Are errors noticed in AV translations? : viewer reactions to subtitle quality

(1)

UNIVERSITY OF EASTERN FINLAND PHILOSOPHICAL FACULTY

SCHOOL OF HUMANITIES English language and translation

Tatu Aapo Juhani Lehto

ARE ERRORS NOTICED IN AV TRANSLATIONS?

VIEWER REACTIONS TO SUBTITLE QUALITY

MA thesis

March 2017

(2)

Faculty

Philosophical Faculty

School

School of Humanities Author

Tatu Aapo Juhani Lehto Title

Are errors noticed in AV translations? Viewer reactions to subtitle quality

Main subject Level Date Number of pages

English Language and Translation

Pro Gradu Thesis x

29.3.2017 85 pages +

Appendix of 3 pages Second Subject

Thesis

Bachelor’s Thesis Intermediate Thesis Abstract

Translation errors have been the subject of relatively few studies. Determining errors is difficult and often depends, for example, on the purpose of the translation. Audiovisual (henceforth AV) research has investigated, for example, the perception and reception of humor or cultural references, and these studies have mainly included dubbed, not subtitled material (see e.g. Tuominen 2012). The present study investigates the reception of subtitles. What do the informants notice in subtitles and, more specifically, what type of negative elements draw attention? In the light of previous studies persons both more and less familiar with translation tend to perceive subtitles in similar ways and share similar opinions regarding their quality.

The study involved seven informants divided into two groups: one group consisted of students of English and translation and the other contained students from other fields. Both groups observed an animated movie in English with Finnish subtitles that contained numerous flaws from wrong word choices to unidiomatic Finnish expressions. The study was conducted as a focus group study during which the informants discussed their experiences with the lead of the moderator.

This study corroborates the results obtained from earlier studies: views concerning the general quality of subtitles did not substantially differ regardless of differences in educational background. The study also reinforced views concerning the nature of AV translations: they are but one component of a multimedia experience in addition to picture and sound. The quality of the subtitles used in the study could be deemed rather poor, but informants did not pay great attention to errors. However, especially creative, unidiomatic or faulty Finnish expressions and any original source text expressions retained in the subtitles attracted attention among the informants and should probably be only be included with caution.

Keywords

AV translation, subtitles, translation error, reception research, focus group study

(3)

Tiedekunta

Filosofinen tiedekunta

Osasto

Humanistinen osasto Tekijät

Tatu Aapo Juhani Lehto Työn nimi

Are errors noticed in AV translations? Viewer reactions to subtitle quality

Pääaine Työn laji Päivämäärä Sivumäärä

Englannin kieli ja kääntäminen

Pro gradu -tutkielma x

29.3.2017

85 sivua + liite, 3 sivua

Sivuainetutkielma Kandidaatin tutkielma Aineopintojen tutkielma Tiivistelmä

Käännösvirheitä on tutkittu suhteellisen vähän. Määrittely on vaikeaa ja riippuu usein esimerkiksi käännöksen käyttötarkoituksesta. Audiovisuaalinen (tästä lähtien AV) tutkimus on havainnoinut muun muassa huumorin ja kulttuurisidonnaisten käännösratkaisujen vastaanottoa. Nämäkin tutkimukset ovat käsitelleet lähinnä dubattuja, eivät tekstitettyjä materiaaleja (ks. esim. Tuominen 2012). Tämä tutkimus tarkastelee yksinomaan tekstitysten reseptiota: mitä koehenkilöt havaitsevat tekstityskäännöksistä ja erityisesti millaisiin negatiivisiin seikkoihin he kiinnittävät niissä huomiota? Aikaisempien tutkimustulosten valossa katsojat havainnoivat tekstityksiä ja niiden laatua varsin samoin periaattein koulutustaustasta riippumatta.

Tutkimukseen osallistui kaikkiaan seitsemän henkilöä, jotka jaettiin kahteen ryhmään:

toinen ryhmistä koostui englannin ja käännöstieteen opiskelijoista ja toinen ryhmä muiden alojen opiskelijoista. Molemmille ryhmille esitettiin englanninkielinen animaatioelokuva, jonka tekstityksessä oli runsaasti erilaisia virheitä aina kielivirheistä epäidiomaattisiin suomenkielisiin ilmauksiin. Tutkimus suoritettiin ns. fokusryhmätutkimuksena, jossa koehenkilöt kertoivat kokemuksistaan moderaattorin johdolla. Tutkimuksen tulokset tukevat aikaisempia tuloksia: näkemykset tekstityksen yleisestä laadusta eivät eronneet merkittävästi koulutustaustan mukaan. Tutkimus vahvisti myös aikaisempia näkemyksiä audiovisuaalisten käännösten luonteesta: ne ovat yksi osa multimediakokemusta kuvan ja äänen lisäksi.

Tutkimuksessa käytettyjen tekstitysten laatua voi pitää melko huonona, mutta osallistujat eivät kiinnittäneet huomattavaa huomiota virheisiin. Kuitenkin, erityisesti luovat, epäidiomaattiset tai virheelliset suomen kielen ilmaisut sekä tekstitykseen sisällytetyt lähtökielen ilmaukset herättivät huomiota, ja niitä tulisi kenties käyttää harkiten.

Avainsanat

AV kääntäminen, tekstitykset, käännösvirhe, reseptiotutkimus, kohderyhmätutkimus

(4)

1 Introduction

Quality in translation is very relative. Equally relative is to define what counts as erroneous in translations. Such characteristics are always related to the needs and expectations of the recipients of the translations: different translations have different aims and audiences.

Therefore, the quality and errors are defined against norms and normative statements that can vary greatly. Translation itself covers a wide variety of fields from literature to documents and AV materials with each genre having its own characteristics. The focus of this study is the perception of errors in AV translations, more specifically in subtitles.

The concept of error or lack of quality can truly escape rigorous classification. What counts as an error in the eyes of one viewer can be considered a delightful solution in the mind of another. In subtitles, absolute loyalty to the source language can be difficult to obtain. This is because subtitles have very limited space and consequently require certain alterations, such as omissions and reductions and transforming spoken dialogue into written form. There are, of course, obvious and unmistakable errors, such as misspellings and wrong word choices, but the nature of translations errors can generally be difficult to define.

Translation errors and the overall perception of subtitles have generally received relatively little attention in AV studies. Especially studies provided by (2005) concerning translation errors and Tuominen (2012) on the perception process of subtitles have, among others, been useful.

