Probably certain

(1)

Author: Bianca Seydel Supervisor: Magnus Levin Examiner: Fredrik Heinat Term: Spring term 2020 Subject: English

Probably certain

Translating hedges in academic research articles

from Swedish to English

(2)

Abstract

While it certainly can be argued that translation is a quite demanding discipline in general, some areas within this field are, naturally, more challenging than others. One of these is hedging, which serves a broad variety of purposes both with regard to the author and to the intended readership, and hence must be translated accordingly. This paper investigates hedges in scientific research articles, the types and frequency of hedges in the analysed Swedish source text (ST) compared to the English target text (TT), and the methods used for translating these hedges and their distribution by means of a short study conducted on two Swedish runology articles.

The study’s quantitative analysis shows that the Swedish ST has a clear preference for adverbial hedges, and to an extent, also for modal verb hedges, whereas the English TT – while yielding an even higher preference for adverbials and also for lexical verbs – uses modal verbs much less frequently. It becomes evident that adverbials may feature so strongly in translations because they are easier to recall and to use than more complex structures, especially for L2 speakers. This practice does, however, result in a somewhat less flexible translation.

The by far most frequently used translation strategy is faithful translation, particularly for content-oriented hedges. However, a fair number of adaptations (both in modal strength and word class change) and omissions – mostly of modal verbs – as well as numerous additions occurred, initiated by influential factors such as cultural differences regarding natural sounding text, L2 speaker perception of equivalence and/or lack of suitable linguistic equivalents.

Thereby, the English translation showed a tendency toward adapting weaker modals compared to their Swedish ST equivalents, confirming the greater reader-orientation of English research articles.

Key words

adaptation, addition, cultural filter, epistemic modality, equivalence, faithful translation, hedges, hedging, L2 learner, omission, scientific research articles, translation

(3)

1 Introduction

With English being the world’s current Lingua Franca (Mauranen 2018, 2015), most scientific research articles and academic publications are usually either written directly in English, or translated to English in a later phase (House 2013). The two research articles that serve as a basis for this translation study belong to the field of runology. As runology is a mixed discipline where both “hard” evidence (say, a 1000 year old rune stone) can be found, as well as where more or less well founded hypotheses are being made (regarding how the runes should actually be interpreted), it is not surprising that a large number of hedges could be found in these articles (see Vázquez & Giner 2008).

The phenomenon called hedging in English, gardering in Swedish and Heckenausdruck in German comes in many different forms. The most easily recognisable hedges are modal verbs such as may/might or adverbs such as perhaps or possibly, but there are many more. To give an example from the analysed data:

(1a) Av själva inskriften återstod dock inte mer än åtta runor (...), varav de flesta tycks ha varit mer eller mindre oläsliga, men där åtminstone × ra (...) kunde urskiljas. [p. 180 ST 1]

Of the runic inscription itself, however, no more than eight runes remained (…), most of which seemed to have been more or less illegible, but where at least × ra (…) were identifiable.

This sentence contains three consecutive hedges, most of which, seemed, and more or less. The same sentence translated with fewer hedges might read something like this, or any combination thereof:

(1b) Of the runic inscription itself, however, nothing more but eight runes remained at the tail end of the rune animal,

(…) most of which were more or less illegible.

(…) most of which seemed illegible.

(…) which were more or less illegible.

And here the same example without any hedges at all:

(1c) (…) which were illegible.

These versions are remarkably different from each other. While the hedged alternatives present the author’s presumptions very cautiously, as a (likely) possibility rather than a certain fact, the unhedged version is much blunter and more assertive, presenting the statement as an undeniable fact.

(5)

Despite their great variability, the various hedges have one thing in common: they cause a special challenge for the translator, who has the task to translate the research article in question to English (or any other language, for that matter). Therefore, this paper investigates the types of hedges, the strategies for translating them from Swedish to English, and their quantitative distribution in two academic articles on rune stones by Magnus Källström. With this, an attempt is made to fill the prevailing gap for this practical approach and, particularly, for this language pair. In the next section, the aim of this study and the underlying research questions will be presented in more detail, as well as the analysed source material and the intended readership.

Lastly, the translation and analysis methods will be described.

1.1 Aim

The purpose of this study is to analyse, quantitatively and qualitatively, the types and frequency of hedges occurring in the Swedish source text (ST) compared to the English target text (TT), the methods used for translating these hedges and the distribution of these methods. To achieve this, the following research questions will be answered:

1. What different types of hedges are used in the present Swedish runology articles, and which methods/strategies are applied for translating them into English?

2. Quantitatively, what types of hedges are used most frequently in the Swedish original text, and how do those numbers differ from the English translation, if at all?

3. What degree of modality changes can be observed in the translation of modal verbs from Swedish to English, if any?

4. What are possible reasons for adapting/modifying hedges (and, contrariwise, for not adapting them) – particularly in Swedish-to-English translation?

With these questions to guide the study, what results can be expected? As a working hypothesis, it can be assumed that the English translation might contain more hedges than the original Swedish articles, as the Anglo-American rhetorical tradition appears to have a greater need to mark the uncertainty of given propositions by hedging the claims compared to other languages (Hinkel 2005:48). Since hedges have a great impact on the way a statement is presented and received (as will be seen in 2.2), translating them appropriately is very important. For this reason, the underlying motivation for this work is for other translators to gain greater insight into the way hedges function and how to deal with them practically.

(6)

1.2 Material & target audience

Now that the aim of the investigation is established, the next section focuses on the material on which the analysis is based. The first two sections describe the two Swedish texts that have been translated to English, and are followed by some remarks on the intended readership of the respective texts.

1.2.1 Translated source text 1

The first article is ”En återuppstånden runsten från Källa ödekyrka på Öland (Öl 57)”, written by Magnus Källström and published this year (2020) in the journal Futhark. Futhark is the International Journal of Runic Studies based at the universities of Oslo and Uppsala (see www.futharc-journal.com). Its articles are published in English, German and Scandinavian languages, the latter languages including an abstract and/or summary in English, and is mainly addressed to scholars. The author Magnus Källström is a Swedish runologist, archaeologist and antiquarian, who wrote this article as a research project for The Swedish National Heritage Board. The article itself explores a new interpretation of the rune stone Öl 57 and re- combination of different rune stone fragments, which is of great interest not only to Swedish but also international runologists, archaeologists and historians. For this reason, a translation to English would certainly be an asset. One of the article’s most notable stylistic features includes the frequent application of hedges (such as omkring, förmodligen or möjligen), which will be explored.

