A Diachronic Analysis of the Translation of English Sound Symbolism in Italian Comics

: Phonosymbolic elements such as ideophones and interjections test the translator’s ability in various ways. These forms would, in theory, require a complete change of form and substance of the source text but this has not always been possible because of graphical, cultural and linguistic reasons, and this led, in certain cases, to a foreignized target-text environment. Recent research has started to consider the relationship between verbal and visual modes as beneficial and not just as a mere constraint for the translator. This research aims to align itself with this approach in order to analyze how verbal and visual modes in Disney comic books have come together to welcome sound symbolic forms and how translators have dealt with them in Italian Disney comics, in particular. In order to clarify the behavior, function, translation and use of expressive sound symbolic devices in Italian Disney comics, this article will offer a diachronic analysis of these strategies as found in a diachronic bidirectional corpus compiled through extensive archival research. real ‘dilemma’ (2003, 53), between the need for naturalness shown by ideophones and the constrained environment imposed by comics is dealt with, which of the two strategies prevails, and to expound on the possible reasons for this. (4) Italian vs Spanish Translation Strategies: Valero Garcés’s research (2008, 237-49) on the translation of ideophones from English to Spanish in imported American comics is the most recent similar study available. The following points summarize the conclusions reached by her study, which I aim to compare with my own final results following the corpora analysis:

volatile nature and, finally, (4) they are often associated with literary genres that are considered frivolous or 'not that serious ' (comics, picture books, manga, cartoons, advertisement, and/or children's literature in general). Valero Garcés postulated her theory more than twenty years ago. The study of sound symbolism in translation has made some progress since then, and scholarly work has been published on the subject. However, much work remains to be done, particularly when it comes to the use and semantic value of these forms (Dingemanse 2012, 654). Therefore, the postulations of Valero Garcés retain their relevance and utility.
Although there are conventions when it comes to the cross-linguistic equivalence of certain established sounds, see animal cries (Lathey 2016, 96), the same certainty does not apply to the multitude of senses and situations covered by these iconic words. There is an impelling need for translators of sound symbolic forms to get accustomed to dealing with 'standard expressions in both source and target languages before making strategic decisions on whether, or how, to translate such a vivid use of language' (Lathey 2016, 64). At the moment, there is a sense that translators are left on their own when it comes to making decisions regarding these forms. A 'touch of foreignization […] is likely to spark interest in sound and language' (Lathey 2016, 96), so strategies for certain types of media have been adopting this approach. See, for instance, the rising tendency to keep Japanese ideophones intact in manga translations, often with an accompanying gloss translation (Ficarra 2012, 58-9;Inose 2011, 250;Jüngst 2008, 67) --this choice being justified by the growing interest in the Japanese writing system of avid manga readers. On the other hand, scholars agree that translating ideophones from certain African languages will always involve an evident loss of meaning as African ideophones cannot usually be translated but 'may only have their import approximated in annotations and glossaries' (Teilanyo 2001, 220). Lupenga Mphande (1992, 119) emphasizes this point when he uses the expression 'textual genocide' when discussing the translations of African ideophones in the hands of missionaries and disciples. This 'disappearing act' (Casas-Tost 2014, 40) is not as visible in other language pairs. In particular, when it comes to translating English ideophones into Romance languages, despite some retention of source text material, there are generally scattered efforts to localize ideophones for the target audience (Valero Garcés 1996, 229). and acts as a 'graphical support' to the expressive forms (De La Cruz & Tejedor 2009, 57). The long-term use of this repetition strategy can indeed lead to positive outcomes: the reader starts self-educating about the meaning of foreign words --words that sometimes start being employed in spoken every-day language and can successfully be comprehended by readers and speakers, fostering linguistic awareness and cultural exchange. The reader then starts appreciating these foreignization strategies that lead to the creation of a sectorial language of comics. Examples are the various English localized forms that have invaded Italian Disney comics or the retained Japanese ideophones in the translation of manga into Italian. This is to say that foreignization strategies applied to sound symbolic forms are not necessarily prone to confuse the reader. Of course, a complete transposition of the sonic experience to the target-text environment might lead to more familiar-looking forms --but it is also due to the image, which acts as a bridge between the foreign expression and the source-text reader, that the untouched form gains some sort of semiotic substance. The process has shown to be quite successful, particularly in Italian and Spanish retention of English ideophones, some of which are now easily understandable and seen as either sectorial language of comics or even as not foreign at all by Italophones and Hispanophones (Valero Garcés 2008, 248).

Methodology
The first phase of this research project involved the creation of a corpus of sound symbolic forms taken from English and Italian Disney comics published between 1932 and 1992. The bidirectional corpus was created through archival work in both Italian and US archives. Firstly, the Italian data was collected in two Italian comic book archives based in Milan and Cremona. A total of 100 Italian stories were surveyed. In order to find the corresponding English source text a 4-month extensive archival work was carried out at the Library of Congress in Washington, DC, between October 2014 and February 2015. The reason for choosing Disney comics specifically relied on the fact that, since the 1930s, imported (American) English Disney comics have been institutionalizing the use of sound symbolic forms, especially ideophones, in Italian mass media, offering an original use of the device that had previously been employed mainly, and relatively infrequently, in Italian poetic and narrative compositions (see its use within the Futurist and Decadent literary movements) rather than paired with images.
