The Neurocognition of Translation and Interpreting 9789027203397, 9027203393

Citation preview

BENJAMINS ■

T R A N S L AT I O N

The Neurocognition of Translation and Interpreting Adolfo M. García

■

LIBR ARY

The Neurocognition of Translation and Interpreting

Benjamins Translation Library (BTL) issn 0929-7316

The Benjamins Translation Library (BTL) aims to stimulate research and training in Translation & Interpreting Studies – taken very broadly to encompass the many different forms and manifestations of translational phenomena, among them cultural translation, localization, adaptation, literary translation, specialized translation, audiovisual translation, audio-description, transcreation, transediting, conference interpreting, and interpreting in community settings in the spoken and signed modalities. For an overview of all books published in this series, please see www.benjamins.com/catalog/btl

General Editor

Honorary Editors

Roberto A. Valdeón

Yves Gambier

University of Oviedo

Associate Editor Franz Pöchhacker University of Vienna

University of Turku & Immanuel Kant Baltic Federal University

Gideon Toury† Tel Aviv University

Advisory Board Cecilia Alvstad

Kobus Marais

Georges L. Bastin

Christopher D. Mellinger

Dirk Delabastita

Jan Pedersen

Daniel Gile

Nike K. Pokorn

Arnt Lykke Jakobsen

Luc van Doorslaer

Krisztina Károly

Meifang Zhang

Stockholm University University of Montreal University of Namur Université Paris 3 - Sorbonne Nouvelle Copenhagen Business School Eötvös Lorand University

University of the Free State University of North Carolina at Charlotte Stockholm University University of Ljubljana University of Tartu & KU Leuven University of Macau

Volume 147 The Neurocognition of Translation and Interpreting by Adolfo M. García

The Neurocognition of Translation and Interpreting Adolfo M. García Laboratory of Experimental Psychology and Neuroscience (LPEN), INCyT, INECO Foundation, Favaloro University / National Scientific and Technical Research Council (CONICET) / Faculty of Education, UNCuyo

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

doi 10.1075/btl.147 Cataloging-in-Publication Data available from Library of Congress: lccn 2019010103 (print) / 2019018852 (e-book) isbn 978 90 272 0339 7 (Hb) isbn 978 90 272 6235 6 (e-book)

© 2019 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Company · https://benjamins.com

Table of contents

Acknowledgments

ix

Foreword

xi

Notes on previous works

xv

Introduction Translation, interpreting, and the brain behind it all I. Stepping into the attic  1 II. Why should TIS be concerned with neurocognition?  3 III. Aims and target audience  5 IV. The contents, at a glance  5 V. Conceptual delimitations  7 Chapter 1 Mind and brain in the study of translation and interpreting 1.1 Minding the brain, braining the mind  11 1.2 Outside the head: Non-neural cognitive approaches  12 1.2.1 Rationalizing translation: Insights from analytical linguistics  13 1.2.2 See but don’t touch: The observational trend  15 1.2.3 Take a look at yourself: Introducing TAPs  17 1.2.4 From product to process: Corpus-based studies  18 1.2.5 Quantifying performance  18 1.3 Within the mind, without the brain: Appraising non-neural cognitive approaches  25 1.4 Not black, not a box: Enter the brain  30 1.5 Historicizing brain-based research on IR  34 1.5.1 Milestones from the mid-twentieth century  36 1.5.2 Milestones from the late twentieth century  37 1.5.3 Milestones from the twenty-first century  38 1.6 A role for neuroscience in contemporary TIS  39

1

11

vi

The Neurocognition of Translation and Interpreting

Chapter 2 The toolkit 2.1 Beyond MacGyver’s knife  41 2.2 A matter of design  42 2.2.1 Single-case designs  42 2.2.2 Single-group designs  43 2.2.3 Between-group designs  44 2.2.4 Pre/post designs  45 2.3 Mind games: A sampler of experimental paradigms  46 2.3.1 Keeping it real  46 2.3.2 Piece by piece  47 2.4 The craft of manipulation  55 2.5 Do it well, do it fast  58 2.6 System breakdown  59 2.7 The brain, in vivo  60 2.7.1 Non-invasive techniques  61 2.7.2 Invasive techniques  69 2.8 How (not) to interpret the data  71 2.9 Final remarks  74 Chapter 3 Prolegomena to the translating and interpreting brain 3.1 Laying the groundwork  75 3.2 A primer on neurology  76 3.2.1 The neocortex  77 3.2.2 Some language-related subcortical structures  80 3.2.3 Two key language-related networks  81 3.2.4 Neurons and synapses  82 3.2.5 Cognitive processing as neuronal teamwork  85 3.3 The verbal brain  87 3.3.1 Tell me where: The functional neuroanatomy of language  87 3.3.2 Electrified words: The neurophysiology of language  92 3.4 It takes two to tango: The prerequisite of bilingualism  94 3.4.1 Linguistic mechanisms in the bilingual brain  94 3.4.2 Executive mechanisms in the bilingual brain  95 3.5 In a nutshell  97

41

75



Table of contents vii

Chapter 4 Building up from breakdown 4.1 Lessons from lesions  99 4.2 Disruptions of IR  101 4.2.1 Compulsive translation  101 4.2.2 Inability to translate  104 4.2.3 Paradoxical translation behavior  112 4.2.4 Translation without comprehension  114 4.3 Charting the territory  115 4.3.1 Taking sides  116 4.3.2 A thing unto itself  117 4.3.3 Coming and going  119 4.3.4 Of words and concepts  120 4.3.5 The unit determines the network  122 4.4 Piecing it all together  123 4.4.1 A neuroarchitectural model of translation routes  123 4.4.2 A neural model of the systems subserving simultaneous interpreting  126 4.5 Interpretive remarks  128 4.6 From static maps to dynamic pictures  129 Chapter 5 The dynamics of directionality 5.1 A sense of direction  131 5.2 Multidimensional signatures of directionality  134 5.2.1 Functional neuroimaging evidence  134 5.2.2 Electrophysiological evidence  140 5.2.3 Psycholinguistic evidence  146 5.3 Back and forth  149 5.4 In the right direction  152

99

131

Chapter 6 Process is as unit requires 153 6.1 The process’s raw material  153 6.2 Conceiving translation units  154 6.3 Spatiotemporal correlates of lexical and sentential translation units  157 6.3.1 Functional neuroimaging evidence  157 6.3.2 Electrophysiological evidence  163 6.3.3 Psycholinguistic evidence  168 6.4 Uniting it all  172 6.5 From unitary to unit-sensitive  176

viii The Neurocognition of Translation and Interpreting

Chapter 7 The interpreter’s brain 7.1 The art of self-sculpting  177 7.2 Simultaneous interpreting, or extreme bilingual processing  179 7.3 En route to expertise  181 7.3.1 So different, so fast  186 7.4 Keep the change (and make it broader)  189 7.4.1 Brains interpreting interpreting brains  200 7.5 The plastic nature of IR systems  204 Chapter 8 A story in the making 8.1 The tale of the attic  205 8.2 Q&A 206 8.3 The good…  209 8.4 … and the bad  211 8.5 Needs assessment  214 8.5.1 More, better science  214 8.5.2 An institutional architecture  217 8.6 Parting words  218

177

205

About the author

221

List of figures and tables 

223

List of acronyms and abbreviations

227

References

229

Index

263

Acknowledgments

Since the start of my career, I have been fortunate to work for outstanding institutions, with generous publishers, and alongside leading experts from various fields. This book stems from such synergies. It is uplifting to look back and recognize the contributions of those who made it possible. At an institutional level, I appreciate the enduring support of the National Scientific and Technical Research Council (CONICET), the Institute of Cognitive and Translational Neuroscience (INCYT), the Institute of Cognitive Neurology (INECO), and the Faculty of Education at the National University of Cuyo (UNCuyo). It is my honor to have conducted most of my research with these excellent affiliations in Argentina, my home country. More particularly, I would like to thank the Benjamins Translation Library, in the figure of its General Editor, Professor Roberto Valdeón, for trusting in this project and prompting its execution. I also acknowledge the several publishers who kindly granted me permission to reproduce previous materials. In addition to its flaws, I am told, the present book has a number of merits. While the former rest entirely with me, the latter build on lessons from many friends and colleagues over the years. The following lines are a meager sample of my gratitude to them. First and foremost, I am deeply thankful for the scientific ride I share with Agustín Ibáñez, Lucas Sedeño, and all the members of the Laboratory of Experimental Psychology and Neuroscience (LPEN-INCYT). Not a day has passed in which I have not learned something new by working with them. My thinking has also been greatly nurtured by the ideas and support of professors Michael Halliday, Ruqaiya Hasan, Sydney Lamb, and Michel Paradis. The exciting exchanges we have often held at our homes, in distant hotels, and through the Internet have left an indelible mark, both professionally and affectively. I also treasure the cooperation received in multiple projects from countless excellent collaborators at institutions in Argentina (Universidad Nacional de Mar del Plata, Universidad Nacional de Córdoba), Australia (University of Sydney, Macquarie University), Brazil (Universidad Federal de Minas Gerais), Chile (Universidad de Santiago de Chile, Universidad Adolfo Ibáñez), China (University of Macau, University of Electronic Science and Technology of China), Colombia (Universidad de Antioquia, Universidad del Valle, Universidad Javeriana, Universidad de Los

x

The Neurocognition of Translation and Interpreting

Andes), Germany (Johannes Gutenberg University), Spain (Universidad de La Laguna, Universidad de Oviedo), the United Kingdom (Durham University), and the United States (New York University, Florida Hospital). My appreciation also goes out to my work groups at the Inter-American Development Bank and the Translation, Research, Empiricism, and Translation (TREC) network. It is a privilege to constantly feed off this vast research web. Finally, this book, much like any other project I have undertaken, would hardly have been completed without the company, love, and patience of Raúl, Elba, Pancho, and María. To these unsung heroes behind everything that bears my name, I say thank you, a thousand times, and then a thousand times over.  

Adolfo M. García Buenos Aires, December 20, 2018

Foreword

Over millennia, translation and interpreting have been studied from a myriad of perspectives. Their scientific inquiry, however, is a much more recent endeavor. Locating that inquiry in the human brain from an integrated, multidimensional perspective is still uncharted territory and it has recently been called by Tymoczko (2012) “a known unknown” in the field of translation and interpreting studies. Therefore, it is not surprising that a book entitled The Neurocognition of Translation and Interpreting might raise some skeptical eyebrows. So far, the most prominent approaches in cognitive translation and interpreting studies should be considered as non-neural. In other words, these approaches seldom draw on brain-informed data and rely predominantly on linguistic and behavioral results. The majority of these approaches have built extensively on an empirical orientation heavily influenced by the information-processing paradigm. Recently, however, a new trend has emerged in cognitive translation and interpreting studies, advocating in favor of 4EA cognition, namely, a view which considers human cognition, and indirectly the act of translating and interpreting, to be embedded, extended, embodied, enacted, and affective (Muñoz Martín 2017; Risku 2017). When confronted with the present volume, it is then only natural that readers versed in mainstream approaches within 4EA cognition would ask: why should translation and interpreting studies be concerned with neurocognition at all? To that remark one could add an even stronger question: why is it important to locate translation in the brain when cognitive translation and interpreting studies seem to be moving away from a strict experimental paradigm towards a view of cognition which is situated and relies on contextual factors surrounding cognitive aspects related to the act of translating and interpreting? It is a quite bold step for an author to take a preeminently subject-centered and biologically grounded outlook within this current trend and state loud and clear that, indeed, looking at translation in the brain offers a cornerstone for an embedded, situated view of cognitive translation and interpreting studies; a view which is necessary in terms of complementarity and relevant due to its elucidating potential. As Adolfo García shows in this book, rationalist, observational, introspective, corpus-based, and quantitative approaches have offered an enormous contribution to this still fledgling field. From an ontological and a methodological viewpoint, these approaches are all “non-neural.” Yet, as García reminds us, “their

xii The Neurocognition of Translation and Interpreting

underlying research procedures do not necessarily entail a rejection of the neurological basis of the observed phenomena.” Nevertheless, none of these approaches is capable of elucidating the intricacies related to what happens in the human brain when translation and interpreting tasks are carried out, thus failing to delve into a critical component of 4EA cognition. This is perhaps why there has been some recent interest in looking beyond the black box and what García terms the “known unknown” fallacy, with scholars in cognitive translation and interpreting studies taking an active interest in brain-based research. The prospects of embracing a neurocognitive approach to study translation and interpreting are now becoming clearer and this is the ambitious and wellaccomplished goal of The Neurocognition of Translation and Interpreting. By revisiting paradigms that have informed cognitive translation and interpreting studies so far, García paves the way for discussing the toolkit used in the field in terms of experimental designs, the role of verbal and non-verbal paradigms, the use of non-invasive techniques, and ways to interpret data. The acknowledgement of what has been achieved by research drawing on verbal report protocols as well as key-logged and eye-tracking data in behavioral studies also throws light on what still remains inside the black box as a “known unknown.” Without striving further in an attempt to answer these questions, “a massive endeavor could be rendered irreparably flawed if basic elements are not considered from the outset.” This quote from García’s book calls for the need to complement information acquired through behavioral tasks with findings which can only be obtained by looking at the physiology of the human brain to make a case for approaching cognitive processes related to translation and interpreting as “neuronal teamwork.” The Neurocognition of Translation and Interpreting does it by building up from breakdown, i.e., by learning from brain-related lesions and how they can inform brain-related processes concerning translation and interpreting. The book also looks at the dynamics of directionality and at the intricacies of cognitive segmentation in translation and interpreting tasks. Additionally, it illuminates the ongoing debate led by researchers interested in 4EA cognition by complementing and even extending the discussion about how a brain-related approach can inform a situated view of translation as a cognitive activity. In short, it’s all about connections, units, time, and distance that make up what Schilperoord (1996) calls “cognitive rhythm” in behavioral terms. Without a shadow of doubt, looking at cognitive rhythms in the translator’s and interpreter’s brain definitely contributes to a better understanding of a story in the making. The Neurocognition of Translation and Interpreting will easily convince interested readers that it is worth looking at translation and interpreting as activities embedded in the brain. The book does show that it is possible to tell the story from a different perspective and, thus, boost our understanding of brain-related

Foreword xiii

phenomena as far as translation and interpreting are concerned. Instead of dwelling on “known unknowns,” García is able to convince readers that the neurocognition of translation and interpreting constitutes a set of “unknown knowns.” By doing this, it throws light on a story in the making; a story about which it is worth writing and reading.  

