![v
i
i
i
First published 2022
by Routledge
605 Third Avenue, New York, NY 10158
and by Routledge
4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2022 selection and editorial matter, James Luke Hadley, Kristiina Taivalkoski-
Shilov,
Carlos S. C. Teixeira, Antonio Toral; individual chapters, the contributors
The right of James Luke Hadley, Kristiina Taivalkoski-
Shilov, Carlos S. C. Teixeira,
Antonio Toral to be identified as the authors of the editorial material, and of the
authors for their individual chapters, has been asserted in accordance with
sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
With the exception of Chapter 3, no part of this book may be reprinted or reproduced
or utilised in any form or by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying and recording, or in any information storage
or retrieval system, without permission in writing from the publishers.
Chapter 3 of this book is available for free in PDF format as Open Access from the
individual product page at www.routledge.com. It has been made available under a
Creative Commons Attribution 4.0 license.
Trademark notice: Product or corporate names may be trademarks or registered trademarks,
and are used only for identification and explanation without intent to infringe.
Library of Congress Cataloging-
in-
Publication Data
Names: Hadley, James (Researcher in literary translation), editor.
Title: Using technologies for creative-text translation /
edited by James Luke Hadley [and 3 others].
Description: New York, NY : Routledge, 2023. |
Series: Routledge advances in translation and interpreting studies |
Includes bibliographical references and index. |
Identifiers: LCCN 2022012937 (print) | LCCN 2022012938 (ebook) |
ISBN 9780367555818 (hardback) | ISBN 9781003094159 (ebook)
Subjects: LCSH: Machine translating. |
Translating and interpreting–Technological innovations. |
Literature–Translations–History and criticism. | LCGFT: Essays.
Classification: LCC P308 .U53 2023 (print) | LCC P308 (ebook) |
DDC 418/.020285–dc23/eng/20220506
LC record available at https://lccn.loc.gov/2022012937
LC ebook record available at https://lccn.loc.gov/2022012938
ISBN: 9780367555818 (hbk)
ISBN: 9781032333847 (pbk)
ISBN: 9781003094159 (ebk)
DOI: 10.4324/9781003094159
Typeset in Sabon
by Newgen Publishing UK](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-13-320.jpg)
![16 Hadley, Taivalkoski-Shilov, Teixeira, Toral
1
6
References
Bahdanau, Dzmitry, Kyunghyun Cho, and Yoshua Bengio. 2014. “Neural
Machine Translation by Jointly Learning to Align and Translate.” arXiv
preprint:1409.0473.
Besacier, Laurent, and Lane Schwartz. 2015. “Automated Translation of a
Literary Work: A Pilot Study.” In Proceedings of the Fourth Workshop on
Computational Linguistics for Literature, 114–
122. Denver, CO.
Brown, Peter F., John Cocke, Stephen A. Della Pietra, Vincent J. Della Pietra,
Frederick Jelinek, Robert L. Mercer, and Paul S. Roossin. 1988. “A Statistical
Approach to Language Translation.” In Proceedings, 12th International
Conference on Computational Linguistics (COLING-88). Budapest, Hungary,
August 1988, 71–
76.
De Saussure, Ferdinand. [1959] 2011. Course in General Linguistics. Translated
by Wade Baskin. Columbia University Press, New York.
Ece, Ayşe. 2015. “The Trajectory of Literary Translation: From Interpretation to
(Re) Writing and a New Life.” Synergies Turquie, 8.
Evangelista, Stefano, ed. 2010. The Reception of Oscar Wilde in Europe.
Bloomsbury Publishing, London and New York.
Gambier, Yves. 2018. “Institutionalization of Translation Studies.” In A History
of Modern Translation Knowledge, edited by Lieven d’Hulst and Yves
Gambier, 131–142.
Genette, Gérard. [1987] 1997. Palimpsests: Literature in the Second Degree.
Translated by Channa Newman and Claude Doubinsky. Lincoln: University
of Nebraska Press.
Glick, Thomas F., and Elinor Shaffer, eds. 2014. The Literary and Cultural
Reception of Charles Darwin in Europe. Vol. 17. Bloomsbury Publishing,
London and New York.
Harper, Kenneth E. 1955. “Translating Russian by Machine.” Journal of
Communication 5 (2): 41–
46.
Joyce, James. 1922. Ulysses. Paris: Shakespeare.
Lennon, Brian. 2014. “Machine Translation: A Tale of Two Cultures.” In A
Companion to Translation Studies, edited by Sandra Bermann and Catherine
Porter, 136–
146. Oxford: Wiley.
Leonardi, Vanessa. 2007. Gender and Ideology in Translation: Do Women and
Men Translate Differently? A Contrastive Analysis from Italian into English.
Bern: Peter Lang.
Macdonald, Neil. 1954. “Language Translation by Machine: A Report of the
First Successful Trial.” Computers and Automation 3 (2): 6–
10.
Mandal, Anthony, and Brian Southam, eds. 2007. The Reception of Jane Austen
in Europe. Vol. 13. AC Black, London and New York.
Mason, Ian. 1994. “Discourse, Ideology and Translation.” In Language,
Discourse and Translation in the West and Middle East, edited by Robert de
Beaugrande, Abdullah Shunnaq and Mohamed H. Heliel, 23–
34. Amsterdam:
John Benjamins.
Miller, Tristan. 2019. “The Punster’s Amanuensis: The Proper Place of Humans
and Machines in the Translation of Wordplay.” In Proceedings of the Human-
Informed Translation and Interpreting Technology Workshop (HiT-IT
2019), 57–65.](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-33-320.jpg)
![Collecting literary translators’ narratives 19
1
9
The characterisation of translation as a form of Human-
Computer
Interaction (HCI) by O’Brien (2012) ultimately highlighted the need to
account for both material and immaterial aspects when investigating
contemporary translation. In this respect, Littau (2016) defines materi-
ality as the non-
human element, e.g., digital tools, machines, computers.
Conversely, immateriality is embodied by creativity, spirit, mind, con-
science, soul, and anything associated with being quintessentially human
(Littau, 2016). In her view, both material and immaterial elements
interact symbiotically and reciprocally shape and influence each other to
the point where “[technologies] are active in effecting the ways in which
we think, read, write and translate” (Littau 2017, 100). Furthermore,
Cronin notes how the constant stream of information, incessant digital-
isation of materials, automation of processes, and speed of communica-
tion that are typical of the Information Age contribute to an overall “sense
of confusion” (Cronin 2013, 1). This “sense of confusion” characterizes
translators as they try to give meaning to this new order of things (Cronin
2013, 1). Thus, it is paramount to include translators in the conversa-
tion and, by giving them a voice, perhaps discover new ways for literary
translators to exist in this new socio-
technological landscape, as well as
co-
exist with new technologies.
This chapter is based on a 2018 study aimed at exploring the dynamic
between human (immaterial) and non-
human (material) factors in literary
translation, recognising materiality as central to contemporary transla-
tion practice and trying to bring literary translators’ voices back into the
conversation. More specifically, literary translators were asked to share
their attitudes towards technology and perceptions of their role in society.
The study’s main research question was “what is the dynamic between
humans and technology in literary translation?”. Two sub-
questions were
formulated to assist and guide the research process, respectively (a) “how
do literary translators perceive their role in society?” and (b) “what are
their attitudes towards technology as related to literary translation?”.
Respondents’ narratives were collected via means of a questionnaire
that registered 150 responses, mostly from Europe. Overall, the study
adopted an interpretivist, constructionist, and mixed-
methods approach.
The theoretical framework and data analysis were informed by the Social
Construction of Technology (SCOT) framework as theorized by Pinch
and Bijker (1984). Although the overall relationship between self-
image
and technology will be briefly discussed in this chapter, its main focus will
be on how participants constructed the notion of technology as related to
their professional practice. The chapter will first provide a review of the
literature on literary translator status and the application of technology
to literary translation workflows. It will then introduce the study’s meth-
odology and present the research findings on literary translators’ attitudes
towards both general and translation technology tools. Finally, it will dis-
cuss the results and suggest a way forward for research on the topic.](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-36-320.jpg)
![24 Paola Ruffo
2
4
SCOT has been identified by Olohan (2017, 2019) as an ideal frame-
work for the study of technology in translation from a sociological rather
than historical standpoint. In fact, looking at technological development
through a SCOT lens could help foster a better understanding of the
social and power dynamics underlying technology acceptance, its rejec-
tion, and the consequences of its introduction in professional translation
practice (Olohan 2017). In this study, the SCOT framework was adopted
at a conceptual level to guide data collection and analysis, with some of
its tenets having been amended to serve the study’s research questions.
In Pinch and Bijker’s (1984) words: “our model is not used as a mould,
into which the empirical data have to be forced, coûte que coûte. […]
Its function is primarily heuristic—
to bring out all the aspects relevant
for our purposes” (Pinch and Bijker 1984, 419). More specifically, this
study is not retrospective in nature, as its object is not the ex post facto
exploration of a tool that has already reached closure. Instead, SCOT is
here adopted proactively, in order to address controversies as they arise
in the present and give voice to literary translators as a relevant social
group. Furthermore, instead of identifying all relevant social groups and
analysing one specific artefact, this study focuses on literary translators
only and on technology in general. This is due to the lack of previous
studies in this area and to the limited timeframe of the project, which
led to prioritising literary translators as a social group whose livelihood
is more likely to be affected by socio-
technological changes in their pro-
fession. Furthermore, the study is not ethnographic in nature—
as SCOT
research usually is—
due to literary translators’ technology use being an
under-
researched area, which thus calls for the need to survey this aspect
before proceeding with an ethnography of specific tools. The overall aim
when employing SCOT was to pre-
emptively identify emerging issues in
the relationship between materiality and literary translation, as well as
devise potential solutions.
Methodology
The study adopted an interpretivist, constructionist, and mixed-
methods
approach. It is interpretivist, in that it “prioritizes people’s subjective
understandings and interpretations of social phenomena” (Saldanha and
O’Brien 2014, 11–
12)—
in this case, literary translators’ narratives of the
technologisation of their profession. In doing so, it also recognizes their
role as “a constructing and constructed subject in society” (Wolf 2007,
1) and as agents of sociocultural change. An interpretivist and social con-
structionist analysis thus allows insights into the way practitioners inter-
pret and assign meanings to their professional reality, which is a pivotal
step for a sustainable technological development. As far as the mixed-
methods approach is concerned, this was adopted to maximize the poten-
tial of a large data set, with quantitative elements helping to support
and contextualize the qualitative nature of participants’ narratives. Both](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-41-320.jpg)
![Collecting literary translators’ narratives 25
2
5
elements were combined in a questionnaire by using closed and open
questions.
The questionnaire consisted of 31 questions and was divided into 6 main
sections, namely Background Information, Language Skills, Professional
Practice, Confidence with Technologies, Open Questions on perceptions
of role and attitudes towards technology, and Final Comments. The lan-
guage used in the questionnaire was English and all answers received
were in English except for one written in German, which was translated
by the researcher. The sampling frame was defined using non-
random
sampling and by targeting literary translators in UK translation associ-
ations and online translation communities. The questionnaire was sent
out to a total of 13 UK general and literary translation associations and
68 online translation communities. The latter included online forums,
Facebook and LinkedIn groups, and mailing lists aimed at both general
and literary translation. Due to the non-
random nature of the sample,
results cannot be statistically generalized; instead, theoretical and logical
generalisation was sought. In this respect, Luker (2008) notices how, even
when sampling issues do not allow for statistical generalisation, it is still
possible to work on a logical and theoretical level of abstraction. When
theoretically generalising, findings are compared to previous theories
and studies “to see how [they] illuminate, contradict, extend, or amplify
existing theory” (Luker 2008, 127).
The questionnaire was live on the Online Surveys (formerly BOS) plat-
form for six weeks between September and October 2018, and it was
completed by 150 respondents. This chapter presents and discusses data
related to participants’ professional and educational background, their
levels of confidence with technology, the technology tools used in their
practice, and their overall attitudes towards technology in literary transla-
tion. It is worth mentioning that the questionnaire did not include closed
items about specific tools. Instead, it only distinguished between general
and translation-
specific technology. General technology was defined in
the questionnaire as “any technology tool that is not translation-
specific
(e.g., online dictionaries, a time management app, a text-
editor software,
etc.).” Translation-
specific technology was defined as “any technology
tool that is translation-
specific (e.g., Translation Memory systems, ter-
minology management software, Machine Translation systems, etc.).”
Given the exploratory nature of the study and the lack of previous
research in this area, respondents could indicate the tools they use in
their daily practice and express their attitudes with as few restrictions
from the researcher as possible. In turn, results could form the basis for
future research on specific tools.
The data analysis consisted of three main phases. During the first stage,
a thematic analysis of the open questions was performed. The coding pro-
cess was supported by the use of NVivo, a software package for quali-
tative and mixed-
methods data analysis. The second phase focused on
quantifying and collating results from the closed questions; no statistical](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-42-320.jpg)
![30 Paola Ruffo
3
0
reworking of texts” and facilitates “the breadth and depth of research
needed for a good literary translation to be done from home.” Technology
is also appreciated when it improves efficiency—
for example, by reducing
time spent typing or proofreading and editing a text. Dictation and
voice-
recognition software, Translation Memory (TM) and autosuggest
features all make an appearance in this category. By assisting translators
with “the housekeeping aspect”, this type of technology can “[take] out
the donkey work” and allow them “more time to focus on the creative
part of [their] work.” Technology was also praised for its ability to com-
pensate for human shortcomings, especially in relation to memory. In
fact, several responses mentioned TM being effective in “[reminding] the
translator of previous renderings (perhaps also the need to avoid them)”
and providing “a searchable record of [one’s] decision-
making process.”
Overall, an appealing tool is one that “helps [one] concentrate on trans-
lating instead of having to fiddle around with complicated systems.” To
add to this, one respondent notes: “I would welcome technology […] that
enabled me to make better quality choices […]—
quality being judged in
my subjective experience of freedom and self-
expression.”
Ultimately, literary translators welcome technology that enhances their
practice by assisting with all aspects surrounding the act of translation—
be it in terms of accuracy, consistency, reducing editing time, or accessing
previously translated content—
rather than interfering with the act of
translation itself. This becomes even clearer in the next section, which
looks at their narratives of unappealing aspects of technology.
Unappealing aspects of technology
While appealing aspects were mainly associated with general technology
and TM, the discourse on unappealing aspects is almost exclusively
focused on translation technology (Table 1.4). In this respect, TM
resurfaces here as a hindrance to human translation, being described as
causing memory to grow lazy and language to become standardized.
According to respondents, TM is in direct opposition to literary
Table 1.3
Appealing aspects of technology (selection)
Appealing aspects of technology Count %
Research 58 39%
Efficiency 33 22%
Assistance to human 27 18%
Accuracy and consistency 24 16%
None 23 15%
Networking 14 9%
Total respondents: 150](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-47-320.jpg)
![Collecting literary translators’ narratives 31
3
1
translation’s essence. In fact, it “takes away from the need to artfully turn
a phrase over and over until it is right, accurate, and artful.” Furthermore,
others expressed concerns regarding the impoverishment of language,
with one respondent noting: “I doubt we’d agree on a voice.” This being
said, stronger attitudes are expressed in regard to MT and AI, which are
associated with a potential worsening of working conditions and wanting
to substitute human translators. What is unappealing about this type of
technology is that it “tries to bypass the human understanding of lan-
guage and its nuances in order to save costs” and, according to another
participant, “attempts to push the boundaries of technology within an
essentially contemplative profession which requires an unfashionable
degree of isolation and respect for experience.” Another theme that
emerges here is that of translation technology disrupting the translation
process and diminishing translator autonomy. CAT tool segmentation
and MT are often mentioned as examples of this, with one respondent
simply stating: “I don’t want a machine singing my part of the duet for
me.” In addition to this, narratives of replacement surrounding some of
these tools are also a source of frustration for literary translators, who
simply do not believe MT to be compatible with literary translation. The
uneasiness that participants feel in this respect is engendered by what they
perceive as a misunderstanding of what literary translation entails and
the misuse of technology in the name of saving time and costs. The fear
is that “all the talk by technology buffs who claim that [this] art/
profes-
sion will be obsolete” will “[lead] publishers to think that all they need is
a good translation tool and a skilled editor.” Finally, other unappealing
aspects mentioned include accessibility of translation technology tools in
terms of costs and learning curve, complicated user interfaces, and ineffi-
ciency caused by too many or wrong inputs.
Ultimately, when talking about unappealing aspects of technology,
the focus shifts to tools developed specifically for translation. In par-
ticular, MT and TM are perceived as a hindrance to the translator, as
well as causing disruption to the translation process, with the discourse
surrounding them threatening translators’ livelihoods in the future.
Table 1.4
Unappealing aspects of technology (selection)
Unappealing aspects of technology Count %
Hindrance to human 41 27%
None 30 20%
Disruption/
Loss of autonomy 27 18%
Outsiders’ narratives 21 14%
Usability and access 17 11%
Inaccuracy and inconsistency 9 6%
Total respondents: 150](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-48-320.jpg)
![Collecting literary translators’ narratives 33
3
3
technology as helpful or expressed ambivalent attitudes, while 15%
stated that there is no relationship between the two or that technology is
less or not helpful for literary translators.
The differences between general and translation technology first
noted when looking at appealing and unappealing aspects of technology
re-
emerge even more clearly in the answers to this question. In particular,
the ambivalence originates from thinking of general technology as helpful
for research, terminology, and networking, while translation technology
once again is a source of concern in terms of disruption of the trans-
lation process. One participant thinks such tools “discourage freedom
in interpreting text on larger scales,” while another states: “I wonder
how much they interfere with my originality.” Corpora, online search
engines, and dictionaries are often referred to as helpful, together with
any technology that aids in “sorting thoughts, terminology, and other
things you’d have to keep in your head otherwise.” Translation tech-
nology is also at the centre of participants’ more negative attitudes. These
narratives see translation technology as not relevant or less helpful for
literary translation than for other kinds of translation and revolve around
the incompatibility of such tools with the complexity of the literary trans-
lation task. Epitomising this viewpoint, one respondent states: “I have
the impression those [translation technology tools] are for people in a
hurry. I work slowly and carefully”. Finally, a few literary translators
mentioned being unwilling to adopt translation technology regardless of
its usefulness and feeling that technological innovation in literary transla-
tion is an imposition from above rather than a response to practitioners’
actual needs.
When linking attitudes towards technology with respondents’
backgrounds, those aged between 18–
25 have the most positive rela-
tionship with technology: 50% thought of technology as helpful for lit-
erary translation, while the rest described the relationship between the
two as either necessary (25%) or harmonious (25%). Generally, those
aged 46+are more inclined to think of technology as being unrelated
to literary translation. In fact, the “No Relationship” category barely
appears in respondents below 45 years of age (it does not appear at all
in the 18–
25 group); however, it occurs for all 46+respondents (14% of
the 46–
55 group, 23% of those aged 56–
65, and 15% of the over 65s).