The study of translation errors, i.e. questionable quality and lack of adherence to norms, is essential for the AV translation industry owing to the fact that the surrounding conditions can be challenging: the working conditions and fees of AV translators have traditionally been very poor and underrated (see e.g. Vitikainen 2013, Mäkelä 2013) and the conditions have been a topic of discussion in the industry. The fluctuating fees, the hectic working pace and lack of professional organization can cause problems (Laine 2007: 265–271). Furthermore, AV-translators are sometimes required to construct so-called “second translations.” This means that the original material has been time-cued to another language, and the Finnish translation has to follow the “foreign rhythm.” This can easily lead to strange and unidiomatic translations (Mäkelä 2013). Such conditions undoubtedly can have their bearing on the quality delivered. For example, someone translating for a cable network has to work

(7)

sometimes without a script or even without visual tape. Furthermore, translation students are often recruited for the job, because experienced translators would not price themselves as low (Abdallah 2007: 276). Translations are often given to the lowest bidder and the pace can be very hectic. Consequently it may not be worthwhile for the translator or there may not be time to do proper research to ensure better quality in translations (Abdallah 2007: 284). However, the working conditions in connection to the production of translations for the purposes of this study are not known. Therefore, there is little point in hypothesizing why such and such an error has come to be. Consequently, I will focus on noting and analyzing observations provided by my informants.

I first studied the perception of errors in subtitles in my Bachelor’s Thesis. More specifically, I observed, noted and categorized perceived errors in AV materials in an attempt to shed light on their character. The material contained both numerous unmistakable errors and a variety of questionable solutions. However, the study concentrated on errors from the point of view of a single person. Moreover, the reception process was far from natural as the material was observed repeatedly and any instances perceived subjectively as erroneous were meticulously noted. Consequently, this thesis sets out to expand the subjective notions of my earlier thesis in an attempt to discover what can be considered erroneous in subtitles. To achieve this, I assembled two informant groups and subjected them to one of the animated films used for my Bachelor’s thesis. The groups differed as to their familiarity with subtitles and translation. My primary aim was to investigate the nature of any perceived errors and the possible differences between informants with various backgrounds.

The thesis is structured as follows: in chapter two I will explain some key concepts of AV translation, strategies for AV translations and means of their evaluation. In chapter three I will present studies on the general quality and norms in translations and in AV translation as well as previous studies on translation errors. In the fourth chapter I will present some essential studies in the field of subtitle research and audience perception. In chapter five I will present the AV material used for this study and my methods of research. The sixth chapter contains the analysis of the translation errors perceived by me in the research material. The errors are categorized according to the assumed reason of each error and illustrated with detailed examples. In chapter seven I will present the focus group setting including the individual profiles of informants and notions obtained. In chapter eight I will present the conclusions of my study.

(8)

2 AV translation: dubbing and subtitling

AV translation can be divided into two different modes: dubbing and subtitling, which contain spoken language and written language, respectively. Dubbing consists of “replacing the original track of a film’s (or any AV text) source language dialogues with another track on which translated dialogues have been recorded in the target language” (Chaume 2012: 1). The remaining tracks with music, special effects and images are left untouched.

As (Chaume 2012: 1 4) points out, in addition to complete dubbing, certain varieties exist: in partial dubbing, the main character(s) are dubbed whereas the remaining cast is interpreted via voice-over. This means that translated dialogues are recorded and both the original soundtrack and the translation can be heard. This is particularly widespread in Eastern Europe and it is also known as single-voice-translation. Voice-over is also used, for example, in documentaries and interviews. Other forms of dubbing include simultaneous interpreting as well as certain less formal practices, such as so called free commentary and “fandubbing”, which may be used for humoristic purposes as well. Videogames can also be released in localized versions with dubbed content.

Dubbing is not particularly widespread in Finland or in the Nordic countries in general. It is often limited to children’s programs whereas subtitling is preferred in television and in filmic productions. Instead, the Nordic countries, including Finland, have had a strong tradition and preference for subtitles as means of transferring the message in the case of foreign imports.

Subtitles could, in general terms, be defined in the following way (Díaz Cintas and Remael 2007: 8):

Translation practice that consists of presenting a written text, generally on the lower part of the screen, that endeavours to recount the original dialogue of the speakers, as well as the discursive elements that appear in the image (letters, inserts, graffiti, inscriptions, placards and the like), and the information that is contained on the soundtrack (songs, voices off).

In addition to translated subtitling for television programs, there are subtitles for video films, translated and even multilingual subtitles for the cinema and for the opera to name some examples. There are monolingual subtitles and so-called open subtitles for the deaf and hard of hearing bearing the dialogue and additional information. Other forms of subtitles include open subtitles that are accessible by choice, live subtitles for live programs and even subtitles

(9)

for the blind with the use of a synthesizer (for various forms of subtitles see e.g. Ivarsson and Carroll 1998).

2.1 The technical restraints of subtitling

Subtitling practices vary between countries and cultures and even inside nations as different television networks have different guidelines and resources. In any case, subtitling has certain technical restraints. First, the length of subtitles is limited. According to Díaz Cintas and Remael (2007: 9), the length of the subtitles may vary between 32 to 41 characters per line in a maximum of two lines. However, according to Vertanen (2007: 133), in Finland, the guidelines also vary between different networks: the Finnish Broadcasting Company, Yle, has approximately 30 characters in use per line, whereas on commercial channels, such as MTV3 and Nelonen, the average length is some 32 characters.

The figures described above apply only for subtitles for television. Subtitles for the cinema often consist equally of two lines, and in Finland, one line is reserved for each national language, Finnish and Swedish. According to Hartama (2007: 192–193), there have traditionally been 40 characters available per line for cinema subtitles in Finland. This creates special conditions as subtitles have to be concise and synchronized (ibid: 189). In an emergency, a few more may be added but this is not recommendable if one wants to avoid damaging the readability of the subtitles.

Second, there are time restrictions. Here recommendations and practices vary. According to Vertanen (2007: 151), in Finland, the minimum duration for television subtitles is limited to one second and the maximum duration for one full-sized line is 30 seconds. The duration of subtitles varies between two to three seconds for one full-sized line and four to five seconds for two full-sized lines. Pedersen (2011: 19) gives the estimated value of approximately three seconds for one full line and six seconds for two lines based on the views by several authors.

Similar estimations have been provided by Díaz Cintas and Remael (2007: 97). According to (Hartama 2007: 193), the reading time for cinema subtitles may be slightly shorter than that of television subtitles with some two to two and a half seconds per one full line.

As these examples point out, no unified code exists. This should not be surprising as different languages have different characteristics. Due to these inherent restrictions, it is essential to

(10)

decide how subtitles are to be constructed and which aspects should be given priority. In fact, because of such restrictions the very status of subtitles has sometimes been questioned: oral text is transferred into written form and synchronization with the multimodal whole with pictures has to be taken into account. It is also necessary to alter the original messages by means of condensation or omission. For this reason, some have claimed that subtitling is a form of adaptation rather than a translation proper (see e.g. Díaz Cintas and Remael (2007: 9).

