Abstract

Making terminology accessible to a non-expert audience is a challenging task for specialized writers. Indeed, they need to strike the balance between finding the right level of popularization for the intended addressee and keeping maximum precision. This task is even more daunting when specialized content is conveyed with the use of graphic elements. This paper presents some reflections drawn from a recent cooperation with the Swiss Federal Statistical Office (FSO) aiming at finding new ways for providing accessible institutional information within an action research paradigm. The analysis focuses on selected chapters of “Statistical Data on Switzerland”, a publication addressed to the broad audience that describes the country by means of statistics. It is compared with a new, provisional version drafted according to the standard of plain language, as well as with the “Statistical Yearbook”, that is its counterpart for a specialized audience. Moving from the French versions of the aforementioned publications, the focus is placed on the use of single- and multi-word terms in both textual and graphic parts. They were identified with the help of a term extractor and validated by a field expert at the FSO. The analysis shows that the FSO tries to adjust terminology according to its public, thus controlling the quantity and level of technicality of the terms used. Furthermore, adding definitions seems a useful way for ensuring precision in specialized texts and explaining unavoidable terminology in popularized texts. Finally, controlling the level of technicality and the presence of terms in graphic elements, as well as ensuring text-image coherence is key when enhancing the level of accessibility.

Keywords: plain language, terminology, graphs, popularization, accessibility, statistics, switzerland

1. Introduction

Switzerland has a long-standing tradition of clear institutional writing, above all in legal matters. At the beginning of the 20^th century, Eugen Huber, the father of the Swiss Civil Code, first introduced the concept of popular legislation (1914), which is still used today. This ideal has been recently enshrined in the law; not only civil servants should strive for clarity in institutional writing but they should also receive appropriate training.[1] Over the last decades, several initiatives were taken to address quality in institutional texts, such as guidelines and seminars.[2] As regards the legislation, it is worth mentioning the thorough legal and linguistic revision work carried out by the Internal Drafting Committee (Höfler 2015).[3] This is in line with the many plain language initiatives that have been undertaken globally by several governments and institutions. However, when it comes to administrative texts on specialized subjects, some efforts still need to be made, as also shown in recent studies on information leaflets for old-age and disability insurance (Felici and Griebel 2019; Griebel and Felici 2021). Following a joint project with the Swiss Federal Statistical Office (FSO), our paper presents a case study to work out effective ways of producing accessible multilingual information for a lay audience.

According to the International Plain Language Federation, “communication is in plain language if its wording, structure, and design are so clear that the intended readers can easily find what they need, understand what they find, and use that information”.[4] This means that the simplification process, which leads from a specialized text to its popularized version, turns metaphorically speaking into a sort of intralingual translation. When it comes to statistics, graphs also play a crucial role in providing accurate and understandable information. Transforming textual content into graphs, and vice versa, may be seen in turn as a form of intersemiotic translation. Although the Swiss institutional context presents a fertile ground for investigating interlingual translation in accessible communication, for reasons of space, this paper focuses exclusively on the French source version of selected multilingual publications.

The paper is structured as follows. After providing an overview of our institutional partner, which also serves as the contextual background of our study (§2), we draw our attention to the simplification of terminology for the general public (§3). Data and methodology follow (§4) together with a discussion on the main findings (§5). Finally, our concluding remarks highlight a continuum in the popularised versions of the FSO and shed light on future steps of the analysis (§6).

2. A case study: the Swiss Federal Statistical Office

This study is part of the MACSI project (Multilingual Accessible Communication in Swiss Institutions).[5] Drawing on an action research approach (Saldanha and O’Brien 2013: 174), we carried out our investigation in close collaboration with and also to the benefit of our research partners. This participatory approach to research “is distinct in emphasizing interaction and involvement of the subjects of inquiry to effect change” and the “research goals involve not only understanding and describing but also changing and improving a situation” (Mellinger and Hanson 2022). More specifically, this project aims to further spread a culture of accessibility within Swiss institutions, while gaining precious data on processes, hurdles and best practices in drafting multilingual plain texts.

