Article body

1. Introduction and rationale

The exclamative is generally recognised as one of the four main sentence types, alongside the declarative, interrogative and imperative. These sentence types are easily distinguishable from each other due to their clear individual syntactic form and semantic purpose. Exclamative sentences are characterised by question words cross-linguistically (Michaelis 2001: 1042) and English and French are no exception. The only universally accepted exclamatives in English are limited to those fronted by interrogative words what (1) and how (2) (Huddleston 1984; Quirk, Greenbaum, et al. 1985: 833; Biber, Johansson, et al. 1999). French reference grammars show more variation in exclamative constructions than in English, with exclamative constructions fronted by quel, que, combien, comme, ce que, and qu’est-ce que (Jones 1996; Grevisse and Goosse 2008; Riegel, Pellat, et al. 2009).[1]Quel, the determiner in question here, is described as both interrogative and exclamative (Grevisse and Goosse 2008: 505; Riegel, Pellat, et al. 2009: 688).

The main aim of this paper is to investigate how what and quel exclamatives are translated into French or English respectively. Much recent work in the fields of Translation Studies and the translation of political or parliamentary discourse has focused on the cultural and ideological implications of translation (Baumgarten and Gagnon 2005; 2016; Schäffner 2010/2016; 2012/2016; Gagnon 2021). The current work aims to advance research on stylistic and linguistic elements of political translation, thus contributing at the “linguistic micro-level” (Schäffner 1997: 2). The French quel exclamative has been said to correspond to the English exclamative what [a] (Jones 1996: 519) and most bilingual and bidirectional dictionaries offer it as a translation (Collins Online[2]; Cambridge Dictionary Online[3]; Larousse[4]). Indeed, syntactically, exclamative what and quel constructions look very similar. Both allow for full (as shown in examples 1 and 3), reduced (or clause fragment) (examples 5-6), and embedded realisations (examples 8-9) and for the what or quel element to act as subject (as shown in example 3), object (examples 7 and 10), complement or prepositional complement (example 9) (see Quirk, Greenbaum, et al. 1985: 833-834; Riegel, Pellat, et al. 2009: 404-405). However, there are a couple of key differences between the two constructions.[5] Exclamative what acts as a predeterminer with singular count nouns, and so these what exclamatives are distinguished by the appearance of the indefinite article, a[n] (Bolinger 1972: 60; Quirk, Greenbaum, et al. 1985: 834). French quel exclamatives have no such feature. They may, however, be used with negation, as shown in example 11, unlike exclamative what, which was not found used with such negation in this data.

Exclamative constructions, such as those fronted by what or quel, are set apart by how they are used. They are generally said to convey an emotional reaction by the speaker (Huddleston and Pullum 2002: 922; Collins 2005: 4). This reaction has been said to generally be to a quality or quantity that exceeds expectation (Rett 2008; 2011; Ghesquière and Troughton 2021: 145-6). In this vein, exclamatives have further been qualified in terms of “mirativity,” the grammatical encoding of surprise (DeLancey 1997; 2001; Krawczak and Glynn 2015; Unger 2019; Ghesquière and Troughton 2021: 146). This paper follows Krawczak and Glynn’s (2015: 354) operationalisation of mirativity as a simultaneous activation of a high degree of both functional performativity and conceptual incongruity. The exclamative construction certainly appears to conform to this definition. It is, by definition, emotionally charged and thus performative, and it communicates the fact that the speaker’s expectations or conceived notions of normality have been exceeded or deviated from in some way. In example (1), “what a scandal” indicates that something about the speaker’s notion of what constitutes a scandal has been exceeded or deviated from in some way, which is evident in context, provided in example (12) below. In this case, Van Orden uses two what exclamatives in a row. The first indicates his perception of an extreme, and thus unusual and unexpected, degree of disgracefulness in the inaction or approval of the other African leaders. This is followed by a second exclamative (example 1), where the speaker expresses that the events in Zimbabwe and the reactions to them have exceeded or deviated from his conception of what constitutes a scandal.

