Burrowing into Translation: Character Idiolects in Henryk Sienkiewicz's Trilogy and its Two English Translations

While computational stylistics and text analysis seems to be in constant quest for the just right set of criteria (see e.g. David Hoover's presentation at the 2002 ALLC/ACH Conference in Tübingen), this presentation will try to apply what has already become a standard in statistical stylistic analysis to a (relatively) novel material. Taking for granted — very unoriginally — the usefulness of John Burrows's method that has been around since his 1987 Computation into Criticism, in discerning stylistic differences between individual characters in works by the same author, I will try to see if the same or similar patterns of similarity and difference travel across linguistic boundaries; if differences between characters' "idiolects" are preserved in translation.

In a typically Polish approach to the matter, I have chosen as my material the trilogy of historical romances by Poland's first literary Nobel Prize winner, Henryk Sienkiewicz, written between 1882 and 1888, and its two English (or, more precisely, American) translations by Jeremiah Curtin (completed between 1890 and 1893) and W.S. Kuniczak (1991). The reasons for this choice have been manifold. First, Sienkiewicz's three novels, With Fire and Sword, The Deluge, and Pan Michael, although set in the turmoil of 17th-century Poland, remain to this day a major classic of Polish literature and the country's most popular reading. Second, a trilogy, the subsequent parts of which share some of their characters, seems ideal for Burrowsian analysis (a fact confirmed by the interesting coincidence of John Burrows's undertaking of a study of Beckett's bilingual trilogy in a much later paper). The final reason was the difference between the two translations, evident both in their being separated by an entire century and, what follows, in the entirely different approaches and results obtained by the two translators: the largely word-by-word transcoding by Sienkiewicz's contemporary and the highly adaptative and "free" method of the modern Polish-American writer.

Faithfully maintaining the original Burrows model, the study of distances between the "idiolects" of the major characters has been based on relative frequencies of the 30 most frequent words in the dialogue of each version of the trilogy. The resulting correlation matrices were then used to produce two-dimensional multidimensional scaling charts of distances between such "idiolects."

This procedure has yielded, in Sienkiewicz's original, a very consistent influence of the personae's social status and ethnic background, especially in the first part of the series. It is particularly visible in idiolects of 'enemy' (non-Polish) collective characters, usually plotted at some distance one from the other. Curtin's translation is notable for 'de-clustering' idiolects of various Polish gentry characters, making their idiolects much less alike. There is also a visible tendency in Curtin to limit the distances between rival collective characters and making them markedly similar rather than divergent as in the original. Idiolects in Kuniczak are even more evenly distributed, with a general trend towards greater distances and less clustering observable in the graphs. The very high resemblance between the idiolects of two major characters, Zagłoba and Wołodyjowski, in all three versions of The Deluge (almost identical in Sienkiewicz) is one of the most consistent traits of this portion of the analysis — an interesting illustration of the fact that the two personae's function of keeping the three volumes together becomes evident in the second part of the series. Sienkiewicz's social/ethnic idiosyncrasy has been confirmed in a plot for idiolects of characters involved in the Polish-Ukrainian conflict in With Fire and Sword, a feature slightly visible in Kuniczak and almost not at all in Curtin.

As perhaps the most consistent effect of all, the peripheral situation of female idiolects is a constant element in almost any configuration. The study of more detailed and thematic configurations of characters is also the source of interesting insight. Plots for female characters exhibit a tendency to group together young (and marriageable) Polish women; Helena's Ukrainian provenience is highly visible in Sienkiewicz, while the ethnic element is indiscernible in both translations. Among characters involved in each novel's eternal triangles, both of the above aspects are clearly visible in all three versions; differences between characters in the same triangle are quite considerable.

In an examination of characters that recur throughout the series, a good consistence has been observed between the three idiolects of the Polish Falstaff, Zagłoba, the most inveterate talker of the series, in each version separately: best in Sienkiewicz, worst in Curtin. Another character, Wołodyjowski, is much more of a developing character, which agrees well with the evolution of the persona in the course of the series, from a humble officer to the hero and spearhead of Sienkiewicz's ideology. This has been confirmed in a separate plotting of idiolects of those two characters.

A number of joint plots for Curtin's and Kuniczak's translations (based on frequent words common in both versions) have been made to investigate if there is a constant pattern in their respective differences. In agreement with some "intuitive" assessments as to the decreasing differences between Curtin's and Kuniczak's versions (mainly due to Kuniczak's gradual abandonment of adaptative procedures, especially on the microstructural level), the patterns become more consistent with time: fairly chaotic movement for the first part of the trilogy has become more ordered in the second and almost uniform in the third. The "stylistic drift" observed between idiolects in Curtin and in Kuniczak — divided, apart from their contrasting approaches to translation, by an even more significant difference of a whole century — is a vindication of Burrows's 'tiptoeing towards the infinite:' that visible and uniform shift in the configuration of the most frequent words in English texts with time.