The second phase of this research project involved analyzing the English to Italian translations of ideophones in the collected Disney comics as found in the compiled corpus. This investigation is situated within the field of product-oriented descriptive translation studies (Holmes 2004, 184-90) as it aims to provide a large-scale diachronic analysis. The results feed into what Holmes calls 'the theoretical branch' of translation studies and attempt to produce a historical analysis of the translation strategies relevant to the specific phenomena considered (i.e., sound symbolic forms).
Despite the fact that in the majority of cases English sounds were retained, which is one of the reasons for the strong presence of English sound symbolism in Italian today, scholars (Dovetto 2012;Mioni 1990;Pietrini 2009) have pointed out that some efforts to translate ideophones can indeed be detected, particularly in the early years of the importation process. The examination of the alterations of the translation strategies in regards to ideophones will allow for analysis of an under-researched area within the discipline of translation studies. As well as clarifying the ideophonic interconnections of the two specific languages considered here (Italian and English), the analysis of the corpus will try to theorize the evolution of translation strategies for this device throughout the eight decades of its occurrence.
Each of the translation strategies was catalogued according to Dirk Delabastita's classification as shown below (1989; also cited in Kaindl 1999, 275). This classification was originally conceived for subtitling localization but it has proved to be effective in other texts where pictorial features are present.
(2) Deletion: form removed. (3) Addition: form has been left but small spelling and phonetic changes have been applied. See the English 'beep beep' that becomes bip bip in Italian.
(4) Partial Substitution: form has been localized and the same type of ideophone/interjection has been used. E.g., English lexicalized ideophone 'crawl crawl' becomes striscia striscia thus keeping the lexicalized form in Italian as well.
(5) Total Substitution: form has been localized but a different type of ideophone or interjection has been used. This includes instances in which there has been a semantic change during the process.
(6) Creation: Form was absent in the source text and was therefore consequently added by the Italian translator.
(7) English alternative: The English form of the source text was translated using a different Anglophonic term in the Italian text.
The first part of the corpus analysis involves a close examination of each of the translation strategies described above and a report of their occurrence and variability across different decades and situations. I analyze how strategies for translating English ideophones evolved over time, focusing particularly on whether specific sounds, emotions or settings tend to be localized more often than others. The main analysis is organized according to different 'time windows', each chosen by virtue of the specific linguistic, historical and cultural vicissitudes that define their boundaries. This approach is considered particularly fruitful since the Topolino ('Mickey Mouse') magazine, which features the majority of these Disney stories, is a rare example of a publication that has been published almost without interruption throughout its eightyfour years of existence. This peculiarity makes it a very revealing medium, as it allows scholars to gain access to historical data on the diachronic evolution of the language featured in its pages. As Verda (1990, p. 58) comments, Topolino is an important 'graphical and linguistic vehicle that faithfully follows the course of time' (translation mine) and perfectly embodies the interchange of terminology typical of those mass media that were propagated in the decades around the two wars. Media that served as tools readily available to experiment with language in ways that other more established means of communication (newspapers and radio above all) might not have allowed at the time.
The second part developed after an extensive review of similar studies including those in respect of other languages. As well as allowing an integration of my study with current research trends, this crosslinguistic approach permits the creation of theories based on the behavior of other languages in analogous circumstances (i.e., when translating English ideophones). In particular, studies on this particular topic are available for the linguistically similar Spanish language (Balteiro 2010;Castillo Cañellas 1997;De La Cruz Cabanillas & Tejedor Martínez 2009;Garcia de Diego 1968;Gasca & Gurbern 2004;Gubern & Gasca 2009;Martinez Fuentes 2003;Valero Garcés 2008). Therefore, an assessment of the correspondences and/or discrepancies of my results with these inquiries can clarify whether the way languages create and translate ideophones might be influenced by their linguistic typology. My study might thus confirm that the inflecting nature of certain languages (e.g., Italian and Spanish) and their lack of morphological flexibility might in fact not provide the 'elasticity' in word formation needed in order to create sound symbolism, as suggested by various scholars in the field (Santoyo 1984;Martinez Fuentes 2003;Valero Garcés 2008, 239;Pischedda 2012;Dovetto 2012, 208).