I believe that it is justified to call subtitles translations as they do their best in conveying a proper message to the audience. But it is indisputably true that the reality of AV translation necessitates adaptive strategies, such as condensation and omissions. I will discuss these strategies in the following section.

2.2 Translation strategies for subtitles

According to Ivarsson and Carroll (1998: 85 88), subtitles will sometimes have to be condensed in order to include the most important information following the translator’s personal discretion. Short dialogues can be merged and only essential information needs to be conveyed from, say, someone’s muddled speech, as well as content that carries no essential information, such as repetitions, tautologies, conversational words, can be omitted.

Similarly according to Vertanen (2007: 152–153), any non-essential elements or something that is already known by the spectator must be omitted, such as names, titles and adjective attributes describing obvious or visible factors. Georgakopoulou (2009: 27–28) gives an extensive list of omissible elements (partially rephrased):

Repetitions

Names in appellative constructions

False starts and ungrammatical constructions Internationally known words, such asOK

Various expressions: salutation, politeness, affirmation, negation, surprise etc.

Exclamations

Anything empty in semantic load, such as you knowand well (see also Karamitroglou (1998).

(11)

Karamitroglou (1998) also adds that easily recognizable responsive expressions such as yes, no and ok are so widely known that they can afford to be omitted if clearly uttered. Other forms, such as slang variations of these expressions should be retained. Furthermore, such tautological cumulative adjectives and adverbs as great big, super extra and teeny weeny could be replaced by expressions such as e.g.huge,extremely andtiny.

Such means of condensation and omission, indeed, render the message more concise and relevant for the viewer. Ivarsson and Carroll (1998: 89) also point out that subtitles should take into account the heterogeneity of any given population, for instance its differing level of education. For such reasons complicated terminology and syntax may call for simplification.

In some countries this fact is also included in legislation concerning public television. In practice this means providing subtitles with concise structure and with more generally known vocabulary.

There have also been views against these measures. For example, Ivarsson and Carroll (1998:

93) claim that omissible details referred to above, such as exclamations and forms of address, should in fact be included in the subtitles as omitting such content fails to take the deaf and hard of hearing into account. But with time and space so limited, this consideration may be impossible to fulfill in every situation.

As the present study is based on translated subtitles, the presence of culture-specific elements, or Extralinguistic Cultural References (ECR), as Pedersen (2008) calls them, is particularly valid as the source text may include expressions that have a close connection to the source culture, but they may be of little or no use to a member of the target audience unfamiliar with such cultural subtleties. Pedersen (2008: 102) defines ECR’s as “expressions that refer to entities outside language, such as names of people, places, institutions, food and customs, which a person may not know, even if s/he knows the language in question.”

Pedersen (2008: 103, 2011: 76) divides translation strategies for ECR’s into six categories:

Retention, Specification, Direct Translation, Generalization, Substitution and Omission. He also discusses the use of an Official Equivalent, which he does not consider to be a strategy, as such equivalents have a very special and established status in the target language (Pedersen 2011: 74 76; see also Karamitroglou 1998 for a similar list). The following list presents Pedersen’s division. It is partially rephrased and I have added examples:

(12)

Retention: source text ECR is retained unchanged or slightly adapted

Specification: More information is added and the subtitled ECR is more specific by completing a name or an acronym (such as NATO) or by adding semantic content, such as adding someone’s occupation or an evaluative adjective

Direct Translation: only language is changed without semantic alterations (government agencies etc).

Generalization: making the subtitle less specific than the source text ECR by using either a Superordinate term (a higher-order concept: cat animal) or a Paraphrase (expressing same meaning in different words)

Substitution: the ECR is replaced by another, either from the source culture or the target culture or with something totally different

Omission: for example ECR’s that are unimportant (Pedersen 2011: 160)

Official Equivalent (a well-established target text equivalent. For example, Donald Duck is known in Finland asAku Ankka).

Pedersen (2011: 159 161) points out that such strategies do carry certain considerations as for their practical use. According to him, for example retention, whether complete or adjusted to target language, may not be appropriate in the case of monocultural ECR’s as target audiences may not understand them. However, sometimes retention may be more appropriate if the ECR is visually available. In such a case, retention may not technically be an error, but it may render the source text difficult to understand. According to Pedersen (2011:79 99), specification may explain, but it is space-consuming and could even be seen as patronizing.

Direct translation could lead to unnatural or unidiomatic translations. Generalization uses less space than specification and is more acceptable as there is no breech of reference to the original language. Substitution may break the illusion involved in the subtitles as something foreign is replaced with something more familiar and the viewer is thus reminded of the presence of the subtitles. As for official equivalents, a translator may opt for several strategies, such as retention, direct translation, substitution or specification.

As these problems and potential opportunities in constructing optimal subtitles point out, good quality can certainly be met in a variety of ways and it can be difficult to establish one optimal solution for each and every problem. These evaluations are inevitably linked to what is perceived as good or bad, more or less optimal in translation: they are linked to what could be considered more or less adhering to norms in translation. In the next chapter, I will discuss the nature of translation norms both generally and in specific relation to subtitles as both definitions for good quality in translation and what could be seen as erroneous in them are ultimately linked to what can be considered appropriate or acceptable in a given context.

(13)

3 Quality and norms in translation

Quality is always related to norms. In the case of translations, norms create expectations against which the success of a translation is measured. Such expectations may vary considerably. I will begin this chapter by defining norms in translation in general terms. Then I will discuss norms related more specifically to subtitles.

3.1 General norms in translation

The quality, or lack thereof, of any possible subject under evaluation is its adherence to any existing norms intended to measure the appropriateness of a given action. Bartsch (1987 as cited in Chesterman 1997: 54) defines norms as” orders or prescriptions which are issued by a superior to a subordinate.” Toury (1995: 55) acknowledges that failing to adhere to norms makes a subject liable to sanctions, whether negative or positive. According to him, norms

“are the key concept and focal point in any attempt to account for the social relevance of activities, because their existence, and the wide range of situations they apply to (with the conformity this implies), are the main factors ensuring the establishment and retention of social order.”

Toury (1995: 54) positions norms to a middle position between laws and idiosyncrasies. He acknowledges that some norms are stronger and closer to laws whereas some are weaker and closer to idiosyncrasies. Furthermore, the borders can be diffuse and grading may be subjective.

