CLiP 2006
[Close Print-Friendly Version]
The TEI Internationalization Proposal is a multicultural and multilinguistic initiative undertaken from the TEI Consortium. It represents a “cultural challenge”: to develop an infrastructure through which it will be possible to translate the TEI Guidelines from English towards other languages and to store all the translations together.
The Text Encoding Initiative is a milestone in the Humanities Computing field. It is the most important worldwide initiative for establishing standard guidelines to codify texts through the use of XML tags. However non-native English speakers are disadvantaged in understanding the Guidelines.
There are already a few translated versions of the TEI Lite (in Italian, French, Spanish, Korean, Japanese, Chinese and Russian). Although they are useful documents, we must emphasise that the TEI Lite is just a short surrogate of the real Guidelines. Plus those versions are not yet synchronized and some of them (the Korean, Chinese, Japanese and the Russian ones) have kept the original English examples.
The TEI Guidelines are written in a modular way: each chapter deals with one main topic. Each paragraph contains descriptive prose and examples. There are “theoretical examples” (showing just the tags’ syntax) and “applicative examples” (the tags mark-up real texts). Our analysis will be focused on the latter.
This submission is focused on the issue of translating these “applicative examples” into Italian.
Although it is “relatively” easy to translate the Guidelines' prose, the element and attribute names, and technical descriptions for tag usage, it is more problematic for the examples. We have three options: keeping the English originals, translating them, or looking for corresponding Italian texts. This submission will address the problems and the main methods involved in the last possible solution.
The main aim of cultural localization is choosing an Italian text that preserves as much information as possible from the original one. The decisions by which this is done needs to follow various analytic and selective criteria: tag used, structure, genre, content and length of the text and, if possible, its date.
In order to illuminate these problems we will analyse four examples in particular (three taken from the seventh chapter of the guidelines, one from the sixth), indicating that for each one of them there are different dominant criteria; and that sometimes these criteria overlap in importance. We aim to demonstrate that it is not possible to establish a hierarchy of importance for these criteria. Differences between texts oblige us to privilege different criteria in different cases.
The first text [pdf] (chap. 7.1.3.) we are going to analyse is a passage taken from an English novel. It shows numbered divisions for texts (<div0>, <div1>, etc.) and their responding attributes (n, id, type). Therefore our Italian example should be taken from a novel (genre) made up of two books containing chapters and paragraphs (structure). Concerns about tags overlap with those about genre and textual structure. In this particular case the content is not so important as it is irrelevant for the argument presented in the guidelines.
Sometimes there are complications due to some “additional tags”, that are extraneous to the aim of the example. Here we find the <trailer> tag marking-up the ending part of the first book; it is not necessary to preserve it. We must maintain the <p> and <head> tags as the descriptive prose links to paragraphs dealing with these elements. They are core elements and we will find them in most of the examples.
Our second example [pdf] (chap. 7.1.4.) is an extract from a newspaper. Here the main feature is the org=“composite” attribute. It is very easy to find Italian texts that allow us to use the same tags and attributes related to the journalistic textual genre. Neither content is important, neither chronology. But how do we deal with the <soCalled> tag? In this case we can adapt the translation via rejection. The example aims to show the tag is placed inside the Quotation section. If the descriptive prose had indicated any link about it we would have preserved that element. If not, we would have to modify the prose as well, with all the complications that would involve.
Our third example [pdf] (chap. 6.3.3.) deals with the <soCalled> tag. We run into four different usages of this element. There are three sentences (we could translate and/or re-invent them) and a passage taken from an essay. Here, different usages are related to different communicative purposes of the text. In this case the discriminating criterion is not about structure but it is about content, that is the linguistic context of the sentence.
Looking for general conclusions, we could state that every time specific tags of this kind occur, it is necessary to privilege the content. The length of the text is also very important when there are many instances of the same element.
The fourth example [pdf] (chap. 7.2.1.) considers where chronological issues could be significant. The tag used here is the <trailer> of an ecclesiastical biography of the bishop of Tours. The text is very old, as we can see from the theme, the rhetorical language, and from the headings’ length. The criterion of textual extent means that we need to find an Italian text that is more or less the same age, where it is more likely to have headings and trailers of a similar length.
This research aims to clarify the criteria informing translation of examples. The cultural localization of examples cannot be undertaken by a normal translator. It needs a person with a genuine experience in text encoding and with a good literary background to understand and apply different criteria where necessary. Consequently, an interdisciplinary and intercultural collaboration is necessary in order to ensure the project is successful.