Fabio Alves Belo Horizonte, December 23, 2018

References Muñoz Martín, Ricardo. 2017. “Looking toward the future of cognitive translation studies.” In The Handbook of Translation and Cognition, ed. by John W. Schwieter and Aline Ferreira, 555–572. Hoboken, New Jersey: John Wiley & Sons, Inc. Risku, Hanna. 2017. “Ethnographies of translation and situated cognition.” In The Handbook of Translation and Cognition, ed. by John W. Schwieter and Aline Ferreira, 290–310. Hoboken, New Jersey: John Wiley & Sons, Inc. Schilperoord, Joost. 1996. It’s about Time. Temporal Aspects of Cognitive Processes in Text Production. Utrecht: Rodopi. Tymoczko, Maria. 2012. “The neuroscience of translation.” Target 24 (1): 83–102. https://doi.org/10.1075/target.24.1.06tym

Notes on previous works

Some of the contents in this book are elaborations or reproductions of previous material, as indicated below. Section 2.7, in Chapter 2, reproduces and extends parts of: Adolfo M. García, Ezequiel Mikulan & Agustín Ibáñez (2016). A neuroscientific toolkit for translation studies. In Ricardo Muñoz Martín (ed.), Re-embedding Translation Process Research, pages 21–46. Amsterdam: John Benjamins (with authorization from John Benjamins Publishing). Online: https://benjamins.com/catalog/btl.128 Section 3.2, in Chapter 3, is a modified and extended reproduction of pages 58 through 68 from: Adolfo M. García, William J. Sullivan, and Sarah Tsiang (2017). An Introduction to Relational Network Theory: History, Principles, and Descriptive Applications. London: Equinox. © Equinox Publishing Ltd 2017. Sections 4.2 and 4.3 (including Tables 4.1 through 4.4), in Chapter 4, are a modified and extended reproduction of: Adolfo M. García (2015). Translating with an injured brain: Neurolinguistic aspects of translation as revealed by bilinguals with cerebral lesions. Meta: Translators’ Journal 60(1), 112–134 (with authorization from Les Presses de l’Université de Montréal). Figure 1.1 is an authorized reproduction of Figure 1 from Grounding translation and interpreting in the brain: What has been, can be, and must be done, by Edinson Muñoz, Noelia Calvo & Adolfo M. García, Perspectives: Studies in Trans­lation Theory and Practice, 2018, https://doi.org/10.1080/0907676X.2018.1549575, reprinted by permission of the publisher (Taylor & Francis Ltd, http://www.tandfonline.com). Figure 2.1 is reprinted from Brain Research Bulletin, 72(1), by Roland H. Grabner, Clemens Brunner, Robert Leeb, Christa Neuper, and Gert Pfurtscheller, Event-­related EEG theta and alpha band oscillatory responses during language translation, pages 57–65, Copyright 2007, with permission from Elsevier. Figure 2.3 is reprinted from NeuroImage, 134, by Maxi Becker, Torsten Schubert, Tilo Strobach, Jürgen Gallinat, and Simone Kühn, Simultaneous interpreters vs. professional multilingual controls: Group differences in cognitive control as well as brain structure and function, pages 250–260, Copyright 2016, with permission from Elsevier. Figure 2.4 is reprinted from PLoS One, 7(4), by Jihoon Oh, Mookyung Han, Bradley S. Peterson, and Jaeseung Jeong, Spontaneous eyeblinks are correlated

xvi The Neurocognition of Translation and Interpreting

with responses during the Stroop task, article e34871 (open access), https://doi.org/ 10.1371/journal.pone.0034871. Authorized reproduction under the terms of the

Creative Commons Attribution License. Figure 2.5 is reprinted from Frontiers in Human Neuroscience, 8, by Wonil Choi, Rutvik H. Desai, and John M. Henderson, The neural substrates of natural reading: A comparison of normal and nonword text using eyetracking and fMRI, article 1024 (open access), Copyright 2014, https://doi.org/10.3389/fnhum.2014.01024. Authorized reproduction under the terms of the Creative Commons Attribution License. Figure 2.6 is reprinted with permission from D. Klein, B. Milner, R. J. Zatorre, E. Meyer, and A. C. Evans, The neural substrates underlying word generation: A bilingual functional-imaging study, Proceedings of the National Academy of Sciences, 92(7), 2899–2903, Copyright 1995, National Academy of Sciences, U.S.A. Figure 2.8 is reprinted from Brain Research, 1158, by Douglas J. Davidson and Peter Indefrey, An inverse relation between event-related and time-frequency violation responses in sentence processing, pages 81–92, Copyright 2007, with permission from Elsevier. Figure 2.9 is reprinted from International Journal of Psychophysiology, 57(2), by Sabine Weiss, Horst M. Mueller, Baerbel Schack, Jonathan W. King, Martha Kutas, and Peter Rappelsberger, Increased neuronal communication accompanying sentence comprehension, pages 129–141, Copyright 2005, with permission from Elsevier. Figures 3.1, 3.2, 3.3, 3.4, 3.6, and 3.7 are free media from Wikimedia Commons. All these images were labeled for commercial reuse with modifications. Due credits for these figures are as follows: – Figure 3.1. Patrick J. Lynch, medical illustrator; C. Carl Jaffe, MD, cardiologist (licensed under Creative Commons Attribution 2.5, License 2006). – Figure 3.2. Anonymous (Public Domain). – Figure 3.3. Henry Vandyke Carter and Henry Gray (1918). Anatomy of the Human Body (Public Domain). – Figure 3.4. John Henkel, from the Food and Drug Administration (Public Domain). – Figure 3.6. Blausen.com staff. “Blausen gallery 2014”. Wikiversity Journal of Medicine. https://doi.org/10.15347/wjm/2014.010. ISSN 20018762. – Figure  3.7. Thomas Splettstoesser (licensed under Creative Commons Attribution-Share Alike 4.0 International license). Figure 3.5, panel A, is reprinted from Neuroscience and Biobehavioral Reviews, 80, by Agustina Birba, Indira García-Cordero, Giselle Kozono, Agustina Legaz, Agustín Ibáñez, Lucas Sedeño, and Adolfo M. García, Losing ground: Frontostriatal



Notes on previous works xvii

atrophy disrupts language embodiment in Parkinson’s and Huntington’s disease, pages 673–687, Copyright 2017, with permission from Elsevier. Figure 3.5, panel B, was co-designed with Agustina Birba, who drew and generated the final version of each image in it. Figure 3.8 is reprinted from Cerebral Cortex, 25(12), by Michel Thiebaut de Schotten, Flavio Dell’Acqua, Peter Ratiu, Anoushka Leslie, Henrietta Howells, Emanuel Cabanis, Marie-Therese Iba-Zizen, Odile Plaisant, Andrew Simmons, Nina F. Dronkers, Suzanne Corkin, and Marco Catani, From Phineas Gage and Monsieur Leborgne to H.M.: Revisiting disconnection syndromes, pages 4812– 4827, https://doi.org/10.1093/cercor/bhv173 PMCID: PMC4635921. Authorized reproduction under the terms of the Creative Commons CC BY License. Figure 3.9 is reprinted from Developmental Cognitive Neuroscience, 1(3), by S. Christopher Nuñez, Mirella Dapretto, Tami Katzir, Ariel Starr, Jennifer Bramen, Eric Kan, Susan Bookheimer, and Elizabeth R. Sowell, fMRI of syntactic processing in typically developing children: Structural correlates in the inferior frontal gyrus, pages 313–323, Copyright 2011, with permission from Elsevier. Figure 3.10 is reprinted from Cortex, 49(3), by Elizabeth Jefferies, The neural basis of semantic cognition: Converging evidence from neuropsychology, neuroimaging and TMS, pages 611–625, Copyright 2013, with permission from Elsevier. Figure 3.11 is reprinted from Journal of Neurolinguistics, 20(3), by Jubin Abutalebi and David Green, Bilingual language production: The neurocognition of language representation and control, pages 242–275, Copyright 2007, with permission from Elsevier. Figure 4.1 is adapted from Traductología y neurocognición: Cómo se organiza el sistema lingüístico del traductor, page 252, Copyright 2012, by Adolfo M. García. Córdoba: Facultad de Lenguas de la Universidad Nacional de Córdoba. URL: http://hdl.handle.net/11086/2715. Authorized reproduction under the terms of the Creative Commons NCND 2.5 License (http://creativecommons.org/licenses/ by-nc-nd/2.5/ar/). Figure  4.2 is reproduced from The Neurolinguistics of Bilingualism: An Introduction, page 205, by Franco Fabbro, Copyright 1999. Hove: Psychology Press. Reprinted with permission from Taylor & Francis. Figure 5.2 is reprinted with permission from Denise Klein, Brenda Milner, Robert J. Zatorre, Ernst Meyer, and Alan C. Evans, The neural substrates underlying word generation: a bilingual functional-imaging study, Proceedings of the National Academy of Sciences, 92(7), 2899–2903, Copyright 1995, National Academy of Sciences, U.S.A. Figure 5.3 is reproduced from “Images of shadowing and interpreting,” by Jorma Tommola, Matti Laine, Marianna Sunnari, and Juha O. Rinne. In Interpreting 5(2),

xviii The Neurocognition of Translation and Interpreting

147–169 / 2000. Reprinted with kind permission from John Benjamins Publishing Company, Amsterdam/Philadelphia. [www.benjamins.com] Figure 5.4 is reprinted from Brain Research Bulletin, 59(3), by Valentina Quaresima, Marco Ferrari, Marco C. P. van der Sluijs, Jan Menssen, and Willy N. J. M. Colier, Lateral frontal cortex oxygenation changes during translation and language switching revealed by non-invasive near-infrared multi-point measurements, pages 235–243, Copyright 2002, with permission from Elsevier. Figures 5.5, 5.6, and 5.8 are reprinted from Adolfo M. García, Ezequiel Mikulan & Agustín Ibáñez (2016). A neuroscientific toolkit for translation studies. In Ricardo Muñoz Martín (ed.), Re-embedding Translation Process Research, pages 21–46. Amsterdam: John Benjamins (with authorization from John Benjamins Publishing). Online: https://benjamins.com/catalog/btl.128 Figure 5.7 is reprinted from Language conflict in translation: An ERP study of translation production, by Ingrid K. Christoffels, Lesya Ganushchak, and Dirk Koester, Journal of Cognitive Psychology, 25(5), 646–664, Copyright 2013, reprinted by permission of Taylor & Francis Ltd (http://www.tandfonline.com). Figure 6.2 is reprinted from Current Biology, 20, by Gianpiero Liuzzi, Nils Freundlieb, Volker Ridder, Julia Hoppe, Kirstin Heise, Maximo Zimerman, Christian Dobel, Stefanie Enriquez-Geppert, Christian Gerloff, Pienie Zwitserlood, and Friedhelm C. Hummel. The involvement of the left motor cortex in learning of a novel action word lexicon, pages 1745–1751, Copyright 2010, with permission from Elsevier. Figure 6.3 is reprinted from Journal of Neurolinguistics, 37, by Cornelia D. Moldovan, Josep Demestre, Pilar Ferré, Rosa Sánchez-Casas. The role of meaning and form similarity in translation recognition in highly proficient balanced bilinguals: A behavioral and ERP study, pages 1–11, Copyright 2016, with permission from Elsevier. Figure 6.4 is reprinted with permission from Guillaume Thierry and Yan J. Wu, Brain potentials reveal unconscious translation during foreign-language comprehension, Proceedings of the National Academy of Sciences, 104(30), 12530–12535, Copyright 2007, National Academy of Sciences, U.S.A. Figure 6.5 is reprinted from Brain Research Bulletin, 72(1), by Roland H. Grabner, Clemens Brunner, Robert Leeb, Christa Neuper, and Gert Pfurtscheller, Event-related EEG theta and alpha band oscillatory responses during language translation, pages 57–65, Copyright 2007, with permission from Elsevier. Figure 6.7 is reprinted from Journal of Experimental Psychology: Human Perception and Performance, 30(5), by Wouter Duyck and Marc Brysbaert, Forward and backward number translation requires conceptual mediation in both balanced and unbalanced bilinguals, pages 889–906, Copyright 2004, with permission from the American Psychological Association.



Notes on previous works xix

Figure 6.8 is reprinted from Frontiers in Psychology: Cognitive Science, 5, by Adolfo M. García, Agustín Ibáñez, David Huepe, Alexander L. Houck, Maëva Michon, Carlos G. Lezama, Sumeer Chadha, and Álvaro Rivera-Rei, Word reading and translation in bilinguals: The impact of formal and informal translation expertise, article 1302 (open access), Copyright 2014, https://doi.org/10.3389/fpsyg.2014.01302. Authorized reproduction under the terms of the Creative Commons Attribution License. Figure 7.1 is reprinted from Frontiers in Psychology, 2, by Alexis Georges HervaisAdelman, Barbara Moser-Mercer, and Narly Golestani, Executive control of language in the bilingual brain: integrating the evidence from neuro­imaging to neuropsychology, article 234 (open access), Copyright 2011, https://doi.org/ 10.3389/fpsyg.2011.00234. Authorized reproduction under the terms of the Creative Commons Attribution License. Figure 7.2 is reprinted from Neuropsychologia, 98, by Alexis Hervais-Adelman, Barbara Moser-Mercer, Micah M. Murray, and Narly Golestani, Cortical thickness increases after simultaneous interpretation training, pages 212–219, Copyright 2017, with permission from Elsevier. Figure 7.3 is reprinted from NeuroImage, 114, by Alexis Hervais-Adelman, Barbara Moser-Mercer, and Narly Golestani, Brain functional plasticity associated with the emergence of expertise in extreme language control, pages 264–274, Copyright 2015, with permission from Elsevier. Figure 7.4 is reprinted from Cortex, 54, by Stefan Elmer, Jürgen Hänggi, and Lutz Jäncke, Processing demands upon cognitive, linguistic, and articulatory functions promote gray matter plasticity in the adult multilingual brain: Insights from simultaneous interpreters, pages 179–189, Copyright 2014, with permission from Elsevier. Figure 7.5 is reprinted from Brain Research, 1317, by Stefan Elmer, Martin Meyer, and Lutz Jäncke, Simultaneous interpreters as a model for neuronal adaptation in the domain of language processing, pages 147–156, Copyright 2010, with permission from Elsevier. Figures 7.6 and 7.7 are reprinted from Micaela Santilli, Martina G. Vilas, Ezequiel Mikulan, Miguel Martorell Caro, Edinson Muñoz, Lucas Sedeño, Agustín Ibáñez, and Adolfo M. García, Bilingual memory, to the extreme: Lexical processing in simultaneous interpreters. Bilingualism: Language and Cognition, 2018, https://doi.org/10.1017/S1366728918000378. Reproduced with permission. Figure 7.8 is reprinted from Journal of Memory and Language, 54, by Ingrid K. Christoffels, Annette M. B. de Groot, Judith F. Kroll, Memory and language skills in simultaneous interpreters: The role of expertise and language proficiency, pages 324–345, Copyright 2006, with permission from Elsevier.

xx The Neurocognition of Translation and Interpreting

Figure 7.9 is reprinted from Acta Psychologica, 155, by Julia Morales, Francisca Padilla, Carlos J. Gómez-Ariza, and María Teresa Bajo, Simultaneous interpretation selectively influences working memory and attentional networks, pages 82–91, Copyright 2015, with permission from Elsevier. Figure 7.10 is reprinted from NeuroImage, 134, by Maxi Becker, Torsten Schubert, Tilo Strobach, Jürgen Gallinat, and Simone Kühn, Simultaneous interpreters vs. professional multilingual controls: Group differences in cognitive control as well as brain structure and function, pages 250–260, Copyright 2016, with permission from Elsevier.

Introduction

Translation, interpreting, and the brain behind it all

I. Stepping into the attic The need to broker between languages has been a constant of humanity (Delisle and Woodsworth 1995). Ever since two speech communities first came into contact, bilingual individuals have engaged in countless forms of interlingual reformulation (IR) – namely, translation and interpreting in any of their modalities (García, Mikulan, and Ibáñez 2016; García and Muñoz forthcoming). Think about an early aboriginal enabling trade between two distant tribes, Saint Jerome working by candlelight, a dragoman serving the Ottoman Empire, a religious translator in the Renaissance, a Jesuit missionary operating in China during the Qing dynasty, a conference interpreter at the Nuremberg trials, or an Argentine freelancer managing projects on Trados Studio for a Philadelphia-based agency. Through time and space, people have been using their heads to convey messages from and into their native and non-native languages (L1s and L2s, respectively). Over the centuries, radical changes have occurred in the types of cultural groups requiring cross-linguistic mediation, the sorts of texts involved, and the technological aids available – not to mention the development of formal training and practice settings. Nevertheless, amid this highly dynamic landscape, one factor has remained unaltered: every single instance of IR, as performed by human beings, has always relied on specific patterns of neurocognitive activity. However, that major thread running throughout the history of cross-linguistic brokering has largely escaped the radar of translation and interpreting studies (TIS). This is clearly not because of theoretical isolationism or epistemological narrowness. Since its very inception, TIS has incorporated multiple tools and concepts from diverse scholarly areas (Holmes 2000 [1972]), so that it has shaped itself as an “interdiscipline” (Snell-Hornby 1992, 2006) or, as some would have it, “a house of many rooms” (Neubert and Shreve 1992). Yet, even despite early calls for interaction between TIS and psychologically-oriented sciences (Holmes 1988), and although broad-gauge cognitive frameworks have become productive catalysts for research, the cerebral basis of IR has been recently deemed “a known unknown” in the field (Tymoczko 2012). As it were, in the vast edifice of TIS, the neurocognitive room so far amounts to little more than a dark, forlorn attic.