Chesterman (1997: 55–56) also distinguishes norms as in a middle position between laws and conventions as conventions do not provoke “generally justified criticism.” Following Bartsch (1987), he gives three conditions for norms: first of all, they merely state a common practice or a condition as such. Secondly, they add to the point that a practice is accepted as being valid and desirable. Thirdly, norms must exist in intersubjective consciousness: they are known and they can be talked about. As for the validation of norms, they may, according to Chesterman, be validated by a norm authority, by someone deemed prestigious enough to make such a judgment or simply by their very existence: they are automatically acknowledged to be valid if acknowledged to exist.

(14)

It could be argued that norms do indeed provide insight into what is deemed more and less acceptable in a given community or a given instance. But it is also evident that defining and grading norms does not escape personal judgment and intersubjective consensus may be difficult to establish. But deviation from norms is always a possibility and norms change over time. Thus deviations from norms may eventually lead to new ways of doing and thinking, i.e.

to new norms. According to Toury (1995: 64), for example, deviant instances in the behavior of a translator may eventually lead to changes in the system itself. Indeed, one could argue that solutions that were once deviant may become standardized and endorsed.

Toury has done academic groundwork in providing definitions for norms in translation. I will next introduce Toury’s norms followed by norms provided by Chesterman (1997).

3.1.1 Toury’s norms

Toury’s categorization is quite complex and contains three main categories: the initial norm, preliminarynorms andoperational norms.

The initial norm means making a choice between the aim of the translation. More specifically, should the translation follow the norms of the source text or the target text? According to Toury (1995: 56 61), the first option leads to an adequate and the latter to an acceptable translation. However, he does point out that regardless of any initial choice, a translator may still opt for solutions that do not adhere to any initial principle and absolute regularities cannot be expected in any domain.

The preliminary norms contain two considerations, the translation policy, i.e. what should be translated in the first place and secondly, the directness of the translation i.e. whether the translation should be direct or adapted.

The operational norms again fall into two categories: matricial norms including such decisions as the fullness of the translation, its location, distribution and textual segmentation.

Textual-linguisticnorms determine the selection of material for the target text or material with which to replace the original material. In other words, they have to do with the relation between the source and the target text and the nature and degree of changes introduced in the translation.

(15)

Toury’s norms have not escaped criticism: first of all, the initial norm defines what is deemed adequate and acceptable. I would agree to the criticism presented by Hermans (1999, as cited in Pedersen 2011: 35): adequate or acceptable by whose standards? The source or the target culture? Acceptable by the author or the target text readers? Furthermore, according to Pedersen (2011: 35) the terms are somewhat ambivalent. Also, deciding upon a text to be translated as dictated by preliminary norms cannot take place after an initial norm: the text to be translated forms the basis of translation activity. Toury also places equal emphasis on all of the norms. Pedersen (2011: 35) points out that whereas other norms can be relatively easily investigated, the textual-linguistic norms are quite complex and of great importance in the actual process of translation.

I agree that textual-linguistic norms are of particular importance in the formation of a translation. Chesterman (1997) has introduced a norm system that is particularly target- oriented. The target text and community are of great importance and translations are seen as products for a receiving audience.

3.1.2 Chesterman’s norms

Chesterman’s (1997: 63 70) translation norms form two main branches: product norms and process norms. Product norms are also referred to asexpectancy norms, meaning that norms are based on the expectations of the intended readers concerning text type, discourse conventions, style, register, grammaticality, statistical distribution of text features, collocations, lexical choices etc. The process norms are subordinate to the expectancy norms and are subdivided into three norms: following the accountability norm, the translator is responsible for any actions taken and that demands of loyalty are met between all relevant parties, the communication norm obliges the translator to “optimize communication, as required by the situation, between all the parties involved. Finally, according to the relation norm, a relevant similarity is maintained between the source text and the target text.

Albeit Pedersen (2011: 36) claims that Chesterman’s norms contain the weakness of aspiring to be prescriptive rather than descriptive, Chesterman (1997: 65) does acknowledge that quality in translation can be met in a variety of ways and that different countries have different translation traditions. Actual translation can be closer or more distant to any norms.

Chesterman does accentuate the position and role of translations as products that are designed

(16)

to meet any given need. Some translations, such as business letters, could be expected to follow the source text quite closely whereas, for example, literal texts could be expected to preserve some of the “local color” of the original. Similarly, according to Tomaszkiewics (2001: 240), it is sometimes justified to retain certain terms, such as pizza or hamburger in order to keep the “local color” of the original text.

It could seem logical to allow prospective readers to decide whether a translation fits its purpose being thus, by definition, of good and acceptable quality. This approach is also problematic. Colina (2009) criticized the functional-pragmatic model of translation quality evaluation developed by House (1997). In House’s model, the quality of a translation is measured by the equivalence of linguistic-situational factors between the source text and the translation. As Colina (2009: 237 238) points out, reader response may not always be the optimal solution in defining quality as not all texts are necessary reader-oriented (e.g. legal texts). Furthermore, according to Colina, such an approach concentrates merely on one aspect of a translation and the purpose of the translation could, for example, require a different response from the readers of a translation.

Also Toury (1995: 53 54) acknowledges translations as potential products aimed at certain readers and recognizes that “‘translatorship’ amounts first and foremost to being able to play a social role, i.e., to fulfil a function allotted by a community – to the activity, its practitioners and/or their products – in a way which is deemed appropriate in its own terms of reference.” It other words, they must be able to make communication possible in a given community by following any existing norms established in the community in question. As for translations as products, Toury mentions that “translators performing under different conditions (e.g.

translating texts of different kinds, and/or for different audiences) often adopt different strategies, and ultimately come up with markedly different products.”

I agree to these views: translation should, first and foremost, be regarded as products designed to meet the requirement of the prospective clientele. It is of course true that stated norms do not give practical insight into individual situations in the translation process as definition of quality and the needs of readers depend largely on the judgment of whoever is deemed appropriate to make such a judgment. Nevertheless, I agree to Pedersen (2011: 37) according to whom Chesterman’s norms make a difference between the actual translation process and the end product and that such a difference is important in the field of subtitling where a

(17)

complicated process “often results in a product that is effortlessly accessed by the reader/viewer.”

3.1.3 Translation quality, audience expectations and norms

If translations are to be thought of as consumable products, what kind of norms could be seen as definite indications of a high-quality product? Definitions and views vary to some extent.

According to Brunette (2000: 174), the most important factor in defining quality is the logic of the translation: it has to be semantically coherent and textually cohesive in order to achieve an effective communication act in the target community. Gouadec (2007: 7) also points out that translations must be “readable, coherent, logical and (preferably) well-written.” He also stresses the importance of the target culture: any source text or cultural elements will have to be altered or omitted to avoid confusion or misunderstanding and the translation must be adapted to the receivers’ level of technical competence.