Here is a list of the hypotheses I have formulated, the methodology I aim to apply in order to prove (or refute) them and the questions I intend to answer: (1) Constrained approach examined: Scholars have described the translation of comics as being subjected to a 'constrained translation approach' (Zanettin 2008, 20), highlighting the fact that there is less freedom on what can be translated since the visual can influence how much one can write/translate, for spatial or graphical reasons (Zanettin 2013). The apparent 'translation conservatism in comics' (Noss 2003, 51) deeply affects the way translators deal with ideophonic forms. In view of this, I am expecting words Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM inside balloons to be localized more often due to the easiness in changing their content, particularly when compared to the challenges involved when modifying text within images. (2) Localization over time: From a first inspection of early comics I expect ideophones to be localized more often during the first couple of decades of importation (30s-50s) with a consequent exponential retaining of English sounds through the decades. I intend to analyze whether this process was gradual or if it involved a sudden switch to English sounds --to the clear detriment of Italian sound symbolism --and possibly to elaborate on the historical and linguistic motivations behind this choice. (3) Investigating opposite tendencies in translation: Ideophones, and expressive language in general, have a unique role in the system of languages in terms of their reliance on the symbolic and immediate effect on the reader, and as they belong to an 'area of language where the relationship between the word and the auditory experience is close by nature rather than by conscious artifice' (Chapman 1984, 37). A total substitution of the target text's term is therefore a commonly used strategy, particularly since 'adhering strictly to the source text would cause an irreparable loss of meaning' (Eco 2001, 57), and it privileges a target-centered approach in order to reach 'equivalence of effect' (as theorized by Nida 1964, 159) rather than equivalence of lexicon, grammar or form. The naturalness of the target text thus becomes a priority, what Nida calls 'dynamic equivalence', that is the need for the final translation to represent 'the closest natural equivalent to the source-language message' (Nida 1964, 166). One would therefore expect strategies of 'substitution' to be favored. The reality is that often 'the source language form is borrowed and transliterated with the assumption that its meaning is either transparent or can be deduced from the context' (Noss 2003, 51). This practice, together with the limits imposed by the comic format itself, as explained in point 1 of this list, have been forcing translators to accept that images cannot sometimes be changed or translated, thus disturbing the dynamic equivalence imposed by ideophones and creating a tense translation environment constituted by opposite forces. I seek to investigate how this apparent conflict, defined by Noss as a real 'dilemma ' (2003, 53), between the need for naturalness shown by ideophones and the constrained environment imposed by comics is dealt with, which of the two strategies prevails, and to expound on the possible reasons for this. (4) Italian vs Spanish Translation Strategies: Valero Garcés's research (2008, on the translation of ideophones from English to Spanish in imported American comics is the most recent similar study available. The following points summarize the conclusions reached by her study, which I aim to compare with my own final results following the corpora analysis: -Two main strategies can be identified: the ideophone is either translated or left unchanged (Valero Garcés 2008, 241). -English sound symbolism usually tends to be retained, particularly when the ideophone is outside balloons (Valero Garcés 2008, 240). Since the start of the importation of American comics during the 50s the prevailing tendency has been to leave English sounds untranslated, with minor localization of certain types of ideophones. -Ideophones in black and white stories tend to be localized more often (Valero Garcés 2008, 240 The results reached will allow me to offer an overview of the use of expressive sound symbolic forms (both ideophones and interjections) in Italian Disney comics since their advent.
Firstly, I will offer an overview on the linguistic discoveries that came to light through the corpus analysis, with a detailed description of use of ideophones and interjections in Disney comics in the last eighty years. I will highlight major shifts in their use over time, and expound on the reasons for these.
Secondly, I aim to offer an account of the translation strategies used for ideophones and interjections and their diachronic change throughout the eight decades under investigation. In particular, a great emphasis will be given to the ongoing Anglophonic influence on Italian forms.

Results
The following section will investigate the translation strategies used for the ideophones and interjections that make up the corpus of forms that was compiled for this research project. The corpus of translated strings contains 1138 forms, taken from a total of 72 stories produced between 1932 and 1992.
The following chart (1) summarizes the percentage of strategies used within each time-frame, presenting the changing and emerging trends for each of these time windows, and highlighting the main strategies employed within each window. Each strategy has been color coded, as shown in the key.
The most evident trend is the rising repetition of source-text forms through the earlier decades covered by this study. From a rate of just over 20% in the 1932-42 period to one of just over 73% in the 1970s, strategies of 'repetition' were constantly on the rise. This tendency was also noticed by Valero Garcés (2008, 243) when looking at translations from English to Spanish of the ideophones in comics. This phenomenon was due to the growing influence of the English language in Italian media, which in turn caused readers to accept some of the English words as part of the sectorial language of Italian comics. This led to more and more English words being retained without any modifications. However, after 1980, this trend appears to change direction, as strategies of 'repetition' in Disney comics undergo a steady drop throughout the last 12 years covered by this investigation . Though the percentage of repetition strategies in this last time window (roughly 40%) is not as low as the percentage in the first two time-frames, it contrasts starkly with the remarkably high percentage (over 70%) found in the preceding decade. Despite still being the most used single strategy, 'repetition' seems to have been left aside by translators in favor of strategies of 'partial substitution', instances in which the forms have been localized but the same type of ideophone has been used (in terms of it being still lexicalized or non-lexicalized). This demonstrates the general initial trend of trying to localize the forms for the Italian audience, with a rising influence of English language through the decades that led to most translators retaining English forms in the target text. As the number of repetitions grows, the number of partial or total substitutions slowly decreases. After many decades of English influence, Italian translators might have felt that something had to be done about the possibly overwhelming Anglicization of the language and they might have decided to try to devise original Italian alternatives.