1.2.2 Translated source text 2

The second article, ”Gamla och nya vägar till Sveriges runinskrifter. En kritisk översikt av källpublikationerna för svenska runtexter”, was also written by Magnus Källström and published in Historisk tidskrift in 2005. Historisk tidskrift, the “Historic Journal”, is a Swedish research journal in the fields of history and economic history (see www.historisktidskrift.se).

Most of its articles are written in Swedish (or other Scandinavian languages), containing an English summary, but only a few are entirely translated to English – thus, a translation such as the present one would surely serve as a valuable addition. The journal is aimed at scholars and students, and the article at hand explores and critically comments on various sources that contain descriptions of rune stones and other runic material from different parts of Sweden.

Like the previous article, it contains a great variety of hedged expressions that have presented a challenge during the translation process.

(7)

1.2.3 Intended readership

The intended audience for the Swedish original articles appears to be mainly scholars, researchers, and perhaps students. In other words, the readership consists of experts in the field of runology, or in any case readers who already have previous knowledge thereof – they are, to an extent, familiar with reading rune stones and/or about rune stones. This can be deduced from the fact that the articles were published in academic journals aimed at precisely such an

“educated” audience on the one hand, and on the other hand from instances in the text where runic inscriptions such as . . . lit · rita · stai . . . — . . sial. . . (p. 181 ST 1) are left uncommented. For a layman, these letters may not make much sense, but a runologist will recognise the remnants of a standard formula found a hundredfold on rune stones¹ (see e.g.

Jansson 1987). The presence of an expert audience meant for the translation that not many adaptations and/or additional explanations needed to be made, save a few additional translations to English of inscriptions presented in Old Norse, which were legible for the original Swedish audience.

1.3 Method

In the following two sections, the applied general translation method will be presented, followed by the methods used for conducting the analysis.

1.3.1 Translation method

Generally speaking, research articles should be translated in a target-oriented way to make the text sound as natural as possible in the target language (Pisanski Peterlin 2010:173, see Nida 2002:6), so that the new readers may perceive them as if they were originals. However, in order to achieve this, certain adaptations may need to be made, as the conventions and preferences for what type and what number of hedges sound natural to a target audience may differ from culture to culture, between different academic fields, different languages with different statuses, and from translator to translator (as will be discussed section 2.5).

With the previous section in mind, it should now be evident that the intended expert readership facilitated a more straight-forward translation, as almost no additional descriptions or explanations were needed. Since the two articles contain highly specialised material about Swedish rune stones, are rooted in Swedish history and frequently mention Swedish landscapes, the professional, scholarly readers expect to be confronted with foreign names of e.g. Swedish

1 “[…] lät · rista · ste[n dessa] […] [Gud hjälpe hans] själ…”, in English “[Name of a person] had this stone carved [for another person] – [may God help his] soul” (see e.g. Jansson 1987)

(8)

landscapes, the locations of the rune stones. I have also chosen to retain the original Swedish titles of books, essays and institutions in the English target text in addition to providing an English translation², as it may make it easier for researchers and scholars to recognise them and find the originals.

In order to translate the two articles, parallel texts (such as Gräslund 2003 and Jansson 1987) describing other rune stones have been consulted for determining the proper terminology and getting a better understanding for the writing in this field. In addition to that, both English and Swedish corpora have been used (mainly from Språkbanken, BNC, occasionally COCA and Linguee) for determining the frequency of certain terms and finding suitable collocations.

Apart from that, extensive use has been made of various dictionaries such as the OED and the Nationalencyklopedin, and also of rune stone descriptions that could be found on the website of the Swedish National Heritage Board.

1.3.2 Analysis method

First of all, in order to analyse the two articles, the types of hedging used in the Swedish STs were categorised, and the quantitative distribution of the different types of hedges in the Swedish original texts was analysed. These results were then compared to a similar overview for the English TT equivalents. For those instances where several consecutive hedges (sometimes called compound hedges, Bogusława 2018:229), were found in the texts (i.e.

example (1)) all hedges were counted separately, so as to give a more accurate representation of the absolute and relative number of hedges a) in this discipline and b) in the source versus target language. Equally, the occurrence of the same lexical verbs in different tenses were counted separately (i.e. seems, seemed), and also the reoccurrences of the exact same hedge were treated as different instances.

In a second step, the various translational strategies applied for these hedges have been explored by means of a quantitative overview. This data was then compared to the findings of previous studies.

In the subsequent sections (3.3–3.5), each translation method was investigated individually, while qualitatively analysing examples from each category that were of particular interest. Additionally, the degree of adaptation of modal verbs were identified, together with the reasons and the need for adaptation/modification of hedges (where relevant) when translating from Swedish to English.

2 (at least when they are mentioned for the first time)

(9)

Some factors that may have influenced the results are, firstly, the fact that this is only a comparatively small study, conducted on a relatively short source text (4692 words in total), which was then be compared to the results of bigger corpus studies, and may thus not be entirely representative. However, it will hopefully serve to shed some light on the translation mechanisms of hedges from Swedish to English. Moreover, in the course of this work it has become obvious that the classification of hedges can be rather difficult and the boundaries blurry. As hedges express uncertainty in various degrees of vagueness (van Deemter 2010:121), their categorisation appears, at times, equally vague, especially since, as van Deemter puts it, all linguists have their own individual theories about how to categorise hedges (and how to translate them, of course), which makes them “difficult to count” (2010:121). Decisions concerning the categorisation of hedges have been made individually from case to case. The next section will lay the necessary theoretical groundwork for the conducted analysis.

2 Theoretical Background

Before being able to analyse the hedges from the translation at hand, a closer look must be taken at what hedges are and how they can be defined, which will be done briefly in section 2.1. To understand why their appropriate translation is important, the purpose of hedging in research articles will also be discussed in the subsequent sections, in addition to what different types of hedges there are, and what different translation strategies can be applied to them.

2.1 Hedges and modals

The terms “hedge” and “hedging” are quite fitting metaphors for their intended purpose.³ The OED describes the act of “hedging” as “surround[ing something] with a hedge or fence as a boundary, or for purposes of defence.” In a similar manner, an academic “hedge” can serve to shield a given author from e.g. criticism and aids him/her when facing questions that cannot be answered (see Markkanen & Schröder 1997).

As far as studies on hedges, their definition and function are concerned, there are a number of influential and therefore much quoted, but relatively old sources such as Lakoff (1973), Hyland (1988) or Markkanen & Schröder (1997). A distinction is made by the latter between two kinds of hedges, “approximators” (such as sort of), conveying “fuzziness” within the propositional content itself, and “shields” (such as (I) think), modifying “the relationship

3 Alternative terminology for hedges found in literature includes downtoners, weakeners, compromisers et cetera (Hyland 1998:9).