Those who have received academic translation technology training also
view technology more positively. In particular, 70% of those with aca-
demic training believe the relationship between literary translation and
technology to be a positive one, against 45% of respondents without
academic training. The latter present higher levels of torn feelings (27%
versus 7% of those with academic training). The same happens for those
with higher levels of confidence with technology, as respondents with less
or no confidence were more likely to think of technology as irrelevant or
unhelpful for literary translation. For example, 44% of those who are
extremely confident find technology helpful and only 6% think there is](https://image.slidesharecdn.com/22272924-250619144604-7cad6dec/85/Using-Technologies-For-Creativetext-Translation-James-Luke-Hadley-50-320.jpg)
Using Technologies For Creativetext Translation James Luke Hadley
- 1. Using Technologies For Creativetext Translation James Luke Hadley download https://ebookbell.com/product/using-technologies-for- creativetext-translation-james-luke-hadley-44545848 Explore and download more ebooks at ebookbell.com
- 2. Here are some recommended products that we believe you will be interested in. You can click the link to download. Using Technologies For Creativetext Translation James Luke Hadley https://ebookbell.com/product/using-technologies-for-creativetext- translation-james-luke-hadley-44680534 Wordpress For Absolute Beginners Create Websites And Blogs From Scratch Using Wordpress Cms 1st Edition Raymond John https://ebookbell.com/product/wordpress-for-absolute-beginners-create- websites-and-blogs-from-scratch-using-wordpress-cms-1st-edition- raymond-john-33828670 Wordpress For Absolute Beginners Create Websites And Blogs From Scratch Using Wordpress Cms 1st Edition J N Raymond https://ebookbell.com/product/wordpress-for-absolute-beginners-create- websites-and-blogs-from-scratch-using-wordpress-cms-1st-edition-j-n- raymond-33820052 Virtual Storytelling Using Virtual Reality Technologies For Storytelling International Conference Icvs 2001 Avignon France September 2728 2001 Proceedings 1st Edition Eric Badiqu Auth https://ebookbell.com/product/virtual-storytelling-using-virtual- reality-technologies-for-storytelling-international-conference- icvs-2001-avignon-france-september-2728-2001-proceedings-1st-edition- eric-badiqu-auth-1535720
- 3. Virtual Storytelling Using Virtual Reality Technologies For Storytelling 4th International Conference Icvs 2007 Saintmalo France December 57 2007 Proceedings 1st Edition Martyn Daderobertson Auth https://ebookbell.com/product/virtual-storytelling-using-virtual- reality-technologies-for-storytelling-4th-international-conference- icvs-2007-saintmalo-france-december-57-2007-proceedings-1st-edition- martyn-daderobertson-auth-4521090 Virtual Storytelling Using Virtual Reality Technologies For Storytelling Third International Conference Icvs 2005 Strasbourg France November 30 December 2 2005 Proceedings 1st Edition Michitaka Hirose Auth https://ebookbell.com/product/virtual-storytelling-using-virtual- reality-technologies-for-storytelling-third-international-conference- icvs-2005-strasbourg-france- november-30-december-2-2005-proceedings-1st-edition-michitaka-hirose- auth-4604592 Using Multimedia Systems Tools And Technologies For Smart Healthcare Services Amit Kumar Tyagi https://ebookbell.com/product/using-multimedia-systems-tools-and- technologies-for-smart-healthcare-services-amit-kumar-tyagi-47573640 Sustainable Security Practices Using Blockchain Quantum And Postquantum Technologies For Real Time Applications Adarsh Kumar https://ebookbell.com/product/sustainable-security-practices-using- blockchain-quantum-and-postquantum-technologies-for-real-time- applications-adarsh-kumar-56595320 Curriculum Models For The 21st Century Using Learning Technologies In Higher Education 1st Edition Maree Gosper https://ebookbell.com/product/curriculum-models-for-the-21st-century- using-learning-technologies-in-higher-education-1st-edition-maree- gosper-4333536
- 6. i “This is an important and necessary book, on a subject about which I often ponder and speculate and converse, but never know where exactly to turn to deepen my understanding. I suspect that there are many, many others out there in the same position, and they will welcome this publi- cation too.” Polly Barton, Japanese-English prize-winning literary translator (www.pollybarton.net/about-me) “This is a book to be read by anyone who has a practical or theoretical interest in the newly emerging field of the use of machines in the transla- tion of literary and creative texts, be they students of translation, transla- tion scholars or practising literary translators.” Roy Youdale, Spanish-English literary translator and author of Using Computers in the Translation of Literary Style: Challenges and Opportunities (Routledge Advances in Translation and Interpreting Studies) “This volume offers a fresh look at one of the most exciting areas in con- temporary translation studies: computing, creativity and translation, as well as offering a new look at the interaction between technology and the translation of creative texts. With this timely contribution to one of the most exciting areas in contemporary translation studies, Hadley et al. make the case for a closer look at the role of computers in translation, even for creative texts.” Dorothy Kenny, Professor of Translation Studies at Dublin City University
- 7. ii
- 8. iii Using Technologies for Creative-Text Translation This collection reflects on the state of the art of research into the use of translation technologies in the translation of creative texts, encompassing literary texts but also extending beyond to cultural texts, and charts their development and paths for further research. Bringing together perspectives from scholars across the discipline, the book considers recent trends and developments in technology that have spurred growing interest in the use of computer- aided translation (CAT) and machine translation (MT) tools in literary translation. Chapters examine the relationships between translators and these tools— the extent to which they already use such technologies, the challenges they face, and prevailing attitudes towards these tools— as well as the eth- ical implications of such technologies in translation practice. The volume gives special focus to drawing on examples with and beyond traditional literary genres to look to these technologies’ use in working with the larger group of creative texts, setting the stage for many future research opportunities. The book will be of particular interest to students and scholars in translation studies, especially those with an interest in literary transla- tion, translation technology, translation practice, and translation ethics. James Luke Hadley is Trinity College Dublin’s Ussher Assistant Professor in Literary Translation and Director of the College’s MPhil in Literary Translation. His research represents his wide- ranging interests, many of which centre on translation in under- researched cultural contexts. His interests include machine translation and computer- assisted transla- tion research, as well as integrating empirical research into Translation Studies. Kristiina Taivalkoski-Shilov is Professor of Multilingual Translation Studies and Vice Head of the School of Languages and Translation Studies at the University of Turku. Her research interests include literary translation, translation history, and ethics of translation. Throughout her career, she has worked on the notion of “voice” in translation, which she has examined from theoretical, historical, and ethical perspectives.
- 9. i v Carlos S. C. Teixeira is a Localisation Engineer at IOTA Localisation Services and Adjunct Professor, Universitat Rovira i Virgili. His research and publications have focussed on translation process research, not- ably the interaction between translators and technology in profes- sional settings. With Routledge, he has authored a chapter on “Revising Computer- Mediated Translations” in Mossop, Brian. Revising and Editing for Translators, 4th edition (2020). Antonio Toral is Assistant Professor in Language Technology at the University of Groningen. He holds a PhD in Computational Linguistics from the Universitat d’Alacant and has carried out research in the area of machine translation (MT) since 2010. His research interests include the application of MT to literary texts, MT for under- resourced languages, and the analysis of translations produced by machines and humans.
- 10. v Routledge Advances in Translation and Interpreting Studies Lifestyle Politics in Translation The Shaping and Re- Shaping of Ideological Discourse By M. Cristina Caimotto and Rachele Raus Reframing Translators, Translators as Reframers Edited by Dominique Faria, Marta Pacheco Pinto, and Joana Moura Transfiction and Bordering Approaches to Theorizing Translation Essays in Dialogue with the Work of Rosemary Arrojo Edited by D. M. Spitzer and Paulo Oliveira Translating Controversial Texts in East Asian Contexts A Methodology for the Translation of ‘Controversy’ Adam Zulawnik Using Technologies for Creative- Text Translation Edited by James Luke Hadley, Kristiina Taivalkoski- Shilov, Carlos S. C. Teixeira, and Antonio Toral Relevance Theory in Translation and Interpreting A Cognitive- Pragmatic Approach Fabrizio Gallai Towards a Feminist Translator Studies Intersectional Activism in Translation and Publishing Helen Vassallo For more information about this series, please visit https://www.routle dge.com/Routledge-Advances-in-Translation-and-Interpreting-Studies/ book-series/RTS
- 11. v i
- 12. v i i Using Technologies for Creative-Text Translation Edited by James Luke Hadley, Kristiina Taivalkoski-Shilov, Carlos S. C. Teixeira, and Antonio Toral
- 13. v i i i First published 2022 by Routledge 605 Third Avenue, New York, NY 10158 and by Routledge 4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2022 selection and editorial matter, James Luke Hadley, Kristiina Taivalkoski- Shilov, Carlos S. C. Teixeira, Antonio Toral; individual chapters, the contributors The right of James Luke Hadley, Kristiina Taivalkoski- Shilov, Carlos S. C. Teixeira, Antonio Toral to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. With the exception of Chapter 3, no part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Chapter 3 of this book is available for free in PDF format as Open Access from the individual product page at www.routledge.com. It has been made available under a Creative Commons Attribution 4.0 license. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging- in- Publication Data Names: Hadley, James (Researcher in literary translation), editor. Title: Using technologies for creative-text translation / edited by James Luke Hadley [and 3 others]. Description: New York, NY : Routledge, 2023. | Series: Routledge advances in translation and interpreting studies | Includes bibliographical references and index. | Identifiers: LCCN 2022012937 (print) | LCCN 2022012938 (ebook) | ISBN 9780367555818 (hardback) | ISBN 9781003094159 (ebook) Subjects: LCSH: Machine translating. | Translating and interpreting–Technological innovations. | Literature–Translations–History and criticism. | LCGFT: Essays. Classification: LCC P308 .U53 2023 (print) | LCC P308 (ebook) | DDC 418/.020285–dc23/eng/20220506 LC record available at https://lccn.loc.gov/2022012937 LC ebook record available at https://lccn.loc.gov/2022012938 ISBN: 9780367555818 (hbk) ISBN: 9781032333847 (pbk) ISBN: 9781003094159 (ebk) DOI: 10.4324/9781003094159 Typeset in Sabon by Newgen Publishing UK
- 14. i x Contents List of contributors xi Introduction 1 JAMES LUKE HADLEY, KRISTIINA TAIVALKOSKI- SHILOV, CARLOS S. C. TEIXEIRA, AND ANTONIO TORAL 1 Collecting literary translators’ narratives: Towards a new paradigm for technological innovation in literary translation 18 PAOLA RUFFO 2 Dutch literary translators’ use and perceived usefulness of technology: The role of awareness and attitude 40 JOKE DAEMS 3 Human– computer interaction in pun translation 66 WALTRAUD KOLB AND TRISTAN MILLER 4 Bilingual e- books via neural machine translation and their reception 89 ANTONI OLIVER, ANTONIO TORAL, AND ANA GUERBEROF ARENAS 5 Catching the meaning of words: Can Google Translate convey metaphor? 116 ALICJA ZAJDEL
- 15. x Contents x 6 Pragmatic and cognitive elements in literary machine translation: An assessment of an excerpt from J. Polzin’s Brood translated with Google, DeepL, and Microsoft Bing 139 PAOLA BRUSASCO 7 The “Oxen of the Sun” hypertext: A digital hypertext in the study of polyphonic translations of James Joyce’s Ulysses 161 LAURI A. NISKANEN 8 Translating with technology: How digitalisation affects authorship and copyright of literary texts 180 MAARIT KOPONEN, SANNA NYQVIST, AND KRISTIINA TAIVALKOSKI-SHILOV Index 199
- 16. x i Contributors Paola Ruffo is a researcher in the field of Computer-Assisted Literary Translation (CALT). She has recently joined Ghent University as a Marie Skłodowska-Curie Postdoctoral Fellow to work on ‘Developing User-centred Approaches to Technological Innovation in Literary Translation (DUAL-T)’. She has previously worked as a freelance trans- lator and as a Lecturer in Translation Technology at the University of Bristol. Joke Daems is Postdoctoral Research Assistant at Ghent University in the field of machine translation and human- computer interaction. They are one of the editors of Reuniting the Sister Disciplines of Translation and Interpreting Studies (Routledge, 2020), and have contributed a chapter to Translation Revision and/or Post-Editing: Industry Practices and Cognitive Processes (Routledge, 2020). Waltraud Kolb is Assistant Professor of Literary Translation at the Center for Translation Studies, University of Vienna. One focus of her research is on digital tools and machine translation in the literary field and lit- erary translation and post- editing processes. She is a member of the executive board of the Austrian Association of Literary Translators. Tristan Miller is Research Scientist at the Austrian Research Institute for Artificial Intelligence. He is a computational linguist specialising in lexical semantics, language resources and evaluation, and creative lan- guage. He is a consulting editor for Humor: International Journal of Humor Research and a contributor to The Routledge Handbook of Language and Humor. Antoni Oliver González is Associate Professor at the Open University of Catalonia (UOC) and Director of the Master’s Degree in Translation and Technologies. His main area of research is Natural Language Processing, with a special focus on machine translation and automatic terminology extraction. Ana Guerberof Arenas is MSCA Research Fellow at University of Groningen. Her project (CREAMT) looks at the impact of MT on
- 17. xii List of contributors x i i translation creativity and the reader’s experience in the context of lit- erary texts. She is also a Senior Lecturer in Translation and Multimodal Technologies at the University of Surrey (UK), where she is a member of the Centre for Translation Studies. Alicja Zajdel is Predoctoral Researcher at the University of Antwerp, where she is a member of the TricS (Translation, Interpreting and Intercultural Studies) research group. She is currently conducting translation process research on decision- making processes in audio description script writing. Her other research interests include media accessibility, machine translation, and literary translation. She is Secretary to the Editorial Board for the Journal of Audiovisual Translation. Paola Brusasco is Associate Professor in English Language and Translation at the University of Chieti- Pescara. Her research interests and publications are in the areas of Translation Studies, ELT, and Postcolonial Studies. She has translated many contemporary and classic works. Lauri A. Niskanen has PhD in comparative literature from the University of Helsinki and researches James Joyce, literary translation, and inter- textuality. Niskanen wrote his doctoral thesis on the Finnish and Swedish translations of Joyce’s Ulysses and has also published on parody, pastiche, intermediality, polyphony, and the musicalisation of fiction. Maarit Koponen is Professor of Translation Studies at the University of Eastern Finland. Her work addresses the use of translation technology, particularly machine translation. She is one of the co- editors of the volume Translation Revision and Post- editing: Industry Practices and Cognitive Processes published by Routledge in 2021. Sanna Nyqvist is Adjunct Professor (Docent) of Comparative Literature at the University of Helsinki. She is the author of several articles on literary appropriation and copyright. Her contribution to this volume was funded by the Academy of Finland (285279) and the Emil Aaltonen Foundation. newgenprepdf
- 18. 1 DOI: 10.4324/9781003094159-1 Introduction James Luke Hadley, Kristiina Taivalkoski- Shilov, Carlos S. C. Teixeira, and Antonio Toral How to solve the problem of translation The histories of Machine Translation and Translation Studies are funda- mentally intertwined, and not only because both concern themselves with translation. They both developed as focused areas of study in the wake of the Second World War (Tymoczko 2006, 156). At this time, a new awareness of translation as a means by which speakers and writers of other languages can be made intelligible was being led by developments in communications, computational technologies, increasingly mechanised work practices, widespread literacy, and the availability of written materials. During the war, early computers had famously been employed by cryptographers in the race to decode enemy transmissions (Gambier 2018, 132– 133). Fundamentally, these machines were codebreakers that could decipher the cyphers used to encode messages, such that the messages could be decoded in order to make them intelligible. This approach has close parallels with Saussurean linguistic theories, which were preeminent at the time and which had shifted the study of linguistics away from etymology and language change to the analysis and description of linguistic structures, underpinned by the notions of the sig- nified and the signifier (De Saussure 2011, 75). Under this paradigm, the lexical unit used to express something is seen as arbitrary, acknowledging that there is no intrinsic link between a word and the thing it represents (De Saussure 2011, 68). In turn, this notion tends to lead to the conclu- sion that signifiers or words are interchangeable, and, therefore, that one language can be used to indicate the same things as another language, even though the two may have no words in common. Thus, if cryptographical machines could be used to replace one set of signs with another to encode or decode messages, it is reasonable to think that the signs could be replaced by words, and, therefore, languages could be treated as coding systems. Under this paradigm, translation is effect- ively the act of moving between coding systems such that the message is recoded but not fundamentally altered (see Lennon 2014, 137). Kenneth E. Harper (1955, 41), an American Russianist and early participant in experiments in what he calls “mechanical translation”, reasons that
- 19. 2 Hadley, Taivalkoski-Shilov, Teixeira, Toral 2 “since mathematics is itself a language— a set of symbols used to com- municate thought— why can’t computers be used to translate French into English, or Chinese into Portuguese?” Some version of this understanding of translation, that texts in different languages could be “equivalent” to one another in terms of the messages they convey, underpinned much research in Translation Studies for the majority of the second half of the twentieth century. This research could be seen as a search for the solution to the equivalence problem, which made translation a messy, time- consuming, and laborious business. The same understanding informed the early experiments in Machine Translation, which took place in the early years of the Cold War, when American intelligence hoped to develop an automatic tool for the deciphering of Russian materials, essentially seeing the Russian lan- guage as a code to be broken. Early experiments, though crude by today’s standards, appeared to provide a proof of concept for the researchers, who created a system capable of translating over 60 Russian sentences into English and, on the basis of this, assumed that the problem of trans- lation could be overcome in the foreseeable future: “Linguists will be able to study a language in the way that a physi- cist studies material in physics, with very few human prejudices and preconceptions … The technical literature of Germany, Russia, France, and the English- speaking countries will be made available to scientists of other countries as it emerges from the presses” (Macdonald 1954, 8) The experiments and their promised results led to substantial state investment over the following years, though the speed of progress was not as meteoric as had been hoped. Despite early successes in trans- lating simple sentences, training machines to decode messages in one lan- guage and then recode equivalent messages in another was more difficult than had been anticipated. Early systems attempted to imitate language teaching models that relied on rules and exceptions. Thus, grammatical structures were programmed into the systems along with those cases which did not conform to the same structures. The highly complex and labour- intensive nature of this work, coupled with the limitations on storage and processing power available in the mid- twentieth century, led to slow progress. This progress was assessed in 1966 by the Automatic Language Processing Advisory Committee (ALPAC), which determined in its report that the early confidence in Machine Translation’s potential had been overestimated, asserting that “translations of adequate quality are not being provided” (National Research Council 1966, 16). As a result, it recommended that research funding be redirected into more fruitful endeavours. Such endeavours included finding “means for speeding up the human translation process,” the “adaptation of existing mechanised editing and production processes in translation,” and the “production
- 20. Introduction 3 3 of adequate reference works for the translator, including the adaptation of glossaries that now exist primarily for automatic dictionary look- up in Machine Translation” (National Research Council 1966, 34). As a result, funding into Machine Translation- proper was reduced, with the funds channelled into what later came to be known as computer- aided translation. Re-evaluating priorities This move opened the door for the development of rudimentary computer- aided translation features, such as terminology databases, which store previously encountered terms in the source language and their user- defined translations in the target language. The same concept developed into translation memories, which are effectively corpora of previously encountered source language sentences, paired with their previously provided translations. These features developed from the 1990s on into a series of tools that could very well be argued to be indispensable to most professional translators of technical texts by the second decade of the twenty-first century. Meanwhile, Machine Translation had also shifted its focus from rules and exceptions to parallel corpora, entering the Statistical Machine Translation paradigm by the late 1980s (Brown et al. 1988). Instead of relying on manually programmed rules and exceptions, Statistical Machine Translation relies on large bodies of parallel sentences representing both the source and target language. Systems built under this paradigm use statistical inference to determine the most likely parallel to the source text provided by referring to the parallel source- target corpus of sentences. The benefits of these systems are not limited to output quality but also include flexibility and the level of human intervention they imply. With rules- based systems, it is necessary to build one system per language pair, and the effort of programming all the rules and exceptions is very sub- stantial. On the other hand, statistical systems rely on the corpora they are given, meaning that the work associated with building a system for a new language pair focuses on creating the parallel corpus rather than crafting the system itself. Statistical Machine Translation systems improved translation quality, especially for language pairs such as English and French, which have similar grammatical structures and large enough amounts of data for the parallel corpora to be created. However, translation between languages with very different structures, or between languages with less human- translated material to base a corpus on, was still problematic. During this time, Translation Studies too saw a shift of paradigms, from the focus on equivalence that had historically dominated research to a more nuanced examination of translations as sociocultural phenomena. The first steps in this direction had been made several years before, when Toury (1978) and others instigated the shift from prescribing best