The last twelve years show a sudden change in previous patterns: despite strategies of repetition still representing the majority (40%), there is a sudden increase in both partial substitutions (from 4% to 17%) and in English alternatives (from 3% to 10%). This last strategy refers to the use of alternative English words that might be considered more familiar to the reader than the English words in the source text, and was also noticed by Valero Garcés (2008, 247) in respect of the translation of English ideophones to Spanish: '[…] this tendency to consider some English borrowings as part of the Spanish language of onomatopoeia is confirmed by the use of well-known English onomatopoeia to translate lesser-known ones, for example Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM 'sob! snif!' to translate 'bad blubber''. This change provides evidence of a willingness, in the last decade of Italian translated stories, to provide a more familiar textual environment to the reader. This does not exclude the ongoing influence of English forms, as proved by the presence of English equivalents, but it would appear possible that by that point these forms had already affected the Italian language to the extent that they were no longer considered to be foreign but were simply seen as part of the sectorial language of comics. Indeed, a study undertaken in 2010 showed that Italian people who had a beginner or elementary level of English were more likely to think that common English ideophones (in this case 'bang' and 'pant') were actually of Italian origin (Pischedda 2010, 45).
Another reason for this rise in the use of alternative strategies rather than repetition could be the arrival of digital fonts and computers in the comic industry at the beginning of the 1980s (Ficarra 2012, 31). These allowed for new and alternative ways of changing and updating the image. It should be noted that digital lettering was not used for comics until 1993 (Ficarra 2012, 62). Before then, all fonts had to be manually drawn by cartoonists and then imported into a computer to be inserted in the comic strip. Interestingly, Valero Garcés's research discovered a steady rise in the strategies of repetition even in the last 15 years, with 75% of ideophones retained in the comics she analyzed. This clashes with the findings highlighted in the corpus complied for the present study.

Translation Strategies According to Ideophonic Type
Analyzing translation strategies according to different ideophonic types will provide an alternative overview on how the translation strategies used for Disney comics might have been influenced by specific factors. The benefit of having a corpus at our disposal is precisely that different parameters can be used and prioritized to analyze the data. This allows for investigations that take different approaches to the study and offer several points of view. The four different ideophonic types used are a pivotal classification within the corpus as they attempt to divide the data based on their semantic value, i.e, the meaning they are trying to convey. While the (1) 'animal' type encompasses all those instances in which animal cries are represented, the (2) 'human' type includes all the expressions used to symbolize people's feelings (this category also includes interjections). When it comes to describing natural events, two different ideophonic types were chosen: (3) environmental and (4) mechanical ideophones that, respectively, represent environmental acts such as hits, bangs, and falls but also bodies and objects in motion (see kinesthesia), and events originating from technological devices and machines (i.e., computers, televisions, phones, doorbells, guns, etc.). Such a specific investigation will foster research on sensorial inputs and how these are perceived and interpreted by the translator.
The following chart (2) summarizes the translation strategies distribution according to the different ideophonic types: At a first glance, the graph immediately shows that environmental and mechanical events are the ones that are more likely to involve repetitions, followed by animal sounds. The 'animal' and 'human' categories are the only ones for which the sum of strategies of substitution is greater than the number of strategies of repetition. This could be explained by the fact that both these two categories involve events that usually possess an alternative in another language --particularly considering that they are both produced using another vocal system so are unarguably easier to imitate (Assaneo et al 2011, 4). Particularly, in the case of the 'human' category, ideophones that convey feelings and emotions are often easily translatable with interjections. The case for the animal category is evidently similar (also because characters are humanized and personified in Disney comics). Languages tend to own words suitable to represent most animal cries, which are readily available for use by translators. It is interesting that the research of Valero Garcés also noticed exactly the same tendencies: 'sounds produced by animals, unarticulated sounds produced by humans (including interjections) and sounds used to show feelings tend to be translated more often ' (2008, 241-3).
Valero Garcés also shows that mechanical and artificial sounds are the ones more likely to be retained (2008, 243) --and, again, this tendency can be seen in the current study as well, since mechanical ideophones have been shown to have more than a 50% chance of being retained rather than being translated. This is possibly due to the fact that mechanical events have a higher chance of being represented with nonlexicalized ideophones. As they are already often highly expressive and not based on actual lexicalized words included in the language this explains the higher chances of them being left intact. Indeed, sounds produced by machines and tools 'are not produced by another vocal system and therefore imply strong imitative efforts' (Assaneo et al 2011, 4) --efforts that translators might appreciate, making them more inclined to opt, in these cases, for a full repetition of source text material.

Translation Strategies According to Position
Investigating such statistics will help understand if the position of the form itself somehow influences its translation. One would expect forms inside balloons to be localized more often due to the relative easiness in changing the content of balloons compared with changing the environment (image) outside of them. The following chart (3) shows the findings As expected, the data shows that strategies of repetition tend to be employed less often for forms that are originally found inside balloons. These forms also tend to undergo deletion more than twice as much as their counterparts found outside (28% --compared to 11% for forms outside balloons). This data comes as no surprise as other studies (Valero Garcés 2008, 240) have shown that forms outside balloons are more likely to be left unlocalized 'since intervening in the picture would involve greater expense and require a high standard of graphic skills'.

