motifs. The results of the case studies support Philippe’s (2016) claim that variation is a stylistic ideal that tends to be more important for French writers than it is in English fiction. Our data have revealed a more pared-down, sober and compact style in English FY and HIST, which contrasts with the stylistic ideal characteristic of French writers, especially in the 19th century. This stylistic ideal, which seems to have a lasting impact on contemporary novels, is a more elaborate style, enriched by various ‘ornaments’.22 Of course, this conclusion should be put to the test in further studies.
32In more general terms, we can conclude that the phraseological motif is a concept that makes it possible to combine a quantitative, corpus-driven methodology with more traditional approaches from the field of literary studies and literary history. We hope that our case studies will provide fresh impulses for corpus stylistics and literary studies that focus on genres of the French and English novel.