- 21. 4 Hadley, Taivalkoski-Shilov, Teixeira, Toral 4 practice in translation activity to describing observed translation activity. Functionalist approaches had come to see translations as texts that fulfil specific roles in their target contexts, as opposed to simply representing their sources, and the field shifted to assessing those roles and the strategies used by the texts to meet them (Snell-Hornby 2006, 51– 56). As a result of two developments, the field shifted from attempts to make overarching theories of translation in the search for equivalence to more granular assessments of translation activities in context. Thus, the field expanded and diversified exponentially in response to the number of contexts in which translation activity is to be found, the case study became the dominant approach, and the theoretical basis around which Translation Studies had previously gravitated— the search for equivalence— lost most of its meaning. Literature and other creative texts Interest in literary translation as a distinct subdivision of Translation Studies could be said to have begun emerging around this time. However, it is important to note that literature had dominated theoretical and pre- scriptive discussions of Translation Studies since the earliest days of the field. And, to this date, it is not clear whether there is, or could be, a clear divide between literary and non- literary forms of translation. Even while Translation Studies diversified into fields ranging from non- professional translators to translation in crisis scenarios, from publishing practices to audiovisual translation, a substantial substratum of research remained squarely focused on the translation of literature in historical or contem- poraneous contexts. A strong branch of research developed around the production of translation historiography, which very frequently focused on works of literature. For example, a whole series of works entitled The Reception of British and Irish Authors in Europe (www.bloomsbury. com/uk/series/the-reception-of-british-and-irish-authors-in-europe/) has been published by Bloomsbury since 2004, covering figures such as Jane Austen (Mandal and Southam 2007), Robert Burns (Pittock 2014), H.G Wells (Parrinder and Partington 2005), and Oscar Wilde (Evangelista 2010). While this series does include figures such as Charles Darwin (Glick 2014), the vast majority are individuals whose work was either fictional or poetic in nature. As has already been noted, literature had a historically prominent place in theoretical studies on translation. However, while professional training in translation and interpreting existed since before the Second World War in some contexts (Gambier 2018, 133), systematic training specifically for the translation of literature and other creative texts was less widely available. However, the early years of the twenty- first cen- tury saw the development of an increasing awareness of the specific skills and training pertinent to the translation of texts of a primarily aes- thetic, rather than primarily functional, nature. Thus, this period saw an
- 22. Introduction 5 5 increasing number of specialist courses on the translation of literature and other creative texts emerge. Eventually, this growing awareness of what sets creative texts apart, and the training needs of translators specif- ically working on them, led to codification in the form of the PETRA- E Framework for Literary Translator Training. This framework is the first of its kind and was originally the product of a network of eight European partners with specialisms in literary translator training: BCLT (Norwich), CEATL (European network), Deutscher Übersetzerfonds (Berlin), ELTE (Budapest), FUSP (Misano), KU Leuven, Nederlandse Taalunie (The Hague), and Universiteit Utrecht, which, by 2022, has expanded to at least 25. It aims to “set up and strengthen the European infrastructure for the education and training of literary translators” (https://petra- education.eu/about-petra-e). As part of this aim, the framework was first produced in 2014, drawing together the research and pedagogical expertise of the network’s partners. The framework sets out to cata- logue, rather than prescribe, the skills and competencies pertinent to con- temporary literary translators, subdividing these competencies into five levels: Beginner, Advanced Learner, Early Career Professional, Advanced Professional, and Expert (https://petra-educationframework.eu). Many of the skills listed inside this framework, including research and evalu- ative skills, overlap substantially with those expected of translators with many different specializations. When the PETRA- E Framework was first developed, literature was still very much beyond the reach of Machine Translation systems, and this relative incompatibility was reflected by the framework’s competen- cies, in which technology was only mentioned in relation to the ability to search the internet. However, this situation was soon to change, since, at much the same time, Machine Translation was experiencing another paradigm shift with the introduction of Neural Machine Translation systems (Bahdanau, Cho, and Bengio 2014). Like statistical systems, neural systems rely on corpora of existing parallel texts in both source and target languages. But the underlying mechanics of how these systems work differ in that statistical systems “chunk” sentences into smaller units which can be processed as they are. On the other hand, Neural Machine Translation systems process each sentence as a whole. But instead of representing the words as they are, the system represents them as numer- ical vectors, which can be used to calculate mathematical relationships, including the distances between words, leading to an improved level of fluency. Thanks to this approach, Neural Machine Translation systems represent a substantial advance in output quality over statistical systems. However, they still suffer from similar limitations, including some which were previously unseen in Statistical Machine Translation systems. For instance, systems that are intended to work in specific domains of know- ledge work best if the training data they are built on also draw from the same domains. It can also be that there is a payoff between generic
- 23. 6 Hadley, Taivalkoski-Shilov, Teixeira, Toral 6 training data and domain- specific training data, meaning that more is not always better than less. The exception, however, is literature and other forms of creative text. Creative- text translation here refers to the translation of texts from one language to another where the texts them- selves pivot broadly on the human creativity employed in their pro- duction. They rely more heavily on aesthetics for their existence than texts that aim to bring about an outcome directly, as in the case of tech- nical texts. Thus, although literary texts— fictional works: novels, short stories, poems, plays, comics, and so forth— have historically occupied the central focus, the broader category of creative texts includes these and also: • non- fictional texts, such as philosophical works, didactic books, and self-help books; • performative works, such as songs, speeches, films, TV shows, and computer games; and • promotional texts, such as commercials, advertisements, and propaganda. While there appears to be a correlation between the quality of domain- specific technical translations produced using domain- specific training data, some, though not all, creative texts challenge this correlation by being highly internally variable. On one end of the literary spectrum are highly popular recent bestsellers with high readability scores brought about by their short, uncomplicated sentences and use of standard vocabulary. However, on the other end of the spectrum are works such as James Joyce’s Ulysses (1922), which has comparatively low readability and generally very low BLEU Machine Translation quality scores (Toral and Way 2018). One of the reasons that Neural Machine Translation systems tend to work better in the specific domains of knowledge on which they have been trained is because these knowledge domains tend to have formulaic constructions that become recognisable and reproducible when enough training data are introduced. On the other hand, creative texts are, to a large extent, defined by their idiosyncrasy, fitting into one and many national, cultural, temporal, and even personal styles. Neural Machine Translation systems generally require training data of many millions of words, organised as parallel sentences. Thus, training a system to translate legal statutes is fundamentally different from training a system to translate sonates, because while all the legal statutes included in the training data may follow a given tradition, the equivalent number of sonates will likely straddle multiple authors, periods, or traditions. Moreover, whereas, in a statute, each sentence can generally be taken as a distinct unit of meaning, enjambment means that, in a sonate, one line may or may not represent one unit of meaning and may also capitalise on this ambiguity to create further meaning. Thus, because a Machine Translation system needs to
- 24. Introduction 7 7 break up a text into units before processing can begin, where such breaks should be introduced in the context of creative texts is not always clear. Broadening the field At the same time as the Translation Studies’ shift from generalized equivalence- based arguments was allowing for greater consideration of context and a questioning of its historic Eurocentrism, the advance into Neural Machine Translation systems facilitated experimentation with new means of dealing with the issue of so- called low- resource languages, those languages that are generally not supranational and do not have large amounts of material that readily lends itself to the creation of par- allel corpora. Previously, these languages could not readily be included in Machine Translation systems because there was insufficient data to achieve a meaningful result. However, Neural Machine Translation systems open new opportunities for such languages, including so- called transfer learning, in which a system is first trained using a high- resource language then a low- resource language that is related. For example, Spanish, a high- resource language, could be used as the basis for training a system to work with Catalan, a comparatively low- resource language that is closely related. Another issue associated with translating creative texts is the com- paratively high rate of referential consistency they exhibit (Voigt and Jurafsky 2012). Referential consistency describes meaning that ties indi- vidual sentences together, often introducing ambiguity if each sentence is considered in isolation. This issue, like other issues of ambiguity, is often not even noticed by human translators, who have a real- world understanding of the contents of the text that underpins their interpret- ation of it. However, the machine has no recourse to any such knowledge. Therefore, in examples such as “the cat tried to climb into the box but it was too small,” a human intuitively grasps that “it” most likely refers to the box into which the cat attempted to climb. However, for the machine, whether “it” refers to the cat or the box most likely comes down to a statistical operation in the training data that is irrelevant to the specific sentence in question, effectively meaning that the choice informing the translated output is a guess. In a language like English, such a guess is unlikely to have a noticeable effect. However, if translated into a language such as French, Italian, Portuguese, or Spanish, where “cat” and “box” belong to different grammatical genders, the effects could be sizeable. For these and many other reasons, literature specifically, and creative texts more broadly, have traditionally been viewed as fundamentally beyond the ken of Machine Translation systems as well as computer- assisted translation systems, which also function most efficiently in contexts with large amounts of repetition and large numbers of for- mulaic constructions. Among literary translation specialists, this senti- ment has traditionally been expressed with a certain amount of hubris,
- 25. 8 Hadley, Taivalkoski-Shilov, Teixeira, Toral 8 where computer- based systems in general are seen as a threat, but one which is kept at bay by the nature of the material. Conversely, Machine Translation specialists have tended to see literary translation as a high cost– low reward activity when compared to the translation of medical, legal, or other technical documentation. However, traditional sentiments change, and, on both sides of the divide, a new generation of scholars has come to ask new kinds of questions over the past ten years (Voigt and Jurafsky 2012; Besacier and Schwartz 2015). In the world of literary translation, a generation of scholars who consider themselves digital natives has arrived who tend away from the subjective description that has often underpinned much case study research in Translation Studies towards empirical evi- dence. In Machine Translation, challenge- oriented scholars have come to describe literary translation as the last bastion of human translation (Toral and Way 2014, 174). Both camps are converging with their dis- creet skillsets on the textual, societal, economic, legal, and technological issues associated with translating creative texts with machines. The year 2019 saw the first CALT (Computer- Assisted Literary Translation) workshop, which was followed by a workshop at the Machine Translation Summit on Literary Machine Translation and a panel at the EST Conference on Technology for creative- text translation. In 2020, the Goethe Institut created an online debate on AI and Literary Translation. In 2021, a full conference on CALT was instigated, there was a panel at the IATIS conference on creative texts, technology and ecology, and the PETRA- E conference devoted a whole day to issues surrounding literary Machine Translation and computer- aided literary translation. Over the same period, seminal publications making the first steps towards synthesizing a range of technological solutions with the translation of lit- erary and other creative texts have been appearing, mostly in the form of the journal articles that are heavily cited throughout this book, but also, importantly, in monograph form. The year 2019 saw the appearance of Youdale’s Using Computers in the Translation of Literary Style: Challenges and Opportunities, which combines Translation Studies’ trad- itional translation and commentary approach with a range of electronic tools that can inform the human translator’s work. Thus, it is clear at this stage that interest in the subject is high and growing rapidly, not only among Machine Translation scholars keen to push the boundaries of what is technically possible but also Literary Translation specialists keen to assess the effects of the advancing tech- nology on texts and readers. This synthesis is bringing about new ways of researching translation for both parties. For Machine Translation specialists, it is increasingly clear that seeing a human translation as the monolithic embodiment of the ideal, as has traditionally been the case, is an overly simplistic perspective on a highly variable process. More and more, it is becoming clear that for what and for whom a translation is produced are also important questions to ask when designing Machine
- 26. Introduction 9 9 Translation systems. Equally, for literary translation specialists, it is clear that, without a quantifiable definition, nebulous but fundamental aspects of text production such as style are not easily analysed empirically, and subjective assessments of textual features can fall flat for an unsympa- thetic audience. However, retaining relevance in translation practice is a substantial challenge for Translation Studies as a whole, as it continues to grapple with the palpable divide with the industry, which has tradition- ally viewed “theory” as useless. Crafting a snapshot This book represents a snapshot of research into this emerging topic at this early stage. It is by no means representative of all the work currently underway on synthesizing technology with creative- text translation. However, it demonstrates not only how far the research has already come in a relatively short period but also what kinds of developments we may begin seeing soon. The chapters are arranged to flow from surveys on existing knowledge through new developments in tools for translating. A further examination of tools, this time in the context of analysing existing translations, follows. Finally, the book moves on to consider the legal and ethical implications of machines being more heavily integrated into human creative- text translation workflows. In Chapter One, Ruffo sets out to assess the state of the relationship between technology and literary translators, asking about translators’ perceived roles in society as well as their attitudes towards the use of technology in literary translation. She goes about this assessment by first establishing the basis on which literary translators build their own self- image and the input that literary translators have had in conversations on the technologiation of translation workflows to date. However, at the core of Ruffo’s study lies a survey of 150 literary translation practitioners from 35 countries, designed to capture their positionality relating to the use of technology and correlate this with other aspects, such as their language pairs or level of experience. Building on Youdale’s distinction between general and translation- specific technology, Ruffo’s findings highlight an important point when considering technology in general as far as it relates to translation of whatever kind— that it is not clear where a line should be drawn between technological and non- technological interventions. While few would argue with the statement that Machine Translation is inherently technological, it is, perhaps, less immediately apparent, but no less true, that an online dictionary or archive, or indeed a word processing application, is also inherently technological in nature, as are paper dictionaries, even though the technology in question may not be digital. In Chapter Two, Daems also makes use of a survey method, focusing on emerging technologies pertinent to literary translation workflows. Daems assesses the awareness and adoption of such technologies among 155
- 27. 10 Hadley, Taivalkoski-Shilov, Teixeira, Toral 1 0 literary translators working into Dutch and establishes the factors that impact a translator’s willingness to adopt new tools into their workflows. Her findings indicate that literary translators may be relatively slow to learn about emerging technologies, implying a kind of vicious cycle of technical translators being the heaviest users of such tools, and, there- fore, the group to which such tools are primarily marketed. A minority of Daems’ respondents appear to hold that technology is inappropriate for the translation of literary texts, implying that it might not be technology in general, but rather the technology that exists currently that is not ideally suited to literary translation. Daems further demonstrates that, despite the potential knowledge gap between the tools that exist and the literary translators who might make use of them, an overwhelming majority of the literary translators she surveyed have an interest in knowing more about technological developments pertinent to them. Thus, it may be that tools specifically aimed at literary translation, which are sensitive to the concerns expressed by literary translators, may be met with less resistance than may be assumed. Turning to one of the functions that such creative- text translation- specific tools might focus on, in Chapter Three, Kolb and Miller assess the usefulness of PunCAT, a tool that assists in the translation of puns. Kolb and Miller focus on the English- German language pair on which the system was originally built by Miller (2019). They evaluate translations produced with and without the tool, the latter done by nine graduate students. Their findings demonstrate that tool use is not always straight- forward, particularly in the context of translating. They find that, in some cases, users’ reactions to the outputs provided by the tool are not as simple as reject or accept but are more nuanced than this, serving as ingredients for brainstorming, and ultimately assisting the translator in coming to an ideal solution. Importantly, Kolb and Miller also assess the translators’ emotional reactions to the use of this tool, finding that, while many appreciated the tool as something that provides suggestions which can be ignored or built on, others found the use of the tool stressful and potentially constraining. These findings are very important both for the future development of the field and for tools that may be developed in the coming years. They show that managing expectations is as important as producing a tool that fulfils a given need. It is important that translators are made to feel that their agency is expanded, rather than constrained, by the tool. Or, to put it another way, that the tool provides one or more possible candidates, but these candidates aim to assist, rather than replace, the human translator’s thinking. In Chapter Four, we turn to the use of Machine Translation as a tool for advanced language learning. Oliver, Toral, and Guerberof Arenas dis- cuss the use of a Neural Machine Translation engine in conjunction with the InLéctor collection of bilingual books for the creation of translated works of fiction that are not intended to be read in isolation but are aids for advanced language learners to decipher the work in the original
- 28. Introduction 11 1 1 language. The underlying principle is that there is a balance to be struck between the speed of Neural Machine Translation and the quality of its outputs. If a language learner is sufficiently advanced to be able to read the work primarily in the original language and only requires reference to a translation as a means of support, the quality of the output may be sufficient to serve this purpose, and the speed by which the output can be produced may make its availability highly attractive. Oliver, Toral, and Guerberof Arenas’ findings show that readers, especially those with a high level of proficiency in the target language, can benefit substantially from the presence of the machine- generated outputs. Specifically, the readers of the bilingual editions, as opposed to monolingual counterparts, found the reading experience easier and more enjoyable. At this stage, it remains to be seen whether these findings transfer into increased learning on the part of the readers or whether finding the answer instantly may hamper retention. Nonetheless, this experiment does stand in very good company with, for example, the Loeb Classical Library, which has been publishing works of classical literature with facing English gloss translations for pedagogical purposes for over 110 years (www.hup.harvard.edu/collect ion.php?cpk=1031). Moreover, the experiment highlights the importance of not seeing translation, whether machine or human, in monolithic terms but as a highly nuanced practice with different requirements depending on intended readership and use. Naturally enough, comprehension works on multiple levels, particu- larly in the case of literary and other creative texts, which may make use of idioms and other devices that problematise understanding through gloss translation. In Chapter Five, Zajdel asks about the specific case of metaphor, comparing the translation of a work of literature into Spanish by a Machine Translation system with the same work translated by human translators. Zajdel subcategorises metaphors into four types, along with idiomatic expressions, and assesses the translation procedures used by a Machine Translation system on 50 of these metaphors found in a single work of literature. She then compares these procedures with their counterparts in two human translated versions of the same text. Zajdel’s findings underscore the importance of not necessarily perceiving a human translation as the zenith of translation quality, as the procedures employed by the two human translators in question vary somewhat. Indeed, this variability is of note, since one of the biggest dividers between the human and machine translators in Zajdel’s findings is the range of procedures employed by each when encountering metaphors. Whereas the Machine Translation system tends to translate each metaphor with a metaphor, the human translators exhibit a wider range of procedures, such as extrapolating metaphors or replacing them with alternative metaphors. Zajdel also finds, to some surprise, that idiomatic expressions tend not to be well translated by the machine in this case, despite such idiomatic expressions presumably finding their way into training data. This finding may be pertinent to future research on idioms and puns as far as training
