From French to English and back again

How a grammatical approach to corpus data can be useful to translation studies

By Noëlle Serpollet (Lancaster University)

Abstract

Corpus linguistics is a methodology which is now frequently employed in translation studies. It has become common practice to use bilingual parallel corpora to throw light on particular translation problems and to train translators. The translation problems this paper is focusing on are the translation(s) of the French subjunctive in English, and the translation(s) of the occurrences of mandative constructions in French. In order to carry out this analysis, I will analyse bilingual parallel texts extracted from the INTERSECT corpus (Salkie, 1995), using ParaConc (Barlow, 1995). I will first report on the analysis of two general categories of a grammatically tagged corpus of British English (FLOB). Then I will compare these results with English target texts extracted from the INTERSECT corpus. Finally, I will describe the types of constructions that are translations of mandative use of constructions in French. I will conclude by indicating that the knowledge of recurrent patterns for problematic translations within specific text types provides effective solutions for translators.

Keywords: corpus linguistics, corpus-based translation studies, corpora paralleli, parallel corpora

©inTRAlinea & Noëlle Serpollet (2002).
"From French to English and back again How a grammatical approach to corpus data can be useful to translation studies"
inTRAlinea Special Issue: CULT2K
Edited by: Silvia Bernardini & Federico Zanettin
This article can be freely reproduced under Creative Commons License.
Stable URL: https://www.intralinea.org/specials/article/1681

1. Introduction

This paper presents work in progress within the framework of Corpus Linguistics, using a grammatical approach to corpus data. My current research [1] focuses on the contrastive analysis of the French mandative subjunctive and its translations in British English (henceforth BrE) on the one hand, and of English mandative constructions and their equivalents in French. The analysis uses for that purpose bilingual parallel corpora.

I intend to explore the impact of corpus linguistics on translation studies, applying insights from linguistics to the practice of translation. One possible application of this corpus-based contrastive analysis is to enable translators to improve the final product in the target text by having a better awareness of the translation of a grammatical concept in a specific target language.

Through the study of a problematic area of grammar, i.e. mandative constructions which can lead to difficulties when one tries to translate them, I will examine how the contrastive analysis of two linguistic systems can make a significant and positive contribution to the practice of translation.

In this paper, I will first report on a study in which I systematically investigated and analysed two general categories – Press and Learned Prose of two grammatically tagged corpora of modern British English (LOB & FLOB). Then I will work from French to English, analysing the mandative constructions in the Press category of the bilingual INTERSECT corpus and will compare these results with the findings obtained from extracts of my reference corpus FLOB equivalent in size and categories. Finally, I will go back to French, working this time on the Learned Prose category of INTERSECT. I will discuss the results of this second bilingual study and describe the types of constructions that are translations of a mandative use of should and of the subjunctive. I will conclude by indicating the possible applications of my research.

2. Theoretical background and motivation

2.1. Corpus linguistics and translation studies

      Translation studies is nowadays an expanding and maybe still emerging multidiscipline (or field of study, Baker, 1993: 234) including both pure and applied translation.

      Corpus linguistics is another discipline which has been growing fast in recent years. Following Leech (1992: 106), I would define it as a powerful methodology, a tool, rather than a subject matter, which was traditionally used to examine and verify in ‘real context’ the validity of theoretical linguistic hypotheses. It employs for that purpose a corpus (plural corpora) which is a collection of written texts or spoken material held in computer-readable form and capable of being analysed automatically.

      Bilingual parallel corpora or translation corpora are “original source language-texts in language A and their translated version in language B” (Baker, 1995: 230); and they are now more and more used as practical tools

·      to verify ‘in real context’ the validity of hypotheses expressed by translation theories

·      to throw light on particular translation problems

·      and to train translators by providing databases of translation patterns.

      Therefore, corpus analysis can, not only link translation and linguistics - presenting very interesting research opportunities - but also bridge the gap between different aspects of translation studies, as indicated in figure 1 below.

Figure 1: Corpus studies as a ‘link’ (Adapted from Holmes, 1988:71).

2.2. Motivation

      I started with the contrastive analysis of Le Monde (1992 & 93) translated in The Guardian Weekly. The articles were extracted from the INTERSECT corpus, [2] aligned at the sentence level (Salkie, 1995), tagged with part-of-speech tags for the French part with Cordial 6 Universités and analysed using ParaConc (Barlow, 1995). All these terms will be defined later on.

      The analysis of the French subjunctive in its mandative use showed that there was of course no one-to-one translation between the French and English mandative subjunctive. The different possible translations that I obtained in the specialised Press corpus are listed below:

·      Mandative subjunctive (example 1)

·      Should (2)

·      Infinitive (3)

·      Indicative (4)

·      Modal (5)

·      Past participle (6)

(Example 1)

La formidable menace que présente la prolifération nucléaire et balistique dans le monde exige, en effet, que toutes les précautions soient prises. (Le Monde, 1993)

The daunting threat of nuclear and missile proliferation in fact requires that every possible precaution be taken. (The Guardian Weekly, 1993)


(Example 2)

M. Pinto de Andrate croit “indispensable qu’un autre parti ou une alliance puisse recueillir quelque 30% des suffrages et jouer le rôle d’une minorité de blocage”. (Le Monde, 1992)

He feels it is vital that a third party or alliance should be able to muster about 30 per cent of the votes and act as a blocking minority. (The Guardian Weekly, 1992)

(Example 3)

“Le gouvernement français réclame que le gouvernement américain reconnaisse officiellement la paternité des scientifiques français dnas la mise au point du test du diagnostique du sida” […]. (Le Monde, 1993)

In the light of the US findings the French government now wants the US administration to make a statement “officially acknowledging that the paternity in developing the Aids diagnostic test belongs to French scientists” […].(The Guardian Weekly, 1993)