(10)

between the propositional content and the speaker” (1997:5). More recent studies mainly by Kranich define hedges and epistemic modality more in terms of “the degree of commitment by the speaker to what he/she is saying” (Kranich & Gast 2015:3). This is a more pragmatic definition I would like to apply for this study, as it is more concerned with the practical strategies of translating hedges on the whole, and less with the theoretical re-definition of them (although evidently, distinctions will be made where useful).

That being said, it should be pointed out that there is some confusion with regard to the concept of modality and that of hedges. Vázquez & Giner summarise the problem thus as either modality being the wider concept to include hedges or vice versa, making hedging the umbrella term and epistemic modality a part of it (2008:173). The main reason for the confusion lies in the fact that modal verbs usually are polysemic, carrying both a deontic root meaning, and an epistemic meaning (ibid.). It is the latter which is important in the context of hedging. Hyland describes modals (with the help of Halliday) as “the area of meaning that lies between yes and no” – including “either yes or no” and “both yes and no” (1998:3). Within that area, modals are gradable: on a scale from forbidden to permitted to obligatory in the root meaning, and from impossible to possible to certain or necessary in the epistemic meaning (Vázquez & Giner 2008:173). It is that gradeability, together with differing cultural preferences and writing conventions in different academic fields, that can make modals tricky to translate. Along that gradable scale, hedges and modals differ in their degrees of explicitness with regard to their indication of a source of evidence (Kranich & Gast 2015:4). Generally speaking, constructions containing e.g. lexical verbs, adjectives or adverbs are more explicit than modal auxiliaries (Kranich & Gast 2015:4). The reason for this is that the latter are more or less restricted to said binary division between the possible and the necessary (Kranich & Gast 2015:5).

One way to determine whether one modal verb constitutes a weaker hedge than another is described in Kranich (2011:87–8). If a proposition is of low modal strength, that is to say, if it represents only one of many possible options with a 50% or lower chance of actually happening, it is negatable without resulting in a contradiction. This can be determined by a negation test. Here is one example for low modal strength: “You might have to take these drugs all your life” vs. “you might not have to take them all your life” (Kranich 2011:88). These two statements do not contradict each other – as opposed to the following example of high modal strength: “You probably have to take these drugs all your life” vs. “you probably don’t have to take them all your life.” (ibid.). Another common structure of similar high strength would be

“X can happen” vs. “X cannot happen”. In these cases, the two respective statements contradict and exclude each other. Further examples for this will be provided in the analysis.

(11)

To summarise, hedges can more practically be defined as an author’s conviction of or commitment towards a proposed content (Kranich & Gast 2015; Markkanen & Schröder 1997).

Modal verb hedges (i.e. modals with epistemic meaning) are gradable from high to low modal strength, and appropriate translations can be determined by means of strength- and negatability- tests (Vázquez & Giner 2008; Kranich 2011). They are generally less explicit than lexical constructions.

2.2 Hedges in research articles

2.2.1 The function of hedges in research articles

Hedges in research articles serve a variety of different purposes, which can roughly be divided into semantic and pragmatic functions. From a semantic point of view, hedges in their core meaning distinguish between factual and non-factual assertions; they mark the author’s uncertainty concerning the truth of a proposition p (Kranich 2011:77), indicating that s/he does not show full commitment to it. From a pragmatic – and for my purpose more relevant – standpoint, hedges are a communicative strategy that can have both content- and reader- oriented functions in social discourse (which will often overlap; see e.g. Kranich & Gast 2015).

In the case of content-orientation, hedges are often used due to lack of absolute evidence – the author is “not sure” (Kranich & Gast 2015:3, Vázquez & Giner 2008:181). By weakening the strength of the proposition – as opposed to “boosters” (e.g. clearly, obviously) which enhance the strength of a statement – they allow the author to present even uncertain information that otherwise could not be shared (Vázquez & Giner 2008:186), or claims that may only last temporarily (Hyland 1998:6). In the case of reader- (or addressee-) orientation, on the other hand, hedging serves to present challenging propositions and debatable information in a way that makes it easier for the reader to accept them, essentially by leaving room for different opinions, allowing them to disagree (Kranich 2011:84, Vázquez & Giner 2008:172– 4, Hyland 1998:7). As a consequence, more hedges mean more interaction between author and readership (see also Hyland 1998:10, Pisanski Peterlin 2010:174–5, Vázquez & Giner 2008:173–4). Low-intensity hedges have even been found to enhance the reader’s memorisation of the text (Hyland 1998:9), which may serve to make the results of a research paper more memorable. Other usages of hedges involve the notion of politeness when making suggestions (e.g. perhaps, Aijmer 2018:145) without threatening the face of the reader (Aijmer 2018:152) – even when the presented data is certain – and without appearing too bold or even boastful (Kranich & Gast 2015:3).

(12)

As far as the readership is concerned, reports have shown that texts containing more hedges feel more engaging and motivating to the readers (Kranich 2016:134). The author may be perceived as humbler and more honest (Vázquez & Giner 2008:172). The downside is, however, that an author using many hedges may be perceived as less confident and less believable (Kranich 2016:134, Hyland 1988:2). The distance that a hedge is putting between the author’s actual opinion and the words he says, may shelter him or her from “strong criticism” (Vázquez & Giner 2008:177–8) but may also be perceived as a strategy to avoid responsibility vis-à-vis the reader (see Vázquez & Giner 2008:175). In more drastic words, hedges are a form of manipulation, aiming to influence the reader to reach the conclusions preferred by the author, by stressing those factors that influence said conclusion on the one hand, and by modulating those assertions deemed risky on the other hand (Vázquez & Giner 2008:174). However, in a reverse scenario where no hedges were being used and a claim turns out to be false, strong disapproval might arise within the author’s research community, and false information may further lead to wrong expectations and/or fears among the public readership (Varttala 2001:261).

To sum up, hedges are highly multifunctional; from a pragmatic point of view, they can further be divided into content- and reader-orientation. The former is linked to a lack of absolute evidence (Vázquez & Giner 2008:181), while the latter aims to present debatable information in a more acceptable manner (e.g. Kranich 2011:84). Given the fact that both hedging and not hedging comes with a set of benefits and drawbacks, it becomes evident why their translation is so difficult – yet so important.