- 29. 12 Hadley, Taivalkoski-Shilov, Teixeira, Toral 1 2 data are concerned. One might conjecture that idiomatic expressions do not become statistically significant in training data until the point that they can be seen as cliché by human readers. Zajdel’s work is important in dispelling any assumption that Machine Translation is simply incap- able of working with metaphor or is restricted to working on the purely superficial level in this regard. Her results illustrate the creativity that can emanate from the use of Neural Machine Translation systems, which could easily prove to be a highly positive attribute as research in the field of literary Machine Translation develops. In Chapter Six, Brusasco focuses centrally on this issue of creativity in Neural Machine Translation systems. She uses three Neural Machine Translation systems to translate the same extract of a literary text in order to assess the procedures that each undertakes and the extent to which cre- ativity is manifest in each case. Brusasco’s analysis assesses the quality of each translation, not only on the basis of creativity but also on the basis of acceptability in the target context. She also raises the important point that it can and possibly should be taken for granted at this stage that the outputs of Neural Machine Translation systems, particularly in the context of translating literary and other creative works, require human intervention in the form of post- editing. While some literary translators may see this shift as a profound one, where the human is demoted to controlling the quality of the machine’s outputs rather than producing their own outputs directly, taken from another point of view, Neural Machine Translation systems as a whole could also be seen as computer- aided translation systems. In other words, since the human post- editor still retains decision- making agency and can choose to alter or overrule the machine’s outputs, in just the same way as in Kolb and Miller’s study, if human post- editing is taken for granted, the post- editor may rise in perceived importance. Brusasco speculates on the possible effects associated with training Machine Translation systems on works of lit- erature, possibly by collecting texts belonging to single genres or even by single authors, and identifies certain potential issues with such a practice. She observes that such an approach could codify idiosyncrasies of style in Machine Translation outputs, which may have the effect of fossilising or stratifying high and popular literature, in a manner reminiscent of the current stratification between high-and low- resource languages. Niskanen shifts our attention in Chapter Seven to the use that machines can have in supporting and augmenting the kinds of descriptive case study research that have become the norm in Translation Studies. His research focuses on intertextuality in four human translations of the same pastiche- laiden text, asking whether the extratextual cues present in the source text are reproduced in each of the translations. Niskanen’s analysis is based theoretically and terminologically on Genette’s (1997) work, codi- fying the hypertext, hypotext, and paratext. The tool he develops uses an electronic version of the text which contains tags that allow a user to gain further insights on intertextual references present within the text and to
- 30. Introduction 13 1 3 assess their treatment in each of the translated versions. Niskanen’s pri- mary aim in this study is to explore the range of new research questions that such a system may make possible to Translation Studies researchers. He finds that, in the process of analysing these intertextual links, it can be observed that some human translators use the translation procedure of drawing on the target tradition as well as, or instead of, the source tradition. While Niskanen’s work immediately opens up new ways for Translation Studies scholars to bring technology into traditional close reading analytical techniques, it also highlights the research element that lies at the heart of much literary and creative translation practice. It is easy to see that, armed with a tool that identifies and elucidates intertextual elements in a literary work, the element of chance that can underpin such work may be reduced. Translators using such tools may be able to work with a certain level of confidence that any intertextual links missed by the human translator will likely be found by the machine. Naturally enough, as seen in Kolb and Miller’s work, the obverse may also be true, that such tools could lead especially emerging translators into a false sense of security that all intertextual links will be identified by the machine, or that the human translator is obliged to act on the links and only those that the machine has identified. Bringing the book to a close are Koponen, Nyqvist, and Taivalkoski- Shilov in Chapter Eight, whose focus falls onto the legal, technical, and ethical issues of copyright and ownership in the context of creative and literary works translated in part or in whole by machines. Koponen, Nyqvist, and Taivalkoski- Shilov set out by assessing the situation of translation in general in the context of copyright, observing the uneasy relationship between a mode of text production that is inherently deriva- tive and a system intended to control the creation of derivative work. Copyright further operates on the assumption that works have named and identifiable originators whose rights can be asserted in the event of derivations of those works being produced. Koponen, Nyqvist, and Taivalkoski- Shilov rightly point out that much computer- aided transla- tion technology, as well as Machine Translation technology, relies on cor- pora of work produced by many individuals whose precise contribution may or may not be identifiable. Even in simple cases such as individual companies’ Translation Memories, the production of the memories’ contents is a collective process, and the assumption is that individual segments will be reused many times in the production of translations. Koponen, Nyqvist, and Taivalkoski- Shilov revisit the assumption that texts produced by Machine Translation systems currently require human intervention in the form of post- editing by pointing out that there are many cases where the copyright for a work has lapsed, leading to the production of new translations in which no such intervention has taken place. This issue, as Koponen, Nyqvist, and Taivalkoski- Shilov point out, is one of quality and reputation from the point of view of authors. They conclude by calling for a reassessment of copyright practices to reflect the
- 31. 14 Hadley, Taivalkoski-Shilov, Teixeira, Toral 1 4 changing landscape of translation in general. Now that the use of tech- nology has come to be integrated into many translation workflows, such legislation should continue to act as a protective measure for text produ- cers in general and not only for those in positions of power. The missing chapters While it cannot claim to be comprehensive in encompassing all research into the use of machines in the translation of creative texts, this book does offer an overview of some of the key aspects of the emerging topic, which may become increasingly prominent over the coming years. In many ways, these topics are tied to the progress not only of technology but also of our understanding of the processes associated with translating creative works. Research abounds in Translation Studies on the interplay between ideology or philosophy and the translation process (e.g. Mason 1994; Leonardi 2007; Tymoczko 2006), the visibility or not of the trans- lator in the final product (e.g. Venuti 2017), and the effects on target readers of the interpretations underpinning translations (e.g. Ece 2015; Vandaele 2002). On the other hand, in Machine Translation, focus has historically fallen squarely on the question of how to produce translations of the highest possible quality. Now that Neural Machine Translation outputs, in the context of high- resource language pairs at least, have reached the stage of being directly comparable with human translations on more than the superficial or grammatical level (see Toral and Way 2018), it is possible to begin knitting these two areas of exploration together to ask how and whether decisions made in the creation of Machine Translation systems go on to have observable effects on the texts produced that fall beyond the scope of quality control. At the same time, it becomes more meaningful than ever before to begin asking questions of a primarily stylistic nature about text- specific features, genre- defining conventions, and author- particular idiosyncra- sies, and how and whether these are rendered by human and machine translators given the same task. With such features in mind, it is no surprise that experiments are beginning in Machine Translation and computer- aided translation specif- ically in the context of highly stylised or formally constrained traditions of text production such as poetry and song. Questions are beginning to be asked about how such constraints can be harnessed in the production of Machine Translation outputs, and, concurrently, how machines can be used to facilitate the work of human translators working with such texts— for example, by identifying rhyme schemes and metrical patterns automatically. Similarly, advances in artificial intelligence mean that it may soon become possible to make computer- aided translation tools in general work not only more efficiently but also more intelligently. At the
- 32. Introduction 15 1 5 same time, work on the experiences of users working with these systems may see changes to interfaces that could assist in familiarising technology to translators who have been historically resistant to it or found it less than useful. In the coming years, it is likely that the pace of research in these and many other aspects will increase, leading to ever more flexi- bility in translating under formally constrained conditions and other situ- ations relevant specifically to creative texts. Another topic that is not directly handled in this book and is likely to attract attention over the coming years is that of voice dictation. As speech recognition software improves in quality, particularly for high- resource languages, it has been integrated with Machine Translation systems, giving a rudimentary workaround for the interpreting of the spoken word. Interpreting is seen by many as a sister skill to translation, with many of the same concerns as well as additional practical constraints, the most obvious of which is possibly the ephemeral nature of the spoken word. Machine Translation systems and CAT tools, on the other hand, have historically only processed written text, meaning that oral speech has needed to be transcribed before it could be translated. In the context of creative texts with oral and other performative components, such as speeches, plays, and many forms of poetry, conceptualizing the material purely in textual form tends to overlook the performative aspect and the textual fluidity that this creates. It is not currently clear how or whether current Machine Translation or CAT tool systems could be adapted to material that is not in a written form. There are fundamental differences between written text and spoken speech that go beyond their two media of communication. Work on copyright and other legal aspects associated with the pro- duction of translations, of the kind seen in Koponen, Nyqvist, and Taivalkoski- Shilov here, is also likely to become increasingly important over the coming years and as the number of works of literature produced primarily or partly by machines rises. The substantial variation in copy- right law in various jurisdictions around the world, coupled with dra- matically different translation and publication norms and expectations globally, will likely mean that issues pertaining to the legal interplay between human and machine in the production of intellectual property is likely to become substantially more complex as the technology advances. Thus, the primary objective of this book is to capture the state of the art of the use of machines in the translation of creative texts at the first stage of its development, when discussing the field in solid, rather than abstract, terms has become meaningful. The book works in full awareness that, in such a rapidly developing field, the gap between the cutting edge and obsolescence is short. However, the thematic range of the research represented by its chapters also goes some way to showcasing the vast opportunities and challenges that are only now being made apparent to us as we take the first steps into this new landscape of research.
- 33. 16 Hadley, Taivalkoski-Shilov, Teixeira, Toral 1 6 References Bahdanau, Dzmitry, Kyunghyun Cho, and Yoshua Bengio. 2014. “Neural Machine Translation by Jointly Learning to Align and Translate.” arXiv preprint:1409.0473. Besacier, Laurent, and Lane Schwartz. 2015. “Automated Translation of a Literary Work: A Pilot Study.” In Proceedings of the Fourth Workshop on Computational Linguistics for Literature, 114– 122. Denver, CO. Brown, Peter F., John Cocke, Stephen A. Della Pietra, Vincent J. Della Pietra, Frederick Jelinek, Robert L. Mercer, and Paul S. Roossin. 1988. “A Statistical Approach to Language Translation.” In Proceedings, 12th International Conference on Computational Linguistics (COLING-88). Budapest, Hungary, August 1988, 71– 76. De Saussure, Ferdinand. [1959] 2011. Course in General Linguistics. Translated by Wade Baskin. Columbia University Press, New York. Ece, Ayşe. 2015. “The Trajectory of Literary Translation: From Interpretation to (Re) Writing and a New Life.” Synergies Turquie, 8. Evangelista, Stefano, ed. 2010. The Reception of Oscar Wilde in Europe. Bloomsbury Publishing, London and New York. Gambier, Yves. 2018. “Institutionalization of Translation Studies.” In A History of Modern Translation Knowledge, edited by Lieven d’Hulst and Yves Gambier, 131–142. Genette, Gérard. [1987] 1997. Palimpsests: Literature in the Second Degree. Translated by Channa Newman and Claude Doubinsky. Lincoln: University of Nebraska Press. Glick, Thomas F., and Elinor Shaffer, eds. 2014. The Literary and Cultural Reception of Charles Darwin in Europe. Vol. 17. Bloomsbury Publishing, London and New York. Harper, Kenneth E. 1955. “Translating Russian by Machine.” Journal of Communication 5 (2): 41– 46. Joyce, James. 1922. Ulysses. Paris: Shakespeare. Lennon, Brian. 2014. “Machine Translation: A Tale of Two Cultures.” In A Companion to Translation Studies, edited by Sandra Bermann and Catherine Porter, 136– 146. Oxford: Wiley. Leonardi, Vanessa. 2007. Gender and Ideology in Translation: Do Women and Men Translate Differently? A Contrastive Analysis from Italian into English. Bern: Peter Lang. Macdonald, Neil. 1954. “Language Translation by Machine: A Report of the First Successful Trial.” Computers and Automation 3 (2): 6– 10. Mandal, Anthony, and Brian Southam, eds. 2007. The Reception of Jane Austen in Europe. Vol. 13. AC Black, London and New York. Mason, Ian. 1994. “Discourse, Ideology and Translation.” In Language, Discourse and Translation in the West and Middle East, edited by Robert de Beaugrande, Abdullah Shunnaq and Mohamed H. Heliel, 23– 34. Amsterdam: John Benjamins. Miller, Tristan. 2019. “The Punster’s Amanuensis: The Proper Place of Humans and Machines in the Translation of Wordplay.” In Proceedings of the Human- Informed Translation and Interpreting Technology Workshop (HiT-IT 2019), 57–65.
- 34. Introduction 17 1 7 National Research Council. 1966. Language and Machines: Computers in Translation and Linguistics (ALPAC). Washington, DC: The National Academies Press. https://doi.org/10.17226/9547. Parrinder, Patrick, and John S. Partington, eds. 2005. The Reception of HG Wells in Europe. Bloomsbury Publishing, London and New York. Pittock, Murray, ed. 2014. The Reception of Robert Burns in Europe. Bloomsbury Publishing, London and New York. Snell- Hornby, Mary. 2006. The Turns of Translation Studies: New Paradigms or Shifting Viewpoints? Amsterdam: John Benjamins. Toral, Antonio, and Andy Way. 2014. “Is Machine Translation Ready for Literature?” Proceedings of Translating and the Computer 36: 174– 176. Toral,Antonio,andAndyWay.2018.“WhatLevelofQualityCanNeuralMachine Translation Attain on Literary Text?” In Translation Quality Assessment, edited by Joss Moorkens, Sheila Castilho, Federico Gaspari and Stephen Doherty, 263– 287. Springer, Cham. Toury, Gideon. 1978. “The Nature and Role of Norms in Literary Translation.” In Literature and Translation: New Perspectives in Literary Studies, edited by James S. Holmes, José Lambert and Raymond van den Broeck, 83– 100. Tymoczko, Maria. 2006. “Translation: Ethics, Ideology, Action.” The Massachusetts Review 47 (3): 442– 461. Vandaele, Jeroen. 2002. “Introduction: (Re- ) Constructing Humour: Meanings and Means.” The Translator 8 (2): 149– 172. Venuti, Lawrence. 2017. The Translator’s Invisibility: A History of Translation. Routledge, London and New York. Voigt, Rob, and Dan Jurafsky. 2012. “Towards a Literary Machine Translation: The Role of Referential Cohesion.” In Proceedings of the NAACL- HLT 2012 Workshop on Computational Linguistics for Literature, 18–25. Youdale, Roy. 2019. Using Computers in the Translation of Literary Style: Challenges and Opportunities. Routledge, London and New York.
- 35. 1 8 DOI: 10.4324/9781003094159-2 1 Collecting literary translators’ narratives Towards a new paradigm for technological innovation in literary translation Paola Ruffo Introduction Research on technological innovation in translation tends to systematic- ally exclude literary translation. This trend was best captured by Toral and Way when they defined the latter as “the last bastion of human trans- lation” (2014, 174). The rationale behind this is mainly twofold. On the one hand, the very nature of creative texts almost implies an inherent degree of resistance to automation. In fact, they are characterized by “vocal multilayeredness and deliberate ambiguity” (Taivalkoski-Shilov 2018, 695), which makes them uniquely inscrutable to the machine’s eye. On the other hand, literary translators’ self- imaging strategies are rooted in the creation of idealized personae, which revolve around their most human qualities and further remove them from the wider discourse surrounding other branches of translation (Sela-Sheffy 2008). In view of this, there would seem to be little to no place for the adoption of tools such as Computer- Aided Translation (CAT) and Machine Translation (MT) for the translation of creative texts. However, despite such tools often being perceived “as either inappropriate or a threat to the skills and livelihoods of literary translators” (Youdale 2019, 199), an increasing number of studies are focusing on the introduction of translation tech- nology to literary translation workflows. These have mainly focused on the application of MT and post- editing to the translation of poetry and prose (Genzel, Uszkoreit, and Och 2010; Greene, Bodrumlu, and Knight 2010; Voigt and Jurafsky 2012; Jones and Irvine 2013; Toral and Way 2014, 2015a, 2015b, 2018; Besacier and Schwartz 2015; Tezcan, Daems, and Macken 2019; Toral, Wieling, and Way 2018; Murchú 2019). This being said, little attention has been given to literary translators as end users of such tools. Furthermore, their voices are consistently missing from the discourse around technological innovation in their profession, with only a few recent studies relating to their attitudes and perceptions, namely Moorkens et al. (2018), Slessor (2019), and Kenny and Winters (2020).
- 36. Collecting literary translators’ narratives 19 1 9 The characterisation of translation as a form of Human- Computer Interaction (HCI) by O’Brien (2012) ultimately highlighted the need to account for both material and immaterial aspects when investigating contemporary translation. In this respect, Littau (2016) defines materi- ality as the non- human element, e.g., digital tools, machines, computers. Conversely, immateriality is embodied by creativity, spirit, mind, con- science, soul, and anything associated with being quintessentially human (Littau, 2016). In her view, both material and immaterial elements interact symbiotically and reciprocally shape and influence each other to the point where “[technologies] are active in effecting the ways in which we think, read, write and translate” (Littau 2017, 100). Furthermore, Cronin notes how the constant stream of information, incessant digital- isation of materials, automation of processes, and speed of communica- tion that are typical of the Information Age contribute to an overall “sense of confusion” (Cronin 2013, 1). This “sense of confusion” characterizes translators as they try to give meaning to this new order of things (Cronin 2013, 1). Thus, it is paramount to include translators in the conversa- tion and, by giving them a voice, perhaps discover new ways for literary translators to exist in this new socio- technological landscape, as well as co- exist with new technologies. This chapter is based on a 2018 study aimed at exploring the dynamic between human (immaterial) and non- human (material) factors in literary translation, recognising materiality as central to contemporary transla- tion practice and trying to bring literary translators’ voices back into the conversation. More specifically, literary translators were asked to share their attitudes towards technology and perceptions of their role in society. The study’s main research question was “what is the dynamic between humans and technology in literary translation?”. Two sub- questions were formulated to assist and guide the research process, respectively (a) “how do literary translators perceive their role in society?” and (b) “what are their attitudes towards technology as related to literary translation?”. Respondents’ narratives were collected via means of a questionnaire that registered 150 responses, mostly from Europe. Overall, the study adopted an interpretivist, constructionist, and mixed- methods approach. The theoretical framework and data analysis were informed by the Social Construction of Technology (SCOT) framework as theorized by Pinch and Bijker (1984). Although the overall relationship between self- image and technology will be briefly discussed in this chapter, its main focus will be on how participants constructed the notion of technology as related to their professional practice. The chapter will first provide a review of the literature on literary translator status and the application of technology to literary translation workflows. It will then introduce the study’s meth- odology and present the research findings on literary translators’ attitudes towards both general and translation technology tools. Finally, it will dis- cuss the results and suggest a way forward for research on the topic.