(Example 4)

Quand les gens viennent me demander mon avis, je me contente de leur dire que je prierai pour que Dieu leur fasse prendre la bonne décision. (Le Monde, 1992)

When people come and ask me what I think, I just tell them I shall pray that God makes them take the right decision. (The Guardian Weekly, 1992)

(Example 5)

[…] et l’on souhaite que le prochain gouvernement de Bangkok mette de l’ordre parmi ses commandants régionaux […]. (Le Monde, 1992)

[…] and the expectation is that the next government in Bangkok will do something about its regional military commanders […]. (The Guardian Weekly, 1992)

(Example 6)

[il] ne veut pas que soient mis en cause ses propres avantages sociaux […]. (Le Monde, 1992)

[he] does not want its own social advantages touched […]. (The Guardian Weekly, 1992)


   
    Under the light of these different translations, I decided to restrict my analysis to both the mandative subjunctive and mandative should in BrE.

      Therefore, the translation problems this paper is focusing on are:

·      not only the translation(s) of the French subjunctive in English

·      but also the translation(s) of the occurrences of mandative subjunctive and should in French.

3. Mandative constructions (non-inflected or morphological subjunctive and periphrastic construction with the modal should)

      The term mandative comes from the Latin mandare, which means to command. Mandative constructions follow, in a that-clause, mandative expressions (verbs, nouns and adjectives that I call triggers) which express a demand, request, intention, proposal, suggestion, recommendation, etc. (see Quirk et al., 1985: 1012-1015).

(Example 7)

We are pleased to announce that the Department will be recommending that X be awarded the degree of Ph.D. [Linguistics Department-Lancaster University]

      The same definition applies for what I also call mandative subjunctive in French. Such a grammatical form is triggered by the expressions mentioned above. Different verb forms can follow mandative expressions:

(Example 8)

She insisted that he leave early.                    [mandative subjunctive]

(Example 9)

Her wish was that he should leave early.            [should + infinitive]

(Example 10)

She was eager that he left early. [3]                        [indicative]

      But in this paper, I will only study two of the three types of constructions (see examples 8, 9 and below) because as Leech (1987: 116) pointed out, in the examples (11) and (13) extracted from my corpora, the should + infinitive construction acts as a subjunctive substitute and is the equivalent of a commanding subjunctive.

(Example 11)

[…] but it is also very important that they should be fair. (LOB Press, B)

(Example 12)

Hence it is important that the project be carried out accurately (FLOB Learned Prose, H)

(Example 13)

The suggestion that Royton should be demolished for the delight of the yuppie mugwumps of Oldham will alarm many Roytoners. (FLOB Press, B)

(Example 14)

[…] although he did suggest that I analyse his Variations for piano in my private studies.(FLOB Learned Prose, J)

(Example 15)

In (2) the Vicious Circle Principle requires that the domain of quantification on the right hand side not include the class of Fs or Gs. (FLOB Learned Prose, J)

      According to Asahara (1994: 2) “the present subjunctive refers to a grammatical form that takes only the base form of the verb regardless of tense contrast, person and number concord”.

      The subjunctive is therefore difficult to identify because it is identical to the base form of the verb; the subjunctive of the verb be is be and the subjunctive of play is play.

      The non-inflected or morphological subjunctive is distinguishable from the indicative only in the following cases:

-      in the 3rd person singular present tense (no –s) (example 8)

-      in past contexts (no sequence of tenses) (example 14)

-      in finite forms of be (base form for all persons and no tense marker) (example 12)

-      in negated clauses (no do-periphrasis and not is placed before the verb) (example 15)

-       

      The verb form which is indistinguishable from the indicative is called the non-distinctive form, as in:

(Example 16)

It might look as if my husband cares about me- insisting that I come along, but I know Gerald better. (FLOB)

      In that case, we can perform a substitution test by putting a third person singular subject in the place of “I”. We obtain: “insisting that he come along” and we can see that we obtain a distinctive subjunctive form.

4. Objectives of the analysis

4.1 Data used and analysed

      I intend to give an account of the actual status of mandative should and of the mandative subjunctive in two texts categories of FLOB (Press and Learned Prose). In order to achieve this objective, I will analyse the evolution of mandative should and of the mandative subjunctive from the 1960s to the 1990s, i.e. I will compare the two corpora LOB and FLOB corresponding respectively to the 2 periods previously mentioned. Then I will analyse the types of constructions that are translations of the French subjunctive in the Press category of INTERSECT and focus on the target texts to compare the findings obtained with the reference corpus FLOB results.

      I will deal with comparable corpora, defined as follows in Baker’s sense (1995: 234): two separate collections of texts [4] in the same language A (BrE), one corpus containing original texts in that language and the second containing translations from a source language B (French) into the language A.

      Finally, I will go back from English to French, analysing the Learned Prose category of INTERSECT in order to see how mandative uses of should and of the subjunctive are translated, and comparing the English part (source texts) with FLOB.

      My aim is to use my parallel corpora as a test bed for translation studies, in the way expressed by McEnery & Oakes (1996: 212): “the corpora can be used to call up sets of words or grammatical features in one language for their examination, and/or for the call up of the foreign language equivalents in the parallel aligned segments”. Or as M. Barlow puts it (1995) “the result of a search can be examined in an attempt to find out how the second language expresses the notion captured by the search term in the first language”.

      I have used ParaConc (Barlow, 1995), a bilingual parallel text concordance program which is used for contrastive corpus-based language research. The drawback of this parallel concordancer is alignment: each sentence has to constitute a separate paragraph, as each unit is delimited by a paragraph return. But it was not a problem for me as I was using INTERSECT which was correctly aligned. Nonetheless some important editing had to be carried out in order to clean the French texts tagged by Cordial 6 Universités. This high-performance software available from Synapse Développement is a grammar and spelling corrector and also a French language tagger/lemmatiser which enabled me to tag the occurrences of the French subjunctive.