As exemplified in (12), exclamative constructions have also increasingly been defined in terms of how they convey degree (Rett 2011). The exclamative word what has specifically been defined as a degree modifier (Bolinger 1972; Ghesquière 2014; Siemund 2017; Ghesquière and Troughton 2021). That is to say that what measures, on an open-ended scale, the degree of a gradable quality or quantity that is either explicit or implicit in the text, indicating that it applies to a very high degree. In examples (13) and (14) below, the qualities explicitly expressed by the noun disgrace and adjective difficult are intensified by exclamative what a to imply an extreme degree of disgrace and difficulty. In (15), the gradable quality of the “end” is not explicit in the noun itself and must be established through the context of the discourse act. In this case, the what exclamative in (15) is preceded by a how exclamative using degree adjectives sad and pathetic, so it is either that these qualities also apply to the noun end in the following what exclamative or that the wh- element intensifies the irony of the situation.

If the French quel exclamative can truly be said to correspond to the English what exclamative, it may be assumed that it implies the same degree of modification. However, while reference works have said that quel “expresses degree or quality” [“exprime le degré ou la qualité”] (Grevisse and Goosse 2008: 506), Marandin qualifies quel as a “non-scalar exclamative word” [“mot exclamative non-scalaire”] (2010: 39). He argues that without a degree noun or modifier, quel does not modify the degree of an implicitly present quality or quantity but rather the construction presents “the ideal or anti-ideal” of the noun in question [“l’idéal ou l’anti-idéal”] (Marandin 2010: 39). Similarly, quel has been described elsewhere as a “type” exclamative word, as opposed to those which can imply “quantity” or “intensity” (Hollerbach 1994: 63). This would imply that in (3), “quel chemin” expresses that the path trodden by the European Parliament has either been the best possible path, or the worst, depending on context, as opposed to quel intensifying an implicit quality about this path (long, winding, treacherous, etc.). By looking at when quel exclamatives are chosen to translate the degree of modification implied by what [a], this paper hopes to shed some light on this ambiguity.

These semantic considerations have had a direct impact upon which structures are included under the term “exclamative” in this study. Some working in this area have drawn a clear distinction between matrix and embedded constructions (the former illustrated in examples 1, 2 and 5-7, the latter in examples 8-10 and 13), arguing that the two are not so similar that they can be included under the umbrella term exclamative, and so refer to these constructions simply as complement or embedded wh- clauses. According to Rett, the exclamative is encoded in terms of its illocutionary force and as this “is a property of an utterance as a whole, not subcomponents,” the term “exclamative” only applies to matrix clauses (2008: 603). For Heine, Kaltenböck, et al. (2020: 216), the term exclamative refers strictly to a distinct sentence type and they argue that as embedded constructions are subordinate clauses by definition, the use of the term exclamative would be inappropriate. They also highlight the impossibility of subject-auxiliary inversion in embedded constructions as a further reason to distinguish between them. They cite Zevakhina who refers to the constructions as “subordinate wh-clauses with exclamative readings” (2016: 766). Zevakina (2016) and Heine, Kaltenböck, et al.’s (2020) reasoning stems from the hypothesis that the matrix exclamative is derived from the embedded construction diachronically through the process of insubordination (see Evans 2007), which indeed has not been ruled out to date (see Ghesquière and Troughton 2021). Nonetheless, interesting as it is, this particular line of argument, terminology from derivation, is not relevant for the study in hand. This paper follows the view that exclamative clauses are subordinated in much the same way as interrogative clauses (Elliot 1974: 233; Quirk, Greenbaum, et al. 1985: 1055; Huddleston and Pullum 2002: 991; Riegel, Pellat, et al. 2009: 501) and so it is taken as a given that these constructions also constitute exclamatives (see Michaelis and Lambrecht 1996; Michaelis 2001; Zanuttini and Portner 2003; Collins 2005). This decision is grounded in the aforementioned degree modifying properties of these constructions. It is assumed that the extreme degree of scandal implied in (1) is much the same as the extreme degree of disaster in (8), and so constructions like the latter cannot be ignored. Furthermore, the use of the indefinite article with what, a feature unique to the exclamative, is not affected by the embedding of a construction. It is also hoped that, by including these constructions, this paper will contribute to the discussion of their status. Indeed, decisions made by translators when faced with these constructions may be indicative of any differences between independent and embedded constructions (disregarding any potential diachronic links). If a translator opts to translate independent exclamative constructions with corresponding independent exclamatives in the target language, but chooses another construction entirely for embedded constructions, this could suggest that the line taken in this paper needs to be reconsidered. It could also reflect differences in the use and understanding of these constructions in either French or English, hence the utility of bi-directional corpus study.