The FSO has mainly an informative function. It surveys and describes the economic, social, territorial, and environmental situation of the Swiss Confederation, thus providing solid data upon which decision makers can discuss their policies and citizens can be informed about the evolution of their country and society. In this respect, the FSO strives to popularize the results of its activity.

Our cooperation started in December 2020 with explorative interviews, focus group discussions and interviews with staff from both the publishing unit and the internal translation service, followed by a larger-scale survey. The objective was to explore the perspective of our partner and identify their needs in terms of accessible communication, as well as potential internal barriers. In September 2021 and November 2022, we organized two workshops, which were attended by around 20 writers and translators[6] and, more recently, we supported the FSO in creating internal plain language guidelines for their text producers. The results confirmed Maaß’ insight, in so far as the producers of texts in plain language

are mostly domain experts that are given the additional task of writing in a comprehensible way. Some are specially trained, but this training seldom goes beyond a very limited number of hours and does not imply a consistent monitoring of the text practice (2020: 180-181).

This is the case of the FSO, where texts are produced mostly in French or German by subject matter experts or statisticians with little or no training in communication and text writing. The texts are then translated into the Swiss official and working languages by translators who face the challenges of interlinguistic transfer of both specialized and plain contents. In this respect, specialized writers are our main working partner concerning the intralingual translation dimension mentioned in Section 1.

The object of our analysis is a selected publication, the Statistical Data on Switzerland[7] (SD), a 52-page document addressed to a large lay audience and issued every year in five languages on a range of different topics (Swiss population, education, economy, transport, social security, and so on). This text was selected with our institutional partner for its essentially informative nature and because its main addressee is the general public. As part of a complete revision of this publication in 2023, the aim is to further improve its accessibility to the lay public, in terms of content selection, language and use of visual elements.

The main difficulty for writers is to conceive texts addressed to “everyone” and to weight up previous knowledge of such a wide and undefined target audience. As Maaß (2020: 181) points out, plain language is often too subject-matter-oriented rather than user-oriented. To enhance user-oriented content, four personas instead of a “general public” were defined in the revision process of SD. These personas are prototypes of the potential target readers with specific profiles regarding their sex, age, origin and educational background. The FSO defined two women and two men in different age groups, with a Swiss or international profile, different educational backgrounds, and previous knowledge in statistics.

This approach is widely used in other domains, such as software development or marketing, but, to our knowledge, it is new in the field of plain language research and practice. By defining concrete personas instead of an undefined group of target readers, text producers can more easily control for potential barriers in language and content presentation while writing.

In this paper, we focus on a specific aspect related to this intralingual translation process, that is, how to cope with specialized terminology and strike the balance between precision and comprehensibility for a wide public.

3. Languages for specific purposes: from terminology towards its popularization

The first models for defining and classifying languages for specific purposes (LSPs) date to the 1950s and 1970s (Arntz and Picht 2014: 11-37). What all these models have in common is that the boundaries between standard and specialized language are fluid, LSP itself also represents a continuum of varieties that are closer or more distant to the standard language, and that there are many LSPs depending on the subject area.

Hoffmann (1985) proposes an LSP model with five levels of technicality with a decreasing degree of abstraction. “A” corresponds to the highest level of abstraction, for example through the use of artificial symbols and mathematical formulas, whereas “E” corresponds to a very low level of abstraction and technicality. According to Hoffmann, each level of abstraction is associated with a specific target reader. Texts belonging to level A are meant for expert-to-expert communication, while texts at level E are used for communication between experts or professionals and laypersons. If we rely on Hoffmann’s model, we can safely state that the FSO communicates on several levels of technicality and addresses different audiences with different levels of expertise. However, even at lower levels of expertise, writers must cope with the heterogeneity of their audience, which is reflected in the personas approach presented in Section 2.