<objectives

 

5. Methods and Results

5.1. Mandative should and subjunctive in LOB/FLOB

Table 1. Concordance of should in LOB and FLOB

</p>
<p>
	                                                                                                                                                                                                                                                                                                                                   </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr align="center">
			<td bgcolor="#D9D9D9" colspan="5" valign="top">
				  SHOULD</td>
		</tr>
		<tr>
			<td valign="top" width="196">
				<b>CATEGORIES</b></td>
			<td valign="top" width="97">
				<b>LOB</b></td>
			<td valign="top" width="116">
				  FLOB</td>
			<td valign="top" width="116">
				<b>Difference </b><b> (abs)</b></td>
			<td valign="top" width="115">
				<b>Difference </b><b> (%)</b></td>
		</tr>
		<tr>
			<td valign="top" width="196">
				<b>PRESS</b> (A-B-C)</td>
			<td valign="top" width="97">
				<b>285</b></td>
			<td valign="top" width="116">
				<b>185</b></td>
			<td valign="top" width="116">
				  ‑ 100</td>
			<td valign="top" width="115">
				<b>- 35.1</b></td>
		</tr>
		<tr>
			<td valign="top" width="196">
				<b>FICTION</b> (K-L-M-N-P-R)</td>
			<td valign="top" width="97">
				  214</td>
			<td valign="top" width="116">
				  250</td>
			<td valign="top" width="116">
				  + 36</td>
			<td valign="top" width="115">
				  + 16.8</td>
		</tr>
		<tr>
			<td valign="top" width="196">
				<b>GENERAL PROSE</b> (D-E-F-G)</td>
			<td valign="top" width="97">
				  472</td>
			<td valign="top" width="116">
				  330</td>
			<td valign="top" width="116">
				  ‑ 142</td>
			<td valign="top" width="115">
				  - 30.1</td>
		</tr>
		<tr>
			<td valign="top" width="196">
				<b>LEARNED PROSE</b> (H-J)</td>
			<td valign="top" width="97">
				<b>330</b></td>
			<td valign="top" width="116">
				<b>382</b></td>
			<td valign="top" width="116">
				  + 52</td>
			<td valign="top" width="115">
				<b>+ 15.8</b></td>
		</tr>
		<tr>
			<td valign="top" width="196">
				<b>TOTAL</b></td>
			<td valign="top" width="97">
				<b>1301</b></td>
			<td valign="top" width="116">
				<b>1147</b></td>
			<td valign="top" width="116">
				  ‑ 154</td>
			<td valign="top" width="115">
				<b>- 11.8</b></td>
		</tr>
	</tbody>
</table>
<p>
 
 
  A = reportage, B = editorial, C = reviews; K = general fiction, L = mystery & detective fiction, M = science fiction, N = adventure & western fiction, P = romance & love story, R = humour;  D = religion, E = skills, trades & hobbies, F = popular lore, G = Belles Lettres, bibliography, essays;  H = miscellaneous, mainly government documents, J = learned & scientific writings.
 
        This table shows that the overall number of occurrences of should has decreased between the 1960s and the 1990s. However this trend is not generalised to all text categories: we can note a decrease in the press and general prose categories but an increase in fiction and in learned prose.
 
        The analysis reported below involved developing complex queries (using Xkwic, a software part of the ISM Corpus Workbench, Stuttgart, Christ, 1994) to retrieve only the relevant instances of the modal. I used two totally comparable, grammatically tagged and computerized corpora of British English and therefore, I could run exactly the same retrieving queries in both corpora (LOB & FLOB).
 
Table 2. Concordances in LOB and FLOB [A-B-C] with verbs, nouns and adjectives as triggers

</p>
<p>
	                                                                                                                                                                                                         </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td colspan="4" valign="top" width="284">
				<b><i>Should</i> in LOB (Press)</b></td>
			<td colspan="5" valign="top" width="284">
				<b><i>Should</i> in FLOB (Press)</b></td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  verbs</td>
			<td valign="top" width="71">
				  nouns</td>
			<td valign="top" width="71">
				  adj.</td>
			<td valign="top" width="71">
				  total</td>
			<td valign="top" width="71">
				  verbs</td>
			<td valign="top" width="71">
				  nouns</td>
			<td valign="top" width="71">
				  adj.</td>
			<td colspan="2" valign="top" width="71">
				  total</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  19</td>
			<td valign="top" width="71">
				  9</td>
			<td valign="top" width="71">
				  6</td>
			<td valign="top" width="71">
				<b>34</b></td>
			<td valign="top" width="71">
				  12</td>
			<td valign="top" width="71">
				  7</td>
			<td valign="top" width="71">
				  1</td>
			<td valign="top" width="68">
				<b>20</b></td>
			<td width="3">
				 </td>
		</tr>
	</tbody>
</table>
<p>

  We can see in table 2 that mandative should has decreased in the Press category from the 1960s to the 1990s after the three types of triggers, which is the general tendency of all the uses of should in Press. It has decreased from 34 to 20 occurrences, which means going from 11.9% of the total number of occurrences of should to 10.8%.
 