A study into the constructions used to translate what and quel exclamatives and also into how degree and non-degree nouns and/or modifiers are translated will demonstrate both how similar these constructions are and provide insight into the translation of degree. This will in turn allow for a discussion of the interpretation of embedded exclamatives and the impact corpus-based translation studies can have on the definition of linguistic phenomena. The translation of such an emotive construction is particularly interesting in this instance as it is a question of the translation of parliamentary discourse on an international level. The translation of the power struggles and negotiations which are inseparable from international politics is of course of vital importance to international relations (Bànhegyi 2014). The translation of the emotion-laden exclamative will help show the conventions of political discourse in the respective languages.

The structure of this paper is as follows. The methodology used for this study will be explained in the following section, as will the choice of corpora and search criteria. This is followed by the analysis of the parallel data, that is to say an analysis of how these exclamatives are translated. The constructions used to translate what and quel exclamatives will be quantified and compared with the source exclamatives to draw any possible links. The translation of non-degree or degree nouns and/or their modifiers will then also be analysed to ascertain to what extent degree is retained. The same will then be carried out for the translation of French quel exclamatives before a discussion of these findings and concluding remarks.

2. Methodology

This paper presents an exhaustive study of the English to French and French to English directional sub-corpora of the Europarl-direct corpus (Cartoni and Meyer 2012). Europarl-direct (Cartoni and Meyer 2012), an adaptation of Version 6 (V.6) of the Europarl corpus (Koehn 2005), is made up of speeches given in the European Parliament that have been transcribed, and the translation of these transcriptions. This allows for both a study of what and quel exclamatives in original source data and a study of the construction(s) chosen to translate them. Cartoni and Meyer’s (2012) adaptation was chosen for this study for several reasons. Primarily, the original Europarl corpus proves problematic for this type of translation study as it does not take directionality into account (see Ustaszewski 2019). Version 7, the latest version of the Europarl corpus is available via SketchEngine[7] and does allow for parallel concordance. An initial search on this platform, however, returned no hits for the exclamative what a when the text type was limited to English as the language of the original speaker. This may be due to a large amount of the data remaining untagged for language of speaker. A preliminary search of Cartoni and Meyer’s (2012) directional sub-corpora using AntConc[8] showed that V.6 promised sufficient data and so was chosen for this study.

It is important to note here that it is arguable as to whether transcriptions and their translations constitute oral or written language. Firstly, it is possible that the transcription process allowed for some “cleaning-up” of the original speech, as pauses and redundant repetition do not appear in the data. This may be said to detract from the authenticity of these corpora as spoken data. In addition to this, some politicians are known to read from a script in parliament. Parliamentary tradition suggests this is more likely to be the case for French MEPs (Ilie 2016: 135-136) but, at the time of writing, it is unclear as to whether this is the case for the data in question. In light of this, this data certainly cannot be counted as spontaneous speech. However, it is undeniable that the original speeches were intended to be given orally and to be heard and so likely to make effective use of the highly expressive exclamative.