2

The Neurocognition of Translation and Interpreting

Admittedly, this scenario is not entirely surprising. TIS has long been rooted in a humanistic tradition, which has allowed for substantial advances without the need for biological considerations. In fact, even process-oriented trends have made great strides building on non-neural approaches (see Chapter 1, Section 1.2). Also, although relevant aphasiological evidence has been available since the early and mid-1900s (García and Muñoz forthcoming), seminal publications appeared exclusively in somewhat obscure, inaccessible journals. Moreover, the fundamental techniques of cognitive neuroscience were established only in the late twentieth-century (Cooper and Shallice 2010). In fact, it was not until recently that they began to be consistently employed to explore translation and interpreting processes (Muñoz, Calvo, and García 2018). But now, come the twenty-first century, the scenario proves quite different. Several trends within TIS have fully embraced experimental cognitive approaches (Ferreira and Schwieter 2015; García 2015a; Göpferich, Jakobsen, and Mees 2008; Jakobsen 2014; Schwieter and Ferreira 2017). In addition, pertinent literature from multiple fields has become widely available (Duffy 2000; Esposito 2016), and so has the methodological toolkit sustaining the growth of neuroscience (Yeung, Goto, and Leung 2017). Furthermore, and perhaps more crucially, increased attention to IR on the part of brain researchers (García 2013a, 2015b) has been echoed by greater interest in neural evidence from the TIS community (Diamond and Shreve 2010, 2017; Elmer 2012; García 2012a; García and Muñoz forthcoming; García, Mikulan, and Ibáñez 2016; Moser-Mercer 2010; Muñoz, Calvo, and García 2018). Thanks to these breakthroughs, we can finally turn on the light-switch in the neurocognitive attic and discover that, more than a “known unknown,” the neuroscience of IR constitutes a set of “unknown knowns” (García and Muñoz forthcoming). Such an endeavor requires framing translation and interpreting from a psychobiological perspective. Multiple sources of insight need to be integrated to understand how relevant mechanisms are organized in the brain, how they operate under different circumstances, and how they change as a result of field-specific training. No less importantly, all this knowledge needs to be linked to the main constructs of process-oriented trends, so that neurological data can seamlessly inform the overarching agenda of cognitive TIS. The project is ambitious, no doubt, but the time is ripe for it to come to fruition. Indeed, dozens of studies shed direct light on anatomical, functional, plastic, and behavioral aspects of IR. The available evidence spans a range of approaches, including neuropsychological assessments of brain-lesioned bilinguals, hemodynamic techniques – e.g., functional magnetic resonance imaging (fMRI), positron emission tomography (PET), functional near-infrared spectroscopy (fNIRS) –, electroencephalographic (EEG) methods – e.g., event-related potentials (ERPs), measures of oscillatory dynamics and functional connectivity –, invasive and



Translation, interpreting, and the brain behind it all

non-invasive brain stimulation methods – e.g., direct cortical electrostimulation, transcranial direct current stimulation (tDCS) –, and, of course, pertinent measures of cognitive performance – indexed through calculations of accuracy and response time in controlled experiments. Owing to these tools, our newly-opened attic teems with pages documenting multiple results and theoretical formulations. This book aims at compiling, organizing, and jointly interpreting them to reveal a coherent, evolving story. In pursuing this aim, I will capitalize on several lines of work I have forged with numerous research teams over the past decade. My experience as a practitioner and teacher of translation will here coalesce with my background in (i) TIS (e.g., Bender, García, and Barr 2010; García 2008, 2009, 2011, 2012a, 2014a, 2015a, 2015b; García and Arrizabalaga 2013), (ii) subfields of bilingualism and multilingualism (e.g., García 2016; García and Suárez Cepeda 2016; García 2014b; García et al. 2014; Santilli et al., 2018), (iii) different areas of linguistics (e.g., García 2010, 2012b, 2012c, 2013a, 2013b, 2015c; García and Ibáñez 2016a; García, Sullivan, and Tsiang 2017), and (iv) various branches of cognitive science and neuroscience (e.g., Abrevaya et al. 2017; Baez et al. 2017; Birba et al. 2017; García-Cordero et al. 2016; García and Ibáñez 2016b, 2016c; García et al. 2016a, 2016b, 2017a, 2017b, 2017c, 2018; Ibáñez and García 2018; Ibáñez, Sedeño, and García 2017; Melloni et al. 2016; Santamaría-García et al. 2017; Yoris et al. 2017). Achieving a harmonious synthesis of key contributions from all these trends is not only my immediate challenge in the following chapters, but also the greatest problem facing the neuroscience of IR in years to come. Fortunately, we are in a good position to move in that direction. II. Why should TIS be concerned with neurocognition? By now you may share the view that this endeavor is timely and probably feasible. Yet, a more pivotal question remains: is it worth the effort? As it happens, neurological evidence has proven unnecessary for characterizing multiple aspects of the translating and interpreting mind. Decades of research show that we can infer cognitive determinants of IR by extrapolating insights from other disciplines, observing professionals and students at work, studying subjective impressions about the operations involved, describing the final products to which they give rise (translated texts), or even measuring implicit physiological markers (e.g., pupil dilation) or task-related behavioral manifestations (e.g., keyboard actions). Unaided by brain-based methods, these approaches have illuminated the differences between backward translation (BT, from L2 to L1) and forward translation (FT, from L1 to L2) (de Groot, Dannenburg, and van Hell 1994; García et al. 2014a), the demands imposed by different translation units (Carl and

3

4

The Neurocognition of Translation and Interpreting

Kay 2011), the allocation of attentional resources between source text (ST) reading and target text (TT) production (Hvelplund 2011), the relative contributions of form-level and conceptually-mediated processes during IR (Kroll and Stewart 1994; García 2015a), and the role of numerous metacognitive strategies (Shreve 2009), to name but a few topics. Yet, while neural insights are not indispensable for cognitive TIS, their incorporation can foster progress in several ways. Allow me to support this claim by briefly anticipating some arguments deployed in Chapter 1. First, mental systems and operations hold consistent, predictable relations with neural structures and processes, as vastly evinced in lesion studies and imaging research (LeDoux 2002). Indeed, as argued by leading figures in the study of language (Pulvermüller 2002), memory (Kandel 2006), and emotion (Damasio 1994), mind and brain are necessarily integrated at all times. More particularly, in the case of IR, such a synergy can allow process-oriented researchers to make substantial discoveries at several levels. For example, the functional organization of key cognitive systems can be aptly studied considering translation and interpreting performance in brain-damaged bilinguals (García 2012a, 2015b) and its real-time neural signatures in healthy subjects (García 2013a). Also, precise characterizations of the inner unfolding of IR can be obtained through electrophysiological methods, with a temporal resolution that exceeds the possibilities of any other approach in the field (Luck and Kappenman 2012). Moreover, the interaction of different cognitive mechanisms (including their integration and segregation) can be explored via functional connectivity metrics (Friston 2011). At the same time, these and other methods can tap on the cognitive particularities of experts in IR. Importantly, too, causal (as opposed to correlational) evidence can be garnered through brain stimulation paradigms (Parkin, Ekhtiari, and Walsh 2015). The ensuing findings could then be used to test or formulate cognitive models and hypotheses, and eventually inspire innovations for the teaching and practice of various modalities. As a corollary, a new, profitable pathway towards interdisciplinary fertilization could be consolidated for present and future generations of TIS scholars. Of course, despite its “seductive allure” (Fernandez-Duque et al. 2015; Weisberg et al. 2008), neuroscientific research is far from perfect. Actually, it is plagued with theoretical and methodological problems (Munafò et al. 2017). However, its imperfections do not annul its contributions,1 and we could certainly do worse than incorporating the latter in the quest to understand IR. Needless to say, brain-based evidence is neither better nor more critical than that offered by behavioral, qualitative or otherwise humanistic approaches; but the opposite is equally false. The phenomena targeted by TIS are just too complex for us to afford the luxury of 1. And, in all honesty, can any scientific field claim to be free of theoretical and methodological shortcomings?



Translation, interpreting, and the brain behind it all

eschewing lines of insight merely because of tradition, dogmatism, or dualistic prejudgments. Without a doubt, any effort to expand current data sources is worth pursuing, and all the more so if the prospects are as powerful and manifold as those immanent in neurocognitive research. To put it briefly, unsuspected riches may be found even in a long-neglected attic. III. Aims and target audience Building on the above premises, The Neurocognition of Translation and Interpreting seeks to achieve five general, interrelated aims, namely: (a) introducing neurocognitive research on IR vis-à-vis other cognitive approaches; (b) describing the methodological toolkit employed so far in the literature; (c) presenting key notions of neurology, the neural basis of language, and the neurocognitive particularities of bilingualism as fundamental constraints to examine findings about IR proper; (d) compiling, organizing, and interpreting neuropsychological, neuroscientific, and behavioral evidence on highly prominent topics for TIS; and (e) discussing the present and future of the field, with emphasis on its accomplishments, strengths, weaknesses, and requirements. As implied in its objectives, the book is mainly intended for aspiring and professional researchers in cognitive TIS. Notwithstanding, given its breadth and the variety of sources it incorporates, it should also appeal to the broader translation and interpreting community, including teachers, students, and practitioners. Moreover, it ought to prove relevant for scholars working on bilingualism and neurolinguistics at large. Importantly, no previous knowledge of neurology, neuropsychology or neuroscience is needed to fully access this text. On a more general note, the volume could be said to be “retrospectively introspective,” for it seeks to provide the information and explanations I would have appreciated to have as I struggled to complement my humanistic background with a neuroscientific mindset. Hopefully, its readers will be able to swiftly address (or maybe even become aware of) many of the conceptual and methodological queries that typify such an intellectual adventure. IV. The contents, at a glance The volume comprises eight chapters. While readers who are new to neurocognitive research would do well to consult them all, those with a background in the neuroscience of language and bilingualism might wish to skip specific sections of Chapters 2 and 3. Whichever the case is, the contents will prove maximally approachable if perused in the proposed order.

5

6

The Neurocognition of Translation and Interpreting

Chapter 1 locates neural approaches to IR within the broad territory of cognitive TIS. First, an account is offered of non-neural trends and models, ranging from early insights rooted in formal linguistics to empirical approximations informed by observational, introspective, and experimental methods. Next, the history of brain-based research on IR is traced from pioneering clinical reports in the early twentieth century to contemporary works couched in neuroscientific and behavioral frameworks. To conclude, a series of epistemological notes is offered concerning the nature of data and conclusions in the field. Chapter 2 introduces the methodological toolkit used for investigating neurocognitive aspects of IR. Basic principles of experiment and stimulus design are followed by an overview of the most widely used tasks and an account of the key methods with which they are combined. The latter include behavioral measures, lesion models, functional neuroimaging, analyses of EEG signals, and even invasive techniques. The chapter closes with reflections on the transition from results to interpretations. Chapter 3 consists of fundamental prolegomena for newcomers to cognitive neuroscience and neurolinguistics. First, it addresses basic notions of neurology, both at macroanatomical and microanatomical levels. Second, it summarizes well-established findings concerning the neural basis of language, such as the functional organization and temporal dynamics of sublexical, lexical, morphosyntactic, semantic, and pragmatic mechanisms. Finally, it discusses the neurocognitive particularities of bilingualism as a prerequisite for translation and interpreting. Chapter 4 deals with disorders of IR in brain-lesioned bilinguals as an empirical framework to understand the broad organization of relevant pathways. The review spans all four translation neuropathologies documented to date: compulsive translation, inability to translate, paradoxical translation behavior, and translation without comprehension. Evidence from over twenty case studies is jointly assessed to shed light on the lateralization, functional autonomy, and internal specializations of the neural systems subserving IR. Chapter 5 focuses on behavioral and neuroscientific evidence on directionality. Whereas previous accounts explicitly or tacitly assumed that BT and FT relied on identical cognitive operations, several reports demonstrate that the two directions differ in terms of the main cortical and subcortical sites they engage, their degree of inter- and intra-hemispheric connectivity, their underlying electrophysiological modulations, and even their overall speed. Chapter 6 addresses the construct of translation units. Contrary to models which uniformly characterize IR processes irrespective of the nature of the ST segment, neuropsychological, neuroscientific, and psycholinguistic evidence indicates that quite diverse mechanisms are implicated depending on such a variable. Data gleaned in numerous studies show that different neurocognitive resources



Translation, interpreting, and the brain behind it all

are recruited for processing words, as opposed to sentences; cognates, as opposed to non-cognates; and concrete items, as opposed to abstract items. Experimental findings are interpreted as support for a dynamic, unit-sensitive conception of IR. Chapter 7 brings together results from more than thirty reports showing that sustained practice of simultaneous interpreting entails various neurocognitive adaptations. The chapter introduces the “interpreter advantage hypothesis” and shows that the formal development of interpreting skills involves neuroanatomical and neurofunctional changes in cortical and subcortical regions subserving verbal and non-verbal functions, some of which are also characterized by behavioral enhancements. An integrative account of the evidence is set forth thereupon, including a discussion of the speed, scope, and specificity of the plastic patterns detected so far. Finally, Chapter 8 offers a critical take on the present and future of the field. Key achievements, problems, and requisites are identified across methodological, theoretical, and institutional levels. By way of conclusion, neurocognitive research on IR is claimed to represent a major avenue for the development of TIS in the twenty-first century. In sum, The Neurocognition of Translation and Interpreting is an invitation to venture into an exciting yet overlooked area of TIS. Its contents are meant to consolidate the work done so far while inspiring further inquiries within and beyond its target topics. If, upon turning the final page, the reader is left with as many questions as answers, the mission is likely to have been accomplished. V. Conceptual delimitations A few conceptual delimitations are in order before delving into the contents. As stated in the title, this book is about translation and interpreting, not (just) about translators and interpreters. Certainly, professional practitioners have been targeted by a considerable number of pertinent studies – specially those assessing the impact of expertise in simultaneous interpreting, as detailed in Chapter 7. A good part of the evidence, however, comes from bilinguals without field-specific training. Importantly, the latter data prove crucial for present purposes, not only because of their quantity and quality, but mainly because several neurocognitive aspects of IR can be presumed unalterable irrespective of expertise. Mutatis mutandis, this is true regarding the existence of both shared and partially independent systems for L1 and L2, the relative functional independence of IR mechanisms relative to others involved in single-language tasks, the partial dissociations between the networks subserving BT and FT, the possibility of recruiting both form-level and conceptually-mediated routes for cross-language processes, or the functional specializations of gross brain areas for specific sub-functions (e.g., letter recognition,