2.2.2 Hedging in popular science vs. science proper

It is commonly thought that the guidelines for academic discourse encourage a neutral, objective, generally impersonal and very direct style of writing, without first person pronouns, metaphors – or hedges, for that matter (Vázquez & Giner 2008:176–7). It has even been said that “if a person writing for publication has to hedge, it is questionable whether he or she is truly ready to publish” (Varttala 2001:54). However, this does not actually hold true, and research articles are not as “objective” as one might think. Academic discourse is a world full of uncertainties and non-final conclusions (Vázquez & Giner 2008:172) where hedges are used frequently. In her study from 2011, Kranich discusses the number of hedges used in popular scientific articles as opposed to their equivalents in science proper, with the result that hedges seem to be more common in science proper, in part for reasons such as showing respect for the views of the author’s professional science colleagues (Kranich 2011:83, Hyland 1998:6). Ergo,

(13)

hedges are less common in popular scientific articles, which are less “peer-addressed” than science proper (Kranich 2011:83–4), albeit still necessary for presenting controversial ideas.

As the two translated articles at hand are aimed at an expert audience and appear to be more

“science proper” than simpler, popular scientific articles (as described in section 1.2), a high number of hedges is to be expected.

2.2.3 Hedging in “hard” vs. “soft” science

In another study conducted by Vázquez & Giner (2008), the epistemic modality markers and hedges in research articles of different fields have been researched. They come to the conclusion that what they call “soft sciences” such as Marketing, which are not mathematically verifiable, strongly influenced by external factors and prone to contradictory opinions and the interpretation of the audience, use of a large number of hedges.⁴ On the other hand, “hard sciences” such as Mechanical Engineering will use fewer hedges.⁵ There are also some in- between cases, such as Biology, with a mixture of certain and uncertain research grounds⁶.

In the case of the runology articles translated and analysed in this study, it could be argued that runology is exactly such a hybrid between “hard” and “soft” sciences, with a tendency towards the soft end of the spectrum. On the one hand, “hard” evidence can be found, such as rune stones or other objects with runes carved on them (of which there are, after all, more than 2500 in Sweden). On the other hand, such evidence is often at least a thousand years old and their meanings can, in many cases, only be guessed (and are, therefore, abstract data and “soft”

speculation). As a consequence, research conducted in this field and the results that are presented are often of uncertain nature. It is not surprising, then, that in the articles analysed in the present study, a high number of hedges and epistemic modality markers were found.

2.3 Types of hedges

There are several different types of hedges that can be classified as follows, after Hyland (1998) and Pisanski Peterlin (2010). First of all, hedges can be divided into lexical and non-lexical hedges. Non-lexical hedges are usually phrases that refer to either a lack of knowledge (e.g. one cannot exclude; we don’t know whether; etc.), to a limit of an applied model or theory (e.g. in spite of its shortcomings; in the context of the proposed model; etc.) or for instance to the – less than ideal – experimental conditions (e.g. under these conditions, it is difficult to conclude;

4 (2.5 hedges per 100 words, 2008:180)

5 (0.7 h. per 100 words, ibid.)

6 (1.1 h. per 100 words, ibid.)

(14)

etc.) (Hyland 1988:141–8). These hedges draw attention to the shortcomings of an experiment, the conditions under which a research was conducted, the difficulties encountered, and the possibility of alternative interpretations (Hyland 1998:4). They are a less obvious hedging type and may include e.g. if-clauses (If…, then…), rhetorical questions, or contrastive markers (although; ibid.). This category remains somewhat blurry, however, and it seems that some overlapping with lexical hedges may occur. Therefore, and in order to limit the scope of this study, I have, not included non-lexical hedges in my quantitative analysis, but I will provide some of the more obvious examples in section 4 for the sake of completeness, since non-lexical hedges do occur in the Swedish ST.Lexical hedges, on the other hand, contain certain verbs (Ia+Ib), adjectives (II), adverbs (III) and nouns (IV) that will be described in more detail below.

The examples are collected from Hyland (1998:103–41) and Pisanski Peterlin (2010:175).

(Ia) The first big category consists of lexical verbs. Lexical verb hedges can contain performative/speculative verbs (e.g. suggest, indicate, assume), conclusive/deductive verbs (e.g. conclude, deduce), quotative verbs (e.g. X showed, Y claimed), perceptive/sensory verbs (e.g. X appears to/seems to indicate) and narrators (We sought to investigate/attempted to gain insight,…). It should also be noted that passive verb constructions can also be counted as hedging.

(Ib) The other form of verbal hedging, which is, perhaps, the most easily recognisable form, consist of modal verbs in epistemic use (such as may/might/could etc).

(II) The second large category consists of modal adverbs functioning as hedges, (such as apparently, probably, around, almost), including downtoners (e.g. quite, fairly).

(III) Thirdly, some hedgy adjectives can also be found, for example (un/likely, approximate, most, somewhat, potential).

(IV) Lastly, modal nouns (albeit rare) exist as well (e.g. possibility).

In short, hedging items can be classified as non-lexical or lexical hedges, the latter of which encompasses a wide range of modal and lexical verbs, adverbs, adjectives and nouns. The distinctions are, however, not always clear-cut. The next section will look at how those different forms of hedging can be translated.

(15)

2.4 Translating hedges

Research articles are, in principle, either translated by the author himself (who, of course, knows exactly what his intentions are), or, as is more often the case, by a professional translator with knowledge about the target culture (Pisanski Peterlin 2010:172). In that case, it is the task of the translator to choose a meaning for the often vague and multifunctional hedged expressions (see Aijmer 2018:142), including the corresponding degree of modality in the target culture. It should be noted, however, that even experienced translators and their work occasionally suffer from the effects of interference (or “shining through”), which influences their choice of hedges (Kranich 2011:77). Moreover, especially novice translators may be reluctant to make changes to the ST (Pisanski Peterlin 2011:173, Kranich 2011:79), and hence tend to translate more directly.

That being said, when translating hedges, one of five general strategies can be chosen by the translator, regardless of source and target language (adapted after Kranich 2011 and others, with examples from my translation⁷):

(I) Faithful translation.⁸ This is the most commonly applied translational strategy (Kranich 2016:147). It is, according to Kranich, often chosen when dealing with content- oriented uses of hedges, as they concern the author’s actual lack of knowledge (2011:93). This method includes straight-forward translation (i.e. a translation that is formally close to the ST, Kranich 2011:79) or, in case of modals, also a translation with a modal of the same strength (Kranich 2016:174), for example:

(2) (...) det kunde röra sig om delar av (...) [p. 187, ST 1]

(…) they could be parts of (…)

(II) Adaptation. the application of a “cultural filter” and finding ”functional equivalence”. This

“cultural filtering” entails the adjustment of a TT to the new target audience’s “communicative conventions” (House 2013:290), in order to make the text seem like the original (ibid.). This method is preferably used when translating reader-oriented hedges according to Kranich (2011:93). In the case of modals, this includes applying a higher or lower modal strength than the original (Kranich 2016:147, type 3a in my analysis):

(3) (...) stenytan som kan utgöra resterna av (...) [p. 188 ST 1]

(…) stone surface that might be the remnants of (…)

7 More detailed examples and explanations will be provided in the analysis.

8 Translating ”faithfully” or ”true to the ST” might imply that a deviating translation (e.g. adaptation) is a ”bad”

translation, which is, of course, not the case.