Translation Strategies according to Strip Colour
Color can have a significant influence when localizing ideophones. Eva Martinez Fuentes's research (2003) into the translation of onomatopoeia in the US comic strip 'Calvin and Hobbes' brought to light that 75% of the onomatopoeia in color were retained in their English form while only half of the onomatopoeia in black and white were retained. Martinez Fuentes states that the color of the strip represents an important factor when deciding whether to opt for a translation or a repetition of onomatopoeias in the source text.
The following chart (4) summarizes the distribution of translation strategies according to strip color: These findings show that, as expected, there is within the corpus of this study a prevalence of repetitions in colored strips, with roughly half of the strips in color (51%) involving cases of repetition. This confirms the findings of Martinez Fuentes (2003) and those of Valero Garcés (2008) regarding the translation of English onomatopoeia to Spanish and, once again, confirms the strong similarities between the approaches adopted for translating ideophones from English in the Spanish and Italian contexts.

The Whole Picture
'Repetition' is the largest single category throughout the eight decades under investigation, though outweighed by a combination of other strategies. If one considers the total percentage of the different strategies that include a modification of the source text in some ways --partial and total substitution, addition and English equivalent, but without considering deletions and creations --the total amount of these localization strategies adds up to 53%, which is higher than the number of repetitions. Nevertheless, strategies of repetition play a considerable part, as the following chart (5) demonstrates. When translation strategies used for ideophones (and ideophonic series) are compared with the ones used for interjections, the results are obviously very different for each of the two features. As mentioned earlier in this chapter, interjections, due to their frequent position within the balloon and due to their easily translatable nature are more likely to be localized. Statistics in charts 6 and 7 below show that while the majority of ideophones (49%) are repeated, in the case of interjections repetitions only represent the 11% of the total, surpassed by, in order, total substitutions (15%), additions (21%) and partial substitution (40%). This means that more than three quarters of strategies used for interjections involve a degree of localization of the source text.

Summative Remarks
The present paper looked closely at the cross-linguistic adaption of sound symbolic forms in Italian Disney comics throughout an eight-decade time span. The conclusion to this chapter will attempt to bring together the hypothesis formulated in the methodology with the outcomes resulting from the analysis of the translation strategies in diachroneity to confirm or refute the postulated assumptions.
First of all, the analysis of the translation strategies confirmed the highly non-lexicalized and interjectional nature of Italian expressive forms. The main evidence of this is the fact that the majority of total substitutions involved a change from a lexicalized form in English to a non-lexicalized one in Italian. Only four cases --3% of the total --involved an English non-lexicalized form turning into a lexicalized Italian form. This outcome allows the present research project to align itself to the under-researched, emerging and highly specialized field of studies on the status of Italian ideophonic forms in comic books (D'Achille 2010, 100; Dovetto 2012, 203;Mioni 1990Mioni & 1992. What these studies have suggested is that, when compared to Italian, English seems to possess a ready-made 'real and extensive' ideophonic base (Feist 2013). This is because the language has a malleable morphology that naturally welcomes neologisms; a large number of monosyllabic words; and a tendency to easily convert words from one class to another (Bueno Perez 1994, 25;Mioni 1992, 92;Newmark 1996, 54). Italian, on the other hand, does not show the same ease when dealing with sound symbolic forms. This constraint is mirrored in the presence of repetitions that have constantly been used through the decades, particularly in the case of ideophones: the only difference being that, while Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM in the first two decades their occurrence is outweighed by strategies of addition and substitutions, later decades (1950s-70s) boast extremely high rates of repeated source text material.
Four different hypotheses were offered in the methodology and will be hereby discussed and supported by the outcomes of the present chapter: (1) Constrained approach examined: the first hypothesis suggested that words inside balloons and black and white strips are usually expected to involve a higher localization rate due to the ease with which the content of the cartoon can be changed. Words outside balloons and colored strips, in contrast, usually involve a Chart 6: Translation strategies distribution for ideophones and ideophonic series.

Chart 7: Translation strategies distribution for interjections.
Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM higher rate of repetition. This is in line with the 'constrained environment' imposed by the comic's format, a factor often mentioned by scholars in the field (Grun and Dollerup 2003;Noss 2003, 51;Zanettin 2008, 20). Although recent scholarly works (Borodo 2015 andZanettin 2014, particularly) started highlighting the comic's visual and verbal characteristics more as opportunities for creativity rather than as constraints, it is also apparent that the layout of the Disney comic book has had a decisive influence and impact on the translation of the words within its pages. In the term 'layout' different aspects are included: in particular, limitations imposed by balloon size, lettering, images, captions, color, and pagination (Zanettin 2014, 1) are all to be considered when evaluating possible changes to the source text. These limitations were obviously particularly debilitating during the first decades in which comics were translated, while technological advance brought a certain level of freedom from an editorial point of view --evidence of which revealed itself through the rise in the use of alternative strategies rather than repetition around the beginning of the 1980s (Ficarra 2012, 31).