Terminology is one of the most investigated aspects in research on language for special purposes (LSP). With reference to the standard DIN 2342 (2011), it includes the totality of terms and their designations in a subject area. Accordingly, the concept of “terms” refers not only to defined–standardized or not–and technical expressions which are pragmatically agreed upon (Arntz and Picht 2014: 27), but also to technical phraseology consisting of nouns and verbs, and/or prepositions, and so on (ibid.: 34-37). The degree of texts’ technicality is determined by the communicative situation, the senders and receivers as well as the text type. Recent trends in terminology have shifted the focus from the Wüsterian prescriptive approach to corpus-based terminology with terms being studied in the context of a communicative situation. Sager pushes the boundaries of terminology and maintains that “one concept can have as many linguistic representations as there are distinct communicative situations” (1990: 58), thus legitimizing synonyms. Along the same lines, Cabré (2003: 183) defines the terminological units as representing cognitive (the concept), linguistic (the term) and communicative (the social context) dimensions.

Our contribution intends to address the “communicative” dimension of terminology, thus questioning its boundaries when introducing highly specialized topics and terms to the general audience. The use of terminology is also explored in graphs, where the alleged accessibility of visual elements is challenged or needs to be consistent with the terms used.

4. Data and methodology

As stated in Section 2, this study is based on the publication Statistical Data on Switzerland in its original version (SD-original) and in its preliminary further simplified version for the 2023 revision (SD-new). As a term of comparison, we also took into account the Statistical yearbook (SY), a technical publication that presents the same content overall but in a more extensive and comprehensive way. The SY is primarily addressed to an expert target audience and includes specific terminology.

Table 1: Selected publications

The three versions display a similar macro-structure and are divided into thematic sections, which provide information on a specific topic through statistical data, both in textual and graphic form.

In this study, we focus on three sections:

Table 2: Sample tokens measured with SketchEngine

From the terminological perspective, it is interesting to note that only SY systematically uses definitions. Each thematic section is followed by a generous glossary defining the main terms used in the previous pages. In SD-original, despite its broad audience, there are no glossaries, supposedly because the publication is deemed to display a sufficient level of simplification and to avoid difficult terms. For SD-new, it was decided to add definitions of unavoidable terms alongside the core text, in boxes. However, at this stage of the simplification process no “terminological boxes” have been added yet to the text.

In general, using glossaries can be a good way of making necessary terminology comprehensible for lay readers. At the same time, it is important to consider their level of usability in terms of ergonomics (Lutz 2015: 152). If readers need to go back and forth several times between the main text and the glossary, the reading and comprehension process is constantly interrupted and the mental representation of the content may be hindered. In this respect, the concept used for SD-new seems appropriate.

As far as graphs and visual elements are concerned, all three publications make use of them. Once again, graphs are not yet available for SD-new, or only in a preliminary version, therefore our analysis will be mostly centered around SY and SD-original for the time being. The idea is to work not only on simplifying content, terminology and language, but also visuals. The results of this preliminary analysis are expected to facilitate the visual simplification process.

Before attempting our analysis, we extracted terms and multi-words via SketchEngine, by means of the Keywords function. Keywords are words that are significantly more frequent in a text sample or in the focus corpus than would be expected in a large general reference corpus (Scott and Tribble 2006). They are used to identify what is “key” and prominent to the focus corpus in comparison with the reference corpus.

SketchEngine calculates keyness with the simple maths formula (Kilgariff 2009), where the relative frequencies of the focus corpus (FC) are divided by the relative frequencies of the reference corpus (RC) and a smoothing parameter N is added to both frequencies.[8] This parameter is used to solve the “zero” problem when certain words are present in the focus corpus but absent in the reference corpus. Moreover, choosing different values for the “add-N” parameter will highlight different frequency ranges (ibid.). Low N values will return unusual words in the reference corpus, higher values will focus on more common words.