Another issue that may be raised here is that it is problematic to study the translation of an element that is more frequent in oral language, and even then, rare overall (Siemund 2015). Quantitively speaking, written fiction would presumably be more fertile ground for the examination of the translation of exclamative constructions; however, the lack of large, freely available parallel corpora of this type does not allow this avenue to be explored at present. Furthermore, the Europarl-direct corpus allows for an exploration of how this element is made use of in a specific context, parliamentary discourse in a European context, and thus also translation in a very specific context. Proceedings were translated so as to be accessible to every member state and, as such, it may be presumed that there would be less of a stylistic impact from an individual translator than there may be in other text types, such as literary fiction.

As the corpora in question are not tagged for parts of speech, data extraction involved using the simple word queries what (15,523 hits) and quel (994 hits) in AntConc, which, while time-consuming, ensured complete results. Siemund (2015) observed that while English exclamatives are more frequent in spoken language, they are rather rare overall. This observation initially appears to be sustained by this study. Only 140 exclamatives were found in the 1,410,121-word English-to-French directional sub-corpus and just 102 in the 1,179,530-word French-to-English sub-corpus. While this may seem a small number, Siemund’s (2015) quantitative study of English exclamatives recorded just 703 what exclamatives in the BNC, a one hundred-million-word corpus. Relativised this equates to 9.9 what exclamatives per 100,000 words in the Europarl-direct English to French subcorpus, compared to 0.7 in the BNC. Although it must be acknowledged that Siemund’s (2015) study did not include embedded exclamatives or those without an exclamation mark, these frequencies still suggest that parliamentary discourse is relatively fertile ground for research into the exclamative construction.

These small datasets were further reduced as information on the speaker was also manually verified to make sure all data was spoken by a native speaker. Twelve English what exclamatives had to be excluded from the data. Eleven of these were spoken by non-anglophones and the remaining one had no language or speaker tag and so was impossible to verify. Ten French quel exclamatives also had to be removed from the data, eight for being uttered by non-Francophones, one for which the speaker information was too vague to identify the speaker’s mother tongue, and one which was described as “shouted out,” thus apparently not uttered by the speaker and so very difficult to verify. This resulted in a total of 128 English what exclamatives and 92 French quel exclamatives to be analysed.

The exclamation mark was not used as a criterion for exclamative status or as a search term in either language, as it is not a reliable indicator of clause type but an unreliable indicator of illocutionary force at best (see Trotta 2000: 120). Any clause type can be given exclamatory force by means of an exclamation mark, as illustrated in examples 16-18, and true exclamative clauses do not necessarily end with an exclamation mark. Indeed, in the original English data, only 31% of independent what exclamatives appeared with exclamative punctuation (embedded examples are excluded here for the interference from the matrix clause).

There were a few exceptions to this rule. In (19), the exclamation mark was taken as a reason to include an otherwise ambiguous construction in the data. A translator’s choice to retain or change punctuation is more deliberate however, and as such will be discussed in the results of this study.

3. Exclamatives in Translation: Results and Discussion

This section will discuss how what and quel exclamatives are translated into French and English by first looking at the different syntactic structures used, followed by how degree elements are translated.

3.1 What Exclamatives in Translation

Initially, French quel exclamatives, as illustrated in examples 20-21, appear to be the preferred construction, with 41% (53/128) of what exclamatives translated as such, the highest proportion of any exclamative construction present in the translated data. This would seem to confirm that these two constructions communicate the same thing and are a natural choice for a translator. However, it does still leave 59% that were translated using another construction. As shown in examples 22-24, what exclamatives were also found to be translated into French using other exclamatives: combien (4%), comme (2%), and que exclamatives (2%). 9% were not translated as a syntactic exclamative but used the expression à quel point to express degree, as in example (25). Interestingly, the largest proportion of what exclamatives in these data (albeit by a small margin - 43%, 54/128) were in fact omitted in translation. Omitted in translation is here taken to mean that no exclamative construction or other marker of degree, such as à quel point or a degree adverb (tel, for example), was used, as illustrated in example (26). Only 5 out of these 54 added a compensatory degree element, such as the word éminemment added in the translation in (27).