(16)

Although this method is used by translators with increasing tendency, resulting in generally more assertive texts, “shining through” still occurs (Kranich 2016:146, 153). Other forms of adaptation include the change of word class (type 3b in my analysis). This category is called transposition after Vinay & Darbelnet and involves a change of word class without a change in meaning (Vinay & Darbelnet 1995:96):

(4) (...) en (…) inskrift som ska ha funnits i (...) [p. 180 ST 1]

(…) an (...) inscription which had supposedly existed at (…)

(III) Omission (zero-translation). Not all hedges are translated, as they can sometimes be left out in the TT. This can occur for reasons of target culture and/or text type conventions (i.e. to sound natural), lack of linguistic equivalents, or insufficient space (see e.g. Aijmer & Altenberg 2019). As a rule, English translators are more likely to omit one of the many Swedish pragmatic particles than vice versa (Aijmer & Altenberg 2019:22). For example:

(5) (...) inskrifter som visserligen behandlas i Sveriges runinskrifter, men (...) [p. 305 ST 2]

(…) inscriptions that have Ø been incorporated in Sveriges runinskrifter (…), but (…)

(IV) Addition. A non-existent hedge in the ST is added in the TT, often for cultural or other reasons (see section 3.5.2).

(6) (...) denna är inte lika lättillgängliga (...) [p. 304 ST 2]

(…) is not quite as easily accessible (….)

(V) Other. A translational choice called “creative” by Kranich, where the hedge, i.e. the

“element of uncertainty” is represented by a creative lexical choice (often longer, sometimes conditional constructions) and where the negation test is not applicable, making the expression impossible to analyse after this model (2016:150–2; 2011:88).

(7) Den källpublikation som främst brukar utnyttjas är (...) [p. 300 ST 2]

One of the most frequently used source publications is (…)

The common practice of adapting/modifying, adding or omitting hedges can have many reasons, three of which are discussed below: cultural preferences, the status of the source and target language, and the perception of translational equivalents by L2 translators.

2.5 Modifying hedges

2.5.1 Reason 1: Cultural preferences

The probably most obvious reason for modifying hedges between the ST and TT is the fact that different cultures have different “communicative preferences”, creating the need for applying

“cultural filters” when translating to the target culture (Kranich 2011:77, Aijmer 2002:60, see

(17)

above). Some of these cultural differences discussed in previous studies will be summarised below.

More recent studies on hedging in translation include Kranich (2016 & 2011), particularly on hedging in English to German translation and the comparison between popular science and science proper, Aijmer (2018) on ska/ll in Swedish to English translations and the English modals should/must/have to, Vázquez & Giner (2008) on epistemic modality markers and hedges in research articles of different fields, Pisanski Peterlin (2010) on hedging in Slovene to English translation and, when widening the scope even further, also Bogusława (2018) on hedging in scientific articles in German–Polish translation. However, there appears to be a need for more research on the translation methods and frequencies of hedges, especially between Swedish to English (and reverse) translation.

To begin with Kranich (2011), she discovers the following cultural preferences: English has, overall, a more indirect, addressee-oriented approach when writing scientific research articles⁹ (77). Therefore, more hedges in general (see also Kranich 2016:142), and more specifically weaker hedges (i.e. hedges with a lower degree of probability) are used (94).

Moreover, English original research articles show a (rising) preference for grammaticalized modal verbs¹⁰ and, to a lesser extent, lexical verbs with epistemic meaning (Kranich 2016:137

& 2011:94, Pisanski Peterlin 2010:179). As a consequence, English articles are experienced as more dialogic and interactional with regard to the reader (Kranich 2011:79). Over time, a shift towards more cultural adaptation compared with translations from previous decades was noticeable, and consequentially fewer instances of the “shining through”-effect have been noted (Kranich 2011:92). Concerning the application of cultural filters itself, a shift toward higher assertiveness, fewer modals, and increasing modal strength (Kranich 2011:91) was noticeable.

This is interesting to bear in mind while translating.

A second study conducted by Pisanski Peterlin (2010) on Slovene to English translation of 90 research articles finds that while English original texts usually prefer verbal hedges, translated texts often prefer adverbial hedges – possibly because they are easier to remember as

“standard translations”. She also finds the abovementioned lower flexibility in translations, but

9 here in direct comparison to German

10 One of the suspected reasons for this rise lies in the English modals’ higher degree of grammaticalization as opposed to e.g. German (Kranich 2016:137), resulting in a “loss of meaning” and the “development of new functions” (Aijmer & Altenberg 2019:32). This rise in the use of epistemic modals is particularly interesting with regard to the general, overall decline of modals in English (Kranich 2016:138, see also Leech 2013 and Biber &

Conrad 2009), which consequently is assumed to concern primarily the deontic use modals (Kranich 2016:138).

(18)

explains it with the fact that in her case, the Slovene as source language has less variation to offer compared to English, which lead to limitations in the translation.¹¹ Pisanski Peterlin also discovers that in one third of the cases where a “may” or “might” was used as a hedge in the English TT, there was no hedge at all in the Slovene ST (187). Still, the translated text only contains half as many hedges as a comparable English original text (188).

As a third, if perhaps somewhat less weighty perspective, Kjellström (2019) conducted a similar investigation to mine in her master thesis on the different hedging categories, their occurrences and translation in a non-fictional text translated from English to Swedish. She found that her English ST contained mostly adverbs and adverbials, followed by modal verbs.

Adjectives and lexical verbs fall further behind.

Lastly, Aijmer’s study from 2018 focuses primarily on the translation of the Swedish modal ska/ll to English and discovers that it is only rarely translated with its cognate form

“shall” (144). She then explores the connotations and effects of the English translational alternatives must/have to/ought to/should, depending on the temporal, deontic or epistemic meaning of ska/ll.

As a conclusion, English research articles appear to be more reader-oriented, containing more and generally weaker hedges than for example comparable German articles (Kranich 2011). Original English articles show a preference for verbal hedges, while comparable translated articles (to English) prefer adverbial hedges, are less flexible and contain less hedges than original ones (Pisanski Peterlin 2010). A general shift towards less interference, overall fewer modals (yet a simultaneous rise of epistemic modals compared to other ways of expressing uncertainty) and higher modal strength has also been observed (Kranich 2016;

2011). It will be interesting to see if similar results for the current Swedish-to-English translation can be found, or if they differ significantly.