7

8

The Neurocognition of Translation and Interpreting

phonological processing, lexical access, multimodal semantic processing, L1 syntactic parsing, articulation), among many other examples. Thus, although several properties of the psychobiological basis of IR are undoubtedly sensitive to the effects of formal training (Muñoz, Calvo, and García 2018), many of its broad features can be aptly studied considering acts of so-called “natural translation” (Harris and Sherwood 1978; Malakoff 1992). In short, this book is not (solely) concerned with the neurocognitive particularities of professionals in the fields of translation and interpreting, but mainly with fundamental biological and behavioral aspects of such activities in general. Also, as seen elsewhere in this Introduction, the term ‘IR’ will be systematically used to refer to the mechanisms that are presumably shared across translation and interpreting modalities. This aligns with broad definitions of the term ‘translation’, which is often used to encompass “any way in which a fragment of source language can be turned into the analogous target language fragment, irrespective of input and output modality” (Christoffels 2004: 5); or any goal-driven, interpretive, and communicative activity whereby a text is reformulated in another language, no matter its type, class, modality, or mode (Hurtado Albir 2001). Nevertheless, the term ‘translation’ is also typically reserved for instances in which both STs and TTs, or at least the former, manifest in print – as occurs in written or sight translation, respectively. To avoid confusions, then, modality-neutral claims will be made by reference to ‘IR’, whereas modality-specific statements will explicitly indicate a particular variety therefrom. Of note, the expression ‘IR’ is not merely a terminological convenience, for it also acknowledges that several neurocognitive systems are comparably involved in all forms of translation and interpreting. Undoubtedly, several functions (cutting across auditory, phonological, graphemic, attentional, and mnesic domains, for example) are differentially engaged during translation and interpreting. Yet, several linguistic subsystems intervening between sublexical- and motoric-level processes are likely shared by all cross-linguistic activities. This is arguably the case for most lexical, semantic, and syntactic operations. The alternative would be to postulate the existence of one lexical system for translation and another one for interpreting, one semantic system for translation and another one for interpreting, one syntactic system for translation and another one for interpreting, one set of inter-linguistic connections for translation and another one for interpreting, and so on – a position that lacks in both parsimony and plausibility (García 2012a). Thus, when used in reference to neurocognitive systems, the term ‘IR’ will account for (presumably) modality-general functions. However, no such generalizations will be implied when a process is described by reference to ‘translation’ or ‘interpreting’ in particular. Moreover, the book is exclusively concerned with interlingual processes. This implies that the evidence and conclusions presented should not be a priori



Translation, interpreting, and the brain behind it all

extrapolated to intralingual or intersemiotic translation (Jakobson 2000 [1959]). The latter two modes are characterized by specific neurocognitive foundations that go beyond the scope of the volume. In keeping with this focus on IR proper, only marginal attention will be devoted to studies which did not require actual translation or interpreting – unless these involve comparisons between students and professionals in a given modality. No extensive treatment will thus be offered of evidence from language-switching paradigms, for example. For detailed overviews of such research and its implications for TIS, the reader might wish to consult other works in the literature (e.g., Diamond and Shreve 2017). Additional specifications must be made concerning the neuroscientific viewpoint adopted throughout the book. First and foremost, by characterizing IR as a collection of neurocognitive processes I am not denying the many other dimensions it involves. Put succinctly, a copula is not necessarily exclusionary: in stating that water is transparent I am not implying that it is not a liquid; similarly, in stating that simultaneous interpreting is a complex psychobiological activity I am not implying that it is not a sociocultural, interactive, institutionalized, skopos-driven task. Needless to say, the same is (or should be) true of every other theoretical orientation in TIS. When scholars working from the vantage point of functional linguistics or cultural studies characterize translation as a textual and social activity, are we to assume that they believe the brain has nothing to do with it? If such scholars are not accused of “functional reductionism” or “sociocultural reductionism,” why should comparable denunciations be entertained for those focusing on biological dimensions? In brief, this book conceives neurocognitive features of translation and interpreting as one of the many facets entailed by such complex, heterogeneous activities. The exploration of those features requires a combination of neurological and behavioral evidence. Except for specific research questions, insights into the neural basis of cognition are most informative when complemented with performance data (Krakauer et al. 2017). Thus, throughout the following chapters, neuroanatomical, hemodynamic, electrophysiological, and brain-lesion evidence will be constantly interpreted in combination with output measures. However, when a behavioral pattern is interpreted alongside a neural correlate it does not mean that the latter is necessarily the cause of the former, or vice versa. Most neuroscientific evidence is correlational and, unless the evidence otherwise allows, I will be scrupulous in choosing wordings that capture such a fact. So, when an outwardly measurable process is said to be “related to,” “associated with,” or “linked to” certain neurological patterns, or when these are said to “subserve” the former or be “implicated in” them or “engaged by” them, the reader should not assume a causal relationship in either direction. In the few cases where such conclusions are justified, the text will be explicit about it.

9

10

The Neurocognition of Translation and Interpreting

Another relevant clarification concerns the distinction between structural and functional correlates. In standard neuroscientific jargon, structural correlates are those pertaining to the anatomy of the brain, both macroscopically (at the level of neural regions) and microscopically (at the level of neurons), without considering any type of activation pattern. These aspects can be assessed, for example, through measures of gray matter volume or cortical thickness in a specific area. By contrast, functional correlates are provided by measures of brain activity, either when the subject is performing a specific cognitive task (in active paradigms) or not thinking about anything in particular (in resting-state paradigms). In the present volume, most evidence on functional correlates of IR comes from fMRI, PET, and EEG experiments. Therefore, when a study reports a structural difference between two groups, for instance, the reader should not interpret that result as evidence of greater activation, and vice versa. Note, however, that while most neuroscientific studies address only one type of correlate, all brain properties are characterized by complex interdependencies, often beyond the analytical capabilities of current technologies (Ibáñez and García 2018). Those biological patterns will often be interpreted as signatures of specific ‘domains’, such as word processing, executive functions or specific sub-operations therefrom (e.g., semantic access or inhibitory control, respectively). Although the notion of domain might seem self-evident given these examples, the term is not uniformly used across (or even within) academic specialties, which calls for some definitional orientation. Accordingly, in line with a widely assumed conception in neuroscience, cognitive domains will be defined here as “theoretically isolatable processes associated with particular types of information, partially recurring phenomenological attributes, and sometimes well-defined neurological correlates, such as language, memory, emotion, and so on” (García and Ibáñez 2018: xi). In principle, no other interpretations should be attached to this term as used in the following chapters. Finally, none of the statements set forth throughout the book is meant to constitute an ultimate truth about IR. What follows is a compendium of empirical results and possible conclusions, but neither should be confounded with definite demonstrations. Even at their most assertive, research-based claims must be deemed tentative or falsifiable, with some proving more plausible than others. Any assumption of factuality should be dispelled or, at the very least, taken with reservation. For better or worse, such is the way of science. With these considerations in place, the scene is set to look at translation and interpreting as embedded in the brain. Let us step into a disregarded room of our disciplinary house and learn about one of the few constants cutting throughout the history of these activities. To take the first steps, all you need to do is turn this page – and deciding to do so, paradoxically enough, is a no-brainer.

Chapter 1

Mind and brain in the study of translation and interpreting

1.1

Minding the brain, braining the mind

Somewhere on planet Earth, an asteroid dweller once met a fox. They greeted each other and talked about flowers, hunters, and chickens. A bond was quickly forged between them, but the visitor soon had to leave… The anecdote would prove fully irrelevant here if it were not because of its oft-cited highlight: as a farewell token, the fox left his new friend with a timeless maxim. “L’essentiel,” the animal affirmed, “est invisible pour les yeux.”2 No doubt, the statement had a loftier scope than that of cognitive TIS, but it could readily become an apt motto for the field. As it happens, the processes involved in IR are literally imperceptible: mental operations such as source-segment comprehension, cross-linguistic lexical search, target-segment formulation, or working memory (WM) storage cannot be actually seen, touched, heard or directly measured in any way. However, each of these processes (just like any other cognitive function) leaves measurable traces outside and inside our bodies, including the accuracy of our productions, our own reflections on them, the speed of the accompanying (conscious and unconscious) movements, the dilation of our pupils, and, of course, multiple patterns of brain activity. It is by reference to such observable proxies that cognitive researchers in TIS can characterize the essentially invisible phenomena they study. All the signatures, correlates, and outcomes listed above give hints on the crux of translation and interpreting, that is, “the ability to generate a series of more than one viable target text [and] to select only one […] quickly and with justified confidence” (Pym 2003: 489). In addition to non-empirical modelling (Nida 1964), multiple aspects of this cognitive skill have been inferred through non-participant observation (Seleskovitch 1968, 1975, 1978), think-aloud protocols (TAPs) (Bernardini 2001), corpus-based studies (Alves and Vale 2017), behavioral measures in controlled psycholinguistic paradigms (García 2015a), recordings of keylogging (Jakobsen 2. The quote belongs to Le Petit Prince, published by Antoine de Saint-Exupéry in 1943. In most English versions of the book, it is translated as “what is essential is invisible to the eye.”

12

The Neurocognition of Translation and Interpreting

2014) and eye-tracking (Göpferich, Jakobsen, and Mees 2008) patterns during multi-sentential text translation, and assessments of relevant executive functions (García 2014a). These approaches have long dominated the field, leading to hypotheses and models which rely exclusively on evidence from “outside the head.” Instead, research on brain markers, though arguably more proximally related to the elusive cognitive processes in question, has not been systematically incorporated into the mainstream toolkit of TIS (García and Muñoz forthcoming). A reversal of the latter scenario is certainly desirable. Neural-based approaches offer unique opportunities to characterize core features of IR, such as the functional organization of its putative mechanisms, the inner time course of relevant processes, the interaction of different hubs in distinguishable functional networks during various sub-processes, key aspects of field-specific competence and expertise, and the relationship of all these factors with outward manifestations. As a complement to well-established trends, neural approaches significantly expand the current repertoire of cognitive TIS, opening new possibilities for methodological, theoretical, and interdisciplinary integration. The following sections chart the existing territory of this field, addressing the strengths and limitations of non-neural research trends and highlighting the historical development, current status, and epistemological particularities of neuroscientific approaches. Without further ado, let us discuss how scholars within TIS manage to describe what, metaphors aside, simply cannot be perceived. 1.2

Outside the head: Non-neural cognitive approaches

The most widely used approaches in cognitive TIS could be deemed non-neural, insofar as they are uninformed by cerebral data. Broadly speaking, these could be classified into five categories. First, rationalist approaches seek to model relevant cognitive systems or operations adopting purportedly axiomatic theoretical premises in the absence of any empirical testing. Rationalist models in cognitive TIS have been mainly informed by analytical linguistics, constituting prime examples of “disciplinary importing” (Kaindl 2004). Second, observational approaches rely on data which emerge naturally in non-manipulated translation or interpreting settings. As such, they are empirical, but not experimental. Third, introspective approaches are based on participants’ reflections about their ongoing mental processes as they translate. Thus, they are characterized by reliance on second-order, metacognitive impressions. Fourth, corpus-based approaches offer insights into the translating mind by considering large amounts of annotated text and translation process data. Finally, quantitative approaches make use of controlled designs to measure brain-external patterns and examine the impact of independent variables



Chapter 1.  Mind and brain in the study of translation and interpreting 13

on dependent variables (see Chapter 2, Section 2.4). These frameworks, couched in methods from cognitive and behavioral science, are currently dominant in the field. Note that the classification of these approaches as “non-neural” is based on a methodological, as opposed to an ontological, criterion: their underlying research procedures do not necessarily entail a rejection of the neurological basis of the observed phenomena. Also, while the five approaches are only briefly presented below (mainly as a backdrop against which to flesh out the distinctive features of neurocognitive trends), they all have a prolific history. Although a full treatment of them falls beyond the scope of this chapter, interested readers may wish to consult the sources suggested at the end of each subsection. 1.2.1 Rationalizing translation: Insights from analytical linguistics The first systematic development within cognitive TIS consisted in the adaptation of a rationalist linguistic model. In the 1960s, Eugene Nida turned to Noam Chomsky’s (1959, 1965) generative-transformational grammar as a basis for characterizing mental operations in translation. This pioneering work, consolidated in Toward a Science of Translating (Nida 1964) and The Theory and Practice of Translation (Nida and Taber 1969), represents a milestone as it introduced the very notion of ‘process’ in TIS (Moya 2004). Nida’s procedure in formulating his three-stage model (Nida and Taber 1969) clearly illustrates the rationalist approach. First, he turned to Chomsky’s (1965) formal model and adopted a simplified set of constructs therefrom, namely: the notions of deep structure, surface structure, and intermediate transformation rules. Second, he proposed that, insofar as translation involves two languages, these elements had to be present twice for a model of IR – once for the source language (SL) and once for the target language (TL). Third, he introduced an ad hoc theoretical component (absent in Chomsky’s formulations) to enable a link between the generative processes supposed to operate in each language. The result was a tripartite model comprising (i) an analysis stage, in which deep structures are construed from the ST’s surface structures via operations such as back-transformations; (ii) a transfer stage, in which the allegedly universal elements of the resulting deep structure can be transferred to a TL deep structure without essential losses; and (iii) a restructuring stage, in which the latter deep structure is subjected to specific transformational rules until a satisfactory surface structure emerges in the TL. This process could result in either formal or dynamic equivalence – for details, see García (2011). A similar approach has been followed by Roger Bell (1991). His cognitive model also hinges on the adoption of constructs from analytical theories of language and the addition of a few components allowing for connections between

14

The Neurocognition of Translation and Interpreting

languages. Drawing on systemic-functional linguistics (e.g., Halliday 1985) and relevance-theoretic pragmatics (Sperber and Wilson 1986), Bell posited a model of translation including systems for visual word recognition (during ST processing) and writing (for TT production); lexical, syntactic, semantic, and pragmatic systems for each language; an idea organizer; a cognitive planner; and a component for deciding whether to translate an active representation or not. These components would allow for the ST to be processed in terms of its constituting forms, meanings, speech acts, and textual coordinates, and for the ensuing high-level representation to be synthesized into a TT through a reverse traversal of all the intervening processing levels. Somewhat surprisingly, the model assumes that all translation operations can be characterized by reference to the simple clause as input and output unit – for details, see Hurtado Albir (2001) and Chapter 6. The works of Nida and Bell, as well as others (e.g., Gutt 1991), exemplify the cornerstone of rationalist approaches. The constructs of formal descriptive theories, rooted in logical analyses of explicit or implicit propositions, are taken as pillars to characterize the processing of STs and TTs. Importantly, since none of those basic models makes any provision for inter-linguistic processes, additional operations are then integrated to bridge between the two intervening languages. Crucially, the plausibility of these proposals is not tested against actual processing data (be they behavioral, physiological, introspective, or neural); rather, it is simply assumed to follow from the soundness of the underlying models. In addition, the demonstration of how the models would operate is typically reduced to the analysis of a few carefully chosen sentences. This approach possesses some obvious strong points. Above all, it capitalizes on the achievements of highly relevant disciplines and fosters direct interdisciplinary discussion with them. Also, it allows for constant revisions of the proposed IR models as new developments emerge in the underlying frameworks – though this possibility has rarely been exploited in TIS. Moreover, it turns the complexities of IR into a conceptually approachable matter by interpolating hierarchical components that support useful reflections on single-language processing. Importantly, too, it provides overarching approximations to the cognitive architecture of IR systems, whereas other approaches can tap only on specific sub-operations one at a time – see Section 1.2.5. However, rationalist formulations also face important caveats. For example, at least in these two cases, since the theories adopted do not characterize productive or receptive processes in particular, the ensuing translation models must incorporate bottom-up or top-down processing assumptions not present in the original conceptual premises. In addition, given that living languages differ greatly in terms of their structural and pragmatic properties (Evans and Levinson 2009) and that theory-specific grammars exist only for a small subset of them, the ensuing IR



Chapter 1.  Mind and brain in the study of translation and interpreting 15

models are either relevant for a few language pairs or framed by unwarranted assumptions of universality. Furthermore, given that the linguistic/pragmatic theories informing them typically lack considerations of extra-verbal functions (e.g., executive or attentional mechanisms), these are excluded from the resulting proposals. More generally, the absence of empirical foundations renders these accounts more speculative than scientific.3 In fact, the linguistic/pragmatic theories considered by Nida’s and Bell’s models do not possess a robust experimental tradition, which dramatically limits their explanatory development.4 For more specific treatments of rationalist models, see Gentzler (2001), Hurtado Albir (2001), and Munday (2001). 1.2.2 See but don’t touch: The observational trend The 1960s also witnessed the rise of a second approach to cognitive modeling in TIS. The trend led by Danica Seleskovitch and Marianne Lederer, first hatched in the École Supérieure d’ Interprètes et Traducteurs (ESIT), sprang from the idea that the cognitive processes involved in simultaneous interpreting (as well as other forms of IR) could be aptly inferred by observing professional interpreters at work (Seleskovitch 1968, 1975, 1978; Lederer 2002 [1978]). Though rooted in empirical data, this approach explicitly rejects the value of experimental designs. To quote one of its leading figures, [i]nterpreting is a human performance in which cognitive activity is first and foremost; it therefore leads us into the field of psychology with no need to resort to special experiments; in this field the connection between thinking and speaking can be observed as it materializes with each segment of speech.  (Lederer 2002 [1978]: 131; my emphasis)

In the ESIT framework, those connections were mainly traced through non-participant observation of simultaneous interpreters (SIs) working at conferences (Seleskovitch 1968) or analysis of the notes taken by consecutive interpreters in professional settings (Seleskovitch 1975). The evidence thus obtained, complemented with insights from linguistics and cognitive science, resulted in the Théorie du sens, whose conceptual core has remained unaltered for more than half a century – see, for example, Seleskovitch (1968, 1978) and Lederer (1994).