The large number of omissions could be attributed to different factors. They may be related to the type of data studied. Previous work has shown that other intensifying or emotive features in political discourse are toned down when translated into French. Demissy-Cazailles remarked that the translation of George W. Bush’s very intentional repetition in his speeches was “excessive for the receiver” [abusive pour le récepteur] and that in French “this repeated hammering, which is not standard usage, renders the concept meaningless” [ce martèlement qui n’est pas d’usage vide le concept de son sens] (2007: 145). Schreiber (2015: 711) observed that in a more general sense, French translators tend to omit, or neutralise, strong stylistic features. While these observations are interesting in view of the data at hand, it must be underlined that, unlike the present study, they are based on qualitative research so it is impossible to make any general statements based on these ideas at the present time. The omission of the exclamative element could also be a part of the “normalisation” that underpins the translation process in general. Normalisation is one of the “universals” of translation, features that are said to be common to all translated texts (as opposed to source texts) regardless of language combination (Baker 1993: 243). Baker highlighted the tendency that translators have to “conform to patterns and practices which are typical of the target language, even to the point of exaggerating them” (1996: 176-177). If the French language is less inclined to performative language, at least in the political arena, then omitting them in translation could been seen as normalisation. At this point it is unclear whether these omissions are due to translation or to the nature of French and/or French translators.

Bearing these potential language factors in mind, it is worth looking to see if certain realisations of the what exclamative led translators to use one or the other (figure 1). Indeed, in this data 70% (45/64) of independent (including reduced) what exclamatives were translated as a quel exclamative. 64% (41/64) of what exclamatives embedded in other constructions were omitted. This is statistically significant, (2, N = 128) = 13.7534, p=.000208. It should also be noted that all instances in which a translator used à quel point, the original English was an embedded what exclamative. This primarily suggests that, while attenuation may certainly be behind some omissions, it is actually the translator’s interpretation of an embedded exclamative that leads to the majority of omissions. This in turn suggests that a translator is less inclined to experience the embedded what exclamative as a true exclamative, contrary to the initial line taken in this study. The use of à quel point does, however, suggest that the translator recognises the degree properties implied by the wh- element in these embedded constructions. Logically, it may follow that a translator uses other elements of degree to compensate for not using a wh- construction. In Section 3.3, other degree properties of these embedded constructions will be examined and discussed further. It may be argued, however, that the use of a phrase such as à quel point does not constitute attenuation. Anscombre (2013: 24) actually gives such a construction as that used in the translation in example (25) as an illustration of an “indirect exclamative” and à quel point is listed as a synonym for the exclamative words comme and combien.[9]

Figure 1

Translation of independent and embedded what exclamatives

Translation of independent and embedded what exclamatives

-> See the list of figures

The differences in the translation of full and reduced independent what exclamatives were statistically insignificant ( (2, N = 64) = 4.670, p = .96373), as 65% (15/23) of full and 73% (30/41) of reduced constructions were translated as quel exclamatives. The remaining full independent constructions were slightly more likely to be translated as another exclamative than the reduced constructions as 17% (4/23) of full what exclamatives were translated using other exclamatives (comme, combien and que) and 17% (4/23) were omitted, whereas 2% (1/41) of reduced what exclamatives were translated by another exclamative (que) and 24% (10/41) were omitted in translation.

Looking even more closely at how the independent what exclamatives, full and reduced, were translated using quel did reveal some interesting contrasts (figure 2). Both full and reduced what exclamatives were significantly more likely to be translated using a reduced quel exclamative: 97% (29/30) of reduced what exclamatives and 67% (10/15) of full what exclamatives ( (2, N = 45) = 7.7885, p = .005258). There is no clear indicator in these data as to why a reduced construction would be chosen over a full one. It may be related to the idea that the French language “prefers the level of abstract expression” (Vinay and Darlbelnet 1958/1995: 105). If an exclamative construction can express a feeling or reaction with fewer words, and if the verb is indeed redundant in the communication act, it appears that is the choice a translator into French would make. Theoretically this would also apply to any additional degree phenomena within the exclamative construction, which will be examined in section 3.1.