According to Brunette (2000: 176 177), the purpose of the communication has two components: intention and effect. In other words, communication sought by the author (action aspect) and communication sought by the translator (reaction aspect). Gouadec (2007: 7 8) presents similar views: is the purpose of the translation to assist in something? Explain something? Seduce etc.

Undoubtedly, if translations are to be viewed as products to be consumed by clients or consumers, the client’s point of view becomes critical in determining quality. According to Mossop (2014: 22), “Quality is always relative to needs. There is no such thing as absolute quality. Different jobs will have different quality criteria because the texts are meeting different needs.”

Mossop (2014: 24 26) defines quality as whatever satisfies the communicational need and is fit for the purpose at hand. He points out that even an inaccurate translation could still serve its purpose (see also Hansen 2010: 386). Mossop and Hansen also discuss the position of machine translation: machine translations result in errors which can be accepted in such a communicative situation. Mossop (2014: 23 26) takes the thought even further by claiming that the quality produced by machines could actually lead to new norms: quality in translation is what machines can produce and everything else becomes undesirable. It would merely be up to human translators to eliminate the worst wrong collocations, unidiomatic expressions

(18)

and wrong prepositions. This may or may not be the case at least for some purposes in translation, but in general sense Mossop does acknowledge that “purpose” may be difficult to determine and that the concept of quality may vary between countries, language pairs or even by direction.

Gouadec (2007: 6 8) also agrees that the most important factor is the effectiveness of the communication process, not the ways used to express the message. In order to be effective, the translation has to take into account the linguistic and cultural standards and usages, rules and regulations, terminological and technical standards and physical and functional limitations concerning, for example, screen displays or Internet-sites. Furthermore, the end users’ value systems, their commonly accepted rhetorical and stylistic conventions, the various cultural contexts, such as national or corporal, as well as the most effective way of providing arguments accordingly will need to be taken into account.

To sum up, the most important aspects in defining quality are logical and understandable texts that meet the requirements of the clients and are carefully adapted from the target receivers’ point of view. These requirements may vary from insightful presentations to throwaway leaflets. But the purpose and especially the desired effect of the translation and the various social, cultural, professional etc. contexts that define such characteristics remain elusive and generalizations are difficult to make (see e.g. Brunette 2000: 177).

The factors that emerge from various authors could be called normative and they seem, by their most essential core, to adhere to Chesterman’s (1997: 69) communication norm: “a translator should act in such a way as to optimize communication, as required by the situation, between all the parties involved.” It should be pointed out that such findings have to do with translations in general. In the next section, I will examine any norms and definitions that are inherent in AV translation. This is not to say that AV translation and the evaluation of quality would be separate from other forms of translation but in AV translations there are certain important special aspects, such as considerations for technical limitations.

(19)

3.2 Norms in subtitles

3.2.1 General and technical norms for subtitles

There are essentially two kinds of norms in subtitles: norms concerning the expression and content of subtitles and norms that have to do with technical considerations. For example, as the time and space for subtitles are limited, certain alterations are inevitable, such as condensation of the translation. I shall now discuss such strategies in more detail.

In addition to varying density and level of condensation, Pedersen (2011: 73 74) mentions that it is generally accepted that retention, omission, substitution of cultural items, generalization and specification can be used as subtitling strategies (see also section 2.2).

Following and commenting on Chesterman (1998, as cited in Pedersen 2011: 212 216), Pedersen presents and discusses norms specific for subtitles:

the source and target texts have the same function the content of the translation is reduced (condensation) the text type remains the same with structural alterations

the style is reduced (less stylistic features, such as alliteration, rhetorical tropes and register)

there is little or no source-text revision subordinate status to the source text

acceptable language variant: “good native style” (layman view that is not necessarily shared with individual translators)

In addition to general norms described above, Pedersen (2011: 130, 214 216) presents the following technical norms for subtitles. Some of the norms presented below were discussed in section 2.1. For example, the number of lines and characters, time and space, are limited:

Expected reading speed: the speed in which the reader is expected to read a subtitle.

This is measured by the average number of characters per second of exposure time Subtitle density: the number of translated subtitles measured by the number of subtitles per minute. These are measured by the number of subtitles divided by the length of the target text in minutes

Condensation rate: The quantitative differences between the source text and the target text presented as a reduction rate between the languages

Translation not localized Translation not matched Restricted space

Limited time.

(20)

Naturally, density may vary and some subtitles are more condensed than others. As examples, Pedersen (2011: 138) mentions that the density of subtitles varies between different genres.

The dialogue in television series, for example, tends to be denser than in film subtitles. The development of electrical cueing used in subtitles also tends to reduce the need for condensation (ibid: 146).

As far the actual structure of the subtitles is concerned, it is therefore evident that subtitles are condensed, syntactically restructured and stylistically diminished translations that are being created under limitations both in space and time.Furthermore, the structure and conventions in subtitling vary between countries and especially between countries that have national traditions either in subtitling or dubbing.

3.2.2 International subtitling norms

National subtitling norms are increasingly challenged by international standards. The use of specific template files with inherent norms has made it possible to work more quickly and efficiently. This has resulted in subtitling norms becoming more uniform internationally (Pedersen 2011: 179). It should be noted, however, that Pedersen studied the development and differences and similarities between subtitling standards in Swedish and Danish subtitles thus concentrating heavily on the Scandinavian region. As for international differences, Pedersen (2011: 150) noticed that there are particular technical differences between subtitling and dubbing countries: in dubbing countries, subtitles tend to reproduce more of the original dialogue. Furthermore, the expected reading speeds tend to be higher, less condensation in used and subtitles are denser than in traditional subtitling countries (see above for technical norms in subtitling). According to Pedersen (2011: 210 211), this could be due to imported Anglophone subtitling norms. Traditional dubbing countries also tend to, at least presently, favor subtitling solution that adhere more closely to the source culture owing to the traditional lack of exposure to Anglophone culture. According to Pedersen, these differences are waning as norms are likely to become more unified in the future both in both technical terms and in terms of content.

(21)

It is therefore obvious that subtitles that are perceived as good or less appropriate have to do with both national and international traditions. In the following section, I will present findings aimed at defining what is expected of good subtitles in general terms.

3.2.3 Definitions for good subtitles

In section 2.2, I examined various strategies to be considered in the subtitling process. These strategies and subsequent practical choices have obvious bearing on the perceived quality and success of subtitles. Strategies and recommendations vary and there have been attempts to construct internationally applicable guidelines for optimal subtitles. For example Karamitroglou (1998) offers a wide and extensive list of recommendations concerning the layout and the duration of subtitles as well as notions concerning punctuation and letter case.