3. This is most obvious in the case of Nida’s three-stage model, which inherits central counterfactual idealizations of generative grammar, such as the assumption of ideal speakers/hearers immersed in perfectly homogeneous linguistic communities (García 2010). 4. For more in-depth critiques of Nida’s and Bell’s models, see Hurtado Albir (2001), García (2011), Gentzler (2001), and Muñoz Martín (2007).

16

The Neurocognition of Translation and Interpreting

The theory posits three main processes: comprehension, deverbalization, and re-expression. Comprehension would arise as the meanings evoked by the ST’s units interact with the subject’s encyclopedic knowledge (bagage cognitif) and contextual knowledge (contexte cognitif), via non-serial, bidirectional links among all systems involved (Seleskovitch 1981). The next phase, deverbalization, consists in apprehending a non-verbal meaning to be reformulated in the TL. Seleskovitch (1978) pointed out that, even at an ideal rate of 100–120 words per minute, it would be impossible to memorize all units in the ST so as to transcode them directly – a claim supported by the preeminently ideographical nature of interpreters’ notes (Seleskovitch 1975). Therefore, she maintained, interpreting hinges on the construal of transitory, non-linguistic states of consciousness in which input is deprived of its sensory features to facilitate the retention of core meanings. Such meanings would then feed a process of re-expression, through which the TT is progressively formulated with reference to encyclopedic and contextual knowledge, in a manner similar to that underlying monolingual production (Lederer 1994). The observational approach has the virtue of offering holistic accounts of the processes analyzed, which is beyond the possibilities of most controlled experiments in the field (see Section 1.2.5). Also, it succeeds in satisfying the imperative of ecological validity, as participants’ behavior is assessed without introducing artificial constraints or manipulations. Moreover, as the foremost incarnation of this approach, the Théorie du sens allowed complementing interpreter training with a practically relevant theoretical framework, while asserting interpreting studies as an autonomous subdiscipline within TIS (Gile 2009). Notwithstanding, this trend is not without major shortcomings. The rejection of experimental designs renders it incapable of teasing apart the specific role and timing of fine-grained sub-processes, such as lexical access, inhibitory control or mental-set shifting – for insights into these, see Chapter 7. Moreover, insofar as observations from members of the community under study (interpreters) are both subjective and prone to positivity and confirmation biases, the ensuing conclusions are grounded in speculation rather than scientific findings (Gile 2009; Moser-Mercer 1994). In fact, specific claims of the Théorie du sens, such as the irrelevance of conceptual mediation for translating numbers (Lederer 1994; Seleskovitch 1975), seem falsified by behavioral evidence (Duyck and Brysbaert 2008). More generally, in the absence of experimental hypothesis testing and interdisciplinary rapprochement with well-established scientific methods, the observational approach can result in idealized, intuitive, prescriptive formulations (Gile 1990). For more detailed accounts of this approach, see Lederer (1994), Hurtado Albir (2001), and García (2011).



Chapter 1.  Mind and brain in the study of translation and interpreting 17

1.2.3 Take a look at yourself: Introducing TAPs In the 1980s, a third approach emerged in an attempt to overcome the limitations of deductive normative models. Building on the introspective method resurfaced by Ericsson and Simon (1984) within psychology, a number of translation scholars started applying TAPs to obtain first-hand data on IR processes. In TAP-based research, subjects are asked to verbally describe or explain their actions and thoughts during or after actual translation tasks, so that their tape-recorded or videotaped responses can be scrutinized for indicators of particular knowledge states (Lörscher 1991, 2005). These indicators include analytical maneuvers (e.g., equivalent monitoring, reduction, editing) and involvement markers (e.g., first-person references, emphatic particles, mentioning of verbal processes), in addition to several other elements (Bernardini 2001). Conclusions about cognitive processes are then made via subjective interpretations of what such indicators might mean. Most researchers in this line have employed monologue protocols, whereby a single participant talks out loud with no intervention from the experimenter. However, the approach also lends itself to more interactive frameworks, via dialogue (House 1988) or group (Hönig 1991) protocols. These concurrent forms of data production are often used in combination with retrospective interviews (Kussmaul and Tirkkonen-Condit 1995). In their various versions, TAPs have been mainly used for assessing the strategies employed by professional and prospective translators, identifying various types of problems during the translation process, and deriving pedagogical proposals (Kussmaul and Tirkkonen-Condit 1995). TAPs represent the first systematic tool to circumvent the shortcomings of rationalist and observational approaches. Their main contribution arguably lies in their capacity to tap into issues which cannot be easily assessed via behavioral experiments, such as local and global strategies at a discourse level or other metacognitive processes during text translation. Also, by the end of the twentieth century, they had become a prime source of empirically-driven hypotheses regarding decision-making profiles, problem-solving strategies, and process automatization in professional and non-professional translators (Kussmaul and Tirkkonen-Condit 1995). Yet, TAPs have been harshly criticized on several respects. Above all things, concurrent verbalization has been found to interfere with IR proper, delaying processes by roughly 25% and altering the unfolding of specific strategies (Jakobsen 2003). In addition, verbalizations capture only a portion of relevant operations at best, as only well-structured thoughts (as opposed to more frequent mental procedures) are typically manifested in the protocols – and these may actually constitute illusory, post-hoc rationalizations rather than accurate descriptions of the actual processes (Hönig 1992; Kruger and Dunning 1999; Nisbett and Wilson 1977). Also, the approach is of limited use for research on interpreting, given that oral

18

The Neurocognition of Translation and Interpreting

production is part of this modality per se and thus cannot be used simultaneously for introspective purposes (Ahrens 2017). Finally, the introspective approach has failed to forge a solid unifying paradigm for study design, data analysis, and presentation of results, which casts doubts on the very generalizability and relevance of reported findings (Bernardini 2001). For in-depth treatments, see Kussmaul and Tirkkonen-Condit (1995) and Bernardini (2001). 1.2.4 From product to process: Corpus-based studies Corpus-based studies constitute a data-driven approach to characterize mental aspects of IR through massive collections of translated texts. Traditionally, analyses of parallel corpora (e.g., Hansen-Schirra, Neumann, and Steiner 2007) have been mainly used to detect prototypical patterns in TTs without explicit cognitive commitments. However, more recent extrapolations of this approach involve tracing processes by reference to products. Thanks to digital tools, translation-process data can be queried semi-automatically in search of prototypical text-production patterns (Alves and Vale 2009). More particularly, studying the unfoldment of those patterns in time offers a window into planning and cognitive effort in translation (Alves and Vale 2017). For example, corpus-based data has been combined with other methods to gain insights into translation units and grammatical shifts (Alves et al. 2010) and to identify patterns distinctively related to specific phases (orientation, drafting, revision) of the translation process (Alves and Vale 2017). Corpus-based research has considerable potential for triangulating process and product data. In particular, it can profit from the availability of rich corpora including varied text typologies from diverse language pairs. However, this approach faces serious challenges. For example, a given pattern detected in translation products may have stemmed from widely different mental processes, which problematizes cognitively-oriented interpretations of the results. This scenario is further complicated by the lack of a well-established framework to ascribe TT patterns to particular cognitive operations. Overall, further work is needed to better gauge the possibilities and limitations of this approach – for relevant considerations, see Alves and Vale (2017). 1.2.5 Quantifying performance Despite their differences, the previous approaches share a reliance on logical, deductive, and/or qualitative analyses. None of them employs normative databases for stimulus construction or validation, objective measures for behavioral assessment,



Chapter 1.  Mind and brain in the study of translation and interpreting 19

or numerical data which can be analyzed as dependent variables through statistical methods. Deprived of these procedures, they can only provide general descriptions of broadly defined macro-constructs (e.g., comprehension, re-expression) or weak speculations about fine-grained operations (e.g., which specific mechanisms are recruited for processing specific translation units). These are the major gaps bridged by quantitative approaches. Quantitative approaches were first implemented in the 1960s and 1970s, with groundbreaking measurements of ear-voice spans (Gerver 1976; Oléron and Nanpon 1964), word-per-minute processing rates (Gerver 1969, 1975), and levels of propositional correspondence between STs and TTs (Barik 1975; Gerver 1975) in SIs. Yet, beyond these historical milestones, objective explorations of IR nowadays rely on four well-established quantitative methods, namely: psycholinguistic paradigms, keylogging experiments, eye-tracking studies, and executive-function assessments. 1.2.5.1 Word by word: Psycholinguistic paradigms In the mid-1980s, Potter et al. (1984) reported two unprecedented experiments in which classical psycholinguistic methods were used to assess bilingual processes, including word translation. This trend evolved rather sparsely until the mid-1990s, when Kroll and Stewart (1994) documented clear asymmetrical effects between BT and FT and set forth the widely-cited Revised Hierarchical Model of bilingual memory.5 Thereupon, psycholinguistic studies on IR have grown exponentially in the field of bilingualism, although the ensuing evidence has only recently found its way into TIS proper (García 2015a). In this type of studies, participants view isolated lexical or sentential items on a computer screen and are asked to perform specific tasks on them, such as deciding whether a given letter string is a word (e.g., Bajo, Padilla, and Padilla 2000), establishing whether two stimuli (from the same or different languages) are similar in meaning (e.g., de Groot 1992; Talamas, Kroll, and Dufour 1999), or translating words out loud (e.g., Christoffels, de Groot, and Kroll 2006; García et al. 2014), among several other possibilities. Typically, participants execute their responses by pressing a key or by uttering a target word next to a voice-activated switch. The subjects’ performance is then usually assessed in terms of accuracy (as either correct or incorrect) and response times (i.e., the number of milliseconds that elapsed since stimulus presentation until a response was made). The ensuing outcomes can be interpreted in cognitive terms. For example, if a group is more accurate than another in recognizing translation equivalents, then the former can

5. For a discussion of this model, see Kroll et al. (2010) and Brysbaert and Duyck (2010).

20 The Neurocognition of Translation and Interpreting

be said to possess greater knowledge of equivalence relations. By the same token, evidence of shorter response times could reflect reduced processing effort or more efficient connections within or between relevant cognitive systems (García 2015a). This framework allows building carefully controlled stimulus sets, such that two conditions can be constructed which differ only in one critical aspect. If performance is better or faster for one of those conditions, then that differential aspect can be reasonably proposed as an important underlying factor – for an in-depth treatment, see Chapter 2, Sections 2.4 and 2.5. Under this rationale, psycholinguistic paradigms have been variously used to reveal cognitive particularities of BT relative to FT (e.g., de Groot, Dannenburg, and van Hell 1994), the role of form-level and semantic associations during IR (e.g., Christoffels, de Groot, and Kroll 2006), processing costs associated with different translation units (such as nouns vs. verbs, cognates vs. non-cognates, and abstract vs. concrete words) (e.g., van Hell and de Groot 1998a), and lexical processing differences between professional translators (García et al. 2014) or interpreters (Santilli et al. 2018) and different control groups. The most prominent advantage of psycholinguistic experiments is that they can circumvent the biases of second-order impressions or self-reports, as data from large samples can be objectively analyzed via conventional statistical criteria. No less important is the possibility of constructing stimulus sets that are carefully matched for key variables, so as to explore fine-grained processes or operations related to particular linguistic categories while ruling out potential confounds. Also, this approach easily lends itself to designs in which the main variable under scrutiny is kept tacit to the subject – for elegant examples, see de Groot (1992) and Thierry and Wu (2007). More generally, psycholinguistic experiments on IR have given rise to a solid research program, with multiple systematic and replicable results (García 2015a). Notwithstanding, word- and sentence-level research lacks in ecological validity, as it typically faces subjects with decontextualized, isolated stimuli. Moreover, this approach creates artificial response conditions (typically, pressing of two or three predefined keyboard buttons) which fail to capture the situated, skopos-driven processes proper to discourse-level IR. Moreover, although psycholinguistic experiments have yielded a systematic and expanding empirical corpus, it is not clear how all available findings could be integrated in an overarching model of translation or interpreting. For a general and critical overview of relevant research, see García (2015a). 1.2.5.2 Type your mind away: Keylogging experiments As mentioned above, despite all their benefits, classical psycholinguistic paradigms possess low ecological validity, given that they involve decontextualized stimuli and highly artificial response settings. However, the introduction of keylogging in the



Chapter 1.  Mind and brain in the study of translation and interpreting 21

1990s offered a valuable opportunity for TIS researchers to partially overcome those limitations while preserving objectivity in their measurements. Keyloggers are computer programs which record and time all or most of a subject’s actions on a computer keyboard as he or she performs a particular task. In the case of translation, the program’s workspace is typically divided in two halves, such that the ST is presented on the upper part of the screen and the TT is typed on a lower section. The main tenet of this approach is that units of text production during typing are indicative of ongoing operations within the participant’s mind. In the words of Jakobsen (1998: 74), the process of writing a translation constitutes behaviour that can be studied quantitatively – across time – and interpreted as a correlate of mental processing. The assumption is further that it will be possible to triangulate qualitative and quantitative data and test hypotheses derived from analyses of qualitative data against quantitative data, and vice versa.

While various keyloggers have been available for decades,6 researchers in the field have mainly resorted to Translog (Jakobsen 1995) and Translog-II (Carl 2012), which were specifically designed to examine translation processes. In particular, Translog-II allows drawing inferences about cognitive processes by assigning a timestamp to each insertion, deletion, cursor movement, and additional keyboard and mouse operations performed by the subject. Such indicators have been used to examine multiple topics, including the distribution of cognitive effort across translation phases (Carl and Buch-Kromann 2010), the allocation of attentional resources between ST reading and TT production (Carl and Kay 2011), the relation between ST difficulty and the size of the translation unit (Dragsted 2005), and the impact of time pressure on translation (Jensen and Jakobsen 2000). Keylogging technologies are unique in that they can be used to explore production processes during IR in a non-intrusive fashion, combining the objectivity of quantitative measures with the naturalness of realistic task settings. In addition, the data captured by relevant activity indicators can be correlated with assessments of TT quality, partially bridging the gap between the process and the product of translation. Moreover, like classical psycholinguistic designs, they offer the option of manipulating multiple subject-, text-, and task-related variables. On the negative side, the temporal resolution of Translog-II and other keyloggers may not be sufficiently high to tap fine-grained processes locked to particular actions or ST items. Furthermore, this methodology leaves researchers blind as to what certain indicators (e.g., writing pauses) might actually be reflecting. In fact, 6. Some of the best-known ones are TraceIt/JEdit (www.nada.kth.se/iplab/trace-it), Scriptlog (www.scriptlog.net), and Inputlog (www.inputlog.net).