This first hypothesis was confirmed. The data offered earlier in the chapter showed that forms that were originally inside balloons tend to feature not only more repetitions but also to undergo deletion more than twice as often as their counterparts found outside (28% --compared to 11% for forms outside balloons). This particular constraint involving balloons also had another impact on the translation rate of interjections, which are often included within balloons and are therefore more often likely to undergo strategies of substitutions and addition rather than repetition. This has allowed translators to localize interjections as much as possible for the Italian readers and, although English ideophones expressing emotions have increased in frequency through the years, typical Italian interjections are still present. So, in a way, both their consistent presence within the balloons and their easily translatable nature allowed Italian interjections to 'survive' through the years and decades of pressing Anglophonic supremacy (roughly between 1953 and the early 1980s). Chart 7 showed that, in the case of interjections, repetitions only represent the 11% of the total, surpassed by total substitutions (15%), additions (21%), and partial substitutions (40%).
In the case of Disney comics published in Italy, layout-imposed constraints were evident from the start. The decision to accompany each comic cartoon with rhyming couplets meant that somehow these couplets were, until they were abandoned in July 1935, the main story tellers, while the image was somehow considered secondary or as a simple accompaniment to the words below it. The evidence of this was the higher rate of repetitions recorded in the first stories published (roughly until the end of 1935).
The findings also show that there is a prevalence of repetitions in colored strips, with roughly half of the strips in color (51%) involving cases of repetition.
Finally, the change in format, from a newspaper size to a smaller size, that took place in 1949 also had an effect on the translation of the Disney stories. The decision allowed the Disney industry in Italy to survive times of financial crisis but at the same time this meant that the large cartoons originally designed for publication in American newspapers had to be reduced in size, which could be one of reasons for the rise of repetitions following the format change.
(2) Localization over time: this second hypothesis emerged after noticing, on first inspection, that, in early comics, forms seemed to be localized more often when compared to later stories. Hence, an exponential retaining of English sounds through the decades was hypothesized. One of the subsequent aims was then to discover if this process happened gradually or if it involved a sudden switch to English sounds to the detriment of Italian sound symbolism. The historical and linguistic motivations behind these choices would have been part of the analysis.
A pivotal switch noticed through the corpus analysis happened between 1953 and 1957, a period after which English markers started being retained rather than localized as in the past. This small time window marks the beginning of a central Anglophonic influence that shaped the world of comic books published in Italy and starts legitimizing English as a major player in the language of the Disney comics industry in Italy. Post-WW2 and after the creation of the 'iron curtain', Italy indeed entered, economically and culturally, the American sphere of influence. The censorship era brought in by Fascism was over, and editors and translators were able to use strategies of foreignization, which will have had an extreme impact on the use Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM of English ideophones in Italian comics. As a matter of fact, the most common English ideophones started appearing in Italian dictionaries during the 1960s.
The analysis carried out in this chapter provided significant insight into the diachronic changes in these forms. From a rate of 20% in the 1932-42 period to 73% in the 1970s, strategies of 'repetition' have been constantly on the rise --a tendency also noticed by Valero Garcés (2008, 243) in regards to English to Spanish translations of ideophones in comics. The time-frame which showed the highest percentage of repetitions was the 1970s, which coincides with the period of crisis for the Disney comics industry in Italy (Tosti 2011, 117). Post-1980s, this trend is totally overturned as strategies of 'repetition' have undergone a steady drop over the last 12 years covered by this investigation and have experienced a 30% drop (from 70% to 40%) in favor of strategies of 'partial substitution'. This finding was unexpected, particularly considering the tendencies detected in the previous four time-frames. The drop in the rate of repetitions seems to happen in favor of two main strategies: partial substitution, which increases from 4% to 17%, and the use of English alternatives, which increases from 3% to 10%. The reason for this drop is difficult to identify at this stage but the technological advances experienced since the end of the 1980s have, most certainly, had an important role in this.
In summary, this second hypothesis was only partially confirmed as the translation shift in the last 12 years had not been predicted.
(3) Translator's dilemma examined: in the third hypothesis it was suggested that although strategies of substitution are expected to be favored in the case of expressive forms in order to recreate the whole aesthetic effect through phonological and syntactical means familiar to the target text culture (Ippolito 2008, 253), strategies of repetition were expected to be the most used throughout the eight decades analyzed. This is due to (1) layout constraints and (2) linguistic and cultural reasons that caused English forms to feature more prominently as the years passed. Indeed, in the case of ideophones this hypothesis was confirmed, as chart 5 showed that, when considering the use of strategies in all the stories regardless of time-frames, repetitions have the highest occurrence rate (49%). As expected, in the case of interjections, the strategy occurred with a percentage of 11%.
(4) Italian vs Spanish translation strategies: the last hypothesis suggested that some of the findings of this research project could be compared to the ones proposed by Valero Garcés (1996; in her project on the translation of onomatopoeia in comics from English to Spanish.
The following section will analyze all the main outcomes suggested by Valero Garcés and state whether these are also applicable to the current study.
(1) Valero Garcés identified two main strategies: the ideophone is either translated or left unchanged (Valero Garcés 2008, 241). This outcome is applicable to Italian Disney comics in translation. Indeed, when looking at the total distribution of strategies for ideophones only, the majority of these involve either a translation or a repetition. Creations and deletions only make up for a total of 7%.