- 37. 20 Paola Ruffo 2 0 Literary translators’ self- imaging strategies Literary translators are often depicted as having a different perspective on their profession than that of their readership, clients, or colleagues in other branches of translation (Sela- Sheffy and Shlesinger 2008; Katan 2017; Ruokonen and Mäkisalo 2018). In particular, this seems to be a direct response to the way their profession is depicted by outsiders, which usually clashes with the way literary translators see themselves. In this respect, their self- imaging strategies can be said to be based on an anti- professionalisation discourse through which literary translators elude any form of standardisation and institutionalisation in order to affirm their professional identity (Sela-Sheffy 2005, 2010, 2016). According to Sela-Sheffy (2008), literary translators adopt three main idealized personae, mainly to oppose their profession’s low status: cus- todian of language, cultural ambassador and innovator, and artist. As custodians of language and cultural ambassadors, literary translators por- tray themselves as gatekeepers of entire cultural and linguistic systems, in that they determine both what enters the translated literature ecosystem and how, effectively shaping the literary landscape in which they operate (Sela-Sheffy 2008). In a similar way, they are innovators of said systems, introducing new works of literature to an audience who would be unable to access them otherwise (Sela-Sheffy 2008). Lastly, their ability to bring creative texts to life in another language is often described with words belonging to the semantic fields of artistry and craftmanship, highlighting the creative effort involved in their work (Sela-Sheffy 2008). These accounts are further probed by other research on the topic, where literary translators’ identities have been found to be deeply related to their perceived professional status. The dynamic between identity and status results in a tendency to amplify traits associated with personal qualities and circumstances and to place emphasis on characteristics that are hard to quantify in terms of professionalisation, such as voca- tion and creativity. In this regard, the literary translation career path is often depicted as the result of a natural inclination, an almost inevitable occurrence, more than a professional choice (Sela-Sheffy 2005, 2008, 2010, 2016; Sapiro 2013; Voinova and Shlesinger 2013). According to Heino (2020), literary translators prioritize social and cultural cap- ital over economic capital. Furthermore, the line between writing and rewriting is often blurred, and literary translators emerge as agents of art- istic creation, often assuming the role of directors as well as performers in the obscure process of the translation and dissemination of literary works (Jänis 1996; Sela-Sheffy 2008, 2016; Sapiro 2013; Voinova and Shlesinger 2013). Ultimately, literary translators’ symbolic capital is structured around the need to actively respond to outside narratives of low professional status, which, if not perceived as threatening, are at least viewed as unrepresentative of their lived truths. In order to oppose these narratives,
- 38. Collecting literary translators’ narratives 21 2 1 literary translators construct their self- image by emphasising their unique traits and by creating distance from the outside world, a quality that Voinova and Shlesinger call a “strange outsiderness” (2013, 41). Literary translation and technology As mentioned in the introduction, despite translation technology being an integral part of contemporary translation workflows, its use in literary translation and by literary translators is understudied. This has started to change in the past few years, with an increasing number of research projects exploring the potential uses of MT, post- editing, and CAT tools for literary texts. The first of these studies focused primarily on MT, and results highlighted how it struggled to preserve both meaning and form with literary and poetic texts (Genzel, Uszkoreit, and Och 2010; Greene, Bodrumlu, and Knight 2010; Jones and Irvine 2013). This being said, sub- sequent studies found that factors such as predictability of the text and relatedness of the language pair could improve MT output, thus making it more suitable for post- editing (Toral and Way 2014, 2015a, 2015b). In this respect, Besacier and Schwartz (2015) found that including post- editing in the translation workflow halved translation time, although at the expense of quality. The introduction of Neural Machine Translation (NMT), which, unlike its predecessors, uses artificial neural networks to predict translations, is rapidly changing this. In fact, most of the studies cited above used Statistical Machine Translation (SMT), which works at phrase level and uses probability to determine its output. NMT, instead, considers both the source and content that has already been translated in the target text. Recent studies that employed NMT noted an increase in both productivity and MT output quality, with 17– 34% of output being evaluated as equivalent to human translation (Toral and Way 2018). This being said, the studies mentioned so far seem to focus primarily on improving productivity and reducing costs, while practitioners’ wants, needs, and support are rarely considered. In this respect, Youdale (2019) opted for taking the spotlight off MT and post- editing to leave more space for the exploration of an alternative technological workflow revolving around the literary translator. In doing so, Youdale (2019) introduces the close and distant reading (CDR) approach. This leverages corpus linguistics and text- visualisation tools to support and enhance the process of literary translation while respecting the translator’s work- flow and prioritising their point of view. With a similar premise, Youdale and Rothwell (forthcoming) challenge the notion that CAT tools are inherently incompatible with literary translation, investigating ways and situations in which their functions might indeed assist the trans- lator and enhance their work. This might be the case for retranslation, for example, whereby the co- presence of the source text and previous translations could not only highlight connections between them that may otherwise have been lost but could also give the texts a multidimensional
- 39. 22 Paola Ruffo 2 2 character with the potential of shedding new light on both the source and its retranslation (Youdale and Rothwell, forthcoming). Despite an evident shift towards human- centred approaches to technological innovation in creative- text translation, the voices of lit- erary translators are still virtually absent from research accounts of technological innovation in the profession. This is especially surprising given that literary translators are arguably the ones who are most likely to be affected by the introduction of such tools in their daily practice. At the time of writing, the only exceptions to this are Moorkens et al. (2018) and Slessor (2019). In the former, literary translators indicate a marked preference for translating from scratch rather than post- editing MT output, which they feel hampers their creativity and leads towards a more literal rendition of the text (Moorkens et al. 2018). Slessor (2019), instead, reports the findings of a survey on literary translators’ attitudes towards technology. Results indicate that literary translators employ sev- eral standard tools and electronic resources, while translation technology is almost absent from their practice (Slessor 2019). Furthermore, when they do employ translation technology, they do so in unique ways, which suggests a need to review technology training for literary translators, considering their distinctive approach to technology adoption (Slessor 2019). These studies indicate that the reasons behind translators’ rejec- tion of technology might not always be straightforward. In non- literary translation, Koskinen and Ruokonen (2017) found that translators are not averse to technology as such but rather to poor usability and tool malfunctions that hinder efficiency and productivity. They ultimately propose user- centred translation technology design and development as a possible solution (Koskinen and Ruokonen 2017). Overall, research on literary translators’ relationship with technology has highlighted a discrepancy between the focus of translation technology research and practitioners’ attitudes. In this respect, Taivalkoski-Shilov (2018) shines a light on how many of the studies on MT in literary trans- lation seem to neglect narrative aspects of literary texts and separate con- tent from form when evaluating translation quality. The separation of content and form is also inherent in the MT+ PE pipeline itself, which seems incompatible with literary translation, in that it separates struc- ture and content (Taivalkoski-Shilov 2018, 694). In fact, it prevents translators from working on the text as a narrative whole since “the segment- by- segment or sentence- by- sentence translation made by the machine cannot but alter the meaning and structure of the source text” (Taivalkoski-Shilov 2018, 694). This being said, MT and other transla- tion technologies could still be useful to the translation of literary text, provided their introduction is the result of sustainable development involving all stakeholders (Taivalkoski-Shilov 2018). From this perspec- tive, it is not the nature of translation technology itself that should be criticized but the discourse surrounding it and the lack of inclusion of all interested parties in the innovation process.
- 40. Collecting literary translators’ narratives 23 2 3 Since the introduction of translation technology in non- literary trans- lation workflows, translators have reported feelings of devaluation and dehumanisation, as well as a worsening of working conditions and pay (O’Brien 2012). In order to prevent this from happening in literary trans- lation, it is paramount to proactively explore the boundaries of Human- Computer Interaction in literary translation and identify areas for improvement, collaboration, and change. As suggested by Large (2018), by attending to the more mechanical tasks, translation technology could ultimately encourage more creativity for the human involved. One of the broader aims of the study reported in this chapter was indeed to explore this newfound interaction between material and immaterial elements in literary translation and identify possible solutions to reduce the gaps between research, practice, and development in the process of techno- logical innovation by centring literary translators’ viewpoints. Some of these possible solutions will be presented in this chapter when discussing results. The next sections will delve deeper into the study’s methodology and the data collected on participants’ attitudes towards technology and technology use. The SCOT framework The theoretical and methodological structure of the study was supported by the Social Construction of Technology (SCOT) framework, which was theorized by Pinch and Bijker (1984) to study technological innovation from a sociological standpoint. It is characterized by being a multidirec- tional model, in that it takes into consideration not only the final version of a technological artefact as resulting from a linear development pro- cess but also all of its variations before it reached the stage of closure. This allows the researcher to lead a retrospective social constructionist analysis of technological innovation by accounting for the problems and solutions that emerged from contrasting meanings assigned to it by different social groups. The framework refutes technological determinism, in that it is not technology that determines society. Rather, in order to be accepted in society, every new piece of technology goes through a process of variation and selection until all issues raised by relevant stakeholders are agreed upon, solved, or a compromise is reached. In practice, a SCOT- informed analysis allows for a retrospective analysis of this process and consists of three main stages. In the first stage, all social groups relevant to the devel- opment of a certain technological artefact and their varying interpret- ations of it are identified. The objective of the second stage is to identify any conflicts that arose from these differing interpretations and how these were ultimately solved and stabilisation reached. This is usually achieved by devising what Pinch and Bijker (1984) call an appropriate “closure mechanism”. During the third stage, research findings are reported to the wider sociocultural context.
- 41. 24 Paola Ruffo 2 4 SCOT has been identified by Olohan (2017, 2019) as an ideal frame- work for the study of technology in translation from a sociological rather than historical standpoint. In fact, looking at technological development through a SCOT lens could help foster a better understanding of the social and power dynamics underlying technology acceptance, its rejec- tion, and the consequences of its introduction in professional translation practice (Olohan 2017). In this study, the SCOT framework was adopted at a conceptual level to guide data collection and analysis, with some of its tenets having been amended to serve the study’s research questions. In Pinch and Bijker’s (1984) words: “our model is not used as a mould, into which the empirical data have to be forced, coûte que coûte. […] Its function is primarily heuristic— to bring out all the aspects relevant for our purposes” (Pinch and Bijker 1984, 419). More specifically, this study is not retrospective in nature, as its object is not the ex post facto exploration of a tool that has already reached closure. Instead, SCOT is here adopted proactively, in order to address controversies as they arise in the present and give voice to literary translators as a relevant social group. Furthermore, instead of identifying all relevant social groups and analysing one specific artefact, this study focuses on literary translators only and on technology in general. This is due to the lack of previous studies in this area and to the limited timeframe of the project, which led to prioritising literary translators as a social group whose livelihood is more likely to be affected by socio- technological changes in their pro- fession. Furthermore, the study is not ethnographic in nature— as SCOT research usually is— due to literary translators’ technology use being an under- researched area, which thus calls for the need to survey this aspect before proceeding with an ethnography of specific tools. The overall aim when employing SCOT was to pre- emptively identify emerging issues in the relationship between materiality and literary translation, as well as devise potential solutions. Methodology The study adopted an interpretivist, constructionist, and mixed- methods approach. It is interpretivist, in that it “prioritizes people’s subjective understandings and interpretations of social phenomena” (Saldanha and O’Brien 2014, 11– 12)— in this case, literary translators’ narratives of the technologisation of their profession. In doing so, it also recognizes their role as “a constructing and constructed subject in society” (Wolf 2007, 1) and as agents of sociocultural change. An interpretivist and social con- structionist analysis thus allows insights into the way practitioners inter- pret and assign meanings to their professional reality, which is a pivotal step for a sustainable technological development. As far as the mixed- methods approach is concerned, this was adopted to maximize the poten- tial of a large data set, with quantitative elements helping to support and contextualize the qualitative nature of participants’ narratives. Both
- 42. Collecting literary translators’ narratives 25 2 5 elements were combined in a questionnaire by using closed and open questions. The questionnaire consisted of 31 questions and was divided into 6 main sections, namely Background Information, Language Skills, Professional Practice, Confidence with Technologies, Open Questions on perceptions of role and attitudes towards technology, and Final Comments. The lan- guage used in the questionnaire was English and all answers received were in English except for one written in German, which was translated by the researcher. The sampling frame was defined using non- random sampling and by targeting literary translators in UK translation associ- ations and online translation communities. The questionnaire was sent out to a total of 13 UK general and literary translation associations and 68 online translation communities. The latter included online forums, Facebook and LinkedIn groups, and mailing lists aimed at both general and literary translation. Due to the non- random nature of the sample, results cannot be statistically generalized; instead, theoretical and logical generalisation was sought. In this respect, Luker (2008) notices how, even when sampling issues do not allow for statistical generalisation, it is still possible to work on a logical and theoretical level of abstraction. When theoretically generalising, findings are compared to previous theories and studies “to see how [they] illuminate, contradict, extend, or amplify existing theory” (Luker 2008, 127). The questionnaire was live on the Online Surveys (formerly BOS) plat- form for six weeks between September and October 2018, and it was completed by 150 respondents. This chapter presents and discusses data related to participants’ professional and educational background, their levels of confidence with technology, the technology tools used in their practice, and their overall attitudes towards technology in literary transla- tion. It is worth mentioning that the questionnaire did not include closed items about specific tools. Instead, it only distinguished between general and translation- specific technology. General technology was defined in the questionnaire as “any technology tool that is not translation- specific (e.g., online dictionaries, a time management app, a text- editor software, etc.).” Translation- specific technology was defined as “any technology tool that is translation- specific (e.g., Translation Memory systems, ter- minology management software, Machine Translation systems, etc.).” Given the exploratory nature of the study and the lack of previous research in this area, respondents could indicate the tools they use in their daily practice and express their attitudes with as few restrictions from the researcher as possible. In turn, results could form the basis for future research on specific tools. The data analysis consisted of three main phases. During the first stage, a thematic analysis of the open questions was performed. The coding pro- cess was supported by the use of NVivo, a software package for quali- tative and mixed- methods data analysis. The second phase focused on quantifying and collating results from the closed questions; no statistical
- 43. 26 Paola Ruffo 2 6 analysis was performed. During the third stage, data related to age, edu- cational background, and professional status was cross- referenced with data on perceptions of role and attitudes towards technology to delve deeper into the relationship between respondents’ background and the way they construct their self- image and approach technology. The SCOT framework informed the structure and analysis of the Open Questions section of the questionnaire. In particular, questions about literary translators’ feelings and attitudes towards technology and its appealing and unappealing aspects were designed to uncover respondents’ inter- pretations as a social group and any emerging controversies in relation to other stakeholders. Results and analysis Respondents’ profile The questionnaire attracted 150 respondents from 35 countries and all age brackets, providing a large and varied sample. The majority of respondents belong to the age groups 36– 45 (23%) and 46– 55 (28%), while the youngest respondents (18– 25) are the least represented, making up 6% of the total. Three- fourths of respondents work in Europe. They are mostly based in the UK (24%), and almost half have English as their first language. As far as respondents’ academic background is concerned, 63% hold an academic qualification in translation— 40% have a Master’s degree and 9% a PhD— however, only 20% have received translation technology training as part of said qualification. A quarter of respondents have received non- academic training in translation technology. In terms of professional background, there is an almost equal number of respondents with 1– 5 years of experience (27%) and those with over 20 years of experience (26%). Almost all of them work as freelancers (87%) and define their status as “professional literary translator” (83%), while more than half work part- time (58%). Finally, 65% are members of a translation association and 88% of an online community. Confidence with technology Participants were asked to indicate on a Likert scale their levels of confi- dence with general and translation- specific technology, respectively. The great majority of respondents indicated being either “Confident” (44%) or “Extremely confident” (35%) with general technology, with only 3% being “Not confident at all” (Figure 1.1). The situation appears considerably more complex when looking at data for confidence with translation technology (Figure 1.2). In fact, levels of confidence drop considerably, with a quarter of respondents indicating they are not confident at all with translation technology. When compared with the previous question, values are halved for the answer
- 44. Collecting literary translators’ narratives 27 2 7 “Confident” and reduced by two- thirds for the answer “Extremely confi- dent”. Notably, 11% mention that the question is not applicable to them, suggesting they are either not aware of or not using any translation technology tool. The youngest respondents (aged 18– 25) are the most confident with translation technology, while the least confident are the oldest, aged over 65. Overall, levels of confidence start decreasing for everyone aged over 36. Conversely, age does not affect confidence with general technology. Educational background has a bearing on confidence with both general and translation technology, as those with a postgraduate academic quali- fication in translation tend to be the most confident. When looking at translation technology, 45% of those with a Master’s degree and 39% of those with a PhD are either “Confident” or “Extremely confident”. These valueslowerto29%and21%respectivelyforthosewithanUndergraduate degree and those with no academic qualifications. Furthermore, levels of confidence rise considerably for those who have received translation tech- nology training, both academic and non- academic. In particular, 63% of respondents who received academic technology training were either “Confident” or “Extremely confident”, as opposed to 28% of those who did not receive academic training. Values are similar for non- academic technology training, with 50% of those trained being either “Confident” or “Extremely confident” versus 29% of respondents without technology training. Finally, professional status appears unrelated to how confident literary translators are with translation technology. Figure 1.1 Confidence with general technology. Figure 1.2 Confidence with translation technology.