According to Ivarsson and Carroll (1998), subtitles should, among other things, be coherent and self-contained entities that take into consideration idiomatic and cultural nuances.

Subtitles should also be semantically straightforward and the register should be appropriate and corresponding. Such lists, particularly the one provided by Karamitroglou, are quite extensive and cover a wide variety of considerations, such as idiomatic and cultural nuances, structure, grammar, duration, correlation between image and the editing process. What is also noteworthy of these two particular normative lists is that both attempt to give international recommendations that would be applicable across Europe. However, the practical application of such recommendations, at least in their entirety does not seem feasible. As Suojanen et al.

(2015: 87) point out, people are used to how things are in their own culture and such harmonization could be contested if viewers would be required to adjust their viewing habits as a consequence. There are further difficulties with these recommendations as some items on the list do seem problematic, especially if designed to be adopted internationally. For the purposes of this study, my aim is not to present such extensive lists in minute detail. Rather, the examples mentioned above serve to illustrate the difficulties of providing guidelines that would take every possible aspect into consideration, especially internationally as cultures and languages with distinct characteristics can vary greatly (for further subtitling guidelines, see e.g. the BBC guidelines, Vertanen 2007 and Georgakopoulou 2009).

(22)

In spite of the various difficulties described above, some interesting notions have emerged from various writers concerned with good quality in subtitles. These notions include the following:

The language register must be appropriate and correspond with the spoken word. The language of the subtitles should be grammatically correct as they serve as model for literacy (Ivarsson and Carroll 1998: 157).

A translation should yield the intended interpretation without exposing the audience to unnecessary processing effort (Gutt 2000: 107).

Professionals need to provide viewers with the shortest possible subtitles and spare them unnecessary shades of meaning that hinder the process of image reading (Hajmohammadi 2004).

Subtitles are said to be most successful when not noticed by the viewer. For this to be achieved, they need to comply with certain levels of readability and be as concise as necessary in order not to distract the viewer’s attention from the programme (Georgakopoulou 2009: 21).

When the subtitles are synchronized with rhythm, picture and sound, there is an illusion of understanding when the viewer might not even acknowledge reading a translation at the same time (Vertanen 2007: 132–133).

They should be effortless to process: subtitles are inserts that have to be processed through the visual channels. This makes it more difficult to relax and enjoy the program (Georgakopoulou 2009: 21).

Reading subtitles is different from other conventional reading practices. Any presence of complicated language or details that require specific attention can be a source of distraction or be missed entirely (Tuominen 2012: 272).

Subtitles should be concise and understandable at one, immediate glance (Tuominen 2012: 272).

[S]imple, easily readable subtitles where the words are not too obscure or too innovative, and language that does not deviate too far from a standard register (Tuominen 2012: 280).

Based on the views described above it would appear that subtitles considered of good quality are concise, immediately understandable without effort, uncomplicated devoid of excessive innovation and deviations from standard language i.e. as “invisible” as

(23)

possible. To this could be added optimal segmentation and synchrony (see Di Giovanni 2016). Consequently, it would therefore be logical to claim that what could be considered erroneous in subtitles would be the opposite of the views stated above. In other words, errors could be perceived as items and solutions in translations that hinder the “effortless invisibility” of subtitles. In the next section, I will discuss translation errors both in general terms and in the case of subtitles.

3.3 Translation errors

The notion of translation error has not been widely studied as such. According to Vehmas- Lehto (2005: 49), translation errors have often been merely perceived as the opposite side of good translations. Naturally, as we have seen in the previous chapter, errors and their definitions depend on whatever norms and circumstances are present.

For example, Sigrid Kupsch-Loseleit (1985, as cited in Nord 1997: 73) defines translation errors as offences against:

the function of the translation the coherence of the text the text type or text form linguistic conventions

culture- and situation specific conventions and conditions the language system.

Such a definition is sound, but rather broad and fails to consider the vastness of various situations and contexts as expectations and contexts may vary. According to Hansen (2010:

386), errors of any kind can be acceptable in certain cases, as in the case of machine translation or in order to create a humorous effect. Different translations obviously may have different aims and functions. In general terms, Vehmas-Lehto (2005: 53) defines translation errors as any element in the translation that alters some factor that is essential to the receiver or prevents the transmission of such factors. Also less serious errors such as spelling or stylistic issues may upset the reader.

(24)

Translation errors have also been studied according their apparent or more disguised presence in translations. House (1997: 45 46) defines errors depending on their perceived visibility into overtly and covertly erroneous errors. According to House, overtly erroneous errors include omissions, additions, wrong selections for translations as well as breaches of the target language system in the form of grammatical errors and odd use. Covertly erroneous errorsrepresent mismatches between the:

[S]ituational dimensions of the ST and the TT in cases where the cultural distance between the respective communities and the differences between the SL and the TL are not too big, and where no secondary function has been added to the TT.

In a similar fashion, Vehmas-Lehto (2005: 64 65) divides errors into overtandcovert errors, According to her, overt errors are deviations from the source text that are not justifiable with, for example, pragmatic reasons. Secondly, overt errors violate the target language system and grammatical norms and thus either alter the text or distract the reader. Covert errors represent a deviation from language that native speakers would likely use. Covert errors ruin authenticity and naturalness rendering the translation unidiomatic and difficult to understand.

Furthermore, since many covert errors are quantitative and connected with the syntactic structure of the translation, they are, according to Vehmas-Lehto (1989: 213), not as easily detectable as, for example, morphological and lexical errors and they have been often neglected.

Nord (1997: 75–76) presents a functional top-down classification for translation errors according their perceived severity. Nord does acknowledge the importance of the context and situation in defining such severity especially in the light of translator training. She categorizes errors as:

Pragmatic translation errors, caused by inadequate solutions to pragmatic translation problems such as lack of receiver orientation

Cultural translation errors, due to an inadequate decision with regard to reproduction or adaptation of culture-specific conventions

Linguistic translation errors, caused by an inadequate translation when the focus is on language structures

Text-specific translation errors, which are related to a text-specific translation problem and, like the corresponding translation problems, can usually be evaluated from a functional or pragmatic point of view.

(25)

According to Nord (1997: 76), pragmatic problems are often not difficult to solve once identified, but their consequences may be severe and, as a result, they are the most important group. The definition of the severity of errors may also vary according to what is sought after:

if the translation is to test language proficiency, linguistic errors are more severe than cultural errors as even a misplaced comma may prove to be serious. Whether cultural knowledge is important, cultural errors weigh more than pragmatic errors.