  Figure 3

(Example 17)
 
  The suggestion that Sadler’s Wells opera should join the National Theatre on the South Bank entirely changes the whole picture. [LOB Press, B]
 
  (Example 18)
 
  All that is known is that Sheffield proposes the funds should be spent in a wide area across Attercliffe [...]. [FLOB Press, A]
 
Table 3. Concordances in LOB and FLOB Press with verbs, nouns and adjectives as triggers

</p>
<p>
	                                                                                                                                                                 </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td rowspan="2" valign="top" width="73">
				 </td>
			<td colspan="4" valign="top" width="255">
				<b>Mandative subjunctives and ND </b><b> in LOB (Press)</b></td>
			<td colspan="4" valign="top" width="262">
				<b>Mandative subjunctives and ND </b><b> in FLOB (Press)</b></td>
		</tr>
		<tr>
			<td valign="top" width="58">
				  verbs</td>
			<td valign="top" width="66">
				  nouns</td>
			<td valign="top" width="66">
				  adject.</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>total</b></td>
			<td valign="top" width="66">
				  verbs</td>
			<td valign="top" width="66">
				  nouns</td>
			<td valign="top" width="66">
				  adject.</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>total</b></td>
		</tr>
		<tr>
			<td valign="top" width="73">
				<b>Subjunc-tive</b></td>
			<td valign="top" width="58">
				  2</td>
			<td valign="top" width="66">
				  2</td>
			<td valign="top" width="66">
				  0</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>4</b></td>
			<td valign="top" width="66">
				  4</td>
			<td valign="top" width="66">
				  1</td>
			<td valign="top" width="66">
				  0</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>5</b></td>
		</tr>
		<tr>
			<td valign="top" width="73">
				<b>Non-distinctive</b></td>
			<td valign="top" width="58">
				  4</td>
			<td valign="top" width="66">
				  0</td>
			<td valign="top" width="66">
				  1</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>5</b></td>
			<td valign="top" width="66">
				  4</td>
			<td valign="top" width="66">
				  0</td>
			<td valign="top" width="66">
				  0</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>4</b></td>
		</tr>
		<tr>
			<td valign="top" width="73">
				<b>total</b></td>
			<td valign="top" width="58">
				<b>6</b></td>
			<td valign="top" width="66">
				<b>2</b></td>
			<td valign="top" width="66">
				<b>1</b></td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>9</b></td>
			<td valign="top" width="66">
				<b>8</b></td>
			<td valign="top" width="66">
				<b>1</b></td>
			<td valign="top" width="66">
				<b>0</b></td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>9</b></td>
		</tr>
	</tbody>
</table>
<p>

This table presents the results of concordances on the base form carried out in the Press category of the two corpora. In that precise case, the total number of mandative subjunctives and non-distinctive forms is identical in the two corpora with 4 subjunctives and 5 non-distinctive forms in LOB and the reverse in FLOB. You can see for yourself the different triggers used and their respective numbers. So it seems that the number of subjunctive forms (including non distinctive) is constant in the Press category.

 

Figure 4.

(Example 19)

Russia insisted that the Western powers take immediate measures […]. [LOB Press, A]

Table 4. Concordances in LOB and FLOB [H-J]

</p>
<p>
	                                               </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td colspan="4" valign="top" width="284">
				<b><i>Should</i> in LOB (Learned Prose)</b></td>
			<td colspan="4" valign="top" width="284">
				<b><i>Should</i> in FLOB (Learned Prose)</b></td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  verbs</td>
			<td valign="top" width="71">
				  nouns</td>
			<td valign="top" width="71">
				  adj.</td>
			<td valign="top" width="71">
				  total</td>
			<td valign="top" width="71">
				  verbs</td>
			<td valign="top" width="71">
				  nouns</td>
			<td valign="top" width="71">
				  adj.</td>
			<td valign="top" width="71">
				  total</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  25</td>
			<td valign="top" width="71">
				  6</td>
			<td valign="top" width="71">
				  15</td>
			<td valign="top" width="71">
				<b>46</b></td>
			<td valign="top" width="71">
				  29</td>
			<td valign="top" width="71">
				  6</td>
			<td valign="top" width="71">
				  1</td>
			<td valign="top" width="71">
				<b>36</b></td>
		</tr>
	</tbody>
</table>
<p>

Table 4 shows that mandative should has decreased in the Learned Prose category, but only after the triggering adjectives, whereas should as a whole had increased in this genre). The construction has decreased, going from 13.9% to now 9.4% of the total number of occurrences of should.

Figure 5.

(Example 20)

[…] we consider it more appropriate that our report should be submitted in the names of the members from outside the Government service who take responsibility for it. [LOB Learned Prose, H]

(Example 21)

[…] but this was removed when the environment lobby persuaded the government that an open barrier with gates should be built instead. [FLOB Press, J]

Table 5. Concordances in LOB and FLOB Learned Prose with verbs, nouns and adjectives

</p>
<p>
	                                                                                               </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td colspan="5" valign="top" width="328">
				<b>Mandative subjunctives and ND </b><b> in LOB (Learned Prose)</b></td>
			<td colspan="4" valign="top" width="262">
				<b>Mandative subjunctives and ND </b><b> in FLOB (Learned Prose)</b></td>
		</tr>
		<tr>
			<td valign="top" width="73">
				 </td>
			<td valign="top" width="58">
				  verbs</td>
			<td valign="top" width="66">
				  nouns</td>
			<td valign="top" width="66">
				  adject.</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>total</b></td>
			<td valign="top" width="66">
				  verbs</td>
			<td valign="top" width="66">
				  nouns</td>
			<td valign="top" width="66">
				  adject.</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>total</b></td>
		</tr>
		<tr>
			<td valign="top" width="73">
				<b>Subjunc-tive</b></td>
			<td valign="top" width="58">
				  1</td>
			<td valign="top" width="66">
				  0</td>
			<td valign="top" width="66">
				  0</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>1</b></td>
			<td valign="top" width="66">
				  9</td>
			<td valign="top" width="66">
				  3</td>
			<td valign="top" width="66">
				  1</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>13</b></td>
		</tr>
		<tr>
			<td valign="top" width="73">
				<b>Non-distinctive</b></td>
			<td valign="top" width="58">
				  4</td>
			<td valign="top" width="66">
				  0</td>
			<td valign="top" width="66">
				  1</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>5</b></td>
			<td valign="top" width="66">
				  9</td>
			<td valign="top" width="66">
				  0</td>
			<td valign="top" width="66">
				  1</td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>10</b></td>
		</tr>
		<tr>
			<td valign="top" width="73">
				<b>total</b></td>
			<td valign="top" width="58">
				<b>5</b></td>
			<td valign="top" width="66">
				<b>0</b></td>
			<td valign="top" width="66">
				<b>1</b></td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>6</b></td>
			<td valign="top" width="66">
				<b>18</b></td>
			<td valign="top" width="66">
				<b>3</b></td>
			<td valign="top" width="66">
				<b>2</b></td>
			<td bgcolor="#DFDFDF" valign="top" width="66">
				<b>23</b></td>
		</tr>
	</tbody>
</table>
<p>