- 45. 28 Paola Ruffo 2 8 Technology use Two open questions asked participants to list, respectively, all the general and translation technology tools they employ in their literary transla- tion activity. For general tools, 82% of respondents mentioned online dictionaries, which top the list, while 38% mentioned, respectively, text- editing software and internet search (Table 1.1). Other popular tools are digital glossaries, terminology databases and thesauri (11%), and task- and time- management apps (10%). Additionally, 9% of respondents said they use no general technology whatsoever. Overall, 397 responses were recorded, and the tools mentioned were grouped into 29 categories. The great variety of tools revealed how, while most respondents agree on online dictionaries, text- editing software, and internet search, both their definition of general technology and the technological customisation of their workflow present widely differing degrees of complexity. In fact, answers include basic hardware, such as mouse, keyboard, and screen, as well as slightly more complex tools, such as bookkeeping software and web hosting services, and highly specific technology (for instance, speech recognition software, desktop publishing, and alignment tools). This suggests that literary translators’ technological landscapes— and their notion of technology— could be as unique as the translators them- selves. Finally, it is worth noting that, while the number of mentions for each tool gives an indication of the ones that are most widely used, it is possible that some of these are so integral to their workflow that many respondents might not have thought about mentioning them at all. Data on translation technology use also supports the distinctiveness of literary translators’ relationship with technology, with 71% of respondents reporting not using any translation technology tools in their practice. Of all the different tools mentioned by those who employ trans- lation technology in their literary translation practice, the vast majority Table 1.1 General technology tools in literary translation (selection) General technology tools Count % Online dictionary 123 82% Text-editing software 58 39% Internet search 57 38% Digital glossary/terminology database/thesaurus 17 11% Task-/time-management app 15 10% None 13 9% Laptop/PC 11 7% Microsoft Office suite 9 6% Social media/ Online communities 9 6% Corpora 9 6% Speech recognition software 6 4% Total respondents: 150
- 46. Collecting literary translators’ narratives 29 2 9 were CAT tools, with only a few mentioning MT and terminology man- agement tools (Table 1.2). Although only a few participants further commented on their use of translation technology, their answers still pro- vide an insight into potential reasons for tool adoption or rejection and the alternative ways in which literary translators might adapt existing technology to their specific needs. More specifically, four respondents described how they use CAT tools for literary translation: one uses them to compile the first draft of a translation, one only when working with a specific agency that requests it, and the other two are in the process of learning how to use them. Finally, it is worth noting that all 43 respondents who reported using translation technology in their practice have high levels of confidence with technology, being either “Confident” or “Extremely confident”. Furthermore, slightly less than half (n= 21) of those using translation tech- nology have received training (academic or not), and virtually all of them (n= 39) also use translation technology in their non- literary translation work. Having seen in the previous paragraph how training increases con- fidence, it would seem that literary translators who are more familiar with translation technology— be it because of specialized training or having to use it for other types of translation— are also more prone to integrating these tools into their workflows. The following sections will explore this further by detailing how literary translators’ attitudes towards tech- nology play into this dynamic. Appealing aspects of technology One of the open questions asked participants to list appealing aspects of technology, whether general or translation- specific (Table 1.3). The most appealing aspect of technology concerns research. In particular, respondents mentioned online dictionaries, internet searches for context clarification, asking colleagues and experts for advice on translation solutions, and the ability to access a great amount of information instantly via the internet. As some respondents put it, appealing is anything “that allows fast information retrieval and fast working and Table 1.2 Translation technology tools in literary translation Translation technology tools Count % None 107 71% Translation technology 43 29% CAT tools 38 25% MT 10 7% Terminology management tools 7 5% Subtitling software 1 1% Total respondents: 150
- 47. 30 Paola Ruffo 3 0 reworking of texts” and facilitates “the breadth and depth of research needed for a good literary translation to be done from home.” Technology is also appreciated when it improves efficiency— for example, by reducing time spent typing or proofreading and editing a text. Dictation and voice- recognition software, Translation Memory (TM) and autosuggest features all make an appearance in this category. By assisting translators with “the housekeeping aspect”, this type of technology can “[take] out the donkey work” and allow them “more time to focus on the creative part of [their] work.” Technology was also praised for its ability to com- pensate for human shortcomings, especially in relation to memory. In fact, several responses mentioned TM being effective in “[reminding] the translator of previous renderings (perhaps also the need to avoid them)” and providing “a searchable record of [one’s] decision- making process.” Overall, an appealing tool is one that “helps [one] concentrate on trans- lating instead of having to fiddle around with complicated systems.” To add to this, one respondent notes: “I would welcome technology […] that enabled me to make better quality choices […]— quality being judged in my subjective experience of freedom and self- expression.” Ultimately, literary translators welcome technology that enhances their practice by assisting with all aspects surrounding the act of translation— be it in terms of accuracy, consistency, reducing editing time, or accessing previously translated content— rather than interfering with the act of translation itself. This becomes even clearer in the next section, which looks at their narratives of unappealing aspects of technology. Unappealing aspects of technology While appealing aspects were mainly associated with general technology and TM, the discourse on unappealing aspects is almost exclusively focused on translation technology (Table 1.4). In this respect, TM resurfaces here as a hindrance to human translation, being described as causing memory to grow lazy and language to become standardized. According to respondents, TM is in direct opposition to literary Table 1.3 Appealing aspects of technology (selection) Appealing aspects of technology Count % Research 58 39% Efficiency 33 22% Assistance to human 27 18% Accuracy and consistency 24 16% None 23 15% Networking 14 9% Total respondents: 150
- 48. Collecting literary translators’ narratives 31 3 1 translation’s essence. In fact, it “takes away from the need to artfully turn a phrase over and over until it is right, accurate, and artful.” Furthermore, others expressed concerns regarding the impoverishment of language, with one respondent noting: “I doubt we’d agree on a voice.” This being said, stronger attitudes are expressed in regard to MT and AI, which are associated with a potential worsening of working conditions and wanting to substitute human translators. What is unappealing about this type of technology is that it “tries to bypass the human understanding of lan- guage and its nuances in order to save costs” and, according to another participant, “attempts to push the boundaries of technology within an essentially contemplative profession which requires an unfashionable degree of isolation and respect for experience.” Another theme that emerges here is that of translation technology disrupting the translation process and diminishing translator autonomy. CAT tool segmentation and MT are often mentioned as examples of this, with one respondent simply stating: “I don’t want a machine singing my part of the duet for me.” In addition to this, narratives of replacement surrounding some of these tools are also a source of frustration for literary translators, who simply do not believe MT to be compatible with literary translation. The uneasiness that participants feel in this respect is engendered by what they perceive as a misunderstanding of what literary translation entails and the misuse of technology in the name of saving time and costs. The fear is that “all the talk by technology buffs who claim that [this] art/ profes- sion will be obsolete” will “[lead] publishers to think that all they need is a good translation tool and a skilled editor.” Finally, other unappealing aspects mentioned include accessibility of translation technology tools in terms of costs and learning curve, complicated user interfaces, and ineffi- ciency caused by too many or wrong inputs. Ultimately, when talking about unappealing aspects of technology, the focus shifts to tools developed specifically for translation. In par- ticular, MT and TM are perceived as a hindrance to the translator, as well as causing disruption to the translation process, with the discourse surrounding them threatening translators’ livelihoods in the future. Table 1.4 Unappealing aspects of technology (selection) Unappealing aspects of technology Count % Hindrance to human 41 27% None 30 20% Disruption/ Loss of autonomy 27 18% Outsiders’ narratives 21 14% Usability and access 17 11% Inaccuracy and inconsistency 9 6% Total respondents: 150
- 49. 32 Paola Ruffo 3 2 Attitudes towards technology One of the open questions devised to elicit respondents’ attitudes towards technology asked them to share their feelings about the relationship between the latter and literary translation. Overall, 49% of participants expressed positive feelings. These were followed by 20% feeling torn and 13% demonstrating negative attitudes. The rest of the responses mentioned indifference, uncertainty, and the belief that technology is in no way relevant to literary translation. The majority of respondents felt lucky, comfortable, confident, happy, grateful, relaxed, and even excited about technology in literary translation. Among those who demonstrated ambivalence in their attitudes are those who are torn between love and hate, gratitude and anxiety, and, again, hate and thankfulness. These feelings seem to originate from uncertainty regarding the future role of technology in the profession and the nature of some translation tech- nology. In particular, some of the unappealing aspects of CAT tools and MT are reprised here, whereby the former are perceived as “inflex- ible”, while some respondents are “uncomfortable about the rise of machine translation,” which is “good in theory, but potentially abusive.” Conversely, technology that facilitates networking and communication is appreciated, together with the internet, the virtually instantaneous availability of electronic resources, and online dictionaries, which is in line with what was reported earlier in relation to general technology use. Finally, negative feelings were mainly directed at the future of the pro- fession and the potential role of translation technology in it. These were feelings of apprehension, sadness, uneasiness, or anger. One respondent states: “I should not be expected to use MT and if I am, I will probably leave the job to someone else.” With the aim of further uncovering their narratives of technology in their profession, respondents were also asked to describe how they see technology as related to their profession (Table 1.5). Differences with the previous question are immediately evident, in that the results are less polarized. In fact, the largest group of respondents either regarded Table 1.5 Relationship between literary translators’ self- image and technology (selection) Relationship with technology Count % Ambivalent 36 24% Helpful 36 24% Less or not helpful for literary translation 23 15% No relationship 19 13% Resistance 9 6% Imposed 5 3% Total respondents: 150
- 50. Collecting literary translators’ narratives 33 3 3 technology as helpful or expressed ambivalent attitudes, while 15% stated that there is no relationship between the two or that technology is less or not helpful for literary translators. The differences between general and translation technology first noted when looking at appealing and unappealing aspects of technology re- emerge even more clearly in the answers to this question. In particular, the ambivalence originates from thinking of general technology as helpful for research, terminology, and networking, while translation technology once again is a source of concern in terms of disruption of the trans- lation process. One participant thinks such tools “discourage freedom in interpreting text on larger scales,” while another states: “I wonder how much they interfere with my originality.” Corpora, online search engines, and dictionaries are often referred to as helpful, together with any technology that aids in “sorting thoughts, terminology, and other things you’d have to keep in your head otherwise.” Translation tech- nology is also at the centre of participants’ more negative attitudes. These narratives see translation technology as not relevant or less helpful for literary translation than for other kinds of translation and revolve around the incompatibility of such tools with the complexity of the literary trans- lation task. Epitomising this viewpoint, one respondent states: “I have the impression those [translation technology tools] are for people in a hurry. I work slowly and carefully”. Finally, a few literary translators mentioned being unwilling to adopt translation technology regardless of its usefulness and feeling that technological innovation in literary transla- tion is an imposition from above rather than a response to practitioners’ actual needs. When linking attitudes towards technology with respondents’ backgrounds, those aged between 18– 25 have the most positive rela- tionship with technology: 50% thought of technology as helpful for lit- erary translation, while the rest described the relationship between the two as either necessary (25%) or harmonious (25%). Generally, those aged 46+are more inclined to think of technology as being unrelated to literary translation. In fact, the “No Relationship” category barely appears in respondents below 45 years of age (it does not appear at all in the 18– 25 group); however, it occurs for all 46+respondents (14% of the 46– 55 group, 23% of those aged 56– 65, and 15% of the over 65s). Those who have received academic translation technology training also view technology more positively. In particular, 70% of those with aca- demic training believe the relationship between literary translation and technology to be a positive one, against 45% of respondents without academic training. The latter present higher levels of torn feelings (27% versus 7% of those with academic training). The same happens for those with higher levels of confidence with technology, as respondents with less or no confidence were more likely to think of technology as irrelevant or unhelpful for literary translation. For example, 44% of those who are extremely confident find technology helpful and only 6% think there is
- 51. 34 Paola Ruffo 3 4 no relationship between the two. Conversely, most of those who are not confident at all think there is no relationship between literary translation and technology (24%), while 18% find technology helpful, and 14% less or not helpful. Eventually, while literary translators are generally positive about tech- nology, their attitudes become more nuanced when this is put in direct relation to their professional character. This manifests in more ambivalent attitudes, which, in turn, consolidate the emerging opposition between general and translation technology tools. The dichotomy between general and translation technology tools and the relationship between literary translators’ self- image and technological innovation that emerge from the findings will be discussed in the next section. Discussion and conclusion The results have highlighted that literary translators are not averse to technology as such. In particular, a dichotomy between general tech- nology and translation technology has emerged when respondents were allowed to define technology in their own terms. Overall, general tools align with narratives of enhancement and support of the literary translator’s character and work. This type of technology is more fre- quently associated with efficiency, quality, and consistency, and it is not perceived as compromising literary translators’ self- image. Conversely, the description of translation technology takes on a tone that is more deeply related to the essence of literary translation than the practicalities of the work. Furthermore, tools such as CAT, TM, and MT are portrayed as imposed from above, incompatible with the very essence of literary translation, and generally interfering with creativity, originality, and freedom. The (perceived) inflexibility of translation technology is where virtually all negative associations with technology converge. Fear, anger, and uncertainty surround the narrative (perceived or real) that transla- tion technology’s aim is to replace the translator, despite it being incap- able of handling the complexities of literary texts, while threatening to impoverish language. Ultimately, literary translators, rather than refuting technology as a whole, seem to inhabit two spaces at the same time, one where technology proves useful to “craft the best literary texts,” and one where its trajectory is in contrast to their notion of a good translation— possibly a sign of the “sense of confusion” highlighted by Cronin (2013, 1). The above are in line with Koskinen and Ruokonen’s (2017) findings that translators reject technology because of its poor usability and nega- tive effect on efficiency and productivity rather than on principle. The complex relationship between materiality and immateriality in literary translation seems to be further exacerbated by what Pinch and Bijker (1984) would term a controversy between different social groups involved in the technological innovation of the field. In this respect, the focus of recent research on MT and post- editing appears to be at odds
- 52. Collecting literary translators’ narratives 35 3 5 with results from this study, as 71% of respondents do not use transla- tion technology for literary translation and only 8% mention MT. These results also confirm Slessor’s (2019) findings regarding the limited use of translation technology by literary translators and support Taivalkoski- Shilov’s (2018) views on MT and post- editing as not aligned with literary translation. Additionally, when looking at participants’ attitudes towards technology, both research and tool development processes appear removed from literary translators’ realities and are mainly perceived as being preoccupied with reducing costs and enhancing productivity rather than accounting for practitioners’ practical needs or reflecting their demands. One of the aims of employing the SCOT framework proactively was to identify potential closure mechanisms (Pinch and Bijker 1984) arising from social groups’ different interpretations of technology in literary translation. Findings indicate that stabilisation could be found by pro- moting collaboration between all social groups involved, paying par- ticular attention to developing tools that consider literary translators’ specific needs and unique ways of employing existing technology and changing the discourse (or how it is perceived) around technology imple- mentation in the profession. Overall, the relation between self- image and materiality in literary translation is complex and warrants nuance. For example, although the association between positive attitudes and gen- eral technology and negative attitudes and translation technology has emerged, some results point towards aspects of translation technology that are not thought of as antithetical to literary translation. This is the case for TM tools sometimes seen as helpful in dealing with recur- rent translation and consistency issues. Thus, a more productive way of reframing the discourse around materiality in literary translation would be to focus on the concept of enhancement (as also suggested by Youdale 2019). In terms of SCOT, enhancement emerges in this study as the link between literary translators and other social groups involved in the pro- cess of technological innovation. In fact, according to the participants, a sustainable tool is one that supports literary translators and empowers them to improve quality and consistency by allowing them to spend less time on more mechanical tasks, freeing up space for an enhanced cre- ativity. While Youdale’s (2019) approach is remarkable in this sense, it appears too complex at this stage, in that it involves different text- visualisation and text- analysis tools and techniques. In this respect, it is worth recalling that levels of confidence with translation- specific tools are generally low: between 65– 75% of respondents did not undertake any translation technology training and only 6% mentioned corpora. Nevertheless, Youdale’s focus on enhancement and offer of an alter- native to MT- centred workflows show great potential for the develop- ment of new technology- inclusive workflows in the future. This study’s results suggest that an ideal tool for literary translators would feature easy access to online dictionaries and internet searches, a straightforward
- 53. 36 Paola Ruffo 3 6 text- editing interface, translation memory functions, and powerful ter- minology, autosuggest, and quality assurance tools that interfere as little as possible with the translation experience. In order to achieve this, a highly customisable interface would seem ideal. Additionally, since most literary translators work in other areas of translation that often require the use of translation technology, a highly customisable interface could promote the development of a single tool that would be able to adapt to different types of texts and areas of translation. Optimising translation technology in this way would also help tackle issues related to tools’ cost and learning curve. Results also showed how technology training positively affects levels of confidence with technology, in addition to being linked to more posi- tive attitudes. This suggests training as another central aspect of potential closing mechanisms. In particular, the active inclusion of literary transla- tion in technology training and a focus on how existing technology can be adapted to the specificities of literary translation could provide literary translators with practical ways of navigating the new socio- technological landscape, as well as improving their confidence. This is also in line with Slessor (2019), who noted the need to account for literary translators’ specific needs when developing technology and training. Findings also show that, for technological innovation to be sustain- able and respectful of literary translators’ self- image, it is fundamental for the latter to be included in the conversation around technological innovation as well as in the tool development process itself, as suggested by Taivalkoski-Shilov (2018). To achieve this, collaboration between all relevant stakeholders should be promoted, with the aim of producing tools that support and enhance literary translators. By rebalancing the relationship between materiality and immateriality, literary translators could eventually “be liberated from the shackles of ‘faithful’ reproduc- tion, of ‘equivalence’ narrowly defined, and freed up to become rather their inner Picasso” (Large 2018, 94). In this respect, this chapter has identified (1) the inclusion of literary translators in the tool develop- ment process and discourse and (2) the development of translation tech- nology training for literary translators as potentially successful closure mechanisms and something future research should focus on. References Besacier, Laurent, and Lane Schwartz. 2015. “Automated Translation of a Literary Work: A Pilot Study.” In Proceedings of the Fourth Workshop on Computational Linguistics for Literature, 114– 22. Denver, Colorado, USA: Association for Computational Linguistics. Cronin, Michael. 2013. Translation in the Digital Age. London: Routledge. Genzel, Dmitriy, Jakob Uszkoreit, and Franz Och. 2010. “‘Poetic’ Statistical Machine Translation: Rhyme and Meter.” In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, 158–66. Cambridge, MA: Association for Computational Linguistics.
- 54. Collecting literary translators’ narratives 37 3 7 Greene, Erica, Tugba Bodrumlu, and Kevin Knight. 2010. “Automatic Analysis of Rhythmic Poetry with Applications to Generation and Translation.” In Proceedings of the 2010 Conference on Empirical Methods in Natural LanguageProcessing,524–33.Cambridge,MA:AssociationforComputational Linguistics. Heino, Anu. 2020. “Finnish Literary Translators and the Illusio of the Field.” New Horizons in Translation Education and Research 5: 141– 157. Jänis, Marja. 1996. “What Translators of Plays Think About Their Work.” Target 8 (2): 341– 364. Jones, Ruth, and Ann Irvine. 2013. “The (Un)Faithful Machine Translator.” In Proceedings of the 7th Workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities, 96–101. Sofia, Bulgaria: Association for Computational Linguistics. Katan, David. 2017. “Translation Theory and Professional Practice: A Global Survey of the Great Divide.” HERMES –Journal of Language and Communication in Business 22 (42): 111– 153. Kenny, Dorothy, and Marion Winters. 2020. “Machine Translation, Ethics and the Literary Translator’s Voice.” Translation Spaces 9 (1): 123– 149. Koskinen, Kaisa, and Minna Ruokonen. 2017. “Love Letter or Hate Mail? Translators’ Affective Responses to Technology.” In Human Issues in Translation Technology, edited by Dorothy Kenny, 8– 24. London: Routledge. Large, Duncan. 2018. “Could Google Translate Shakespeare?” In Other Words 2019 (52): 79– 95. Littau, Karin. 2016. “Translation and the Materialities of Communication.” Translation Studies 9 (1): 82– 96. Littau, Karin. 2017. “Response by Littau to the Responses to ‘Translation and the Materialities of Communication’.” Translation Studies 10 (1): 97– 101. Luker, Kristin. 2008. Salsa Dancing into the Social Sciences: Research in an Age of Info-Glut. Cambridge, MA: Harvard University Press. Moorkens, Joss, Antonio Toral, Sheila Castilho, and Andy Way. 2018. “Translators’ Perceptions of Literary Post- Editing Using Statistical and Neural Machine Translation.” Translation Spaces 7 (2): 240– 262. Murchú, Eoin P. Ó. 2019. “Using Intergaelic to Pre- Translate and Subsequently Post- edit a Sci- Fi Novel from Scottish Gaelic to Irish.” In Proceedings of the Qualities of Literary Machine Translation Workshop, 20– 25. Dublin, Ireland: European Association for Machine Translation. O’Brien, Sharon. 2012. “Translation as Human– Computer Interaction.” Translation Spaces 1: 101– 22. Olohan, Maeve. 2017. “Technology, Translation and Society: A Constructivist, Critical Theory Approach.” Target 29 (2): 264– 283. Olohan, Maeve. 2019. “Sociological Approaches to Translation Technology.” In The Routledge Handbook of Translation and Technology, edited by Minako O’Hagan, 384– 397. London: Routledge. Pinch, Trevor J., and Wiebe E. Bijker. 1984. “The Social Construction of Facts and Artefacts: Or How the Sociology of Science and the Sociology of Technology Might Benefit Each Other.” Social Studies of Science 14 (3): 399– 441. Ruokonen, Minna, and Jukka Mäkisalo. 2018. “Middling- Status Profession, High- Status Work: Finnish Translators’ Status Perceptions in the Light of their Backgrounds, Working Conditions and Job Satisfaction.” The International Journal of Translation and Interpreting Research 10 (1): 1– 17.
- 55. Another Random Document on Scribd Without Any Related Topics
- 56. “The Dartaway is gone!” exclaimed Ned. “So is the sloop! They’ve stolen our boat! They must suspect something!” At first Jerry and Bob could not believe it. Then Ned went over again all that he had heard, telling how he had slipped away to inform them of how matters stood. “They must have known it was our boat,” said Jerry. “Was Bill Berry one of the men?” “I couldn’t see them plainly,” Ned replied. “Their voices didn’t either one of them sound like Bill’s though. What’s to be done?” “We’d better give notice of our boat being stolen,” said Jerry, “and think of some scheme to get home. After that we’ll have to turn detectives and get the Dartaway.” The minstrel show was over when the boys went back to the pavilion, and the crowd was coming away. The dock was thronged with persons seeking their boats to make the run for home. “Hello boys!” called a voice, and they saw Captain Jenkinson, of the Three Bells coming toward them. “You look as if something had happened.” “Something has,” spoke Jerry. “Our boat’s been stolen.” “Stolen! You don’t mean it. How?” Ned related how it had happened, saying nothing however of the conversation he had overheard. “We’ll tell the manager of the place, and he’ll send out a general alarm,” said the skipper of the Three Bells. “It isn’t the first time boats have been stolen from the docks around here. I believe there’s an organized gang. If we had a decent police force here it wouldn’t happen so often.” The manager of the pavilion and amusement resort, who was also a sort of deputy sheriff, promised the boys to do what he could to recover their craft. He said he would have notices of the theft posted at his dock and all the other wharves along the coast.
- 57. “How you boys going to get home?” asked Captain Jenkinson. “Walk I guess,” replied Jerry with a laugh. “It’s only about five miles by the shore road.” “Well I guess you’ll not walk while there’s gasolene in the Three Bells,” was the hearty answer. “Get aboard. We’ll cruise around a bit, and maybe we’ll get sight of those rascals on the sloop. They probably sailed out to sea, towing your boat, and they’ll be likely to hide it in some cove until the affair blows over.” Captain Jenkinson had a party of friends aboard his boat, but there was plenty of room for the boys. There was much sympathy expressed for them, and every one volunteered to be on the lookout for the Dartaway. In the powerful boat quite an area of the bay in the vicinity of the dock was covered, in a search for a sight of the sloop. It was a fairly light night, and a sailing vessel could be made out some distance away. There were several false alarms, and once the Three Bells gave chase to a sloop that seemed trying to get away. But when they got up to her they found it was only a fishing boat getting a start for early morning work, and there was no sign of the boys’ craft. “Too bad!” remarked Captain Jenkinson, as he landed the boys at their dock. “But it’s sure to turn up sooner or later. They’ll not dare to sell her, and can only keep her hidden. There are not many places where a boat the size of yours can be tucked away.” It was kindly and well meant consolation, but the boys never felt in lower spirits than they did that night. Mrs. Hopkins tried to cheer them up, but it was hard work. She even suggested they hire another boat and make a search of the nearby harbors, for the missing one. “I guess that’s a good plan,” said Jerry to his chums. “We’ll do it to-morrow.”