It should be noted, however, that the abovementioned definitions and categorizations, as subjective as they are, mainly have to do with other texts than AV translations. As the notion of translation errors in AV translation can be thought of in equally vague terms, some guidance and categorizations are called for in helping to define such instances.

For the purposes of this study, I have constructed a categorization based on the comparison of the transcript or errors compiled for my Bachelor’s Thesis (Lehto 2013) and categorization provided by Vehmas-Lehto (2005: 56–57) based on the various possible reasons for translation errors. The errors are divided into three main categories: errors due to insufficient acknowledgment of source language or culture norms, errors due to insufficient acknowledgement of linguistic or cultural contrasts between source and target languages and errors due to insufficient acknowledgement of target language norms.

As noted by Vehmas-Lehto, the translator may have insufficient knowledge of the source language, which leads to misunderstandings. In the case of insufficient knowledge ofcultural and linguistic differences, the translator fails to consider the differences between the source and target cultures and what the target audience would require to understand the message. In other words, he lacks cultural and linguistic contrastive competence. If the translator has insufficient knowledge of the differences of the source and the target language, the result will be linguistic interference i.e. the source language affects the target language leading to unnatural target text expressions. Interference may equally occur when translating into one’s own native language. In the case of linguistic interference, the translator may also fail to consider that a word in the source text has several equivalents in the target language. An incorrect equivalent may consequently be chosen. This was particularly noticeable during the course of the present study.

Retained source language, whether full, partial or adjusted, is included in the categories for the reason that such expressions were occasionally difficult to understand and they attracted attention during viewing. It was instantly apparent that they could generally pose difficulties

(26)

and be a source of distraction during viewing. Also Pedersen (2011: 78) mentions that retention may not be the ideal solution in the case of extra linguistic monocultural references.

The categories do not include such potential errors as lack of synchrony and segmentation since these did not appear to be an issue. Furthermore, there were no extensive omissions or reductions and the categories include merely what the film was perceived to contain. The errors contained in the research material are categorized in the following way according to the reasons that may have caused them:

Table 1: categories of errors according to their reason Errors due to insufficient

acknowledgment of source language or culture norms

Errors due to insufficient acknowledgement of linguistic or cultural contrasts between source

and target languages

Errors due to insufficient acknowledgement of target

language norms

Misunderstood words Misunderstood expressions

Retention (full, partial, adapted) Unidiomatic Finnish

expressions

Incorrect Finnish equivalent

Spelling errors Incorrect punctuation Non-existent Finnish word

Some of these errors are of course not necessarily linked to competence: spelling errors, for example, may be the result of negligence or tiredness. They are included in the categories in case they attract viewer attention. Certain errors were also more frequent than others. It should be noted, however, that not all potential errors are necessarily strictly to be placed under one category. For example, unidiomatic expressions contain wrong equivalents which may well be the reason of their unidiomatic status. Such overlapping is of course natural and to be expected. In spite of such considerations, the categories serve to highlight the nature of potential errors to be found in the particular AV translation.

In the following chapter, I will present previous reception studies specifically aimed at studying the various aspects of AV translation. I will return to the perception of errors in AV translations at the end of the next chapter as various tests and their subsequent results have been presented. The results of the focus group studies will be discussed and analyzed in Chapter 7.

(27)

4 Previous studies on the reception of AV products

Vast majority of studies on AV translations have been looking into various segments but errors have not been the subject of particularly heated debate. According to Díaz Cintas and Remael (2007: 9), subtitles have the reputation of “lesser” activity:

[C]oncept of adaptation seems to equate the process to a lesser activity and becomes enough of an excuse to carry out a linguistic transfer that is clearly inadequate but nonetheless justifiable since it is only a case of adaptation.

Perhaps indeed on account of such a status the various aspects of subtitling, including errors, have been overlooked. Indeed, previous studies on AV perception, more specifically how AV products are perceived by audiences, include a fair amount of studies on dubbing, more specifically AV products dubbed in Italian, such as the overall perception of dubbed AV products (Antonini 2008, Antonini and Chiaro 2009), the perception of dubbed humor (Chiaro 2004), the reception of dubbed cultural references (Antonini 2009), the presence of formulaic language in dubbed Italian products (Bucaria 2008). Other aspect that have been under study include general profiling of cinematic audiences opting for subtitled cinema (Widler 2004).

One Italian study, Di Giovanni (2016: 58 78), investigated the perception of subtitles with particular focus on what was perceived to be objectionable in them. The study consisted of three experiments conducted at Italian film festivals. According to results, bad synchronization and segmentation were seen as particularly remarkable. The most important assurance for quality was, not surprisingly, good translation.

4.1 Studies on the perceptibility of subtitles

The supposedly subservient role of subtitles has, however, generally been reinforced by previous studies. Although it has also been suggested that whenever subtitles are present they are also noticed. According to some previous studies by d’Ydewalle et al. (1987,1991, as cited in Lång et al. 2013: 73), subtitles are read, at least to some extent, automatically even when they are not necessary for understanding. This arguably happens because the visual subtitles represent a more efficient channel for information than spoken dialogue. Although Tuominen (2012: 316) also reported findings of the apparent inevitability of subtitles, according to her findings they do not, despite being noticed, control the viewing experience:

(28)

[T]he informants’ commitment to reading subtitles was not very strong, and their expressed reception strategies occasionally minimised the role of subtitles. The story that these focus group discussions tell is of viewers who use subtitles whether they intend to or not, and who also listen to the source text, whether they want to or not.

Subtitles are a tool, a supportive text that helps in absorbing the audiovisual message.

They are not something to be read attentively and remembered afterwards, but something to be read quickly, superficially and accidentally in the multimodal flow of the film.

Also Lång et al. (2013: 77) noticed that the test subjects’ gaze was often directed at the bottom of the screen where subtitles usually appear whenever a character started to speak even when the subtitles were not present. Despite this, according to Lång et al. (2013: 83), the subtitles were not automatically sought by the participants at the beginning of a scene when there was something visually remarkable present until the subtitles actually appeared on screen. Therefore, the authors argued, quite logically, that the context of the subtitles combined with dialogue, sound effects, music and various visual elements all affect the viewing experience.

As the reading of subtitles and indeed the entire multimodal viewing experience are subject to individual variation, some studies have deployed eye-tracking which offers an interesting addition to the perception of subtitles as part of the multimodal context as it has the possibility of revealing unconscious reception processes. As translation errors have not been a frequent subject of study, eye-tracking studies have not extensively investigated them, either.