To extract terminology, namely words that are unusual or rare in the reference corpus, we set the N parameter to 1 and extracted three keyword lists by comparing our three small corpora (SY, SD-original, SD-new) to the French Web corpus 2017 (frTenTen17) available on Sketch Engine. The French Web Corpus is a 5.7 billion words French corpus made up of texts collected from the Internet and is meant to be representative of general language. The corpus belongs to the TenTen corpus family,[9] a set of web corpora built using the same method with a target size 10+ billion words. As a further term of comparison, we also extracted keywords by comparing our corpora with the French Project Gutenberg corpora 2020, also available on Sketch Engine. The corpus is made up of free ebooks available in the Gutenberg database in April 2020[10] and should be also representative of non-technical language. To exemplify the output of this analysis, we report in table 3 below the first 15 candidate terms that are prominent in SD-new compared to French Web corpus 2017. A high score indicates the words that are key in the focus corpus and rare or unusual in the reference corpus.

Table 3: Keyword list of SD-new (focus corpus) compared to French Web Corpus 2017 (reference corpus)

Sketch Engine also allows for the extraction of multi-word expressions (MWE), by matching the multi-words which appear more frequently in the focus corpus than in the reference corpus with the term grammar, a set of rules written in CQL (Corpus Query Language) that define the lexical structure of MWE, typically noun phrases.[11] We set the same N value to 1 for the extraction of MWE and we extracted keywords list for each of our focus corpus (SY, SD-original, SD-new).

Once we extracted the keyword list, we selected terms and MWE by setting a threshold of the first 50 candidate terms or MWE according to the frequency range per each list. As our focus corpora (FC) deal mainly with economics, banking, insurance and politics, we only selected terms and MWE pertaining to these semantic fields. We intentionally removed from the first 50 candidate terms all the names of Swiss parties or institutions (Raiffeisen, UDC, PLR, PDC), which were peculiar to our focus corpus and clearly showed a high rank in our keyword lists. Once we agreed on the candidate terms and MWE, we asked an FSO expert to validate the selected terminology. We then analyzed the extracted terms from a qualitative perspective, by taking into account their contexts of use. During this qualitative phase, we also delved into the visual dimension and reflected on the use of graphs and tables within the three publications.

5. Results and discussion

The validation by an FSO expert provided us with a list of 36 terms and 30 MWE, which will be discussed in Section 5.1 and 5.2, before the analysis of graphs and tables (Section 5.3).

5.1 Terms

Table 4 below shows the key terms that were extracted and validated in our focus corpora, with their raw frequencies.

Table 4: Validated terms across the three focus corpora (SY, SD-original, SD-new)

The terms highlighted in green are common to all three corpora, those in orange are present both in SD-original and SD-new, and those in blue both in SY and in the SD-new. Before attempting any interpretation of table 4, we need to consider that the absence of certain terms from one corpus or the other is due to the threshold of the 50 candidate terms selected from the keyword lists. Thus, a term like revenu (revenue) is indeed present in SD-original, but it has a lower keyness value in our focus corpus because it is ranked after the first 50 terms in both our keyword lists.[12]

As expected, table 4 shows more key terms in SY because it is the expert-oriented corpus and also because of its size. The higher number of terms in SD-new compared to SD-original may be interpreted as a sign of complexity rather than simplification or of increased lexical variety. If lexical variety is more prominent in informative and creative texts compared to specialized ones,[13] we have to point out that SD-new is a rather small sample of less than 1000 words. In SD-new, topics are presented one after the other in an abridged version, with the intent of providing the maximum information in very few words. This operation makes the simplification task harder. While simplified texts tend to use explanations, paraphrases, notes, and glossary entries, which make the final text undoubtedly longer, SD-new responds to a very marked need for compactness.

Looking closely at terms, référendum (referendum) and votation (voting) are the only words present in all three corpora. They are defined in the SY glossary, but not in SD-original. Indeed, they belong to the domain of politics, which is more familiar to the general public compared to banking and economics. Referendums and public elections are firmly anchored in the Swiss political culture and public consultations may be initiated on many current issues. Therefore, residents of Switzerland will not perceive these terms as technical. However, it is worth mentioning that the FSO will include among the personas someone who does not live in Switzerland. Accordingly, this may lead to reconsider the technicality of the two terms. “Referendum” and “popular initiative” are listed in the glossary of the SY together with conceptual differences between the two terms and explanations on their legal basis. These explanations could also be maintained in the final version of SD-new.