2.5.2 Reason 2: Language status

With English being the world’s current Lingua Franca (Mauranen 2015), many scientific research publications and academic articles are usually either written in English to begin with or translated to English from a different original language (House 2013). As a consequence, English can be said to be a language with a comparatively high status (see e.g. Kranich

11 Another argument could be made that adverbials offer greater flexibility with regard to word order and sentence structure. However, I am primarily referring to short adverbials which become “standard translations” in the translator’s memory (e.g. always using “probably” as a go-to translation, instead of variations with modal verbs or modal-adverb combinations)

(19)

2016:153¹²). Moreover, since many readers with other native languages have (almost) daily exposure to it and are, therefore, used to the language and culture, these readers are much more inclined to accept features from this higher-prestige source-language when translated into their native language (see Kranich 2011:79). It also means that translators generally, upon translating a somewhat higher-prestige language such as English to a somewhat lower-prestige – at least in direct comparison – language like Swedish, are more reluctant to make changes to the original textual features; they are less likely to meddle with the author’s intended commitment to the truth value of his statements by i.e. inserting and/or deleting hedges (Kranich 2011:79, Pisanski Peterlin 2010:173). When translating the other way around, however, from Swedish to English for instance, more adaptations are made. In addition to that, unconscious transfers (i.e. interference) may happen more frequently, especially by translators strongly influenced by the English language and culture (Kranich 2016:153). That being said, it is notable that, while English is clearly leading UNESCO’s ranking of the world’s most translated original languages in the Index Translationum, and is also in the top 5 target languages for translation, Swedish is also in the top 10 – in both categories. Therefore, it would seem that, surprisingly enough, the Swedish language is not that peripheral after all.

2.5.3 Reason 3: L2 speaker perception of equivalence

As briefly mentioned earlier, translating reader-oriented hedges can require adaptation by finding a suitable equivalent. One difficulty with this, however, lies in the different perceptions of bilingual speakers of what those equivalent expressions might be (Kranich 2011:94).

First of all, second-language (L2) users have a tendency towards using considerably fewer hedges than L1 users, and towards simplifying the ones they do use (Hinkel 2005:47, Pisanski Peterlin 2010:179,184) – perhaps out of uncertainty and less familiarity with target culture conventions. In other words, L2 writers express stronger commitment than their L1 correspondents (ibid.). In earlier research, the advice had even been given to students to preferably use no hedges at all (Hyland 1998:8). This means that there is less flexibility in the translations than in the ST, and that more variations of the same or a similar pattern are used instead of drawing on the target language’s great variety: there is a tendency to both underrepresent unique linguistic features that are specific to a target language, and to at the same time overrepresent those features that have “straightforward translation equivalents”¹³ (Pisanski Peterlin 2010:180). In addition to the latter, translators and L2 speakers will often

12 on language status/prestige, see also Even-Zohar’s polysystem theory (e.g. 2005), Munday (2016:172ff.)

13 i.e. a translation that is formally close to the ST (Kranich 2011:79, see above)

(20)

have a so-called “press-the-button”, standard translation (Pisanski Peterlin 2010:184). This could explain why there are often no notable differences in the translation of e.g. downtoners¹⁴ (Pisanski Peterlin 2010:184) as they are rather straight-forward.

As far as modal auxiliaries are concerned, non-native speakers of English show difficulties using modal verbs appropriately (Aijmer 2002:56). This is problematic, as they play a central role in English (57). Especially Swedish (and German) L2 writers show a general overuse of modal auxiliaries and adverbials (55). In addition, they suffer from “register- interference”, that is they tend to transfer modal expressions and patterns which are influenced by speech (58) to their writing (55), and choose collocations which would not be used by native speakers (e.g. can probably; 70). Other potential reasons for the overuse of modals include a transfer from L1 (60), the novice writer effect and the accompanying tendency of learners to use redundant expressions (71–2). House (2013:288) emphasises, however, that non-native writers are very much capable and entitled to work in languages other than their mother tongue, as long as they “adhere to the expectations of native-English readers, rather than those of readers using English as a lingua franca” (ibid.).

To summarise, hedges can be modified for various reasons, most notably due to different cultural conventions concerning natural-sounding text, which suggest more addressee-oriented hedges in English (e.g. Kranich 2011). Other factors, such as the present runology articles representing proper scientific research articles in a hybrid-to-soft field of study, indicate the presence of a high number of hedges. Other potentially influential factors include the status of the ST and TT language, an L2 speaker’s perception of translational equivalence, avoiding repetition and personal preference. Interference plays a role as well, for both experienced and novice translators, with the latter showing a tendency to translate more directly (Pisanski Peterlin 2011:173, Kranich 2011:79).

3 Analysis

The conducted empirical and descriptive analysis of the translation of hedges from the Swedish ST to the English TT will begin with a statistic analysis and comparison of the types of hedges occurring in the ST and TT, followed by a quantitative overview and discussion of the applied translation strategies. The remaining sections focus more closely on each individual translation strategy, while analysing representative examples in more detail.

14 such as SWE “ganska”, EN “quite”, “fairly”

(21)

3.1 Hedging types in the Swedish ST and English TT

Initially, the types of hedges that occur in the Swedish ST will be analysed quantitatively and be compared to the types in the English TT. In the ST, the following types of lexical hedges occur: modal verbs (e.g. kan...ha tillhört), lexical verbs (e.g. verkar inte finnas), adjectives (e.g.

utan större problem), adverbs (e.g. förmodligen), nouns (e.g. möjlighet), other (e.g. bland annat). The number of occurrences of these hedging types has been analysed and compared to the number of occurrences of the TT hedges, which are of the same types, but, as will be noticeable, of different distribution. The ST also contains a number of non-lexical hedges.¹⁵ To give an example from the translation:

(8) Av de uppgifter som hittills har varit kända verkar det alltså som om (...) [p. 183 ST 1]

Of the data known up until now it seems thus as if (…)

Non-lexical hedges also include hypothetical IF-then-clauses, such as for instance:

(9) (...) utrymmesmässigt verkar detta namn kunna passa in om man tänker sig att en avslutande l-runa har haft sin plats strax före bandknuten, (...) [p. 189 ST 1]

(...) in terms of space, this name seems to fit if one is to assume that a final l-rune has had its place right before the union knot.