22

The Neurocognition of Translation and Interpreting

the ensuing data can prove too noisy and hard to interpret; for example, typos and errors of various kinds typically abound in the obtained logs, and their relation to putative cognitive processes cannot be unequivocally ascertained. Also, given that performance differs greatly from subject to subject, it is often hard to generalize findings at the sample level – let alone at the population level. Finally, whereas psycholinguistic designs allow constructing restrictive stimulus lists such that multiple variables can be controlled between source and target words (e.g., García et al. 2014; Santilli et al. 2018), this is beyond the possibilities of text-based keylogging experiments. For further insights into keylogging and its applications, see Carl and Kay (2011) and Hvelplund (2011). 1.2.5.3 Windows to the (translating) soul: Eye-tracking studies Just as TT production can be examined in real time via keylogging, so too ST processing can be assessed via eye-tracking. Based on the same two-window layout used for keylogging, eye-tracking relies on an array of cameras and/or sensors to quantify diverse aspects of ocular activity while the subjects perform translation tasks on a computer (Göpferich, Jakobsen, and Mees 2008). The whole approach rests on the so-called ‘eye-mind’ and ‘immediacy’ assumptions, which, respectively, posit that “there is no appreciable lag between what is being fixated and what is being processed,” and that “the interpretations at all levels of processing are not deferred; they occur as soon as possible” (Just and Carpenter 1980: 331). Insofar as these notions are accurate, changes in eye-specific parameters can be taken as proxies of underlying cognitive operations. Fixations (i.e., how long the eyes rest on a given portion of text) are taken to reflect how effortful a process is or which parts of the ST or the TT are receiving greater attention at a given moment. This can be estimated by considering the total number of fixations as well as their total or mean duration. Moreover, eye-tracking also capitalizes on several other indicators, such as pupil dilation and constriction (changes in the size of the center of the iris), eye saccades (short, rapid movements as the eyes change their object of attention), or regressions (returns to previous chunks of text). Various combinations of these measures have been reported in the TIS literature, often in tandem with keylogging,7 to investigate differences in attentional allocation among the orientation, drafting, and revision phases (Carl and Buch-Kromann 2010); the coordination of ST comprehension and TT production (Carl and Kay 2011; Dragsted 2010); the impact of text complexity during

7. Insofar as the processes tapped by eye-tracking and keylogging cannot be readily disentangled as distinct activities (Dragsted 2010), the joint employment of these methods is both justified and characterized by major methodological difficulties.



Chapter 1.  Mind and brain in the study of translation and interpreting 23

translation (Sharmin et al. 2008); and differences between BT and FT (Pavlović and Jensen 2009), among other issues. Despite the artificiality of the head-mounted devices, eye-tracking largely meets the imperative of ecological validity, especially as it allows working with real(istic) texts. Also, as opposed to keylogging and psycholinguistic paradigms, it provides data for processes which lack overt motor correlates. This approach further stands out by its capacity to tag certain operations as driven by ST or TT processing. Importantly, too, its complementarity with keylogging results in a powerful approach to jointly examine multidimensional receptive and productive operations during IR. Yet, eye-tracking is not without shortcomings of its own. First, individualized calibration can prove cumbersome and unreliable. Second, the technique is so sensitive that it can be affected by several artifacts, such as pupillary adaptations due to environmental light changes, noise induced by head motion, and uncontrollable individual physiological factors (e.g., eye humidity). Indeed, estimations of which text segment was being looked at may yield inaccurate detection rates of up to 35% (Jensen 2008). Third, most indicators are ambiguous as to which particular process they are reflecting; at best, they can be interpreted as indices of coarse-grained constructs (e.g., cognitive effort or attentional allocation) but they prove mostly moot concerning fine-grained operations (e.g., lexical access, semantic activation, form-level cross-linguistic processing). Lastly, serious concerns can be raised regarding the consistency and generalizability of findings: as a tradeoff to naturalness, processing time for the same (portions of) STs differs greatly among subjects – and between-subject comparability is even more complicated when it comes to ocular activity on TTs, which are necessarily different for each participant. For a detailed treatment of eye-tracking research in TIS, see Göpferich, Jakobsen, and Mees (2008). 1.2.5.4 The non-verbal side of IR: Executive-function assessments Although all the approaches described above focus on linguistic materials and processes, the tasks they employ also tax multiple non-verbal functions. However, the very use of verbal stimuli makes it hard to ascertain to which extent the results obtained are driven by language-specific or extra-linguistic (domain-general) mechanisms. Assessments of executive functions represent a well-established approach to delve into the latter. Executive functions comprise a number a interrelated top-down mechanisms for selecting, coordinating, sustaining, and monitoring ongoing processes, including attentional allocation, inhibitory control, short-term memory (STM), WM, and mental flexibility (Diamond 2013). They underlie the unfolding of any type of behavior, irrespective of the task and processing modality at hand. Each executive

24

The Neurocognition of Translation and Interpreting

domain can be evaluated through specific paradigms. For example, the Stroop test has been widely used to study inhibitory skills (Scarpina and Tagini 2017), whereas digit and word span tasks are reliable proxies of STM capacity (Conway et al. 2005) – for an account of these and other executive tests, see Chapter 2, Section 2.3.2.2. As is the case with psycholinguistic paradigms, the subjects’ accuracy (or number of hits) and response times are assumed to reflect the efficacy and the efficiency of the subdomain being assessed, respectively. Though mostly conducted outside the jurisdiction of TIS proper, a considerable body of research has looked at executive functions in connection with IR. In particular, emphasis has been placed on interpreters and interpreting students, given that these functions are hypothesized to be advantaged in them (García 2014a) – see also Chapter 7. For example, building on a pioneering study by Bajo, Padilla, and Padilla (2000), several scholars have relied on relevant tasks to test whether professional and prospective interpreters outperform other populations in their ability to retain (Christoffels, de Groot, and Kroll 2006; Köpke and Nespoulous 2006) or recall (Signorelli, Haarmann, and Obler 2012) information, perform simultaneous tasks (Strobach et al. 2015), shift between different mental schemata (Babcock and Vallesi 2015; Yudes, Macizo, and Bajo 2011a), or handle subcomponents of attention (Morales et al. 2015), to name but a few areas of research. The main contribution of executive tasks to TIS lies in their capacity to focally assess very specific sub-functions which are conflated with, but different from, linguistic operations. At the same time, they offer the possibility of statistically modeling the relation between those operations and verbal mechanisms (Christoffels, de Groot, and Waldorp 2003). Of note, there are myriads of validated instruments available and many of them possess normative data for multiple sociodemographic groups, which allows for population-based comparisons. However, this line of research also features a number of limitations. The use of artificial, decontextualized tasks raises doubts regarding how the obtained results relate to context-rich, real-life IR processes. Also, many of these paradigms were originally conceived for screening patients with neurological disorders and they are thus prone to ceiling effects in healthy subjects. More generally, although each task sheds light on a particular cognitive domain, the compartmentalized view of cognition they offer proves blind to the seamless interaction that naturally occurs among executive functions across multiple mental processes (Ibáñez and García 2018). For an overview of pertinent research in TIS, see García (2014a).



1.3

Chapter 1.  Mind and brain in the study of translation and interpreting 25

Within the mind, without the brain: Appraising non-neural cognitive approaches

Non-neural approaches have signaled the emergence and growth of cognitive TIS. With a history that spans more than half a century, these trends have yielded several achievements, leaving indelible hints on how to foster sustained progress and prompting critical issues that have not been (or simply cannot be) properly addressed with their tools. By making explicit the main milestones hit so far, how they were reached, and which questions remain unanswered, we should be able to pinpoint and appreciate the particularities of brain-based research as a most valuable extension for the field. As is the case with their counterparts in any other discipline, each of the abovementioned approaches is limited and imperfect. Yet, their successive and cumulative contributions have allowed cognitive TIS to thrive as a hotspot of insights into IR. Thanks to them, process-oriented studies can now build upon a set of solid attainments. Non-neural approaches have endowed cognitive TIS with a set of fundamental constructs to guide ongoing and future investigations. Perhaps the most basic one is the identification and validation of three overarching macro-phases, consisting in ST processing, cross-linguistic (or non-linguistic) processing, and TT processing (for details, see Chapter 5, Section 5.1). Of course, each of these has been variously defined and dissected depending on both the theoretical assumptions and methodological possibilities of each framework; yet, beyond such different guises, this tripartite conceptualization has afforded a workable epistemic structure to organize and interpret research. The same is true of other issues, such as the distinction between form-level and conceptually-mediated operations (with early precursors in the Théorie du sens and empirical formalizations in the psycholinguistic tradition), the recognition of BT and FT as cognitively distinct tasks (an observation that was mostly absent in models from the 60s, 70s, and 80s), the identification of distinct verbal and non-verbal processes (including particular sublexical, lexical, syntactic, semantic, and pragmatic mechanisms, and diverse functions related to attention, monitoring, inhibition, mental-set shifting, and WM, respectively), the acknowledgment of specific translation units as key determinants of cognitive load (a breakthrough mainly attributable to quantitative approaches), and the establishment of field-specific expertise as a general modulator of fine- and coarse-grained mental operations (ranging from lexical access to attentional allocation to strategy deployment). Importantly, too, the field has grown mature enough to abandon notions that proved too restrictive or implausible. For example, the macro-phase of cross-linguistic (or non-linguistic) processing is no longer framed in universalist and purportedly

26 The Neurocognition of Translation and Interpreting

algorithmic terms, as Nida and Taber (1969) once aspired. Instead, its operations are now understood by reference to more context-sensitive, embodied, cognitively plausible frameworks (Muñoz Martín 2016b). Also, psycholinguistic discoveries have shown that, during that macro-phase, form-level and conceptually-mediated processes are not mutually exclusive, but rather simultaneous, interactive, and driven by task-sensitive dominance (García 2015a). Likewise, owing to quantitative approaches, exclusively serial conceptions of the macro-phases have been replaced by more dynamic and realistic accounts which recognize that ST comprehension and TT production may occur either successively or in an overlapping fashion (Carl and Kay 2011; Dragsted 2010; Jensen 2008; Ruiz et al. 2008).8 These and numerous other breakthroughs have given rise to a plethora of models, each of which emphasizes specific aspects and neglects others. Without any claims to exhaustiveness, the list includes generative, functionalist, pragmatist, connectionist, computational, effort-based, cognitive-behavioral, and semantically-oriented accounts – for critical reviews of the main models representing these categories, see Hurtado Albir (2001), García (2011, 2015a), Ahrens (2017), and Carl and Schaeffer (2017). The result is a profuse collection of theoretical frameworks, marked by myriad forms of complementarity and incompatibility, variously oscillating between empirical and speculative foundations, and ranging from the exceedingly general to the exceedingly narrow – all of them being guided (if not biased) by ultimately axiomatic premises. Yet, the operative word in the previous sentence is profuse. A young field tasked with the characterization of phenomena as complex as those inherent in IR could hardly be expected to progress if its conceptual pool were restricted to only a few, necessarily partial options. In this sense, theoretical plurality is yet another merit of non-neural approaches. A further asset of such frameworks is that they created new demands on the editorial arena, leading to the consolidation of a vast specific literature. Continuous advances in cognitive TIS have achieved massive circulation through books (e.g., Alves 2003; Carl, Bangalore, and Schaeffer 2016; Ferreira and Schwieter 2015; Muñoz Martín 2016a; Schwieter and Ferreira 2017; Shreve and Angelone 2010), well-established journals (e.g., Perspectives, Target, Translation and Interpreting Studies, Meta, Translation Spaces), and even newer journals exclusively devoted to the area (e.g., Translation, Cognition & Behavior), not to mention dozens of conferences, workshops, and bootcamps held regularly across the globe. As a result, cognitive TIS has entered a virtuous circle of growth, to the extent that it now constitutes an autopoietic academic space. 8. The reader might wish to refer to relevant conceptual dyads, such as vertical and horizontal translation (Ruiz et al. 2008), sequential and integrated coordination (Dragsted 2010), serial and parallel production (Jensen 2008), or alternating and divided fixation units (Carl and Kay 2011).



Chapter 1.  Mind and brain in the study of translation and interpreting 27

From a broader perspective, these accomplishments have led to the expansion of TIS at large. Nowadays, in addition to philosophical, linguistic, textual, sociological, post-colonial, interactionist, pedagogical, and institutional frameworks, scholars interested in translation and interpreting can opt to analyze their objects of study from a cognitive stance. Moreover, this possibility has opened highly enriching points of contact with long-established fields which were once presumed to be irrelevant for the study of IR. As it were, extending a specific room of a house necessarily extends the house as a whole. An added value of these milestones is that they are instructive: we can deconstruct the initiatives, maneuvers, and principles behind their crystallization and take them as lessons to implement similar lines of development. This is a vital point, as the present book is intended to represent an additive, non-rupturing extension of current approaches to cognitive TIS. Crucially, as we shall see next, the incorporation of brain-based studies is rooted in the same tacit procedures identifiable across the productive history of non-neural approaches. Generally speaking, cognitive TIS has grown due to successive acknowledgments of the virtues and gaps in its state-of-the-art, decade after decade. With a few exceptions, researchers in the abovementioned approaches have been neither oblivious to the achievements nor blind to the limitations of their predecessors. Valuable hypotheses, constructs, and methods have been resurfaced or updated throughout the years, so that the field could proceed cumulatively and expansively rather than reinvent itself anew with each novel trend. For example, the proposal of keylogging as an alternative to TAPs (Jakobsen 1998, 2003) helped to better identify the types of question that could and could not be validly addressed through the latter approach, with three main consequences: a number of previous conclusions were challenged, others were highlighted as plausible, and several new research avenues emerged that had not been systematically (if at all) explored thus far. Importantly, this process required sustained academic update on the part of scholars, who constantly faced the need to become acquainted with new concepts, data types, and analytical viewpoints. The epistemological widening just mentioned fostered progress in two major ways. First, new perspectives were forged to examine long-standing questions, with apparently satisfying answers often requiring several years to appear. At the same time, novel queries were brought to the scene, sometimes driven by specificities of the methods involved, sometimes motivated by conceptual apparatuses imported from other disciplines. The combination of recapitulated and sui generis research topics has consistently inspired new ways of reflecting on, and learning about, IR. This implicit ethos was nurtured by unprejudiced methodological openness. Time and time again, forward-thinking scholars in cognitive TIS demonstrated that continual expansions of their empirico-theoretical toolkits were indispensable

28

The Neurocognition of Translation and Interpreting

to increase understanding of their complex object(s) of study. As seen throughout Section 1.2, this was mainly done by incorporating relevant techniques from other areas and adapting them to field-internal needs (consider, for example, the origins of TAPs, psycholinguistic paradigms, keylogging and eye-tracking studies, or executive-function research). An ever-greater repertoire of indicators, traces, and markers was thus progressively consolidated, augmenting the number of ways in which hypotheses could be tested. The lesson, here, is that the adoption of new methods has always constituted a cornerstone in the evolution of cognitive TIS. The above tendency has naturally called for dynamical interactions between researchers operating within and outside TIS proper. Scholars trained in classical translation and interpreting spheres soon became aware of the need to study works from external areas (such as psychology and psycholinguistics) and to establish collaborations with experts in varied specialties (including statistics, programming, and research design, let alone specific branches of cognitive science). The result has been an invigorating exchange of knowledge, untainted by dualist or narcissist prejudgments. Put succinctly, another key element behind the progress of cognitive TIS is that it was never constrained by disciplinary chauvinism. Finally, the most successful approaches are those that have made explicit not only their foundations, assumptions, procedures, and contributions, but also their problems and obstacles, thus recognizing the provisional and partial nature of their postulations. Therein lies the difference between a dogmatic school and a research program: whereas the former rests on zealous followers who are basically restricted to repeating untestable dictums, the latter can thrive on critical thinkers capable of empirically challenging previous beliefs and inaugurating productive lines of inquiry.9 By embracing a full-disclosure philosophy alongside basic scientific practices, emergent approaches can aspire to become forward-looking, debate-friendly, international research arenas, as opposed to self-asserting platforms rooted in a priori maxims. Taken together, these lessons constitute guiding principles for cognitive TIS to broaden its horizons, which is certainly desirable if not downright necessary at this juncture. As a matter of fact, for all the progress made so far, numerous core issues remain underexplored or directly unexaminable in the scope of non-neural approaches. Consider the following outstanding questions:

9. The difference between one and the other can be readily ascertained. To this end, it usually suffices to access a foundational publication from a given approach and compare it with a recent one: if no substantial changes are found in the methodological and theoretical formulations, and no significant novel findings are introduced in the latter, the reader is probably facing little more than an enduring dogma.



