Poetry Translating as Expert Action
In run-through RT1 (24tu/2m) of Draft 2, Irene quickly re-read the source poem
to put its text world into working memory. This prepared her for RT2 (639tu/49m):
a long, painstaking revision of her English version from Draft 1 into another hand-
written, pure-line version. This was briefly checked in RT3 (35tu/3m).
In RT1 of Draft 3, Irene typed up Draft 2’s final handwritten version, revising
as she went (109tu/8m). In RT2 (53tu/4m), she quickly evaluated this version for
rhythm and debugged a couple of lexical problems. In RT3 (61tu/5m) she re-read
the version to check its correspondence with the source poem, showing how con-
cerned she was that source and target text-worlds should be as close as possible:
TU174 that
TU175 was when time set in, he means. #
TU176 I mean otherwise it’s too
TU177 much of my interpretation going on in there
TU178 entered then into time^12 , that was when time set in.
TU179 And I vaguely remember that that’s what the poet himself told us when we met.Though other translators’ runs-through differed in what they tackled when, how
Irene used repeated runs through the text to manage workload within each draft
was typical of all five translators.5.3.4 Macro-sequences and Lines
Each run-through contained, on average, 5.1 macro-sequences. Each macro-se-
quence averaged 2m (29tu) in length, but this again concealed huge variations:
between 0.1m and 14m (1 and 184tu), even excluding my 1h 9m (907tu) macro-
sequence tackling Stanza I in Draft 2. There was no significant tendency for some
drafts or translators to produce shorter macro-sequences than others^13. This indi-
cates that translators managed their work similarly at this level too.
This similarity is confirmed by the transcripts. In Drafts 1 and 2, all translators
tended to divide the longer ‘slow reading + writing’ runs-through into macro-se-
quences corresponding at the smallest to Lines or clauses, and at the largest to stan-
zas. This enabled them to work systematically though the poem, focusing their at-
tention on one Line or clause after the other – as in Figure 36, which lists Irene’s
Draft-2 macro-sequences in the run-through producing Figure 37’s pure-line Ver-
sion. When clauses span two Lines, as in Stanzas III-V, clause-based macro-se-
quences are preferred to Line-based macro-sequences (e.g. Figure 36: Dr2/
Ma2.8–2.9). At intervals, translators back-track and revise the stanza just tackled or
the poem so far – whether to check how the Lines or clauses just processed fit into a- Strike-through denotes the translator’s deletion of text from a version.
- One-way ANOVA: F 1.42, p 0.23 (translators); F 1.19, p 0.31 (drafts).