In both SD corpora, the word ménage stands out, meaning exclusively household in the context of our texts, that is ménages monoparentaux (single-parent households) and ménages de personnes de 65 ans ou plus (households of people over 65). This is in contrast with the reference corpus, where ménage encompasses less technical meanings like “cleaning lady” (femme de ménage), “housekeeping” (ménage, repassage, vitres), or figurative senses of the word “union” (Ostéopathie et rationalité scientifique feraient-elles un bon ménage?). Given the specialized focus of our corpora, the term is used with reference to demographic and economic topics. In SD-original, it collocates above all with the words revenu (revenue), dépenses (expenses), budget, which are modified by the term ménage:

Example 1: Term ménage in SD-original

The sentences are not complex, but the specialized nature of the term’s collocates may hamper full understanding. Looking at SD-new, ménage occurs in very short sentences as a noun or as the head of a prepositional phrase and typically refers to generic households or families, which are qualified as “households of people in working age”, “households with debts” and so on. Only in two occurrences, it works as a modifier of the word budget.

Example 2: Term ménage in SD-new

The other two terms, dépense and revenu, are visually exemplified in a pie chart in both versions. The preliminary chart in SD-new shows the symbols of the type of expenses, such as for houses, food, clothes, and so on, thus making more evident what the expenses are for (example 3).

Example 3: Visual representation of composition of the household’s expenses in SD-new

We also find an explanation of certain expenses, like the dépenses obligatoires (mandatory expenses). The new version explains that elles doivent être données à la société (they must be paid to the government and they may be used to pay for schools, roads, and pensions). Along the same line, revenu is accompanied in SD-new by an explanation of peoples’ different earnings. Endettement (going into debt) also receives a kind of gloss in SD-new, being contextualized as a source or cause of poverty. In SY, which is a publication primarily aimed at experts in the field, the two terms are widely scattered across the text, often modified by another noun or adjective (dépenses obligatoires, d. de transfert, d. de consommation, revenu brut, r. disponible, r. equivalent, r. de ménage). Their meaning seems to be taken for granted, but at the end of the publication they are listed in a glossary with their explanations.[14] The glossary is quite comprehensive and provides explanation even for the common word pauvreté (poverty), which is described in socioeconomic terms as the rate of poverty and as the poverty risk rate threshold.

Our terminological extraction did not return many prominent acronyms in SD-new.[15] PIB (produit intérieur brut), which stands for GDP (gross domestic product), is not explained and is assumed to be well understood. On the other hand, our keyword list returned high specialized acronyms in SY: Libor (London Interbank offered Rate) and Saron (Swiss average rate overnight), which are explained in the glossary. They belong to the monetary domain and find no mention in SD-original and SD-new.[16]

5.2 Multi-word expressions (MWE)

Table 5 shows the MWE and their raw frequencies that met the 50 candidates’ threshold in our keyword lists. In this case, the tool returned no MWE that was common to the three corpora. We highlighted in blue the MWE prominent in SY and SD-new, and in orange the MWE prominent in SY and SD-original. In the abridged version (SD-new), each MWE constitutes a hapax because of the shortness of this document.

Table 5: MWE and raw frequencies

As already mentioned above, the term dépense (expense) is postmodified by adjectives, thus also resulting in MWE. In SY, the expression is explained by referring to its destination: for instance, dépenses obligatoires are expenses used to pay for social services. As regards the dépenses publiques, it is stressed that they require “our contribution”, meaning the general public. The relevance of this “public” and “social” contribution to the dépenses publiques is further highlighted with the example of pensions. As life expectancy has increased, “we” need to pay for pensions to ensure people a better future during retirement. On the other hand, the expression recettes importantes (important revenues) is not explained. It refers to the revenues generated by the Swiss banks abroad, but the text does not offer an interpretation.