Chapter 1.  Mind and brain in the study of translation and interpreting 29

A. Which functional systems can be identified as independent from one another within the overall architecture of the translating and interpreting mind? For example, are there single, all-purpose systems for BT and FT, for form-level and conceptually-mediated IR processes, or for translating different unit types? Or do such operations recruit specific, partially autonomous mechanisms? B. How critically do those systems participate in different forms of IR? Does their level of activation necessarily correlate with differences in outward performance? C. What are the inner temporal particularities of IR? In other words, how do critical cognitive processes unfold before the onset of a production gesture (e.g., finger movement to press a key in translation, vocal articulation in interpreting)? D. What types of interaction take place between (potentially discernible) cognitive systems during IR? Which mechanisms cooperate and which ones do not when translating different units, or when doing so in backward or forward direction? E. How does sustained practice of certain modalities modulate fine-grained cognitive domains? And how soon after the onset of field-specific training does a student’s mind begin to manifest significant changes? F. And, more generally, can cognitive TIS enter in fruitful, reciprocally informative dialogue with the natural sciences? Fortunately, all these issues can be examined through a brain-based approach, which, it must be noted, entails a commitment to all the virtuous principles supporting the growth of the field so far. In fact, a neurocognitive perspective can profit from the advances of previous approaches and bridge some of their gaps, illuminating both long-standing and original questions through the appropriation of external methods, forging collaborations across multiple academic specialties, and laying the foundations for an explicit, dialogical, non-dogmatic research program. In this sense, neuroscientific investigations into IR should not be conceived as a clean-slate re-invention of cognitive TIS; rather, with their strong and weak points, they represent yet one more source of insights in the cumulative, co-constructive tradition that has enabled the sustained expansion of this academic space. Also, as will be shown below, neither should brain-based research on IR be seen as a caprice or wishful thinking: there are solid reasons to justify the relevance of neurological data when exploring mental phenomena, and a host of concrete findings have already been produced in dozens of studies focused on translation and interpreting. To verify such statements, let us finally venture into the neurocognitive basis of these captivating forms of cross-linguistic processing.

30

The Neurocognition of Translation and Interpreting

1.4

Not black, not a box: Enter the brain

Echoing a popular metaphor often favored by behaviorists, Gideon Toury once wrote that […] translating processes, i.e. those series of operations whereby actual translations are derived from actual source texts […] are a kind of ‘black box’ whose internal (Toury 1985: 18) structure can only be guessed, or tentatively reconstructed.

At the time these words were written, the sources of evidence employed in cognitive TIS (extrapolations of formal linguistic models, non-participant observation, TAPs, psycholinguistic paradigms) were perhaps too far removed from the most intimate operations of the mind, both in time and space. Indeed, the measurable, analyzable indices of at least some non-neural approaches (e.g., utterances based on introspection, keyboard actions) manifest considerably after their underlying cognitive processes have been initiated and deployed. Moreover, their physical triggers (e.g., fingers in psycholinguistic paradigms and keylogging studies, eyes in eye-tracking experiments) are not necessary for a mind to exist: in fact, translation and interpreting can be successfully performed by mute persons, hand amputees or blind individuals. The black-box trope, as resurfaced by Toury, could be reasonably upheld against this background. However, the situation is dramatically different when one considers neurological factors. To put it bluntly, translation and interpreting are simply impossible without a brain. In fact, this organ might well be the only source of evidence in whose absence there would be no mind at all. By studying cerebral phenomena, then, we can examine the very structures and signals which critically support cognition (Damasio 1994; Kandel 2006), partially circumventing some of the constraints immanent in the black-box metaphor. In short, brain-based data grant us access to the topological and temporal coordinates of the biological events without which IR would not even exist.10 Of course, this does not mean that translation and interpreting are fully reducible to neural processes, nor that extra-cerebral indices are irrelevant. On the contrary, both ontologically and methodologically, the brain is but one of the multidimensional ingredients that converge for the emergence and exploration of the inner workings of IR. Still, the point is that there is a necessary bidirectional connection between all such experiences and neural phenomena. Therefore, the term 10. To conceive of a mind without a brain one would need to invoke some sort of artificial intelligence (but surely scholars in cognitive TIS are not aiming to explain how computers work). On the other hand, to conceive of a brain without a mind, one is left with few options other than to think of a corpse (and surely neither is the goal of those scholars to forge a career in a thanatology).



Chapter 1.  Mind and brain in the study of translation and interpreting 31

‘neurocognition’, as used in the title of this volume, is an ontological tautology but a disciplinary necessity, given that several approaches allow exploring the mind in the absence of neurological evidence. Albeit vaguely, the fundamental links between IR and neural tissue have long been acknowledged by translation scholars from diverse orientations, even those that did not integrate neuroscientific research into their works. For instance, in Stylistique comparée du français et de l’allemand, Malblanc (1965) briefly stated that translation is rooted in cerebral structures. The same notion can be found in the Théorie du sens, whose main figures have cited neuropsychological works (e.g., Barbizet 1964; Barbizet and Duizabo 1977) and defined core concepts, such as bagage cognitif, as “a deverbalized whole contained in the brain” (Lederer 1994: 37; my translation). Similarly, in her widely read Traducción y Traductología, Hurtado Albir (2001) acknowledges brain research as a source to understand cognitive aspects of translation, even though that empirical corpus falls outside the scope of her textbook. Much more emphatically, Tymoczko (2005: 1092–93) first maintained that “[b]iologists interested in language, language acquisition, and bilingualism will become central players in translation studies,” as part of “research teams […] that bring together translation scholars, cognitive scientists, literacy and language experts, and neurophysiologists,” and then predicted that “[s]ome of the most exciting advances in translation studies in the near future will result from its intersections with neuroscience” (Tymoczko 2012: 98). In the realm of interpreting, Pöchhacker (2004: 79) also referred to the brain as an essential dimension of research, stating that “the material substrate of mental processes can be targeted with models of cerebral organization and brain activity at the most fundamental, neural level of inquiry.” But beyond these general enunciations, what exactly is to be gained by including neuroscience in the agenda of cognitive TIS? What specific and distinct contributions can result therefrom? The answer to these questions, and to those listed in the previous section, is manifold. Let it be stated, at the outset, that mental processes are non-randomly related to identifiable neurological phenomena (LeDoux 2002). For example, focal cerebral lesions can trigger specific and replicable cognitive, affective, and motor dysfunctions. Also, brain genetics contributes to delineating psychological and behavioral traits. Furthermore, neurochemical modulations, be they natural or otherwise, can induce, alter, or attenuate particular mental states. By the same token, multiple cognitive tasks yield systematic hemodynamic and electrophysiological patterns across subjects, in proportion to ongoing demands and with predictable variations linked to individual characteristics, such as field-specific expertise. Because of these and other reasons, cognitive and neural processes are literally consubstantiated. Nobel Prize laureate Eric Kandel (2006: xii) maintains that

32

The Neurocognition of Translation and Interpreting

“mind and brain are inseparable” and that “mind is a set of operations carried out by the brain, much as walking is a set of operations carried out by the legs, except dramatically more complex.” Similarly, in his groundbreaking Descartes’ Error, Antonio Damasio (1994: 90) states that “having a mind means that an organism forms neural representations which can become images, be manipulated in a process called thought, and eventually influence behavior by helping predict the future, plan accordingly, and choose the next action.” More suggestively still, leading neurolinguist Friedemann Pulvermüller (2002: 9) posits that [t]he brain machinery is not just one arbitrary way of implementing the processes it realizes, as, for example, any hardware computer configuration can realize almost any computer program or piece of software. The claim is that, instead, the hardware reveals aspects of the program […] In other words, it may be that the neuronal structures themselves teach us about aspects of the computational processes that are laid down in these structures.

Insofar as these positions are accurate, studying neurological systems and processes in relation to particular tasks means studying critical aspects of the very mental operations involved. This is no less true for IR than for any other human cognitive function. More particularly, brain-based research can contribute to TIS in ways that largely escape the possibilities of other approaches. Let us consider some of them (for details, see Chapter 2). First, studies on brain-lesioned bilinguals shed light on the functional organization of critical mechanisms – i.e., the existence of partially autonomous but interconnected systems, each critically responsible for a specific set of cognitive operations. Particularly, the establishment of single and double dissociations can reveal that certain skills remain preserved even when others become impaired (Dunn and Kirsner 2003). Complementary evidence can be obtained through online recordings of task-related neural activity (see Chapter 3). Such observations offer important constraints for models seeking to identify distinct processing components involved in IR (e.g., García 2012a, 2015b). Second, electrophysiological methods are unique in that they can reveal the inner time course of relevant processes (Luck and Kappenman 2012). Unlike response-time measures, which tap into the overall duration of a mental operation, techniques such as ERPs allow establishing that two experimental conditions (e.g., BT and FT) involve different temporal dynamics before an external response occurs, even if behavioral outcomes are similar or absent in both cases (see Chapter 2). In this sense, neuroscience seems better equipped than any other field to illuminate covert cognitive effects. Third, functional connectivity studies (Friston 2011) can show how different hubs in distinguishable functional networks interact during IR and its intervening



Chapter 1.  Mind and brain in the study of translation and interpreting 33

processes. Specifically, functional networks are established whenever two or more regions operate in a mutually dependent fashion through the coordinated activity of their respective neurons. The coupling and decoupling of multiple neural systems can thus be objectively studied, offering a more integrative view of target phenomena. Importantly, this method allows partially overcoming the focus on isolated mechanisms that typifies most approaches within cognitive TIS. Fourth, various neuroscientific methods can reveal key aspects of field-specific competence and expertise (Bilalić 2017). Several techniques rooted in other traditions can demonstrate that professional translators and interpreters outperform students and other bilinguals in the quality of their TTs or the speed with which they produce them. However, when no such differences are observed, it may still be the case that the underlying processes were more efficient or distinguishably modulated in the more skilled subjects. By revealing structural or functional brain differences between subjects with and without field-specific training, hemodynamic and electrophysiological methods are uniquely apt to reveal otherwise undetectable signatures of translation and interpreting competence and/or expertise. Fifth, brain stimulation studies (Parkin, Ekhtiari, and Walsh 2015) rank among the best extant alternatives to obtain causal evidence in cognitive TIS. Tools such as tDCS allow boosting or inhibiting specific functions through controlled manipulations of electrochemical patterns in predefined neural sites. Ensuing data can thus motivate conclusions that go beyond the associative patterns revealed by most other techniques – which, nonetheless, are often wrapped in an appealing but unjustified causal jargon. In light of these considerations, brain-based evidence proves highly useful for testing (and thus supporting, falsifying or extending) cognitive models and hypotheses in TIS. Indeed, among other things, available findings already suffice to corroborate a long-standing distinction between form-level and conceptually-mediated routes for IR (see Chapter 4), challenge particular constructs related to directionality (see Chapter 5) and translation units (see Chapter 6), and elaborate on recent relevant proposals, such as the so-called “interpreter advantage hypothesis” (see Chapter 7). Furthermore, all such extant and potential contributions could inspire pedagogical and practical innovations. Indeed, neuroscientific breakthroughs have already been used to inform the teaching of reading (Dehaene 2009) and the design of interventions for enhancing discourse-processing skills (Trevisan et al. 2017). In the long run, the consolidation of a neurocognitive trend within TIS could also give rise to appliable11 approaches for translation and interpreting. 11. The term ‘appliable’, as coined by Michael Halliday in the field of linguistics (Halliday 2007 [2002]; Mahboob and Knight 2010), refers to theoretical formulations which can be used to solve problems of particular communities through a combination of reflection and action.

34

The Neurocognition of Translation and Interpreting

More generally, a brain-based approach to TIS can expand the field’s interdisciplinary avenues by forging a link with the natural sciences (García 2012a). Neuroscience, in particular, has become a fruitful platform to characterize the myriad connections between molecular, genetic, neural, cognitive, emotional, and behavioral aspects of human experience, as attested by several multi-billion-dollar initiatives (e.g., the Blue Brain Project in Sweden, the Human Brain Project in the European Union, or the Human Connectome Project and the BRAIN Initiative in the United States). The novel knowledge that is thus being generated can be profitably channeled into a new productive space for translation and interpreting scholars. Now, perhaps the most fascinating aspect of the above paragraphs is that they are not merely rooted in hope or promise. In fact, they are firmly supported by decades of research that, for one reason or another, has passed unnoticed for most investigators in TIS. What follows is a broad overview of the main milestones that have already been reached in the (for many, surprisingly long) history of brain-based research on IR. 1.5

Historicizing brain-based research on IR

In 2012, Maria Tymoczko published a position paper underscoring the need for translation studies to integrate neuroscientific findings. In her own words, neuroscience […] is an area that should be tracked in the field of translation studies as a whole, if only because the neuroscience of translation is one of the most (Tymoczko 2012: 98) important known unknowns of the discipline.

While the present book stands in full alignment with the above recommendation, it does not endorse its justification. To claim that the neuroscience of translation constitutes a “known unknown” means that there is a particular gap in knowledge of which the TIS community is fully aware. However, the situation seems to be diametrically different: insights into the neural bases of IR have been accruing for decades, without the TIS community ever taking heed of them. Therefore, a syntactic rearrangement seems in order. Rather than a “known unknown,” we are in the presence of multiple “unknown knowns” (García and Muñoz forthcoming). In fact, empirical evidence on the translating brain began to surface in the late 1920s and it has been growing considerably since the turn of the century. This evolving history, spanning roughly 90 years, encompasses several neuropsychological, neuroscientific, and theoretical milestones, all illustrated in Figure 1.1 (Muñoz, Calvo, and García 2018). As proposed elsewhere (García and Muñoz forthcoming), relevant works can be periodized in three loosely defined stages.

Chapter 1.  Mind and brain in the study of translation and interpreting 35



First reports of paradoxical translation behavior

1982

First special issue on the topic

1989

First compendium of translation disorders

1984

1990

First studies on lateralization

1994

First flowchart model of SI

1980

First FC study 1970

1960

1951 First report of inability to translate

1950

1990

2000

2010

1995

First PET study

1997

First position paper

1999

First anatomical model

2002

First fNIRS study

2004

First ERP study

2005

First fMRI study

2010

First NIBS study

2012

First DCS study First review focused on SIs First book on the topic

1940 First report of translation without comprehension 1931 1929

2020

2013

First review focused on translation

2014

First structural imaging study

2016

First iEEG study

1930

First report of compulsive translation

Figure 1.1  Milestones in the research of neurocognitive aspects of translation and interpreting. The timeline depicts only the first occurrence of each empirical or theoretical development in the literature. Neuropsychological milestones are indicated in blue; neuroscientific milestones are signaled in red; theoretical milestones are marked in green. DCS: direct cortical stimulation; ERP: event-related potentials; FC: functional connectivity; fMRI: functional magnetic resonance imaging; fNIRS: functional near-infrared spectroscopy; iEEG intracranial electroencephalography; NIBS: non-invasive brain stimulation; PET: positron emission tomography; SI: simultaneous interpreting; SIs: simultaneous interpreters. Reproduced with permission from Muñoz, Calvo, and García (2018).