  In the Learned Prose category, we are shown that both the numbers of mandative subjunctives and non-distinctive forms have greatly increased respectively from 1 to 13 and from 5 to 10. This would tend to prove that the subjunctive is not dying in BrE, but on the increase from 1961 to 1991.
 
  (Example 22)
 
  […] usually by recommending that politicians or administrators introduce incentive […]. [FLOB Learned Prose, J]
 
 
 
  Figure 6.
 
        Recent studies using new corpus resources now widely available (Övergaard, 1995; Asahara, 1994; Hundt, 1998) have presented the following findings: there is a remarkable increase of the use of the mandative subjunctive in British English, especially in late 20th-century. Apparently this use of the subjunctive, although not very frequent is far from becoming extinct, whereas the use of the modal should as a periphrastic alternant to the non-inflected subjunctive seems to be decreasing both in British and American English.
 
        Therefore, the trend identified by previous research has been verified in my analysis, i.e., mandative should is decreasing in the two categories analysed and represents about 10% of the uses of the modal; on the contrary, the mandative subjunctive is increasing or if not staying stable.

5.2 Bilingual analysis: From French to English

5.2.1. Analysis of the Press category in INTERSECT

As I have explained in 4.2., I used Cordial 6U to tag the French part of the corpus and ParaConc to carry out parallel concordances and obtain the different translation of the French subjunctive in BrE (see examples 1 to 6).
 
Table 6. Press (NEWS [Le Monde & The Guardian Weekly 1992, 93])

</p>
<p>
	     </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<p>
	 </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td colspan="3" valign="top" width="284">
				<b>French data</b></td>
			<td colspan="3" valign="top" width="284">
				<h4>
					English data</h4>
			</td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  Subjunctive (Total)</td>
			<td colspan="2" valign="top" width="142">
				<b>Mandative subjunctive</b></td>
			<td valign="top" width="142">
				<b>Mandative Subjunctive </b></td>
			<td colspan="2" valign="top" width="142">
				<b>Others </b></td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  157</td>
			<td colspan="2" valign="top" width="142">
				<b>21 </b> (=100%)</td>
			<td valign="top" width="142">
				<b>4 </b> (=19.0%)</td>
			<td colspan="2" valign="top" width="142">
				<b>17 </b> (=81.0%)</td>
		</tr>
		<tr>
			<td rowspan="5" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				  Verbs</td>
			<td valign="top" width="71">
				  15</td>
			<td rowspan="5" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				<b><i>Should</i></b></td>
			<td valign="top" width="71">
				<b>2 </b> (=9.5%)</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  Nouns</td>
			<td valign="top" width="71">
				  4</td>
			<td valign="top" width="71">
				<b>Infinitive</b></td>
			<td valign="top" width="71">
				<b>5</b> (=23.9%)</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  Adjectives</td>
			<td valign="top" width="71">
				  2</td>
			<td valign="top" width="71">
				<b>Indicative</b></td>
			<td valign="top" width="71">
				<b>6 </b> (=28.6%)</td>
		</tr>
		<tr>
			<td colspan="2" rowspan="2" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				<b>Modal auxiliaries</b><b> (will-root should)</b></td>
			<td valign="top" width="71">
				<b>2 </b> (=9.5%)</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				<b>Past participle</b></td>
			<td valign="top" width="71">
				<b>2</b> (=9.5%)</td>
		</tr>
	</tbody>
</table>
<p>

We are shown that 13.4% of the total number of French subjunctives are in fact mandative subjunctives. They are translated for 19% by English mandative subjunctives, and for 81% by other constructions. Amongst these other constructions 9.5% are mandative should, 9.5% other modals, 9.5% past participles, 23.9% are infinitives and 28.6% are indicatives.

5.2.2. Comparison FLOB/INTERSECT

      Now, to be able to compare FLOB Press (original texts) with INTERSECT Press (BrE target texts), which constitutes a comparable corpus, I had to carry out searches on relevant triggers in INTERSECT because this corpus is not grammatically tagged. This enabled me to retrieve both mandative subjunctive and should:

Table 7. Press in (LOB), FLOB and INTERSECT (comparable corpus)