Figure 2

Full constructions vs. reduced translated using a quel exclamative

Full constructions vs. reduced translated using a quel exclamative

-> See the list of figures

In this data, it appears that a translator faced with an English what exclamative is most likely to either use a quel exclamative or to omit the element entirely, replacing it with a simple noun phrase. This would suggest that the quel exclamative does indeed express the same degree modification and mirativity inherent in the what exclamative but also raises the question of why there are so many omissions. Furthermore, the data suggests that this choice is influenced by whether the exclamative is independent or embedded, either confirming that embedded exclamatives are interpreted differently, or suggesting that embedded quel exclamatives do not function in the same way as embedded what exclamatives. If this is an exclamative characteristic or a particularity of the French language will be discussed later in view of the results of the French-to-English directional data.

3.2 Quel Exclamatives in Translation

Similarly to the English-to-French directional data already discussed but to a greater extent, what exclamatives appear to be the construction of choice for translating quel exclamatives, with 62% (57/92) of them being translated as such, as is the case in example (28). Fewer were omitted, as in example (29), in the French-to-English data, just 16% (15/92), and it must be noted that four of these had an element of degree added, such as the adverb very in example (34), compensating for the removal of the wh-element. Nearly as many, 13% (12/92), were translated using a form of the English how exclamative, as in example (30), 2% (2/92) using degree modifying such, example (31), 2% (2/92) by a rhetorical question, example (32), or 1% (1/92) by an exclamation (of a different sentence type to the exclamative), example (33). It is worth stressing the very small proportions of these last constructions, as they demonstrate that while a range of alternative translations are possible, what exclamatives are the translation of choice by a wide margin.

At this stage it is worth remembering that 75 out of 92 original quel exclamatives are reduced constructions and so it is not possible to make exactly the same contrasts as were made for what exclamatives. It is, however, worth taking a closer look at how these dominant constructions were translated. Indeed, in line with the general trend in the data, 64% (48/75) of reduced quel constructions were translated as a what exclamative. Full, independent quel exclamatives (7/92) were exclusively translated by what exclamatives or omitted, but leaned towards a what exclamative translation with six out of the seven translated in this way.

As for the less common embedded quel exclamatives, seven out of the ten attested to in the data were found translated as a how exclamative (70%). This provides an interesting directional contrast as, while not the apparently immediate counterpart construction, the translation remains an exclamative and thus shows effort to retain this element in translation even in an embedded construction. Furthermore, in only one of these embedded constructions was the exclamative element omitted in translation. This would suggest that it is simply that exclamatives are a more natural part of English than French, at least in this context, rather than the embedded construction being interpreted differently. Vinay and Darbelnet (1958/1995: 231) remark that exclamations are used freely in English, as opposed to rhetorical devices which are more common in French. Nevertheless, the small number of embedded quel exclamatives prohibits any conclusions being drawn at this time.

3.3 Degree Phenomena in Translation

Both sets of directional data were also analysed in terms of their degree content, outside of the wh- element itself. This was primarily to assess whether the degree content of an exclamative has any impact on the construction used to translate it, but also allowed for a more general exploration of the translation of degree. This includes investigating whether a degree noun remains degree in translation and the translation of intensifying modifiers (adjectives in this case). The English to French data is discussed first, followed by the French to English.

English what exclamatives involving an additional element of degree, be that a degree modifier or noun, were translated by a variety of constructions including the quel exclamative and indicated no patterns. On the other hand, albeit a minor category (just 6 out of 128), four out of the six instances where the what exclamative featured an unmodified, non-degree noun (what a year or what a position) were translated using a quel exclamative. This would suggest the perceived importance of maintaining the element of degree in translation, but it is not possible to draw any firm conclusions from such a small number of these instances.