- 58. “What ought we to do about warning Mr. Hardack?” asked Ned, as they got ready to retire. “Oh, I almost forgot about that,” came from Jerry. “I guess there will be time in the morning. Besides, we want to think it over a little more.”
- 59. CHAPTER XXIV THE SECRET OF THE COVE The boys made inquiries the next morning and found it would be impossible to hire another motor boat. The season was at its height and all the craft were engaged. Several of their newly-made friends offered them the use of their boats for a part of each day, but the boys did not want to take them on this condition. Besides, they wanted to be away by themselves, as they had much to talk about. “Hadn’t we better warn Mr. Hardack?” asked Ned, when they had exhausted the possibilities of hiring a boat. “I’ve been thinking of it,” Jerry said, “and I don’t see what particular good it would do. We only know half the story. Besides, those fellows might only have been talking to hear themselves talk. It’s a mighty risky thing to wreck a vessel and tamper with a lighthouse. They wouldn’t dare do it.” “But it wouldn’t do any harm to tell the keeper what we know,” Ned insisted. “Only to get laughed at for our pains, in case nothing happened. Besides, how are we going to get to the lighthouse without a boat?” “We can walk around by the shore, it’s not more than ten miles,” went on Ned. “Love will find a way,” sang Bob, and he had to jump back to escape a playful blow Ned aimed at him. “I suppose we could,” remarked Jerry. “Maybe it wouldn’t do any harm to go around there and see how the land lays. We can sort of give him a hint, and warn him to be on the lookout for Bill Berry.”
- 60. “That will be a good plan,” agreed Ned. “When can we start? The sooner the better.” “Not until after dinner, I vote,” from Bob. “I can’t walk on an empty stomach.” “I don’t believe you ever tried,” Jerry said. The boys inquired the best way to get to the lighthouse by going along the shore, and learned it was about eight miles. They started after lunch, and hard enough walking they found it, as most of the way lay through deep sand. “This is awful,” groaned Bob as he toiled along. “Dry up, Chunky,” from Ned. “It will do you good. Besides, think of what a service you may be doing.” “I’m thinking about my chances for supper.” The boys had gone about half the distance and were going along a series of high sand dunes, covered with a growth of tall, rank grass, when they were suddenly halted by a voice calling to them: “Where you fellows going?” They looked up, to see a roughly dressed man barring the path. “I don’t know’s it’s any of your business,” Jerry replied hotly, for he resented the man’s tone. “This is a free country.” “That’s all right, my young friend,” the man said with a sneer, “but this is private property, and we don’t allow any trespassing on it. So you’d better be off if you know what’s good for you.” “Can’t we go along the beach?” asked Ned. “No, sir, you can’t go anywhere along here. We own for two miles back, and if you try to pass here you’ll get into trouble. So be off!” The boys hesitated. There was no way of proving that what the man said was not so, and they did not want to get into trouble. To get past meant walking two miles back in shore. As they stood
- 61. there, wondering what to do, another man came from somewhere in the tall grass and said to the first: “The boat’s in the cove. Floated in at high tide.” “That’ll do!” was the quick retort, as though the first man feared his companion would say too much. “I’m warning these chaps off our land.” “Yes, and they’d better go if they know what’s good for them,” said the other. “Oh, we’ll go,” came from Jerry. “We don’t care about walking on your property. I guess we can manage without doing so.” The three lads turned and began to walk inland, across the waste of sand, which was hot with the afternoon sun. The coast at this point was rather high, there being a series of bluffs, which sloped abruptly down to the beach. “What shall we do?” asked Jerry when they were out of earshot of the men. “Nothing to do but to go around their place,” said Ned. “It will mean about four miles more.” Bob groaned. “It’ll get some of that fat off you, Chunky,” Jerry remarked with a laugh. “It’s all right enough for you to joke,” growled the fleshy youth. “But I can’t help it.” “I wonder what he meant by saying the boat was in the cove?” came from Ned. “I was thinking about that myself,” Jerry put in. “I didn’t know there was a cove along here.” “Let’s take a look,” suggested Ned. “How can we, without going back and meeting the men?”
- 62. “Easy enough,” came from Jerry. “We’ll walk along for a mile or so, then cut down along parallel to the coast and walk back toward the beach. We ought to come out right back of the cove, if there is one, or very near to it.” This was voted a good plan to follow, and, with no further objections from Bob, the boys trudged along. It was hot and hard work, but they were very anxious to find out the secret of the cove, as they believed the men had some object in not allowing them to pass. The locality was a deserted one. It was half way between two summer resorts, and there was not a house in sight. It was about as lonesome a place as one could find in the midst of civilization. Nothing was to be seen but sand and rank grass. “Do you s’pose these men had anything to do with stealing our boat and with the lighthouse plot?” asked Bob, as he and his chums walked along. “I’m not good at riddles,” returned Jerry. “First we’ll see what the cove holds.” They kept on for an hour and began to work their way in toward the shore again. They kept a careful watch for the men but saw no one. They came to a place where the weeds and grass were quite high. It was tangled together by the wind and they had to struggle to get through it. Jerry, who was in the lead, emerged on a clear, sandy place. He gave one look down and uttered a low cry. “What is it?” called Ned. “The secret of the cove!” exclaimed Jerry. He pointed to a small body of water below them as they stood on a high sand bank. As the boys looked they saw a sailing vessel and another craft floating near a small dock. “There’s the sloop!” cried Bob.
- 63. “And there’s our boat!” exclaimed Ned. “We have found her again!” “Easy!” whispered Jerry as he sank down, pulling his companions to the earth. “There are the men!” As he spoke three roughly dressed men came from a small shack near the dock, and walked to where the sloop was moored. They were carrying boxes and bales aboard. “Looks like the stuff we picked up from the broken-backed steamer,” whispered Ned. “I guess they’re wreckers, who gather stuff that floats ashore,” came from Jerry. “And I’ll bet they’re the men I heard plotting about the lighthouse,” said Ned. “We are on their track!” “But how did they get their boat in here, and how did they float the Dartaway in?” asked Jerry. “This looks like a little lake.” “Isn’t that a sort of creek over there?” asked Bob, pointing to an opening in the midst of the rushes that surrounded the cove. “So it is. That’s what he meant about high tide. They can only get in the cove when the tide is up, and makes an entrance by way of the creek.” “And, for the same reason, we can’t get our boat out until high tide, and that will not be until late to-night,” said Bob. “We’ll have to wait until then.” “Lucky we have the chance,” came from Ned. “I hope the coast will be clear.” “We’ll get our boat, anyhow!” exclaimed Jerry. “I’d like to see those men keep her.” “We’ll wait until there’s water enough in the creek to float her out, and then we’ll sneak down there, get in and start off before they know what’s happened,” spoke Ned. “I hope she’s in running order.”
- 65. CHAPTER XXV AT THE LIGHTHOUSE The boys remained concealed in the high grass for some time. They watched the men moving about on the sloop and near the hut, but the thieves seemed to pay little attention to the motor boat. “I wonder if they’re getting ready for a trip?” said Bob. “That will make it easier. If they leave we can go down there and get something to eat.” “Oh, Chunky!” exclaimed Ned. “You—” but he could think of nothing appropriate to say, and so stopped short. “They’re all boarding the sloop,” Jerry remarked, as he saw four men come from the shack and go on the sailing vessel. “Going out of the cove maybe.” “Can’t, with the water as low as it is.” “I only hope they go to sleep in the shack,” Ned remarked. “It will be easier for us then.” Through the long afternoon the boys waited. The little camp on the shore of the hidden cove seemed deserted. None of the men was to be seen. Toward evening there arose a thin column of smoke from the galley of the sloop. “They’re getting supper,” remarked Bob, with a sorrowful note in his voice. “Never mind, Chunky, you’ll get yours sooner or later,” said Jerry as comfortingly as he could. As it grew darker the boys noticed that the water in the cove was agitated. The sloop, and the motor boat rocked at their anchorages.
- 66. “The tide’s coming in,” said Jerry. “It will soon be time to act. I hope we can get to the Dartaway without being seen.” “We’ve got to,” spoke Ned. “If they see us it means we’ll have a lot of trouble. We must crawl along until we get close to her. Then we’ll get in. I’ll crank up, you can steer, and Bob can use a boat- hook to fend us out from the shore.” “Lucky she’s headed the right way to get out of the cove,” Jerry remarked. “It will save time by not having to turn her.” Thus it was arranged, and the boys, tired and hungry, remained hidden in the grass until it was dark enough to put their plan in operation. They watched the sloop closely. After their supper aboard, the men came on deck and stood conversing a while. The boys could just make out their forms in the dusk. One seemed to be doing the most talking, and he frequently motioned off toward the sea. “Acts as if he was trying to get them to go somewhere,” spoke Bob softly. But in the end the men went ashore, and after looking to the fastening of the motor boat and a small rowing craft tied near it, they went into the shack. Presently lights shone from it, and Jerry said: “I guess we can sneak down now. Go easy, everybody.” Cautiously the boys left their hiding places and began to descend the slope that led from the bluff to the shore of the cove. Every now and then they paused to listen. They could hear the men laughing and talking in the hut. Foot by foot they crept nearer. There was a path leading from the top of the sand dune to the hut, but the boys did not take this, fearing they would be seen. Instead they crawled on their hands and knees through the grass. The process was a painful and slow one, for their arms and legs came in contact with sand burrs, while innumerable insects attacked them. But they suffered in silence.
- 67. “Easy now, we’re almost there,” came from Jerry. At that moment the door of the hut opened, and a man looked out. The boys, with wildly beating hearts, crouched down. They feared they had been discovered. “See anything?” called some one from inside the hut. “No,” was the answer, “I thought I heard some one at the boats, but I guess it was the tide swinging the sloop. Looks like a storm. Hope we’ll get one by to-morrow night. It’ll be just what we need,” and the man re-entering the hut, closed the door. For a few seconds after this the boys remained silent in the grass. “Lucky escape, that,” murmured Bob. “Five seconds more and he’d caught us.” Cautiously they resumed the progress toward the boat. Nearer and nearer they came until Jerry, who was in the lead, was able to step over the side into it. Ned and Bob followed. The latter grasped a boat-hook and stood ready to fend off when the start was made. Ned and Jerry cut the bow and stern lines with which the Dartaway was made fast to the little dock. They worked quickly and silently. Jerry turned on the gasolene, and waited a few seconds to allow it to fill the carburettor, as the boat had not been run in several hours. Then he switched on the spark. “Turn her over!” he whispered to Ned, who was in the engine cockpit. The big flywheel went around under the impulse of Ned’s sturdy arm. There was a sort of cough from the engine. Then came a chug, followed by a splutter, and the motor got into action. “Fend her off! She’s headed into the bank, and I can’t steer her out quick enough!” cried Jerry to Bob. Chunky pushed with all his strength, on the pole, against the bank. Slowly the nose of the boat came out from the shore. The screw was churning the water into foam. Jerry spun the wheel
- 68. around, and headed the craft for the channel, the opening of which he could just make out. At that instant the door of the hut flew open, and in the light which streamed forth several men could be seen running toward the shore. “Hi there! Stop! Bring that boat back!” they called. “Guess not! She’s ours!” Ned called back. “We’re off!” exclaimed Jerry in a low tone. “She’s running like a charm. They’ll never catch us!” There was the sound of feet on the dock. Then came a squeaking of a pulley block, the creak of ropes and the rattle of the boom on the mast. “What’s the use going after them in the sloop?” they heard some one cry. “There’s no wind. Take the rowboat!” The thud of men jumping into the small craft tied near the sail boat could be heard. There was the rattle of oars, and then the splash of them in the water. “They’ll never get out of the channel,” the boys heard one of the men say. “We’ll catch ’em before they strike open water.” “You will, eh?” thought Jerry. “We’ll see about that.” The engine was speeded up. Jerry was beginning to distinguish things better as his eyes became accustomed to the darkness on the water. The channel was a narrow and winding one, but the incoming tide had made it plenty deep enough. The boys could hear the men frantically rowing after them, but it was a hopeless race. The Dartaway was speeding ahead. It kept Jerry busy steering to avoid running into the bank, but presently the channel widened and he had no more difficulty. On sped the craft until the little creek emerged into a small bay, which, in turn, opened into the ocean.
- 69. “We’re safe now!” cried Jerry. “Let’s light the lamps, and put for home.” The men in pursuit had been left far behind. While Jerry held the boat on her course up the beach Ned and Bob kindled the red and green side lights and the search lantern. In about two hours the Dartaway was safe at her dock, and the boys were telling their story to a number of their friends. “We must notify the police and get after those thieves,” said Captain Jenkinson. “They’re dangerous men to have around. It’s a good thing you discovered that cove. They probably have been hiding there a long time.” But the primitive police system of the shore summer resort could not be gotten in readiness for a raid that night, and when some constables did go to the cove the next morning they found the sloop gone and the hut seemingly deserted. The boys found their boat had suffered little damage at the hands of the thieves. Some tools had been removed as had a few of the cooking utensils, but these were easily replaced. “Now I guess we’d better make a trip to the lighthouse,” remarked Ned, the next afternoon, when the Dartaway had been put in shape. “We ought to warn Mr. Hardack.” “And, incidentally, I suppose, Jessica,” added Bob. “I think they’ll give the whole plan up, now they see we are after them,” Jerry added. “I believe they’ve cleared out for good.” “It’ll do no harm to go over and see Mr. Hardack,” Ned insisted. “If we find out there’s no likelihood of the thing coming off, we needn’t say anything.” They got to the lighthouse about five o’clock. Mr. Hardack greeted them warmly. “Come right in,” he said. “Sorry Jessica is not home. She was just wishing some visitors would come, and about an hour ago that
- 70. Nixon chap came along in his boat and took her for a ride.” Ned seemed less happy than when the start had been made. “But come in,” the lighthouse keeper went on. “I’ve got some fresh milk and Jessica baked some cookies this morning.” Bob was the only one who looked pleased. As the boys were getting out of their boat they saw a man coming down toward where the oil lamps were usually filled. At first they thought it was Bill Berry, but a second look showed them it was not. “Got a new helper?” asked Jerry, trying to speak calmly. “Yes, my other one skipped off yesterday. This chap came along and I hired him. Had to have some one in a hurry.”
- 71. CHAPTER XXVI HELD PRISONERS The boys glanced at each other. This was something they had not counted on. Evidently Bill’s companions had told him what had happened, the night the motor boat was stolen, and he had fled, for some reason. It looked as if the scheme of the plotters had fallen through. “Did Bill—er—did your other helper say where he was going?” asked Ned. “Not a word. He was filling the lamps—let’s see—it was yesterday morning—come to think of it. A boat pulled up at my dock, and a man got out and spoke to Bill. I had to go up in the tower, then. When I came down Bill was gone and so was the man in the boat.” “Rather strange,” commented Jerry. “So it struck me,” Mr. Hardack went on. “But then you know these chaps are sort of tramps. They’re here to-day and gone to-morrow. Always roving around. Of course in the winter I have a regular assistant the government provides, but in the summer time, just as at the life saving stations, they take things a bit easier. However, this other man came along, and he seems a lot nicer than Bill Cherry or whatever his name was.” The keeper led the way up the steps to the house, the boys following. “Guess it’s just as well not to say anything,” spoke Ned in a low voice. “They’ve given up the plot. We’d only be laughed at if we mentioned it.”
- 72. His companions agreed with him, glad enough to feel there was going to be no attempt to wreck a ship by means of false lights. The keeper set out a big pitcher of cool milk and a plate of cookies, which, as Bob said, were the best he ever ate, but then Bob was apt to say that about anything in the culinary line. “Yes,” Mr. Hardack was saying, “Jessica would have been glad to see you. Poor girl, she has quite a trouble on her mind. I’ve been hoping things would straighten out, but they don’t seem to. Her father, he—” “Ting-a-ling-ling-ling!” rang the telephone bell. The keeper sprang to answer it. The boys listened idly to the one-sided conversation. “Yes, this is Mr. Hardack.” “What’s that? Kate sick?” “Come over? Yes—er—that is—Yes, I can come. I forgot I had a new helper. I’ll be right over. Anything serious?” “Can’t tell, eh? Well I’ll come as fast as I can,” and he hung up the receiver. “Any trouble?” inquired Jerry. “Looks like it,” the keeper said. “My sister is quite sick. Taken suddenly. They want me.” “Where does she live?” “It’s about six miles back in the country. I guess I can make it and get back here by nine or ten o’clock. I wish I knew whether it would be safe to leave the new man in charge.” “Don’t the regulations provide for it?” asked Ned. “Oh, yes, it’s my day and night off, and I have a right to go. But I sort of hate to leave the light with him. He knows all about it, however, and he’s got a government civil service certificate. He knows just what to do, for he’s been in lighthouses before. I wish I knew what to do.”
- 73. “Let us stay and help him,” suggested Ned. “Will you?” asked Mr. Hardack eagerly. “Sure,” chorused Jerry and Bob. “Then I’ll do it. I want to see my sister. Her health is not very good, and the doctor said she might die in one of her spells. I’d feel safe to go if I knew you boys would stay here and help the new man if necessary.” “We’ll see to things,” exclaimed Jerry. “It will be jolly fun to be partly in charge of the lighthouse.” “Whatever happens, don’t forget two things,” cautioned Mr. Hardack. “What are they?” “The light must be lit at sunset, and it must be kept burning all night. It must revolve regularly, even if it has to be done by hand, and there must be a white flash and two red ones, at proper intervals. But, you needn’t worry about that. The machinery is in perfect order. The man will light the lamp, and start it going. It only has to be trimmed once in a while. I’ll be back before ten o’clock. When Jessica comes, she’ll get supper for you.” Ned said nothing, but he looked as if that would be the best part of it all, while Chunky’s eyes lighted up at the mention of another meal. Mr. Hardack was soon ready to go. He had to walk the entire distance, as there was no conveyance handy, but he said he did not mind that. “I’ll introduce you to the new man,” he said, calling his helper from where he was still busy filling the lamps. “His name is John Elkwood.” The assistant did not seem a very good natured chap. He only nodded to the boys, when Mr. Hardack introduced them, and, as he went back to his work, Jerry heard him muttering to himself.
- 74. “Well, I guess I’ll get under way,” said the keeper as he started off. “I say,” called Elkwood after him. “What is it?” “I don’t need those boys here. I can get along without ’em. They’ll be in the way.” “I want ’em to stay,” was Mr. Hardack’s answer, at which the boys heard the new man muttering again. “Not very friendly,” commented Jerry. “Still we can get along I guess.” The boys spent an hour going over the lighthouse, with which they were now rather familiar. In the meanwhile Elkwood was busy filling lamps, there being a number used in the big tower. He attended to the light in the big glass lantern and spent some time oiling the machinery. “I wonder what time Jessica is coming back?” said Bob, as they sat down in the sitting room. “Was that one thought for her and two for the supper?” inquired Jerry. “It’s about time she should be back, I think,” came from Ned. “He’s only thinking of her, you see, Chunky,” Jerry went on. “No, but it seems to be getting foggy,” added Ned, “and Noddy isn’t any too good a hand at managing a boat. I wish she hadn’t gone out with him.” “Oh, she’ll be all right,” commented Bob. “Tell you what’s let’s do. We’ll get supper and have it all ready when she comes. I guess we can find the things to eat.” “Trust Chunky for that even if he doesn’t have any dishes on the table,” Jerry remarked. “Well, we’ll get the meal and invite Noddy to it.”