</p>
<p>
	                                                                                                   </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td rowspan="2" valign="top" width="92">
				<b>Mandative constructions</b><b> (BrE)</b></td>
			<td colspan="2" valign="top" width="161">
				  LOB</td>
			<td colspan="2" valign="top" width="169">
				<b>FLOB</b><b> (original texts)</b></td>
			<td colspan="2" valign="top" width="169">
				<b>INTERSECT</b><b> (target texts)</b></td>
		</tr>
		<tr>
			<td valign="top" width="77">
				<b>n</b></td>
			<td valign="top" width="84">
				<b>%</b></td>
			<td valign="top" width="84">
				<b>n</b></td>
			<td valign="top" width="84">
				<b>%</b></td>
			<td valign="top" width="84">
				<b>n</b></td>
			<td valign="top" width="84">
				<b>%</b></td>
		</tr>
		<tr>
			<td valign="top" width="92">
				<b><i>should</i></b></td>
			<td valign="top" width="77">
				  34</td>
			<td valign="top" width="84">
				  79.1</td>
			<td valign="top" width="84">
				<b>20</b></td>
			<td valign="top" width="84">
				<b>69.0</b></td>
			<td valign="top" width="84">
				<b>8</b></td>
			<td valign="top" width="84">
				<b>42.1</b></td>
		</tr>
		<tr>
			<td valign="top" width="92">
				<b>Subjunctive</b></td>
			<td valign="top" width="77">
				  4</td>
			<td valign="top" width="84">
				  11.6</td>
			<td valign="top" width="84">
				<b>5</b></td>
			<td valign="top" width="84">
				<b>17.2</b></td>
			<td valign="top" width="84">
				<b>8</b></td>
			<td valign="top" width="84">
				<b>42.1</b></td>
		</tr>
		<tr>
			<td valign="top" width="92">
				<b>Non-distinctive forms</b></td>
			<td valign="top" width="77">
				  5</td>
			<td valign="top" width="84">
				 </td>
			<td valign="top" width="84">
				<b>4</b></td>
			<td valign="top" width="84">
				<b>13.8</b></td>
			<td valign="top" width="84">
				<b>3</b></td>
			<td valign="top" width="84">
				<b>15.8</b></td>
		</tr>
		<tr>
			<td valign="top" width="92">
				<b>Total</b></td>
			<td valign="top" width="77">
				  43</td>
			<td valign="top" width="84">
				  100</td>
			<td valign="top" width="84">
				  29</td>
			<td valign="top" width="84">
				  100</td>
			<td valign="top" width="84">
				  19</td>
			<td valign="top" width="84">
				  100</td>
		</tr>
	</tbody>
</table>
<p>

  We can notice that there are more mandative should in the original texts than in the translated texts, 69% compared to 42.1%, on the contrary there are more subjunctives in the translated texts with 42.1% compared to 17.2%. The number of non-distinctive forms is almost equal in the two corpora. However, we must be cautious with any conclusions drawn as the numbers are very small. I have also placed LOB in the table above and on the chart below as a reference point but I am only comparing FLOB and INTERSECT here.
 
 

Figure 7.

5.3 …and back again (From English to French)

5.3.1. Analysis of the Learned Prose category in INTERSECT

      I worked from BrE to French: the translations of mandative should, genuine mandative subjunctive and non-distinctive forms, have been retrieved as I already said in 5.2.2. by searching for the trigger verbs, nouns and adjectives in INTERESCT using ParaConc.

Table 8. Learned Prose (MISCE [Esprit, ILO] + SCIENT [Telecom])

</p>
<p>
	                                                                                                                                                                                                                                                                                                                         </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td colspan="3" valign="top" width="284">
				<b>English data</b></td>
			<td colspan="3" valign="top" width="284">
				<b>French data</b></td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  SHOULD (Total)</td>
			<td colspan="2" valign="top" width="142">
				<b>Mandative <i>should</i></b></td>
			<td valign="top" width="142">
				<b>Mandative subjunctive</b></td>
			<td colspan="2" valign="top" width="142">
				<b>Others</b></td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  832</td>
			<td colspan="2" valign="top" width="142">
				<b>114 </b> (=100%)</td>
			<td valign="top" width="142">
				<b>81</b> (=71%)</td>
			<td colspan="2" valign="top" width="142">
				<b>33</b> (=29%)</td>
		</tr>
		<tr>
			<td rowspan="5" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				  Verbs</td>
			<td valign="top" width="71">
				  87</td>
			<td rowspan="5" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				<b>Infinitive</b></td>
			<td valign="top" width="71">
				<b>8</b> (=7%)</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  Nouns</td>
			<td valign="top" width="71">
				  8</td>
			<td valign="top" width="71">
				<b>Indicative</b></td>
			<td valign="top" width="71">
				<b>16</b> (=14%)</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				  Adject.</td>
			<td valign="top" width="71">
				  19</td>
			<td valign="top" width="71">
				<b>Conditio-nal</b></td>
			<td valign="top" width="71">
				<b>4</b> (=3.5%)</td>
		</tr>
		<tr>
			<td colspan="2" rowspan="2" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				<b>Nominali-sation</b></td>
			<td valign="top" width="71">
				<b>2</b> (=1.8%)</td>
		</tr>
		<tr>
			<td valign="top" width="71">
				<b>Different construc-tion</b></td>
			<td valign="top" width="71">
				<b>3</b> (=2.7%)</td>
		</tr>
		<tr>
			<td colspan="3" valign="top" width="284">
				  <b>Subjunctive</b></td>
			<td valign="top" width="142">
				<b>Mandative subjunctive</b></td>
			<td colspan="2" valign="top" width="142">
				<b>Others</b></td>
		</tr>
		<tr>
			<td colspan="3" valign="top" width="284">
				<b>42</b> (=100%)</td>
			<td valign="top" width="142">
				<b>22</b> (=52.4%)</td>
			<td colspan="2" valign="top" width="142">
				<b>20</b> (=47.6%)</td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  Verbs</td>
			<td colspan="2" valign="top" width="142">
				  32</td>
			<td rowspan="4" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				<b>Infinitive</b></td>
			<td valign="top" width="71">
				<b>10</b> (=23.8%)</td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  Nouns</td>
			<td colspan="2" valign="top" width="142">
				  2</td>
			<td valign="top" width="71">
				<b>Indicative</b></td>
			<td valign="top" width="71">
				<b>2</b> (=4.8%)</td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  Adjectives</td>
			<td colspan="2" valign="top" width="142">
				  8</td>
			<td valign="top" width="71">
				<b>Present participle</b></td>
			<td valign="top" width="71">
				<b>1</b> (=2.4%)</td>
		</tr>
		<tr>
			<td colspan="3" valign="top" width="284">
				 </td>
			<td valign="top" width="71">
				<b>Nominali-sation</b></td>
			<td valign="top" width="71">
				<b>7</b> (=16.6%)</td>
		</tr>
		<tr>
			<td colspan="3" valign="top" width="284">
				<b>Non-distinctive form</b></td>
			<td valign="top" width="142">
				<b>Subjunctive</b></td>
			<td colspan="2" valign="top" width="142">
				<b>Others</b></td>
		</tr>
		<tr>
			<td colspan="3" valign="top" width="284">
				  4</td>
			<td valign="top" width="142">
				<b>0</b></td>
			<td colspan="2" valign="top" width="142">
				<b>4</b></td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  Verbs</td>
			<td colspan="2" valign="top" width="142">
				  4</td>
			<td rowspan="3" valign="top" width="142">
				 </td>
			<td valign="top" width="71">
				<b>Indicative</b></td>
			<td valign="top" width="71">
				<b>3</b></td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  Nouns</td>
			<td colspan="2" valign="top" width="142">
				  0</td>
			<td valign="top" width="71">
				<b>Present participle</b></td>
			<td valign="top" width="71">
				<b>1</b></td>
		</tr>
		<tr>
			<td valign="top" width="142">
				  Adjectives</td>
			<td colspan="2" valign="top" width="142">
				  0</td>
			<td colspan="2" valign="top" width="142">
				 </td>
		</tr>
	</tbody>
</table>
<p>