Degree nouns appearing in what exclamatives were largely translated into equivalent degree nouns in French. The very few exceptions observed showed degree elements being downscaled in translation. In example (35), the intensifying adjective absolute is removed in the translation process and, in (36), the negative type specifications explicit in the degree noun indictment are only implicitly present in translation and would have to be interpreted through context. No change was observed in the translation of non-degree nouns, including those with gradable modifiers. This potentially points to the French language as either being more subdued than English or quel exclamatives as more performative, rendering additional degree phenomena redundant. Either the extreme level of hypocrisy in example (35) is wholly expressed by the quel exclamative and thus adding an extra adjective is redundant, or having two intensifying elements, especially when one is performative by nature, is too strong in French and so one has been omitted to sound as natural as possible (see Demissy-Cazailles 2007; Schreiber 2015: 711). This attenuation could also be attributed to the translation process of normalisation.

In the French to English directional data, as was the case in the other direction, the majority of instances of non-degree nouns occurring with quel were found translated as a what exclamative. Out of the 18 quel exclamatives involving an unmodified, non-degree noun, 14 were translated as a what exclamative (78%). A further two were translated as a how exclamative and one by a rhetorical question, leaving only one where the wh-element was omitted. Interestingly, examples (37) and (38), could have been excluded from this study if it were not for the punctuation. In (38) a degree modifier was added in translation to compensate for this fact. Modified non-degree nouns and degree nouns were found translated as a variety of constructions.

Very little variation was again observed in the translation of degree words from French to English. The few exceptions in this direction, however, all added an element of degree, the opposite of the exceptions in the English to French data. Some non-degree nouns became degree in translation, as shown in examples (39-42), and gradable modifiers were added to both degree and non-degree nouns in examples (43-44). This supports the above observations that, at least in this political context, the English language is less subdued than the French and tends to use more emotive language. Furthermore, especially when considered in view of the large number of omissions in the English to French data, it suggests that Schreiber’s (2015: 711) observation, that French translators are likely to attenuate stylistic features, is indeed a tendancy of French translators working from English. It remains unclear whether this can be attributed to the language as a whole, or rather to a quirk of English-speaking politicians and/or the nature of British and French political discourse. More comparative study and perhaps a cross-genre translation study would be necessary to help answer this question.

It was hypothesised earlier that if a translator is less inclined to experience the embedded what exclamative as a true exclamative, but does recognise the degree properties implied by the wh- element, that they would use other elements of degree to compensate for the absence of the wh- element. This theory is not supported by the data in this study. Just six out of the 52 embedded what exclamatives that were translated without quel or another exclamative word added an element of degree.

3.4 Exclamative Punctuation in Translation

While punctuation was not used as a search criterion in this study (see Section 2), a brief investigation was conducted into how it was used in translation. While punctuation may be arbitrary, the choice of whether to keep it or not is more likely to be a conscious choice. As such it could be said to reveal something about the status of exclamatives and the use of performative language features in either language.

From English to French very little was changed in translation. In all instances where a what exclamative was transcribed with an exclamation mark, this punctuation was retained. In 12% (13/107) of instances where it was transcribed without an exclamation mark, one was added in translation. These were mainly independent constructions (12/13) with nothing else to consistently mark why these particular constructions would need to have an exclamation mark added.

There was slightly more variation in punctuation in the French to English directional data. Here there were some instances of exclamative punctuation being removed in translation: 24% (16/66) of quel exclamatives with exclamative punctuation were translated without it. On the other hand, 12% (3/26) of those without had exclamative punctuation added in translation.

Perhaps more noteworthy is the significant difference in the use of this punctuation between languages, 21/128 used exclamation marks in English compared to 66/92 in French, which certainly makes punctuation seem a little less arbitrary than perhaps first thought. When considered alongside the slight changes made in translation, this difference could hint towards two things. It may be that exclamative punctuation is simply used more in French or that exclamative constructions are less emotive in French and so require this added signal.