The following table 1 shows an overview of the distribution of hedging types in the two texts.¹⁶ Table 1: Absolute and relative occurrence of hedges in the Swedish ST and the English TT

CATE- GORY

Absolute number of hedges in SWE ST

Hedging categories in

% (of 201 ST hedges)

Hedges per 100 words in ST

Absolute number of hedges in EN TT

Hedging categories in

% (of 204 TT hedges)

Hedges per 100 words in TT Modal verb

hedges

49 24.4% 1.1 36 17.6% 0.6

Lexical verb hedges

27 13.4% 0.6 33 16.2% 0.6

Adjective hedges

17 8.5% 0.4 19 9.3% 0.3

Adverbial hedges

77 38.3% 1.6 80 39.2% 1.4

Noun hedges

24 11.9% 0.5 22 10.8% 0.4

15 As previously mentioned, this type of longer constructions is usually content-oriented, as the hedges are used due to lack of absolute evidence (Vázquez & Giner 2008:181) – as can be the case when trying to interpret a 1000- year old, damaged rune stone. See section 2.3 for a more detailed definition.

16 The following statistics do not include non-lexical hedges, as discussed earlier.

(22)

Other hedges

7 3.5% 0.1 14 6.9% 0.3

TOTAL 201* 100% 4.3 204* 100% 3.6

*excluding non-lexical hedges

To begin with, it is noticeable that the English TT contains slightly more hedges than the ST.

This seems to confirm the previously formulated hypothesis predicting more hedges in English than in Swedish in accordance with Pisanski Peterlin (2010) and Hinkel (2005). However, due to the increase in the overall word count, the relative number of hedges in the English TT appears to be lower (3.6 per 100 words) than in the Swedish ST (4.3 per 100 words). I shall return to this phenomenon later in the analysis. That the absolute number of hedges only rises slightly and not as much as expected may perhaps be explained by the fact that source language interference always influences the translation to an extent, and that especially novice translators tend to show greater reluctance to apply changes to the author’s original text (Pisanski Peterlin 2011:173, Kranich 2011:79). This occurs also in spite of the translation being carried out from a language of a somewhat lower status to one of higher status, where more changes could be expected to happen (see Kranich 2011:79). As a consequence, the translated English text contains, as a rule, fewer hedges than a comparable original English research article usually would (see Pisanski Peterlin 2010:188). That being said, both the ST and TT are so rich in hedges¹⁷, that their relative number amounts to more than three (TT) and almost four (ST) times the expected number of hedges for “hybrid” sciences after Vázquez & Giner (2008:180), and also still noticeably more than their average for “soft” sciences. This seems to indicate that runology is more of a speculative science than initially thought.

Secondly, when taking a closer look at the individual hedging categories, it is noticeable that the Swedish ST strongly prefers adverbial hedges, which constitute the largest category of all occurring hedges, followed by modal verbs which make up about one quarter. These categories are, in turn, followed by lexical verbs, nouns and adjectives (in that order), which share a comparable, roughly one-tenth distribution. While having a somewhat higher percentage than the ST, the English TT also shows a strong preference for adverbial hedges.

Unlike the Swedish ST, however, modal verbs occur far less often, leaving more room for lexical verbs. Like in the ST, nouns come in fourth place in the TT, followed by adjective

17 as they are ”science proper” texts, particularly many hedges were expected (Kranich 2011:83, Hyland 1998:6)

(23)

hedges which occur more frequently than in the ST. Finally, the number of other hedges almost doubled compared to the ST, indicating translational creativity.

To give these results some perspective, we shall compare them to previous studies.

Material in these studies has shown that English original research articles have a preference for modals and lexical verbs (Kranich 2011:94, Pisanski Peterlin 2010:179), which make the text, together with overall weaker hedges, more addressee-oriented and interactional (ibid.).

However, these studies have also found out that, contrary to the original texts’ preferences, translated English articles prefer adverbials instead, and contain a lower number of modal verbs (ibid.) – a phenomenon that also seems to prove itself true in this study. In addition to pointing out the significance of adverbial hedges in translations, Pisanski Peterlin (2010) also notes a consequential lower flexibility in translations. A potential reason for this might be the fact that adverbials are easier to use and to translate (recall what Pisanski Peterlin (2010:184) named

“press-the-button”, standard translations) than, for instance, more complex structures containing a combination of modal and lexical verbs.¹⁸ It is also possible that the choice of hedging type in some cases is a simple matter of a translator’s personal preference, also supported by the higher number of creative translational choices in the ST, which seems to contradict the hypothesis of decreasing flexibility somewhat.

To summarise, the Swedish ST shows a strong preference for adverbial hedges, and to an extent, for modal verb hedges. The English TT shows a slightly higher preference for adverbial hedges and, to a lesser extent, for lexical verbs, while displaying a noticeably less frequent use of modal verbs. These results confirm the conclusions drawn by Kranich (2011) and Pisanski Peterlin (2010), indicating also that adverbials might feature so strongly in translations because they are easier and faster to use than e.g. complex modal verb structures, but perhaps also somewhat lacking in flexibility. The English TT also contains an overall slightly higher absolute number of hedges; yet the rise is not as high as expected, and was perhaps dimmed by common influences such as source text interference and translator novicey (Pisanski Peterlin 2011:173, Kranich 2011:79).

18 This being said, whether modal verb structures are more difficult to use than complex adverbial structures also depends on the language pair in question and on the translator’s native language. The complexity of modals lies in their gradeability; as for adverbials, the greater flexibility in word order can pose difficulties.

(24)

3.2 Quantitative overview of translation strategies

Moving on to how the hedges described above have been translated, an overview will be presented in table 2 over the frequency of the translation strategies (as categorised in the theory section III) used for translating the hedges from the ST to the TT.

Table 2: Translation strategies used for hedging Occurrence of each translation

strategy (absolute)

% of all ST hedges (201*)

% per hedging category

(1) Omission ⁷ ^3.5% 8.2 % of all modal verbs

3.7 % of all lexical verbs 0.0 % of all adjectives 2.6 % of all adverbials 0.0 % of all nouns 0.0 % of all other (2) Faithful

translation

148 73.6% 51.0 % of all modal verbs

85.2 % of all lexical verbs 82.3 % of all adjectives 80.5 % of all adverbials 79.2 % of all nouns 71.4 % of all other

(3) Adaptation ³² ^15.9% 34.7 % of all modal verbs

7.4 % of all lexical verbs 5.9 % of all adjectives 6.5 % of all adverbials 20.8 % of all nouns 28.6 % of all other

(4) Addition ¹⁷ ^8.6% ^-

(5) Other/

creative translations

14 7.0% 6.1 % of all modal verbs

3.7 % of all lexical verbs 11.8 % of all adjectives 10.4 % of all adverbials 0.0 % of all nouns 0.0 % of all other TOTAL 204* (hedges in TT)

(194 translated hedges +17 additions -7 omissions)

108.6%

(100% translated or omitted ST hedges + 8.6% additional hedges)

-

*excluding non-lexical hedges

When looking at the distribution, it becomes obvious that a large majority of ST hedges have been translated by means of a faithful translation (e.g. förmodligen  probably), in all categories except modals. This number is approximately comparable to Kjellström’s findings

(25)

(2019). However, in the current translation, adaptations have often been made to the ST hedges in form of changes in modality strength (e.g. kan utgöra  might be) or word class (e.g. ska ha funnits  had supposedly existed), mostly to modal verbs, to which we shall return in section 3.4.1. Kjellström (2019) reports an even higher number of modifications (26%). In the present study, ST hedges have been omitted (again mostly modals) in only half the number of cases Kjellström (2019) noted. However, more than double that number have been added to the TT (which is more than Kjellström’s (2019) addition of 11 hedges). Finally, a comparable number of creative translations have also been applied, mostly to adjectives and adverbs. Examples for omissions, additions and creative translations are provided in section 3.5.