13.7% of the total number of modals are in fact mandative should. They are translated for 71% by French mandative subjunctives, and for only 29% by other constructions. 7% of the translations by other constructions are infinitives, 14% are indicatives, 3.5% are conditionals, nominalisations account for 1.8% and a translation by a totally different construction account for 2.7%.

      Regarding the translation of the mandative subjunctive, it is almost divided half in half between French mandative subjunctive (52.4%) and other constructions (47.6%: infinitives, indicatives, present participle and nominalisations).

      In the 4 cases reported here, the non-distinctive form of the subjunctive is not translated by a subjunctive but by other constructions.

5.3.2. Comparison INTERSECT/FLOB

      This time, I am comparing INTERSECT Learned Prose (BrE originals) to FLOB Learned Prose (original texts) to check any differences due to the data themselves and not to the translation process.

Table 9. Learned Prose in (LOB), FLOB and INTERSECT (BrE original texts)

</p>
<p>
	                                                                                                                                                                   </p>
<table border="1" cellpadding="0" cellspacing="0" width="550">
	<tbody>
		<tr>
			<td rowspan="2" valign="top" width="102">
				<b>Mandative constructions</b><b> (BrE)</b></td>
			<td colspan="2" valign="top" width="151">
				  LOB</td>
			<td colspan="2" valign="top" width="169">
				<b>FLOB</b><b>(original texts)</b></td>
			<td colspan="2" valign="top" width="169">
				<b>INTERSECT</b><b>(target texts)</b></td>
		</tr>
		<tr>
			<td valign="top" width="67">
				<b>n</b></td>
			<td valign="top" width="84">
				<b>%</b></td>
			<td valign="top" width="84">
				<b>n</b></td>
			<td valign="top" width="84">
				<b>%</b></td>
			<td valign="top" width="84">
				<b>n</b></td>
			<td valign="top" width="84">
				<b>%</b></td>
		</tr>
		<tr>
			<td valign="top" width="102">
				<b><i>should</i></b></td>
			<td valign="top" width="67">
				  46</td>
			<td valign="top" width="84">
				  80.7</td>
			<td valign="top" width="84">
				<b>36</b></td>
			<td valign="top" width="84">
				<b>66.7</b></td>
			<td valign="top" width="84">
				<b>114</b></td>
			<td valign="top" width="84">
				<b>71.3</b></td>
		</tr>
		<tr>
			<td valign="top" width="102">
				<b>Subjunctive</b></td>
			<td valign="top" width="67">
				  1</td>
			<td valign="top" width="84">
				  1.8</td>
			<td valign="top" width="84">
				<b>13</b></td>
			<td valign="top" width="84">
				<b>24.0</b></td>
			<td valign="top" width="84">
				<b>42</b></td>
			<td valign="top" width="84">
				<b>26.2</b></td>
		</tr>
		<tr>
			<td valign="top" width="102">
				<b>Non-distinctive forms</b></td>
			<td valign="top" width="67">
				  10</td>
			<td valign="top" width="84">
				  17.5</td>
			<td valign="top" width="84">
				<b>5</b></td>
			<td valign="top" width="84">
				<b>9.3</b></td>
			<td valign="top" width="84">
				<b>4</b></td>
			<td valign="top" width="84">
				<b>2.5</b></td>
		</tr>
		<tr>
			<td valign="top" width="102">
				<b>Total</b></td>
			<td valign="top" width="67">
				  57</td>
			<td valign="top" width="84">
				  100</td>
			<td valign="top" width="84">
				  54</td>
			<td valign="top" width="84">
				  100</td>
			<td valign="top" width="84">
				  160</td>
			<td valign="top" width="84">
				  100</td>
		</tr>
	</tbody>
</table>
<p>

Here we are dealing with a specialized corpus of government documents and scientific writings which constitute a very particular genre. We can notice

the presence of a very important and almost equal use of mandative should in both corpora (66.7 and 71.3%) and an almost equal use of the subjunctive (around 25%). The number of non-distinctive forms varies between 9.3 and 2.5%. Therefore, we can say that the two corpora belonging to the same genre produce fairly similar results regarding the mandative constructions.