4. Concluding Remarks and Suggestions for Future Research

The main aim of this paper was to investigate the translation of what and quel exclamative constructions used in the Europarl-direct corpus and to find any patterns explaining these choices. This has led to some interesting observations. The idea that what and quel exclamatives are corresponding constructions is not disputed by the data in this study, given that the most common choice by the translator in either language was either to use a what or quel exclamative or, in the case of what exclamatives in translation, to omit. Interestingly however, nearly all exclamative constructions as defined by reference grammars (both English and French) were also attested in the parallel data, albeit to varying degrees. Only the French comment exclamative did not feature at least once. The large proportion of omissions in the English to French data, combined with the lack of compensatory degree additions, has two potentially co-existing explanations: exclamatives are employed less often in French and so are excluded by a translator to “normalise” the discourse (Baker 1993; 1996), or that French translators tend to soften strong stylistic markers (see Demissy-Cazailles 2007; Schreiber 2015: 711). These explanations are sustained by the data, supported by the fact that a large part of what and some quel exclamatives were omitted in translation, especially since, when taking the additional degree phenomena into account, this attenuation appears to be a directional issue, occurring more in the English to French data. Further study of translation data from different mediums (for example fiction or non-political discourse) is needed to ascertain if this is really a consequence of the context of the translations of this study or a characteristic of the French language more generally.

The very nature of the data suggests that further comparative research into stylistics may be of use. It was not possible to take such factors into consideration in this study, but the amount of omissions and alternative constructions used in translation may be attributed to stylistic differences between individual translators (in addition to differences attributed to the target language). While a particular speaker may use the what exclamative rather over-enthusiastically, a particular translator may equally choose to tone this use down, whether this be the result of personal preference or style, or a professional choice. Unfortunately, there is no record of translator in these corpora so this is impossible to verify.

In terms of degree phenomena and the exclamative, the data studied here indicates two things. The first is that upscaling or downscaling through the removal or addition of degree phenomena (the omission of the exclamative wh- element included) seems to be affected by the direction of translation. The second is that the fewer additional degree words there are in an exclamative construction, the more likely an exclamative what [a] or quel construction translation is. This may be said to be self-evident given that to remove the wh- element in these instances would be to remove all elements of degree and so lose the source text’s intention entirely. The latter is of particular interest as it implies that the exclamative construction itself is less important to a translator if there are other elements of degree present. This infers that it is the degree modification implied by the wh- element that is crucial to exclamative meaning, rather than the notion of mirativity often attached to it (DeLancey 1997; 2001; Krawczak and Glynn 2015; Unger 2019; Ghesquière and Troughton 2021: 146), as otherwise the exclamative element would have to be retained in translation. This is arguably supportive of Krawczak and Glynn’s (2015) interpretation of mirativity as a scalar phenomenon. It may be that the more mirative an exclamative is the more likely this construction would be kept in translation. A study targeting the mirative nature of exclamatives in translation specifically would be necessary to further this discussion.

The secondary aim of this paper was to contribute to the discussion of the definition of the exclamative sentence type itself. As outlined in the introduction to this paper, the categorisation of the embedded what [a] construction as an exclamative is not universally accepted (Rett 2011; Heine, Kaltenböck, et al. 2020). Despite the reasoning given for its inclusion in this study, the manner in which these constructions were translated may be argued to support the opposing view. While the small proportion of embedded quel exclamatives in the data prevent a cross-linguistic conclusion from being drawn on this issue, the translator certainly appears to interpret embedded what constructions in a different manner to their clause initial counterpart, with a significant number of these being omitted in translation. This may be argued to be evidence of their different status. However, this could also be attributed to the ambiguity of the embedded quel exclamative, a factor that would in turn explain the small number present in the source French data. Unlike the what exclamative’s distinctive indefinite article, the quel exclamative has no marker to distinguish it from its interrogative counterpart and so is less effective at communicating degree and/or mirativity than its clause initial counterpart.

While the results given in this study do still leave the reader with a number of questions, it is these questions that demonstrate how corpus-based translation studies open doors to many other disciplines. These results not only provide insights into the translation practices and the use of the exclamative in French and English, but also inevitably cross-linguistic comparison of political discourse and the definition of elements of language in general such as the exclamative and the embedded exclamative.