The higher number of additional hedges as opposed to omitted hedges results in a total of 108.6% hedges compared to the ST. However, all the English hedges together make up 3.6%

of all English TT words, as opposed to 4.3% of all Swedish ST words. These two results seem to contradict each other – a higher absolute number of English TT hedges and, at the same time, a lower relative percentage as opposed to the Swedish TT. This apparent contradiction can, however, be explained by a higher word count of the TT.¹⁹

To sum up, most hedges have been translated faithfully, yet a fair number of adaptations have been made to the ST hedges. In multiple instances, hedges have been translated more freely or have been added, and a few have been omitted. Most notable deviations can be observed with modals, which have been adapted and omitted more than any other hedging category. In the following section 3.3, a closer look will be taken at faithfully translated hedging examples.

3.3 Faithful translation

Faithfully translated hedges are those that have been translated more or less straightforwardly, that is to say without adjustments in modal strength and/or changes in word class. According to Kranich, this method is especially preferable when dealing with content-oriented hedges (2011:93). The first example is that of a lexical verb–adjective hedge:

19 5627 words in the TT as opposed to 4692 in the ST. This can partially be explained by the fact that, unlike Swedish, English does not use long compound words and therefore has to split one ST word into multiple TT words, generating a higher word count and thus an apparent lower percentage of hedges when viewed relatively to the overall text. Other reasons include the translational choice of retaining the numerous original names and titles of e.g. books and offering an additional translated title. While it is certainly true that future commissions may contain a word/page limit, none was given at present, and I have, therefore, chosen to translate the text freely and analyse what “happened” in the translation, rather than to adapt the translation to expected results and/or restrictions.

(26)

(10) (...) men det verkar ändå troligt att det handlar om en direkt kopia av [Rhezelius]

teckning. [p. 181 ST 1]

(…) but it seems likely nonetheless that is that is a direct copy of [Rhezelius’] drawing.

In this example, the author Källström does not know with absolute certainty whether Liljegren has directly copied Rhezelius’ work; despite undeniable similarities between their works, he finds that there are also deviations. The hedges have, therefore, been inserted as shields due to lack of certainty with regard to the truth of the propositional content (see Kranich 2011:77, Hyland 1988:5). They allow Källström to present this uncertain piece of information, that otherwise could not be shared (Vázquez & Giner 2008:186, Hyland 1998:6). For reasons of content-orientation, and since English offers suitable straight-forward translation equivalents, the hedges have been translated faithfully. One translation alternative might have been a rendition as “is/seems possible”; however, this choice would have decreased the strength of the hedge slightly, as “troligt” is defined as “sannolik” in Lexin, whereas “possible” shows less commitment and assertiveness than its direct translation “likely” (or probably). The following example further demonstrates content-oriented hedges, this time in form of an adverbial hedge:

(11) De två fragmenten av Öl 57 som Wallman avbildade förvarades vid hans besök förmodligen inne i kyrkan. [p. 186 ST 1]

The two fragments of Öl 57 that Wallman depicted were probably stored inside the church during his visit.

Even here, Källström cannot be entirely sure whether what he assumes actually holds true, as he was not present at the time when Wallman visited Källa Old Church. The fact that such a large majority of adverbial hedges have been translated according to this method, might also be connected to the previously mentioned “press-the-button” translation equivalents for L2 speakers (Pisanski Peterlin 2010:184). It becomes visible here that, as the Swedish adverbs in question often have suitable, straightforward translation equivalents in English, they do not necessarily need to be adapted. Of all adverbial hedges in the Swedish ST at hand, 80.5% have been translated faithfully, making the translation somewhat less flexible but, therefore, easier (see Pisanski Peterlin 2010).

As a conclusion, content-oriented hedges and hedges of the “shield” type have, as a general rule, been translated faithfully, especially when English offered straight-forward, suitable equivalents for ST adverbs in particular. This seems to correspond to Pisanski Peterlin’s (2010) findings regarding more readily accessible translational equivalents.

(27)

3.4 Adaptation

In this section, the second most frequent translation strategy of “adaptation” will be focused on in greater detail, including a more in-depth look at the changes made in modality strength (category 3a) and word classes (category 3b), and the application of cultural filters (Kranich 2011:77). The applied translation methods are thus distributed as presented in an overview in table 3, and in further detail down below.

Table 3: Translation strategy (3) “adaptation” used for hedging

Table 3 shows that all observed changes of modal strength involve decreasing strength of the TT modals compared to their ST equivalents. Kranich’s (2011:91) findings on (GER-EN) translational tendencies towards generally fewer modals and specifically higher modal strength, i.e. higher assertiveness over the last decades come to mind. The more common adaptation strategy is, however, a change in word class from the ST to TT referent. We shall look at both phenomena consecutively.

3.4.1 Change in modality strength

To determine whether one modal verb is a weaker hedge than another, we must return to Kranich’s negatability test (2011:88), according to which a proposition of low modal strength is negatable without resulting in a contradiction. To give one translation example: “there is an edge on the stone surface that might be the remnants of (…) an a-rune” vs. “there is an edge the stone surface that might not be the remnants of (…) an a-rune”. These two propositions do not necessarily contradict each other, thus indicating low modality strength. If the proposition is, on the contrary, of high modal strength, it is not negatable without resulting in a contradiction.

For example: “the latter ones can (…) have belonged to (…)” vs. “the latter ones cannot (…)

ADAPTATION Absolute % of all ST

hedges (201*)

per 100 ST words 3a) change in modality strength

increasing decreasing

8 - 8

4.0%

- 4.0%

0.2 % - 0.2 %

3b) change in word class 24 11.9% 0.5 %

TOTAL 32 15.9% 0.7 %

Probably certain