If we look at SD-original, personne seule (single person) qualifies single people under 65. It is quite straightforward and its meaning may be derived, as is the case with ménages monoparentaux (single-parent households), from the single semantic units. However, revenu disponible équivalent (equivalized disposable income), which is not mentioned in SD-new, deserves to be mentioned for the differences between SD-original and the SY’s glossary:

Example 4: Explanation of the term revenu équivalent in SY and SD-original

In SD-original, this term is introduced to highlight social inequalities in relation to income and it is explained how it is calculated, that is by subtracting the mandatory expenses from the gross household income and dividing the resulting disposable income by the size of the household. It is further explained as an index of people’s standards of living, which indicates that rich people can afford more equivalent disposable income than poor people. However, the paragraph lacks coherence and it may be difficult to reconstruct the definition of the term. The explanation seems to be much more effective in the glossary in SY, where–despite some terminology–we get to know that the term corresponds to the available income of a household, divided by the number of household members, which are equalized to a weight according to their age (a child is weighted less than an old person in terms of expenses).

Most of the other MWE in SY are accompanied by a gloss. However, these entries can themselves be complex and technical, which is appropriate in the case of an expert audience. This does not necessarily mean that expert readers need to receive more explanations, but rather that the FSO is pursuing a greater precision, since much of this data will be available and used internationally for further statistics.

With regard to popularization, the same terms are sometimes also used in the SD versions, but without explanation. The lay reader is not only confronted with an opaque technical term or MWE: the expression also lacks explanation and context. Therefore, glossary entries are indeed necessary, provided they are popularized and they contain only the necessary information for comprehension.

To sum up, our small sample showed an increased level of simplification when it comes to specialized terminology. The simplification is expressed in different ways: use of adjectives functioning as postmodifiers, examples of the use of a particular term/MWE, explanations, glosses, and in certain charts the use of symbols.

5.3 Use of graphs and tables

Visual elements are a central component of scientific writing (Miller 1998) and of specialized texts in general. They support technical communication and influence its degree of technicality. Graphs are often part of the fixed text structure of certain types of specialized texts, for example medical leaflets (Roelcke 2020: 135), and also essential elements of statistical texts. In expert-lay communication, the use of graphic elements makes the text more appealing and increases the readers’ motivation to engage with it (Cutts 2013: 247). They also help explain terms, as already mentioned in the previous section. In this respect, visuals are not only useful in specialized texts, where they fulfill a denotative purpose and respond to the need for compactness and precision; they can also serve in popularized texts to enhance clarity. It comes as no surprise that guidelines and the literature on plain and easy language tend to suggest the use of such elements along with text (Ministère fédéral de la Fonction publique de Belgique 2015: 64-66, Bredel and Maaß 2016: 271-296). However, the multimodal presentation of information can also hamper comprehension if it is not carefully designed. According to Bredel and Maaß (2016: 295-296), the function of the images (duplication, exemplification, explication, expansion, condensation) and the way in which the text-image coherence is established must be determined during the initial stage of text production.

In SD, the FSO aims at describing Switzerland through the lens of statistical data and using graphs and tables can be an effective way of illustrating, for example, proportions or the evolution of variables. Presenting the same content in a text format would sometimes require many more words and numbers and the outcome would be a less attractive and understandable text. Example 5, taken from the “Economic and social situation” section of the SD-original, shows the use of visual elements:

Example 5: Pie chart from SD-original

Pie charts are arguably accessible to a lay public and allow the reader to easily grasp the proportions of different components making up a whole. In this concrete example, both the “whole” and the “components” consist of referents and concepts that are part of the encyclopedic knowledge of common citizens. Déductions obligatoires (mandatory deductions) may be less transparent compared to the other terms, although it is defined in a footnote within the same pie chart. Similarly, there is a footnote for Autres biens et services (other goods and services), in which however the nature of the “sporadic earnings” is not included. The footnotes here may result in a “duplication” or “expansion” of information (Roelcke 2020: 137). However, while they increase precision, they may distract the reader’s attention and make it harder to process the text and the graph.