(2) English sound symbolism usually tends to be retained, particularly when the ideophone is outside balloons (Valero Garcés 2008, 240). Since the start of the importation of American comics during the 50s the prevailing tendency has been to leave English sounds untranslated, with minor localisation of certain types of ideophones. This statement is only partially applicable. When ideophones are found outside balloon they indeed tend to be retained and after the 1950s repetitions start to systematically rise. As mentioned before, comics in Spanish ---or at least the ones analyzed by Valero Garcés --did not seem to involve the revival of substitution strategies post-1980s detected by the current project.
(3) Mechanical or artificial sounds are usually retained while 'sounds produced by animals, unarticulated sounds produced by humans (including interjections) and sounds used to show feelings or attitude' tend to be translated more often (Valero Garcés 2008, 241-43). This statement is applicable. As stated during the analysis of the different strategies used for each ideophonic type, mechanical sounds had the highest Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM percentage of repetitions while those belonging to the 'animal' or 'human' categories had the highest degree of localisation. As explained earlier, this could be due to the fact that mechanical events have a higher chance of being represented with non-lexicalised ideophones and of being placed outside the balloon. Corpus data shows that mechanical ideophones have the higher percentage of non-lexicalised forms while the 'human' category has the lowest.
(4) Ideophones in black and white stories tend to be localised more often (Valero Garcés 2008, 240;also mentioned in Martinez Fuentes 2003). This is also applicable to the present study. As shown earlier, ideophones in coloured stories have a higher chance of being left unchanged. These findings also show that, indeed, there is a prevalence of repetitions in coloured strips, with roughly half of the strips in colour (51%) involving cases of repetition.
(6) Sometimes creative ideophones in English are translated with better-known English ones, which have become part of Spanish tradition (see 'sob' or 'sniff'). This shows that some English ideophones are perceived as being part of the Spanish language (Valero Garcés 2008, 244). As previous research showed (Pischedda 2010), some English ideophones are perceived by Italian people with a lower level of English as being of Italian or unknown origin. Strategies that use English equivalents that are considered simpler to understand have been shown to be gradually employed, particularly after the end of the 1950s. So this statement of Valero Garcés is also applicable to Italian translation of English expressive forms in Disney comics.
(7) Spelling is often adapted. See 'he he' for laughter translated as je, je (Valero Garcés 2008, 243). The use of strategies of 'addition', that is the substitution of certain graphemes in order to provide a more familiar target text environment was found within the early time-frames. 'Addition' is one of those strategies that has shown to have been consistently used throughout the time-frames. Despite showing stable decrease through the time windows (22%-->24%-->17%-->11.5%-->12%), it has always been a fruitful way to localise ideophones and interjections in Disney comics published in Italy.
(8) A higher percentage (75%) of English ideophones has been retained in the last 15 years (Valero Garcés 2008, 243). This statement is not applicable because the current research project does not take into consideration publications post 1992 (Disney comics stopped being translated in the early 1990s). Nevertheless, it would be interesting, for future research, to evaluate whether this statement also applied to the Italian comics market. As things stand now, and following the statistics of the last time windows, strategies of repetitions would have been expected to continue to drop post-1992, but without any data this cannot be confirmed.
(9) A higher percentage of English consonants (particularly the non-Spanish letter /k/) has been retained during the last ten years (Valero Garcés 2008, 245). Again, this is an interesting remark that cannot be confirmed or refuted as the current project did not focus on the last ten years of Disney comics in translation.
The data shows that out of nine outcomes, six appear to be true while one is only partially true and two more are non-applicable as there is not enough data available to confirm them. Taken as a whole, Spanish and Italian do indeed seem to share similar trends both when it comes to translating and creating ideophones and interjections in comics. This phenomenon can be explored from both linguistic and graphical points of view.
Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM From a linguistic point of view, both languages tend to use strategies of 'addition' as the best way to modify expressive forms for the target audience through the addition or removal of Anglophonic markers. On the same note, with reference to these two specific studies (Valero Garcés's and the present one), both languages seem to show a high reliance on English forms that have become widely understood, as evidenced by the growing use of strategies including English equivalents. Furthermore, Spanish and Italian both contain an extensive array of interjectional forms, evidence of which is provided by the higher localization rate of interjectional forms when compared to ideophones. The nature of the sound and what it represents has been shown to influence the strategies used. Therefore, non-lexicalized sounds that are visually and sonically more iconic (i.e., mechanical ideophones) will show a lower degree of localization as they are already instinctively considered symbolic enough for the audience to instinctively grasp what they refer to, especially with the additional help of visual cues.
From a graphical point of view both Spanish and Italian comics seem to accept more localization when the strips are in black and white and if ideophones are within balloons.