Figure 8.

(Example 23)

This is why it is proposed that a more detailed programme of work necessary for the day-to-day implementation of the programme be established […] (Esprit).

C’est pourquoi il est proposé qu’un programme de travail plus détaillé nécessaire pour la réalisation du programme au jour le jour soit établi […] (Esprit).

(Example 24)

[…] it can decide that a specific day should be commemorated at the national level […]. (International Labour Organisation)

[…] pour décider qu’une journée sera consacrée à célébrer un événement ou une personne […]. (ILO)

(Example 25)

However, it is preferable that these high-speed channels should, as far as possible, be placed […]. (International Telecommunication Union)

Toutefois, il est préférable que les voies à grande rapidité de modulation soient dans la mesure du possible, établies […]. (ITU)

6. Conclusion and applications

      I hope to have presented (although on a small scale) an interesting approach to grammar, using corpus data and to have provided translators with a possible grammatical approach to the English and French languages and also to have shown that the knowledge and awareness of recurrent patterns for problematic translations of the modal should, of the subjunctive in English and of the French subjunctive within two specific text types can provide effective solutions for translators.

      On that aspect I quote Wolfgang Teubert (1996:241) who says that “total bi-directional correspondences are extremely rare phenomena” and that therefore we “have to search for second-best matches”, selecting one of several alternatives. In order to do that he suggests that the translators consult a large corpus. My aim is to eventually derive and develop, from the corpora analyzed, a database of translated segments aligned with original texts, i.e. a direct translation database of grammatical categories in French and their equivalents in English and vice versa.

      This is still work in progress and I have presented a pilot analysis with some first results that can be also useful to study the use and evolution of mandative should and the use of the French subjunctive.

References

Algeo, J. (1992). “British and American mandative constructions”. In C. Blank eds.,  Language and Civilisation: A Concerted Profusion of Essays and Studies in Honour of Otto Hietsch, Vol. 2, Frankfurt-on-Main: Lang, 599-617.

Asahara, K. (1994). “English Present Subjunctive in Subordinate That-Clauses”. Kasumigaoka Review, 1-30.

Baker, M. (1993). “Corpus linguistic and translation studies, implications and applications”. In Baker M. / Gill Francis / E. Tognini-Bonelli eds., Text and technology: in honour of John Sinclair,  John Benjamins, 233-250.

Baker, M. (1995). “Corpora in translation studies: an overview and some suggestions for future research”. Target 7, 2: 223-243.

Barlow, M. (1995). “ParaConc: a Concordancer for parallel texts”, Computer and Text, 10: 14-16, Oxford: OPU.

Christ, O. (1994). “A modular and flexible architecture for an integrated corpus query system”. COMPLEX’94, Budapest.

Holmes, J. S. (1988). Translated! Papers on Literary Translation and Translation Studies, Amsterdam: Rodopi.

Hundt, M. (1997). “It is important that this study (should) be based on the analysis of parallel corpora: On the use of mandative subjunctive in four major varieties of English”. In H. Lindquist, et al., eds. The major Varieties of English. Papers from MAVEN 97, Växjö University, 159-175.

Leech, G. N. (1971), (1987, 2° Ed.). Meaning and the English Verb. London: Longman.

Leech, G. N. (1992). “Corpora and theories of linguistic performance”. In J. Svartvik, ed. Directions in Corpus Linguistics: proceedings of Nobel Symposium 82, Berlin & New York: Mouton, 105-22.

McEnery, T. & M. Oakes (1996). “Sentence and word alignement in the CRATER Project”. In J. Thomas / M. Short, eds., Using Corpora for Language Research, London & New York: Longman, 211-231.

Övergaard, G. (1995). The Mandative Subjunctive in American and British English in the 20th Century. Stockholm: Almqvist & Wiksell International, Acta Universitatis Upsaliensis, Studia Anglistica Upsaliensia, Vol. 94.

Quirk, R., S. Greenbaum,  G.N. Leech & J. Svartvik (1985). A Comprehensive Grammar of the English Language. London: Longman.

Salkie, R. (1995). “INTERSECT: a parallel corpus project at Brighton University”, Computer and Texts, 9, 4-5, OPU.

Salkie, R. (1998). “Unlocking the power of SMEMUC”. In Botley S.P., A.M. McEnery & A. Wilson, eds. Multilingual Corpora in Teaching and Research, Amsterdam: Rodopi, 148-156.

Teubert, W. (1996).“Comparable or parallel corpora?”. International jounral of lexicography, 9/3: 238-264.

Notes

[1] Acknowledgement: the research reported here was supported by an award from the Economic and Social Research Council (UK).

[2] I am grateful to R. Salkie for providing this corpus to me.

[3] The examples 8, 9 and 10 are extracted from Algeo (1992: 599).

[4] The corpora have the same length and cover the same genre(s).

About the author(s)

Noëlle Serpollet si è laureata presso l'Università di Poitiers (Francia) e sta svolgendo un dottorato di ricerca presso la Lancaster University (GB) sotto la supervisione del Professor Geoffrey Leech con una ricerca dal titolo “A Corpus-based Approach to Modality and the Subjunctive in English and in French”. I suoi interessi riguardano la linguistica dei corpora, l'analisi contrastiva, la teoria della traduzione, la teoria francese dell'enunciazione e predizione e la storia della lingua inglese.

Email: [please login or register to view author's email address]

©inTRAlinea & Noëlle Serpollet (2002).
"From French to English and back again How a grammatical approach to corpus data can be useful to translation studies"
inTRAlinea Special Issue: CULT2K
Edited by: Silvia Bernardini & Federico Zanettin
This article can be freely reproduced under Creative Commons License.
Stable URL: https://www.intralinea.org/specials/article/1681

Go to top of page