- 75. “Not a bit of it!” exclaimed Ned. “When he sees us here he’ll go back where he came from, fast enough.” The boys found a well-stocked pantry, and, because of their camping experiences had little difficulty in getting a meal ready. By this time it was nearly seven o’clock. Ned kept rather anxious watch of the hours. “Let’s go down to the dock and see if we can get sight of her,” he suggested. “Who?” asked Bob. “Why Jessica. It’s time she was back.” Though he did not say so, Jerry was also a little anxious. The weather looked anything but promising, and he had small respect for Noddy’s ability to manage a motor boat in a calm, let alone a storm. Still there seemed to be no cause for alarm. The craft might have been stalled, but he did not believe Noddy would venture far from shore, and, in the event of a breakdown, he could signal to other boats, as there were several about the harbor. It was still quite light, and would not be dark for another hour. It was no use worrying, Jerry thought, until there was something to get excited over. They all went down to the dock, however, and scanned the sea for a sight of the boat containing the girl and Noddy. Though there were several craft in sight the boys did not notice Noddy’s, which they had come to know from seeing it several times. It was one with a blue hull, distinguishable for some distance. “I vote we eat,” said Bob, as he turned to go back to the house. “It wouldn’t be polite,” suggested Ned. “We’re only visitors.” As they walked up the stone steps leading to the house, the boys were met by Elkwood. The man had a scowl on his face.
- 76. “It’s time you chaps were moving,” he said in surly tones. “I don’t want you hanging around here.” “Why, Mr. Hardack asked us to stay,” put in Jerry. “I don’t care whether he did or not. I’m in charge here. This is government property and I’m the boss. I tell you to go, and don’t lose any time over it, either.” “I guess we’ll stay,” said Jerry coolly. “We told Mr. Hardack we would, and we’re going to.” “And I say you’re not. I order you off. It’s against the regulations for you to be here after dark.” “It isn’t dark yet,” spoke Ned. “None of your lip!” exclaimed Elkwood. “Are you going to leave?” “Not until Mr. Hardack comes back!” “Then you can take the consequences!” Elkwood put his fingers to his lips and blew a shrill whistle. At the signal three men sprang out from behind some rocks that bordered the stairway. They rushed at the boys, who were too surprised to stir. One of the men was Bill Berry. “We’ve got you this time!” their old enemy cried. The next instant the boys were struggling with the men, who endeavored to throw bags over their heads.
- 77. CHAPTER XXVII TRYING TO ESCAPE The struggle was a sharp but short one. The boys were no match for the husky men, and though the lads kicked and punched with all their might, they could not save themselves. In a few minutes they were securely bound, and with the bags tight over their heads, were picked up by the men. “Where you going to put ’em?” they heard Elkwood ask. “The storehouse is a good place,” Bill Berry replied. “They can yell there all night and never be heard. Take ’em to the storeroom!” The boys felt themselves being carried up the steps. Then they could tell, by the muffled footfalls, that they were being taken into some dungeon-like place. “Shall we leave the bags on?” one of the men asked. “No, I don’t want to smother ’em,” Bill replied. “They can’t make themselves heard in here, no matter how they yell. Besides, there’s nobody around. We’ve got Hardack out of the way and he’ll not be back until morning.” “You forget the girl. She may be back any minute.” “I guess not. Noddy has charge of her. He’ll detain her some way or other. Those motor boats have a habit of breaking down, you know.” Then the bags were taken from the boys’ heads, but their bonds were not removed, and they were laid down on the cold stone floor of the storeroom. With sinking hearts they heard the men withdraw and lock the door, leaving them prisoners in total darkness.
- 78. For a few seconds none of the boys spoke. They were so surprised and shocked at the suddenness of it all they did not know what to say. At length Jerry’s voice broke the silence: “Are either of you hurt?” “Only scratched and bruised,” replied Ned. “My wrists are cut by the rope, and my legs hurt,” said Bob. “I’m hun—” “Let up on that!” exclaimed Jerry with a violence he seldom used. “This is no time to think of eating. Boys, it’s a mighty serious matter. These men are going to wreck the ship!” “Do you think so?” inquired Bob. “Of course; what else is it? They have carried their plot into effect, but they did it differently than I expected. Bill Berry’s going away was only a blind, and it fooled us. This new man, of course, is in the game. He came along as soon as Bill left, so no one else would be hired for the place.” “Do you think they got Mr. Hardack away by a false message?” asked Ned. “Of course they did. It was all in the game. Noddy is helping them.” “If I ever get hold of him I’ll make him wish he’d never had a hand in it,” and Ned spoke so sincerely that his companions knew he would keep his word. They thought of Jessica out alone with the bully, who, possibly had purposely disabled the engine to keep her from getting back to the lighthouse. “Oh, if we could only do something,” exclaimed Ned. “We’ve got to!” cried Jerry. “We can’t let the ship be wrecked by them changing the light.” “But how we going to stop ’em?” asked Ned.
- 79. “We must try and get loose,” Jerry replied. “They tied us in such a hurry maybe some of the knots will slip. That’s our only plan. There’s no use calling for help. It’s just as Bill said, no one would hear us. Try and work your hands free.” They all tried but to little purpose. The ropes were firmly tied. Strain as they did they could not loosen the fastenings, and at last they had to stop, as the cords cut into their flesh. “Well, they certainly got us into a trap!” exclaimed Jerry as, once more, he tugged at his bonds. Suddenly Bob uttered an exclamation. “Are you hurt?” cried Ned. “Something cut my wrist!” “What is it?” “A piece of glass, I think.” “Glass! Good!” came from Jerry. “Can you get it in your hands?” “I have it.” “Roll over towards me, and bring it with you.” Bob did so. He came close to where Jerry was still tugging away at the ropes. “How did you find it, Bob?” “I was trying to get the knot loose and something sharp touched my wrist. I felt around until I found the glass.” “What’s your plan, Jerry?” sung out Ned. “I’m going to get Bob to hold the glass and I’m going to saw through the ropes on my hands. Then I’ll set you all free!” “Can you do it?” “I’m going to make a big try.”
- 80. Then in the darkness they began their efforts to escape. Bob stretched out on his face, holding the jagged piece of glass from a broken bottle between his bound hands. By careful feeling Jerry edged his way over to him, until he could bring his wrists close to Bob’s. Then both boys turned on their side, back to back, and Jerry began sawing at the cords that bound him. It was hard work, and more than once they had to stop because their arms ached. Several times Jerry’s hands slipped and the glass cut him, but he did not mind. Back and forth he drew the rope over the keen edge until he could feel the strands giving way. “It’s almost loose,” he said. In another minute he gave a triumphant cry. “I’m free!” “Now to loosen us!” called Ned. Jerry reached into his pocket for his knife. Luckily the men had not searched them, or taken anything away from the boys. With his hands free Jerry soon had the ropes from his legs. Then he cut the bonds of Ned and Bob. Their limbs were stiff, from being tied so long, but vigorous rubbing soon restored the circulation. “Now to escape!” exclaimed Jerry. “We must find a way out of this, and stop the rascals from setting the false lights!” They stumbled about in the darkness. The storehouse was filled with boxes and barrels, over which they fell as they felt around, seeking for some door or window. At last Ned cried out: “Here’s a door!” The other boys made their way toward the sound of his voice. “It’s locked!” said Jerry, as he pushed against the portal. “Can’t we batter it down with a box or a barrel!” Bob said. They searched around in the gloom for something to use, but could find nothing. Everything was too heavy.
- 81. “Maybe we can cut around the lock with our knives,” suggested Ned. In the darkness and silence they toiled. They could hear nothing from the men they knew must be in the lighthouse, working to cause the destruction of the steamer. They felt as if they were imprisoned in a vault. “I wonder if we can get out and be in time?” said Ned. “It must be quite late.” “Don’t talk! Work!” came from Jerry. They redoubled their efforts to cut around the lock. But the door to the storeroom was thick and strong, and the lock was a heavy one. “It’s no use,” declared Bob after an hour’s hacking away at the tough wood. “We’ll have to stay here until they let us out.” “Don’t give up,” Ned spoke. “Hark! What’s that?” asked Jerry. The others listened. “They’ve started the machinery!” cried Bob. “The lenses are turning.” “Yes, and they are the wrong ones! They will get the ship on the rocks!” cried Jerry. “We must escape!” Terror struck to the boys’ hearts. They had tried every means and failed. The plotters had outwitted them. They could do nothing. They beat upon the door with their fists as though by their feeble efforts they could break it down. Ned stumbled aimlessly in the darkness, seeking for something with which to batter down the door. As he passed by a pile of boxes and barrels he uttered a cry. “Have you found anything?” asked Jerry. “Something, yes! A window in the wall! An open window!”
- 82. Bob and Jerry hurried to where they heard Ned’s voice. As they did so he had climbed up on a box. He pressed his face close against the wall. A cool wind fanned his cheek. “There is an opening!” he exclaimed. “But it is too small for us to get out of. It’s only a ventilating window. But wait! Someone is coming!” The boys almost held their breaths. Then Ned called in a loud whisper: “Jess! Jess! Here we are! Let us out! Some bad men are in charge of the place and are going to change the lights! They are going to wreck a steamer!”
- 83. CHAPTER XXVIII JESS TO THE RESCUE “Who are you talking to?” asked Jerry. “Jess, of course,” replied Ned, greatly excited. “She’s outside. Jess! Jess!” he called again. “We are locked in the storeroom!” The boys waited anxiously. Then, from without, came a whisper that sounded loudly through the darkened room. “What has happened? Where is my uncle? Who are you?” “It’s me; Ned,” was the reply, whispered from the prison. “They captured us! Have you a key? Can you let us out? How did you get away from Noddy?” “Oh, this is terrible!” cried Jess. “How did it happen?” She was standing under the small slit in the masonry that served to let air into the storeroom. The light from a lamp in the kitchen of the place streamed out from a window full on her, so Ned could see the girl plainly, though of course she could not see him. “Why you are all wet!” cried Ned. “Did you fall in the water?” “No, I jumped,” came the tense whisper. “What shall I do to let you out?” “Can you get the key to this place?” asked Ned. “If you can, sneak into the house, and open the door, let us out and we’ll call help, and try to prevent the men from changing the light.” “Where is my uncle?” “He was called away, by a false telephone message, we believe, to see his sick sister! The men put up a game to get him away! Quick
- 84. Jess, or it will be too late!” Ned saw the girl step back out of the path of illumination and gaze upward. As she did so she uttered a half suppressed scream. “They are changing the light!” she uttered in a shrill whisper. “And there’s a storm about to break! What shall I do?” As she spoke there came a low rumble of thunder off to the west and a flash of lightning. “Let us out if possible!” whispered Ned. “They are so busy with the light they may not notice you. Get the keys and let us out!” “I will! I will!” exclaimed Jess. “If I can only succeed!” Ned saw her dart around the corner of the house. Then she was out of his line of vision. They could only wait developments now. “Do you think she can do it?” asked Jerry. “She will if it’s possible,” replied Ned. “Only there is not much time. My! But it’s going to storm fierce!” A loud crash of thunder sounded, making the stout lighthouse vibrate. The flashes of lightning showed through the ventilating window, illuminating the small apartment with a weird glow. The wind was howling about the place. “There’ll be a heavy sea on,” said Jerry. “The ship will get upon the rocks and go to pieces. Then these scoundrels will go out and pick up the cargo.” “There may be many lives lost,” exclaimed Bob. “The life saving station is short-handed. They all are in the summer time.” “If Jess can only get the keys!” Ned murmured. It seemed like an hour before there sounded a tapping on the storeroom door. Ned sprang to answer it. “Are you still there, boys?” they heard Jess ask. “Yes! Yes!” whispered Jerry. “Have you the keys?”
- 85. “No, the men must have taken them.” “Then get an axe and see if you can’t break the lock.” “It is too strong. Besides they might hear the blows.” “Where are the men?” “In the lantern tower,” the girl replied. “Wait a minute, let me think of a plan.” Outside the storm was raging. Locked in the storeroom the boys felt like beating at the door with their fists to break it down, so they might get out, change the light, and save the steamer. “I have it!” Jess whispered through the big keyhole. “I will burn the lock out.” “How?” asked Ned. “With a hot poker. I’ll heat it in the kitchen stove. I’ll burn a lot of little holes all around the lock, and then I can knock the piece of door out! The men can’t hear that!” “Good!” cried Ned. “Hurry Jess!” They could hear the girl moving about the kitchen. The rattle of iron on iron came to their ears. Presently there was the smell of burning wood. It grew stronger. Then a dull red point pierced the door, and came through into the storeroom. “That’s the first hole!” whispered Jess. “I’ll burn them as fast as I can.” To the boys it seemed as if there was half an hour between each reappearance of the glowing point of the poker, but it was only a few minutes. There were seven holes burned, when they heard Jess hurry away. Then resounded the tramp of feet in the lower part of the lighthouse. A few seconds later the boys heard voices. “Is it working all right?” a man asked.
- 86. “You bet,” was the reply. “Now you and Bill had better put off in the sloop. She’ll strike pretty soon, and you may pick up passengers with a lot of valuables.” “It’s blowing pretty hard to go out in the sloop,” one of the crowd objected. “Oh, don’t get chicken-hearted,” was the sneering response. “You and Bill have got to go. Me and Jim will stay here and work the light. We can tell when the rockets go up that she’s struck, and then we’ll skip. We’ll meet at the cove.” The voices died away, as though the men had left. The sound of the storm increased. Anxiously the boys waited for Jess to come back. It was several minutes before she did so. Then she whispered through the keyhole: “I had to run and hide when I heard the men coming from the tower. Two of them have gone out, and the others have gone back to the light. We must hurry!” Once more came the smell of burning wood, and once more the dull red point of the poker began to show. But it was slow work, for the door was thick, and of hard material. Then too, the poker would get cool carrying it from the stove to the portal. But Jess worked like an Amazon. Back and forth she went with the hot iron, burning herself several times when it slipped. But she gave small heed to this. She wanted to save the ship and the honor of her uncle, who might be blamed for losing control of the lighthouse. Hole after hole was burned. Now Ned began trying to knock out the piece of door containing the lock. He found a small stone and hammered on the weakened wood. But it was still too strong for the feeble instrument he had. “Ten more holes and I think it will come out,” the girl whispered. Out on the deep, struggling through the storm which had suddenly broken, was a large steamer, laden with a rich cargo. There were not many passengers, as it was from a South American port,
- 87. but these few, as well as the crew, had no warning of the danger that threatened them. In the bow stood the lookout, scanning the expanse of angry water for a sight of lighthouses and headlands that would indicate the channel up the dangerous coast. Suddenly off to his left there shot out two brilliant red flashes. “North light two points off the port bow!” he called to the pilot. “Lookout?” called the pilot. “Aye, aye, sir.” “Are you sure that’s the North light?” “Aye, aye, sir. The south light shows a white flash and two red ones. These were only two red. There they are again, sir.” “Yes, I see them,” as once more the false lights flashed across the sea. “We must have passed the South light while the weather was thicker. I’ll have to put her in a bit.” Then the pilot, deceived by the light, steered the vessel over toward the ledge of dangerous rocks, instead of keeping out, as he would have done, had the two red flashes been preceded by a white one. But in the lighthouse three brave boys and as brave a girl, were striving to aid the ill-fated steamer. Would they be in time? Jess made hole after hole, though her arms ached, her eyes smarted with the smoke, and her hands were burned in a number of places. Again and again Ned beat with his stone on the wood around the lock. The circle of holes was complete at last. “It’s giving away! It’s loosening!” cried the boy. He struck with all his force. The stone flew from his hand, and fell through the opening that suddenly appeared. The lock had been burned away, and the heavy door swung inward. The boys were free.
- 88. “Now to change the lights!” cried Jerry, as, followed by his chums he dashed toward the winding stairs that led to where the big lantern lenses revolved. At that instant the door of the kitchen flew open and Mr. Hardack entered, wild and disheveled, dripping water from the storm which was now raging at its height.
- 89. CHAPTER XXIX THE RIGHT LIGHTS “What has happened!” cried the keeper. “The light is flashing wrong! There is a steamer outside the bar! It will be wrecked! Who did it? Where is my assistant? There’s been foul work here! I was waylaid on my way back when I found my sister was not ill. I just managed to get away from the men. Speak, some of you! Quick!” The keeper was panting from his exertions and from the excitement. His face was drawn and pale, and his eyes were wild, while his hair, matted by the rain, for he had lost his hat, straggled about his forehead. “The scoundrels are in possession of the tower!” cried Jerry. “We must attack them and set the right light!” “Come on!” cried the keeper, seizing the poker Jess had used to burn the door. “Come on! I’ll give ’em battle!” His eyes glared, in the fierceness of his righteous anger, at those who would do so dastardly a deed. “Come on!” cried Ned, seizing a heavy billet of wood. “I’ll call the police on the telephone!” exclaimed Bob, springing for the instrument. “We’ll need help!” “I’ll not wait for the police!” fairly shouted the keeper. “I’ll tackle ’em single handed if need be!” Bob rang up central, and, not waiting to be connected with the distant police station, told the operator what the trouble was, imploring that aid be sent promptly. Then he ran to join his companions. Jess was crying in one corner of the room.
- 90. Mr. Hardack led the way to the stairs which extended up inside the tower to the lantern. He fairly ran up the stone steps, followed by the boys. He was shouting challenges to the men as he ran. “Let me get at you!” he yelled. “I’ll show you how an old man can fight!” Suddenly from above them a door slammed shut. There was the clicking of a lock. Then, as they came to the heavy portal, which gave access to the room where the lantern was, a voice cried: “You’re too late this time, old man!” Too late! The men had shut themselves up in the top of the tower, and could control the working of the light to suit their evil purposes. The keeper could not get in. Mr. Hardack beat upon the door with the poker. Ned hammered it with the block of wood. “Let me in!” cried the aged man. “Let me in! Do you want to send the ship to the bottom?” “That’s just what we do!” was the mocking response. “Get an axe and chop the door down!” cried Jerry. “It would take too long,” replied the keeper, in a strangely calm voice. “It is bound with iron, and is double thick. There is no help for it. The steamer will be lost!” Footsteps were heard coming up the stairs. “Maybe help is at hand,” said the keeper hopefully. Then Jess came into view. In her hand she held something which she extended to Mr. Hardack. “Here is your old horse pistol, uncle!” she exclaimed. “It is loaded with a heavy charge. Fire it through the lock and shatter it! I heard you pounding on the door and knew they had locked it!” “Hurrah for you, Jess!” called Ned, and the girl blushed through her tears.
- 91. Welcome to our website – the perfect destination for book lovers and knowledge seekers. We believe that every book holds a new world, offering opportunities for learning, discovery, and personal growth. That’s why we are dedicated to bringing you a diverse collection of books, ranging from classic literature and specialized publications to self-development guides and children's books. More than just a book-buying platform, we strive to be a bridge connecting you with timeless cultural and intellectual values. With an elegant, user-friendly interface and a smart search system, you can quickly find the books that best suit your interests. Additionally, our special promotions and home delivery services help you save time and fully enjoy the joy of reading. Join us on a journey of knowledge exploration, passion nurturing, and personal growth every day! ebookbell.com