Petra-e Framework

[Nerio Cintron]

Thevery first draft of the Framework distinguished only 55 descriptors for the eight competences. In the course of the discussions during the meetings in Utrecht, Misano, Budapest and Antwerp, the number has risen to 127. This number gives an indication of the complexity of the job of a literary translator. But, on the other hand, it may also complicate the applicability of the Framework as a whole.

The definition and description of the competences of a literary translator resulted in a detailed list of competences comparable to the EMT model. But a list is not yet a framework, let alone a framework of reference comparable to the Common European Framework of Reference for Languages (CEFR), to which our Framework refers in the definition of the language competence. In order to become a framework the competence list had to be complemented by a learning line distinguishing several levels of competence acquisition, and a system of evaluation and assessment that allows for valid and reliable tests that would guarantee a legitimate level shift. From a unilinear (vertical) competence list the model would have to evolve towards a three-dimensional framework that contains

The first tension can be solved by admitting that the learning line operates within a didactic context, and that the necessary level shifts can be legitimated by intermediate tests adapted to the single levels, known from the regular bachelor and master programmes. When it comes to competence evaluation, however, the competences should be considered as final competences. The second tension can be solved by referring to the flexibility and the openness of the Framework; it is no problem when some descriptors are moved to an earlier or later phase of the learning line.

Talking about competences and learning levels presupposes a valid and reliable system of evaluation and assessment. The open and free character of the Framework should eventually be linked to an evaluation system that guarantees learners, wherever they learn their job, valid and reliable information about their competences. The implementation of such an evaluation system, however, could not be realized within the current project. It will be the object of a follow-up project.

The Framework is a practical, not a dogmatic tool. It is not meant to prescribe which way a translator should go to reach her/his goal, nor does it prescribe contents and didactic methods. It can be used as a swift tool for different applications.

For institutes offering programmes for literary translation the Framework is a handy tool for curriculum design. What kind of courses and content must be offered in order to educate competent literary translators on a certain level? For existing programmes the Framework offers an assessment tool: do all the courses in the programme contribute effectively to the education of competent literary translators? On a national and European scale the Framework offers a tool for the comparison of education and training programmes. This would allow learners to decide where to go for the further development of their knowledge and skills. If the Framework succeeds in developing a valid and reliable testing system, it could be used as a basis for the Europe-wide assessment of translation institutes. Every individual teacher can use the Framework as a diagnostic tool for her / his courses: what does the course offer, and on what level? What is missing? etc. In the end the Framework should help teachers and trainers to design better courses and workshops.

The Framework can support networking between several institutes and providers of courses. It can help to discover complementarities between programmes. Probably no single institute will be able to offer all competences. The Framework helps to discover where a learner can find the competences missing in their home institute.

Especially the online version of the Framework can (in principle) offer bibliographical references, tips for courses, exemplary tests, best practices etc. for every single descriptor. It can provide new ideas for didactic methods and training tools.

The Framework can be used by translators in discussions or negotiations with publishers. It can help to make clear what kind of competences are needed to translate certain texts. The Framework may contribute to more professional visibility of literary translators and improve their economic position.

The work on the Framework will probably never be finished. Once it has been applied in the several fields indicated above, new elements will come up and others already present will prove to be of lesser use. So a constant adaptation and refinement of the Framework will be desirable, if not necessary.

The PETRA-E Framework maps out competences and levels in literary translating. It aims at improving the quality of literary translations and the visibility of literary translators in Europe. At present, the framework is available in nine languages: Bulgarian, Dutch, English, French, German, Hungarian, Italian, Portuguese and Spanish.

The Network welcomes new members, academic and non-academic, who are committed to the PETRA-E Framework and its underlying principles, wish to implement the Framework in their own institutions and programmes and contribute to its further development. The first meeting will be held in the Autumn of 2017.

This year we launched a visibility campaign for literary translators on World Book and Copyright Day, the 23rd April 2024, asking you to post photos and stories about your translation work with the hashtag #translatingismysuperpower. And boy, did you answer our call!

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

The human brain is highly organized at multiple scales. At the broadest scale, neuronal populations are structurally connected across large anatomical distances with white-matter fiber bundles, forming a set of interconnected networks. This macroscopic organization can be studied via diffusion-weighted magnetic resonance imaging (MRI), which measures the direction of water diffusion in vivo1,2,3. During childhood the emergence and continual refinement of these large-scale brain networks allows for increasing functional integration and specialization4,5. This process is thought crucial for the growth of complex cognitive processes such as language6 and executive function7,8,9,10,11,12. However, there are individual differences in the organization of these networks across children, and these differences mirror important developmental outcomes. Indeed, differences in macroscopic networks have been implicated across multiple neurodevelopmental conditions13, including ADHD14, autism15,16, and language disorders17.

The generative network model (GNM) can be expressed as a simple wiring equation24,25 (Fig. 1a). If you imagine a series of locations within the brain, at each moment in time the wiring equation calculates which two locations will become connected. It calculates this wiring probability by trading-off the cost of a connection forming, against the potential value of the connection being formed. The equation can be expressed as:

The GNM simulates this process across the whole brain, until the overall number of connections matches those found in the observed brain network. Subsequently, to test the accuracy of the simulation, an energy function, E, must be defined which measures the dissimilarity between simulated and observed networks24,25:

The four measures in the energy equation are good candidates for evaluating the plausibility of simulated networks. They are critical statistical properties of realistic networks and have featured within the most well-documented simulated network models29,30,31. Moreover, these statistical properties have been implicated in a number of neuropsychiatric conditions32,33 in addition to being shown to be heritable34.

Replicating previous work, we find that our simulated networks, optimized via the statistical properties included in the energy Eq. (2) via homophily generative mechanisms, accurately capture these properties in observed networks24,25,36. But do these capture crucial network properties not included in the energy equation, like their spatial embedding? We next examined if the spatial patterning of these network properties arises simply from the generative model.

While small generative parameter differences result in differential network properties, we have yet to show how this variability may occur over the development of the networks. That is, how do differences in parameter combinations across subjects manifest themselves when the network is developing? To address this, we examined how between-subject variability in optimal GNMs emerge at the level of cortical nodes and their connections. This is possible by simply decomposing the optimal simulation into its constituent parametrized costs (Di,j)η, values (Ki,j)γ, and wiring probabilities (Pi,j) at each time point, for each subject (Fig. 6a, b). This allows us to quantify growth trajectories and thus establish which aspects of network emergence vary most in the sample.

Underlying these macroscopic changes in brain organization across time are a series of complex molecular mechanisms. These are partly governed by genetically coded processes that vary across individuals. We next tested whether these processes may steer the brain network toward a particular growth trajectory within our GNMs.

Genes do not act in isolation, but instead converge to govern biological pathways across spatial scales. To move from individual genes to biological processes (BPs) and cellular components (CCs), we performed a pathway enrichment analysis43. Pathway enrichment analysis summarizes large gene sets as a smaller list of more easily interpretable pathways that can be visualized to identify main biological themes. Genes were ordered according to their frequency in being significantly associated with connectome growth across subjects for that component. For example, for nodal values PLS1, top of the list was the gene associated with connectome growth in the most subjects (CHI3L1; significant for 49.4% of our sample), the next was the second most frequent gene (PRKAB2; 36.4% of our sample) and so on. Our list stopped when genes were significant for

3a8082e126