At the same time, effective visualizations presuppose being aware of the potential complexity they can convey. Our case study allowed us to spot some issues that need to be addressed in publications destined to a wide public:

1. Compactness
2. Technicality due to the presence of terminology within the graph
3. Integration of graphic elements within the text

5.3.1 Compactness

As far as compactness is concerned, using too many visuals can overwhelm the lay reader. The number of visuals may make it difficult to understand how they contribute to the content. Table 6 reports the number of graphs and tables used in the selected sections of SY and SD-original:

Table 6: Use of graphs and tables in SY and SD-original

Interestingly, the “intralingual translation” from a specialized (SY) to a popularized text (SD-original) does not lead to a reduction in the use of graphs and tables. On the contrary, we find the same number or more visuals in SD-original. This result is even more surprising if we consider that SD-original is overall more than one third shorter than SY (cf. Table 2), which means that the visualizations/text ratio is much higher in the former than in the latter. Besides these quantitative data, we found that only few graphic elements of the SY are carried over to SD-original without any change and that, overall, visuals used in SD-original tend to display a good level of transparency. They are mostly pie, bar, and line charts.

To the best of our knowledge, there is a lack of empirical evidence concerning both the level of accessibility of graphic elements and text-graphs coherence in plain language research. The aspect of accessibility has been addressed from the perspective of easy language (Poncelas and Murphy 2007; Bock 2019). Furthermore, Rink (2020: 333-338) shows good practices of integrating graphs in easy language legal-administrative texts, although these graphs can sometimes be more difficult to process (Pridik 2019).

5.3.2 Technicality

Another aspect to take into account is the degree of technicality that can be associated with graphs. Visualizations are not mere extralinguistic elements and, as expected, they always contain some brief texts, and often terminology. This means that graphs can entail the same terminological hurdles discussed in Sections 5.1 and 5.2.

Example 6: Line chart and table from the “Bank and insurance” section of SD-original

In example 6, the line chart on the left shows to the lay reader that interest rates have decreased from the 1970s onwards up to the present. However, the three variables could prove difficult to decipher. While the terms hypothèques (mortgages) and dépôts d’épargne (saving deposits) are probably understood by the general public, as they are also used in standard language, obligations de caisse (medium-term bank-issued notes) might be more obscure. This term is not explained in the publication and readers are expected to understand it. The content of the two footnotes does not enhance the comprehension either, as they provide further technical information about how the displayed data were collected. The same considerations hold true for the table on the right. It summarizes effectively the composition of assets and liabilities of Swiss banks in 2019, but it needs more than basic knowledge of economics to be understood.

The “Bank and insurances” section of SD-original does not contain any text, but only one graph and five tables. The function of the visual elements is thus exclusively “condensation”. The graph and tables may be difficult to understand if taken separately. Moreover, the text-image coherence is missing as there is no explicative text.

5.3.3 Graphs-text integration

This last consideration leads us to the third issue identified in the analysis. Stand-alone visual elements might be appropriate for specialized genres, where the readers are able to resort to their field knowledge and fill the informational gap with little inferential effort. Lay readers, however, need to be accompanied by the writer, who should minimize implicit information related to specialized content. Besides the extreme case of the “Bank and Insurance” section mentioned above, the presence of a textual explanation does not always suffice to fully popularize specialized content and terminology.

This difficulty of providing a graph with the appropriate text might be related to the training of specialized writers. As Sancho Guinda (2011: 118) points out, research and practical guidelines on how to effectively address the integration of visuals and texts in specialized communication remain scarce. It is not surprising that the same difficulty occurs when specialized writers create texts for a lay audience, as is the case for the FSO (Canavese, Felici and Griebel 2023). This research finding suggests the need of integrating these aspects in plain language training offered to field experts.

6. Concluding remarks and further perspectives

Making terminology accessible is perhaps one of the main challenges of plain language. Terms are inherently technical and an apparently easy sentence like “household size has changed over time” is incomprehensible if one does not know the full meaning of “household”. This is because terminology delimits and designates concepts within a particular area of specialized knowledge, thus helping professionals to communicate effectively. Substituting terms into everyday vocabulary on the basis of plain language rules promotes indeed accessibility, but it inevitably brings with it some losses. Going back to our example above, “family” would be a good lay synonym for “household”, but the idea of the house occupants regarded as a unit is inevitably lost.