However, some noteworthy observations have been made. Lång et al. (2013: 72 86) studied specifically the effect of badly synchronized television subtitles on the attention allocation of viewers. The study consisted of two groups composed of students and staff of the University of Eastern Finland. A control group observed the material with normal subtitles and a test group was subjected to manipulated subtitles. The study material consisted of scenes from a British television sitcom Absolutely Fabulous subtitled in Finnish. On 32 occasions, the subtitles were purposely either lengthened unnecessarily or shortened to appear simultaneously with spoken dialogue. Specific eye-tracking equipment was installed to record the test subjects’ eye-movements during viewing.

Interestingly, in spite of the rather considerable number of manipulations, most of the desynchronized instances were not noticed at a conscious level. What is more, the subjects did react to erroneous instances unconsciously: lengthened subtitles repeatedly drew attention and

(29)

shortened subtitles that had been synchronized closely with the original dialogue were not read entirely due to lack of time to do so. Even when most of the manipulation affected the gaze paths of the viewers they were not consciously recognized and remained at the level of automatic sensory reactions.

This lack of conscious observations further resulted in another interesting finding: the translation, despite containing numerous errors, was not criticized in any particular manner, even though certain individual erroneous instances were reported. This would suggest that the overall AV experience is fluently perceivable even when individual erroneous instances have been noticed in it. Furthermore, both groups used in the study reported a roughly similar number of noticeable instances with something “odd” present in the subtitles. The difficulty of giving examples was also manifested: although half of the informants in the control group reported having seen something questionable, none of them could name any instances. In the test group, most had noticed something unusual, but only a clear minority (4 out of 11) could name an instance with one particular instance named by all of them.

Such a tendency to overlook problems and the difficulty of naming any particular instances can, of course, be at least to some extent the result of the test situation as members of a control group are not observing the subtitles in a natural environment and the very presence of specific eye-tracking equipment may well have affected the alertness and any notions made by the informants. The study makes a note of this (Lång et al. 2013: 75). However, the supposed tendency to overlook problematic instances in subtitles is further reinforced by a study conducted by Tuominen (2012). As her study explored a relatively arid terrain in translation studies, her findings and conclusions merit close observation as far as the entire AV reception process and individual differences and similarities are concerned. Her study revealed several interesting findings concerning the overall reception process and differences and similarities between informants both more and less familiar with translation. What is particularly intriguing is that these differences do not appear to be particularly noticeable, which is in line with previous studies on AV reception. In the following section, I will present further evidence to support the claim that AV translations, in general and along with their quality, may well pass almost unnoticed or without being noticed too meticulously.

(30)

4.2 Studies on the quality of AV translations

There have been only a few studies available about the specific reception of errors in AV products. As was noted at the beginning of this chapter, previous studies have, rather, focused on other individual subjects. However, these studies have suggested that AV translations tend to go largely unnoticed by subjects exposed to such a multimodal whole.

Albeit these studies have not directly addressed the question of errors in subtitles, the results obtained from general reception of AV products provide notably interesting results. For example, in a study conducted by Widler (2004) on Austrian cinema audiences, the perceived general quality of subtitles was not seen as notably low. In another study, Gottlieb (1995) analyzed the perception of deviations from subtitle standards in Danish television programs subtitled in Danish for the deaf and hard of hearing. The study group consisted of elderly deaf and hard of hearing persons. However, the quality of subtitles was not deemed generally poor.

The subtitles were, on certain occasions manipulated slightly to include certain errors or other alterations. The exact number of manipulated instances does not become apparent in the study. However, as a concluding remark, Gottlieb notices that viewers do not tend to notice faulty subtitles particularly and occasional mistranslations do not seem to be an issue.

Furthermore, in another study concerning the perceived quality of subtitles, Alves Veiga (2006) investigated the perception of subtitled Portuguese television programs among high school students. He does not compare these evaluations to any statistics and reports figures and percentages obtained from the data of the study, but according to his findings, the quality was mainly deemed “good” or “very good” as only a small minority of informants deemed the quality “poor” or “very poor.”

The same tendency to overlook problems is also present in other fields of AV translation.

Several Italian studies have focused on the perception of dubbed AV products. In a manner similar to studies involving subtitles, the overall quality of dubbing was rated as “satisfactory”

or “good.” Similar results regarding the quality of AV translations were also reported by Antonini (2008). In her study, remarkably 75% of test subjects demonstrated positive attitudes towards the quality of dubbed Italian filmic products, whereas only 25% deemed the quality mediocre or poor.

To illustrate further the overall tendency to passively accept whatever translations are encountered by viewers, the study by Bucaria (2008: 149–163) investigated the perception of

(31)

an Italian language variant used in dubbed products also known as “dubbese” in English.

More specifically, the reception of certain expressions not deemed by the author to be part of naturally occurring Italian, such as retained cultural bound references, humor in wordplay or literally reproduced typically English words. The study included three groups, from general audience to persons more familiar with AV translations. Albeit all groups deemed the dubbese expressions unlikely to occur elsewhere, the quality of AV products was not wholeheartedly criticized. According to Bucaria, this could be due to the fact that dubbese is recognized, especially by relevant professionals, as an individual language variant that does not have to imitate everyday speech. Consequently language that could normally be regarded as strange or even erroneous could be deemed acceptable as part of an AV product and, consequently, even unnatural language would not be deemed erroneous if such deviations from standard speech are the norm in dubbed Italian AV products. Antonini and Chiaro (2009: 111–112) reached the same conclusion claiming that dubbese is perceived by Italian audiences as a separate language variant inconsistent with everyday speech. It was, in general, given passive acceptance as long as it would not enter the outside reality. Of course, since Italy is predominantly a dubbing country, this could affect perception. However, the more general implication of this notion is that AV quality or lack thereof is not necessarily a topic that would evoke heated arguments or objection.

There were, however, other problematic issues inherent to dubbed AV products mentioned by Antonini and Chiaro (2009: 109–110). Namely, as the study material consisted solely of dubbed material, many informants failed to understand retained source language materials, such as signs, newspapers and letters, which contained valuable information. Another problematic issue was acronyms containing socio-cultural information, such as BLT- sandwich standing for bacon, lettuce and tomato.

In spite of such individual problematic instances, the study did state that Italian viewers did not generally criticize harshly the quality of dubbed Italian products as the vast majority would judge them “satisfactory” or “good.” On the other hand, the acceptance seems to be rather passive and neutral as only a small number of informants would deem them either as excellent or poor (ibid: 106).

The pervasive tendency to overlook problems and to honor the legitimacy of AV products, whether dubbed or subtitled, tends to suggest that any conceivable translation issues may not be generally disturbing in subtitles. Indeed, several studies suggest that several potential