par la Commission, les autorités néerlandaises ont precise, le 10 février, que les
parties concernées ne s’étaient pas encore complètement entendues sur le plan de
restructuration, dont certains éléments doivent être examines ultérieurement.
Dans ces conditions, toute déclaration sur ce dossier présente pour le moment un
caractère provisoire.
Déjà avant l’effondrement de Daf, la DG IV avait examiné deux cas d’aides
non notifiées concernant un financement à court terme que les gouvernements
néerlandais et flamand avaient accordé. Ces deux gouvernements ont annoncé
qu’ils apporteraient encore leur soutien, à condition que toutes les parties
arrivent à un accord sur un plan de restructuration complet. Etant donné que cela
entraînerait certainement d’importantes aides d’Etat, la DG IV suit l’affaire avec
attention. […]
In designing the test, the first step will then be to draw up a table of
specifications—i.e. what the tester proposes to assess. In addition to the broad
macro-skills of source text processing, transfer and target text processing, the
tester may wish to identify selected features corresponding to skills or
knowledge which have figured in the (part of the) curriculum to which the test
refers. Let us imagine that in our sample case, it is the processing of
intentionality which is under particular scrutiny. What might then feature in the
test specification is ability to relay the veiled remonstrance implicated in the
source text by such elements as:
qui aurait été préparé…absorberait…entraînerait…(conditionals of
allegation)
with
later collocations, jointly indicating a discourse of diplomatic complaint:
les communiqués de presse indiquent…[press communiqués suggest]
A la suite d’une demande adressée par la Commission…[following a
request from the Commission]
Déjà…deux cas non notifiés…[two previous cases which had not been
notified]
La DG IV suit l’affaire avec attention…[the DG IV is following the
matter closely]
The marking grid for the test would then determine the credit to be given for
competent handling of this intentionality, particularly in the case of the
conditional of allegation which initiates this discourse, dynamically intruding
into a more static narrative account. At the level of the text, the assessment grid
might require testers, in respect of the selected criteria, to indicate as a response
to the question: ‘Has criterion X been met?’ either ‘Yes’, ‘Partly’ or ‘No’. Such
forms of rating might accompany a more traditional numerical assessment and
mitigate the relative unreliability of the latter.
ASSESSING PERFORMANCE 175
SUMMARY
In this chapter various recommendations have been made which, although they
are far from sufficient for the purpose of ensuring complete reliability and
validity in translation testing, may assist in promoting a more systematic
approach in which testing is less random. We have suggested (1) that as testers
our first task is to determine what purpose a test is to serve and that formative
assessment should generally be distinguished from summative assessment; (2)
that in formative assessment, discrete-point testing (multiple-choice, cloze) and
discrete-skill testing (e.g. via a commentary which the test taker submits with a
translation) are feasible and can provide useful feedback; (3) that testing
procedures be as explicit as possible; (4) that, for this purpose, useful tools are a
table of translator skills (cf.
Figure 12.1
), a test specification indicating particular
skills/features to be tested, an assessment grid closely geared to the specification
and a set of criterion-related grades which avoid norm-referenced expression
(‘above average’, ‘outstanding’) and define levels of mastery of criteria; (5) that
for purposes of feedback a descriptive profile may be of greater assistance to the
trainee than a numerical score; (6) that the term ‘error’ be restricted to significant
mismatches of denotational meaning or breaches of the target
language system
and that all else in translations be judged in terms of adequacy for intended
purposes.
Taken together, our three pedagogical chapters (
10
,
11
,
12
) make a plea for
greater consideration to be afforded to text-level issues (genre, discourse, text
type) in curriculum design, monitoring of trainees’ output and in testing. We
hope to have shown that organizing principles such as markedness (the static and
the dynamic), evaluativeness (monitoring and managing) and the interrelatedness
of context, structure and texture can be useful in avoiding hit-and-miss
approaches to translator and/or interpreter training.
176 THE TRANSLATOR AS COMMUNICATOR