More data would be needed to investigate the validity of the last two 'non-applicable' statements and, also, more research should be carried out to provide evidence that the results of this current project also apply to other genres within the comic book industry, such as graphic novels, comic books created strictly for an adult audience and picture books including comic strips. Nevertheless, the findings suggest that Italian and Spanish share some attitudes in the treatment of sound symbolic forms in translation and that both place a high degree of reliance on English linguistic material as suggested by various scholars in the field (Bueno Perez 1994, 25;Dovetto 2012, 208;Pischedda 2012;Valero Garcés 2008, 239). This is particularly shown, in the case of Italian, by the presence of truncated forms in Italian comics that attempt to imitate English expressive forms ending in consonants, a phenomenon that does not appear to occur within the Spanish language of comics. Consonant-ending words are readily perceived to be more iconic as linguists have provided evidence that words ending in consonants are, in most languages, somehow considered more expressive as they are characterized by 'an inherent confining and limiting effect which words with a vowel sound do not possess' (Reid 1967, 47). These results would go hand in hand with recent scientific studies (such as Assaneo et al 2011) that have demonstrated that 'onomatopoeia across languages are indeed linked to the sounds they refer to by imitation' (Assaneo 2011, 7) and not just through cultural and linguistic arbitrariness.

Conclusions
The analysis has provided new original data on the use of ideophones and interjections in Disney comics since their first publication in the 1930s. The research has involved a close examination of the way this complex and overlooked set of forms has been created, employed and translated.
When it comes to assessing the place of sound symbolic forms within the discipline of translation studies, one soon realizes that it is not always a mere matter of recreating the same effects on the reader through phonological and/or morphological patterns. It is true that a 'bow-wow' can be easily converted into the Italian bau bau, but the process is not always this straight-forward. The constraints imposed by the format itself have to be considered and, if one adds to the mix the various cultural influences and interferences, the process of creating and translating these forms is hardly a predictable one. The role of English as lingua franca and the Anglophonic monopoly within media studies have a lot to answer for (Chmielewska 2011, 3;Crystal 2007, 308;Fanfani 2010). On top of this, if one considers the English language propensity when it comes to creating and modifying these forms and the evident ease in which this is achieved, it is clear why the English language has had a fairly easy job from the start. Romance languages such as Italian and Spanish, on the other hand, had to try to figure out how to interpret and process these forms in a matter that felt useful and fruitful and that could make the most of the morpho-phonological devices offered by the language. And this apparent 'struggle' is indeed mirrored in the high variability of the translation strategies detected throughout eight decades: an initial even distribution of the three main translation strategies (addition, repetition and substitution) in the pre-war period  is followed by Brought to you by | University of Leeds Authenticated Download Date | 3/5/20 3:11 PM an increase in the rate of repetitions, which reaches its peak in the 1970s, which then drops again in the last time window .
The discoveries regarding the diachronic use of translation strategies allow the researcher to place the translation of sound symbolic expressive forms within the concept of 'transcreation', which has been more and more relevant in translation theory (Melby et al 2014). The term was coined during the 1960s and 1970s to refer to those texts that required a translation that retained the emotion of the source text as much as possible, a requirement often laid down within the marketing and advertising industries. As a matter of fact, the term is a portmanteau for 'translation' and 'recreation' and it aims to stress that the preservation of the message, the style, the images and the emotions of the source text is, in certain fields or media, much more central in order to fully localize the contents (Balemans 2010). This is not to deny that all translations of texts aim to reach this kind of equivalence between the source and target text, but the aim of using the term 'transcreation' is to stress the importance, for certain types of texts, of the need to be expressive and to reach the hearts and not just the minds. This results in some types of texts requiring a higher level of transcreation than others (Balemans 2010). Transcreation is therefore the adaption to the text that is not a 'straightforward transfer of perceived meaning' (Melby et al 2014, 395), a process that scholars have defined as 'reconceptualization', resulting in a translation that is only loosely tied to the original and its details (Ibid.). For these reasons, I posit that the role of the translator of sound symbolic forms in comic books can be defined not just as a 'semiotic investigator' (Celotti 2010, 33) but, rather, as a 'creative semiotic investigator'. The translator of multimodal texts is indeed an interpreter of different signs and how these shape narratives, as his/her aim is that of reaching the 'equivalence of the effect' rather than an 'equivalence of the lexicon, grammar or form' (Newmark 1996, 48). This is particularly pertinent in a context, such as multimodal texts, in which the sensoriality and the interplay between different verbal and visual media is pivotal and represents a primary source of meaning. A rupture of balance between semiotic elements could jeopardize the entire reading experience. The term 'semiotic investigator' stresses the variety of sensorial inputs and perceptions available to the translator when trying to localize comics. Adding the adjective 'creative' stresses even more the need for a total recreation and reconceptualization of the source text material in the case of ideophones and interjections. This is true particularly when one takes into consideration that the translation of these forms is a clear 'problematic task' (Noss 2003, 53) that faces the translator with a 'wide variety of strategies' and options (Ibid.). This variety of options is a result of the intrinsic language-specific nature of these forms. On top of trying to imitate senses, and hence performing 'a referential function' (Casas-Tost 2014, 41), ideophones, in particular, are 'highly expressive words which also have tremendous allegorical potential' (Casas-Tost 2014, 41) since they can evoke specific feelings in the reader. This peculiarity, together with their highly-marked phonological status, lexical and grammatical uniqueness and genre-specific conventions, means the translator is confronted with a real 'translation dilemma' (Noss 2003, 53) that certainly adds a certain level of difficulty to the translation process (Bueno Perez 1994, 23).