Our collaboration with the FSO has shown that there are several issues to be addressed when it comes to the “intralingual translation” of terminology. The most problematic issue is found in the definition of the general public, as for the SD publication, because the audience is very heterogeneous in terms of background knowledge. Moreover, information has to be presented in a very condensed way, while explanations would lead to a larger volume of text. At the same time, “intersemiotic translation” into visual elements such as different types of graphs can enhance accessibility. However, they can also hamper understanding if they contain additional technical terms that are not explained, or if they are not coherently linked with the concurrent text. In this regard, terminology and complex graphs deserve special attention for being at the very heart of specialized texts. Using everyday words and keeping technical features to a minimum requires careful consideration in order to avoid lack of precision and altering the content.

Despite the small size of our sample, the comparison among SY, SD-original and SD-new has shown that the FSO tries to adjust terminology according to its public, thus using different strategies. The SY is addressed to an expert, often international audience, and has many technical terms, which are explained in a glossary at the end of each chapter. SD-original uses a moderate amount of terms because of its general audience. However, except for some footnotes, definitions are scarce for reasons of space. This gap is to be filled with SD-new that is meant to provide explanations of key terms in small boxes.

After the conclusion of this pilot stage, our future investigation will be extended to the whole publication.[17] This will allow us to gain better insights into the popularization of institutional writing. The next steps in research and collaboration with the FSO will involve interlingual translation. As Switzerland is a multilingual country, accessible communication in all its official languages (German, French and Italian), plus English, the international lingua franca, is an institutional duty. This places a special demand on writing and translation in order to maintain the same level of accessibility in all languages. It will also allow us to address accessibility issues multilingually, thus gaining better awareness of expert communication and its degree of technicality at the institutional level. In addition, SD-new will be evaluated by representatives of the personas defined by the FSO in order to consider the receiver’s side. This sort of usability test for SD-new is very relevant from a research perspective, as to our knowledge personas have not yet been integrated into comprehensibility and translation research. It also serves our action-research approach, as one of the FSO’s missions is to inform the entire Swiss population and interested persons at home and abroad about the current state of Switzerland.

About the author(s)

Annarita Felici is Associate Professor of Translation at the Faculty of Translation and Interpreting of the University of Geneva since October 2014. Her fields of special interest include legal and specialized translation, contrastive linguistics, corpus linguistics, institutional communication, plain language and barrier-free-communication. She was previously “Juniorprofessorin” of legal linguistics at the University of Cologne in Germany and spent over ten years in the UK lecturing translation, general linguistics and Italian as a foreign language. She has worked as a translator and as translation project manager in the area of linguistic validation. She is a member of Transius (Centre for Legal and institutional Translation Studies).

Paolo Canavese received a PhD in Translation Studies from the Faculty of Translation and Interpreting (FTI) of the University of Geneva in 2023. He is currently undertaking a postdoctoral mobility at Dublin City University, funded by the Swiss National Science Foundation. He also teaches translation and linguistics at the FTI. His current research focuses on translation in multilingual institutional settings and barrier-free State-to-citizen communication, both from a corpus-based and a sociological perspective. He also works as a freelance translator.

Giovanna Titus-Brianti received a Master’s Degree and a PhD in Italian Linguistics from the Faculty of Arts, University of Geneva. She teaches Italian, Linguistics, and Translation at the Faculty of Translation and Interpretation of the same University. Her research focuses on textual and contrastive analysis applied to translation, as well as the didactics of Italian as a mother tongue and foreign language.

Cornelia Griebel holds a PhD in Translation Studies from the Faculty of Translation Studies, Linguistics and Cultural Studies at the University of Mainz. Her current research focuses on legal translation, comprehensibility and plain language in administrative and institutional communication. Furthermore, she is a postdoctoral researcher in the interdisciplinary research project “Intercepting with interpreters” (University of Neuchâtel). She teaches legal and economic translation at the FTSK and research methodology in the doctoral programme ReMeTIS at the University of Geneva and she has worked as a freelance translator for over 20 years.

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

•  Previous article
• Next article