36

The Neurocognition of Translation and Interpreting

1.5.1 Milestones from the mid-twentieth century The initial stage, cutting throughout the first part of the mid-twentieth century, was typified by the sporadic publication of seminal neuropsychological studies. A series of single-case reports on bilingual and multilingual aphasics offered the first explicit links between (disturbed) translation skills and (damaged) brain regions. The very first paper dates from 1929, when a polyglot was observed to compulsively translate utterances following lesions to left posterior areas (Kauders 1929). The same tendency was observed in other studies (Stengel and Zelmanowitz 1933; Veyrac 1931; Weisenberg and McBride 1935). Notably, in one of them (Veyrac 1931), the patient was also noted to correctly translate expressions despite not being able to grasp their meanings. By the end of this period, an unprecedented disorder was also described in which a patient with left-sided damage was incapable of translating words in either direction despite having preserved word production skills in each language separately (Gastaldi 1951) – for a full treatment, see Chapter 4. Clearly, the evidence in these studies was somewhat weak. Only a few reports described the patients’ translation abilities  – which could mean that translation-specific disorders were rare or, most probably, that they were neglected by the examiners. Also, whenever such disorders were characterized, it was only in an anecdotal style and by reference to very few examples mainly highlighted as curiosities. Data on translation performance were either produced spontaneously by the patients or elicited via ad hoc assessments lacking in experimental control. Moreover, the accompanying descriptions of the patients’ brain lesions and neuropsychological profiles were vague and unsystematic, which further limited the establishment of robust conclusions. Needless to say, the discussion of findings was completely uninformed by adequate theoretical views of translating processes, as these were simply unavailable at the time. However, the contributions of these pioneering works must not be underestimated. They afforded foundational empirical hints to understand the relation between particular brain regions and IR processes, even before TIS (let alone cognitive TIS) had asserted itself as an autonomous scholarly arena. In particular, they may well represent the first scientific indications of the relation between verbal and non-verbal mechanisms during translation – a topic that has been resurfaced in contemporary research (e.g., Christoffels, de Groot, and Kroll 2006; Christoffels, de Groot, and Waldorp 2003; Morales et al. 2015). Also, these studies inspired more detailed examinations of normal and pathological translation processes. In fact, they were noted in groundbreaking works (Paradis 1977) paving the way for the development of formal tests of translation skills in bilingual aphasics (Paradis 1979). Finally, their findings have been incorporated as constraints for the formulation of neuroscientific hypotheses and models of translation mechanisms (García 2012a, 2015a). In sum, the first seeds of the field were planted approximately 90 years ago.



Chapter 1.  Mind and brain in the study of translation and interpreting 37

1.5.2 Milestones from the late twentieth century The second stage, comprising the last quarter of the twentieth century, was marked by significant breakthroughs anchored in methodological and theoretical advances. The neuropsychological corpus inaugurated in previous decades was considerably broadened. Importantly, the release of the Bilingual Aphasia Test (Paradis 1979) allowed for more systematic evaluations of translation performance in brain-lesioned patients. In particular, Part C of this instrument consists of tasks tapping on cross-linguistic equivalent recognition, word translation, and sentence translation, all in both backward and forward direction and in multiple language pairs (Paradis 2011).12 Moreover, throughout this period, even relevant clinical reports which did not use such an instrument offered more meticulous assessments of the patients’ translation skills, including tests for specific word types – e.g., abstract vs. concrete nouns (Nilipour and Ashayeri 1989) – and IR modalities – e.g., oral vs. written translation (Aglioti and Fabbro 1993). Also, the inclusion of relevant tasks led to the discovery of a hitherto unknown translation disorder following posterior brain insult (Paradis, Goldblum, and Abidi 1982). At the same time, relevant neurolinguistic experiments began to be conducted with healthy participants. The first ones relied on behavioral techniques. For example, dichotic listening paradigms (see Chapter 2) were used to assess brain lateralization in aspiring and professional SIs (Fabbro et al. 1990; Fabbro, Gran, and Gran 1991), whereas results from psycholinguistic experiments motivated inferences on the connection strength between different types of cross-linguistic equivalents (de Groot, Dannenburg, and van Hell 1994; Kroll and Stewart 1994; van Hell and de Groot 1998b). Neuroscientific tools entered the scene in the mid-1990s. Profiting from what was an innovative approach at the time, Kurz (1994, 1995) used EEG to examine functional connectivity across both hemispheres during simultaneous interpreting. For their part, neuroscientists like Klein et al. (1995) and Price, Green, and von Studnitz (1999) obtained the first images of hemodynamic brain patterns during word translation. A biologically-oriented perspective on IR could thus begin to be forged on bases other than speculation. Indeed, in 1984, Meta devoted a special issue to the links between translation and the brain. The volume, entitled Cerveau, langage et traduction, featured an overview of translation disorders in aphasia (Paradis 1984) and preliminary reflections on this now approachable connection (Bouton 1984). Relevant theoretical accounts also began to appear in the literature, as seen in Fabbro and Gran’s (1997) position article, focused on simultaneous interpreting. What is more, explicit models of critical mechanisms of IR were built on neurocognitive evidence. For example, 12. The multiple versions of the Bilingual Aphasia Test can be downloaded freely from the following website: https://www.mcgill.ca/linguistics/research/bat.

38

The Neurocognition of Translation and Interpreting

Paradis (1994) posited a neurolinguistic model of inter-equivalent connections, while Fabbro (1999) set forth a neuroanatomical model of the systems mediating simultaneous interpreting (see Chapter 4, Section 4.4.2). In sum, this period witnessed an expansion of the toolkit available to look into the translating and interpreting brain. Before the turn of the century, it was clear that IR could be examined in terms of its structural and functional neural correlates, and that joint efforts from TIS scholars and neuroscientists could reveal valuable aspects of the phenomenon. Thanks to these developments, a new area of research became ready to bloom in the twenty-first century. 1.5.3 Milestones from the twenty-first century The third, ongoing stage began in the 2000s. We are witnessing what might be called an ‘expansion era’ for brain-based research on IR, with developments in several directions. Novel neuropsychological works have been published, including new single-case studies (e.g., Fabbro, Skrap, and Aglioti 2000; García-Caballero et al. 2007) and comprehensive reviews of the literature (e.g., García 2015b). Also, behavioral investigations on the processing of cross-language equivalents have grown dramatically and become of hotspot for bilingualism research (García 2015a). Among other things, ensuing findings have provided critical data to characterize cognitive effects related to directionality (see Chapter 5) and various translation units (see Chapter 6). In parallel, extending the work of Bajo, Padilla, and Padilla (2000), multiple authors have conducted assessments of executive functions in SIs (García 2014a). A flourishing subfield of research has thus been established, with implications that go even beyond the interests of TIS (see Chapter 7). Yet, the most notable developments correspond to neuroscientific research proper. Real-time measures of brain activity during IR have been obtained through varied techniques, including PET (e.g., Rinne et al. 2000), fNIRS (e.g., Quaresima et al. 2002), ERPs (e.g., Proverbio, Leoni, and Zani 2004), fMRI (e.g., Lehtonen et al. 2005), and intracranial recordings (García, Mikulan, and Ibáñez 2016). Also, explorations of the causal role of particular brain regions in the translation of specific units have been performed via tDCS, in healthy subjects (e.g., Liuzzi et. al. 2010), and direct cortical stimulation, in brain-damaged patients (Borius et al. 2012). Moreover, several experiments have looked at how interpreting practice impacts brain structure and function (e.g., Hervais-Adelman, Moser-Mercer, and Golestani 2015; Hervais-Adelman et al. 2017). This empirical corpus has become sufficiently large to invite systematic reviews. Such contributions have begun to yield coherent accounts of the neurocognitive correlates of translation directionality, the mechanisms engaged by different translation units, and the anatomo-functional impact of interpreting expertise (Elmer



Chapter 1.  Mind and brain in the study of translation and interpreting 39

2012; García 2013a). Furthermore, a full book has been devoted to erecting a neuroanatomical model of putative translation routes (García 2012a). Considering this overview, then, if the neuroscience of IR remains a mystery for some scholars in TIS it is less due to a lack of evidence than of awareness. The following chapters are a deliberate attempt to cover such a gap. 1.6

A role for neuroscience in contemporary TIS

A further milestone of the ongoing ‘expansion era’ deserves separate treatment. Looking beyond the paralyzing black-box trope and the “known unknown” fallacy, several scholars operating within TIS have taken an active interest in brain-based research. The prospects of embracing a neurocognitive approach to study translation and interpreting are now becoming clearer to an ever-larger portion of the community. Year after year, collaborations between researchers in TIS and neuroscience are less of a rarity. The interdisciplinary innovations pioneered by Laura Gran and Franco Fabbro at the University of Trieste (Fabbro et al. 1990; Fabbro, Gran, and Gran 1991) may have once been perceived as an eccentricity, as was probably the case with those pursued by Ingrid Kurz and Hellmuth Petsche at the University of Vienna (Kurz 1994, 1995), and by Jorma Tommola and Juha Rinne at the University of Turku (Rinne et al. 2000; Tommola et al. 2001). However, similar interactions are now proving more and more common in institutions across the globe. These include the University of Zurich, in Switzerland (e.g., Elmer, Meyer, and Jäncke 2010; Elmer et al. 2014); the University of Geneva, also in Switzerland (e.g., Hervais-Adelman et al. 2014, 2017); the Laboratory of Experimental Psychology and Neuroscience, at the Institute of Cognitive and Translational Neuroscience, in Argentina (e.g., García 2013a, 2015b; García, Mikulan, and Ibáñez 2016); Durham University, in the United Kingdom (e.g., Zheng et al. forthcoming); and Federal University of Minas Gerais, in Brazil – the latter, in collaboration with the Instituto do Cérebro at Pontifícia Universidade Católica do Rio Grande do Sul (e.g., Alves, Szpak, and Buchweitz 2018). Moreover, new laboratories have been set up which focus primarily on neurocognitive approaches to TIS – consider, for example, the Translation & Cognition Center, at Johannes Gutenberg University Mainz, in Germany (e.g., Hansen-Schirra 2017); and the Centre for Studies of Translation, Interpreting and Cognition, at the University of Macau, in China (e.g., Lu and Yuan 2018). Also, neurocognitive considerations have begun to appear in works from well-established translation scholars, such as Diamond and Shreve (2010), Moser-Mercer (2010), Tymoczko (2012), and Ahrens (2017). This tendency has actually been acknowledged in the specialized literature. For example, Muñoz Martín

40 The Neurocognition of Translation and Interpreting

(2016b: 16) has affirmed that “the community is rooting the cognitive aspects of translating and interpreting in the brain.” The very inclusion of the present book in the Benjamins Translation Library is a most compelling corroboration of that claim. I do not believe, however, that we are in the presence of a radical change of direction. TIS has long fallen prey to the craze of advancing alleged “paradigm shifts” or “new turns” (Snell-Hornby 2006). However, most of such proposed revolutions have amounted to little more than fads, usually framed by catchy metaphors but devoid of the epistemological, empirical, and institutional reconfigurations of a true paradigmatic change (Kuhn 1962). Let us then be explicit about it: the consolidation of a brain-based approach to cognitive TIS does not entail a major break from current tenets, notions, and lines of work. On the contrary, and at the risk of reiteration, it constitutes an addition, an extension, a complement to the various non-neural approaches described earlier. The sooner we recognize the futility of those self-indulgent pronouncements, the greater our chances of reaching the plural, multidimensional framework needed to understand the mental facets of IR. All in all, neurally-oriented research represents yet another valuable set of resources to infer properties of the invisible phenomena targeted by cognitive TIS. The remainder of the book dissects this approach from a variety of angles. A preliminary step consists in characterizing its tools and their rationale. Such is the aim of the following chapter.

Chapter 2

The toolkit

2.1

Beyond MacGyver’s knife

“Well, when it comes down to me against a situation, I don’t like the situation to win.” The quote belongs to MacGyver, the protagonist of the eponymous American TV series first aired in the 1980s. Episode after episode, viewers were amazed by their hero’s resourcefulness to sort out countless complications as he fought crime and injustice. On any given day, the inventive secret agent could find himself picking a lock, disabling a deathtrap, hot-wiring an elevator, creating a smoke cloud to distract dangerous goons, saving a friend from poisoning, or escaping a gas chamber. Now, the most remarkable thing is that, to perform all these feats, he required little else than a Swiss Army knife. In the hands of MacGyver, one tool sufficed to solve virtually any problem. In that aspect, the show stood in stark contrast with the workings of science. Unfortunately, no single instrument offers viable solutions to all the issues worth addressing in a scholarly discipline. Rather, across fields and specialties, scientific progress is tied to the expansion of the available toolkit. Medicine is instructive in this sense. For centuries, infections, venereal diseases, and several other health problems were inadequately treated with multipurpose techniques, such as bloodletting or the application of life-endangering mercury compounds. The scenario was reversed in the twentieth century, thanks to the introduction of penicillin and numerous developments forged thereof. Moreover, the incorporation of new technologies opens new lines of inquiry in academia. By adopting the microscope, for example, several research areas broadened their horizons as diverse types of cells, tissues, and microorganisms could finally be subjected to careful study. As far as science is concerned, then, critical problems can often be properly addressed only by embracing new resources. As seen in Chapter 1, this lesson is not new to cognitive TIS. While the field was first confined to unsystematic ad hoc extrapolations from formal linguistics and impressionistic spectatorial conjectures, TAPs paved the way for studies based on introspective insights from aspiring and professional translators. Then, several limitations of TAPs were overcome with the arrival of keylogging technologies, which, at the same time, made it possible to assess hitherto unexaminable questions. Similarly, as argued elsewhere (García, Mikulan, and Ibáñez 2016), further

42

The Neurocognition of Translation and Interpreting

avenues of progress can be consolidated by integrating techniques from cognitive neuroscience. Building on these principles, the sections below describe the fundamentals of the methods and tools available for brain-based research on IR. Overall, this chapter pursues two goals: first, it seeks to equip readers with key notions to understand and criticize relevant works; second, it aims to raise awareness of crucial considerations for those interested in conducting their own studies. Specifically, an overview is offered of the main types of research designs, tasks, variables, measures, techniques, and interpretive constraints proper to this approach. The result should be one that MacGyver would be pleased with: when it comes down to the readers against a challenging situation in the literature or in their laboratories, the odds should now be in their favor. 2.2

A matter of design

Brain-based research relies on numerous types of experimental design, some of which have proven crucial to investigate IR. So far, the literature consists of single-case, single-group, between-group, and pre/post studies. With their similarities and particularities, these alternatives offer an ample framework to understand the neurocognition of translation and interpreting. 2.2.1 Single-case designs Single-case designs consist in the assessment of one individual or a series of individuals. This is a standard approach in neuropsychological studies, as a particular brain-lesioned patient may offer a valuable model to explore the role of affected and unaffected regions in preserved and spared functions (see Section 2.6). Moreover, on rare occasions, patients with intracranial electrodes implanted for clinical reasons can be individually studied in the quest of spatiotemporally precise links between neural mechanisms and cognitive domains (see Section 2.7.2.2). Typically, the patient is required to perform two or more tasks with a view to revealing which of them proves disrupted given a brain lesion or which physiological patterns they evoke in a region of interest. In either case, a specific relationship can be postulated between the target function and the neural substrate at hand. Although comparisons with other subjects are neither usual nor necessary in these studies, the patient’s performance is sometimes contrasted with that of a healthy individual (Aglioti et al. 1996) or a control group (García et al. 2017c), or even against relevant normative data (e.g., Steeb et al. 2018).



Chapter 2.  The toolkit 43

Single-case designs have been repeatedly used to assess various aspects of IR in bilingual aphasics. For example, numerous reports have documented effects of brain damage on specific translation directions, units, and modalities (see Chapter 4). Also, preliminary findings on brain connectivity differences between BT and FT have been obtained from intracranial recordings during bilingual processing (García, Mikulan, and Ibáñez 2016). Notably, some of the most specific hypotheses concerning the neural organization of IR mechanisms are mainly informed by results from individual neurological patients (García 2015b). The main limitation of single-case designs is their generalizability. Given that no two brains are identical and that patterns of neural damage vary greatly among subjects, even highly significant results in one individual may not be replicated in another. However, the history of neuropsychology abounds in single-case reports whose observations prompted major breakthroughs and motivated robust hypotheses that could be refined in time (Thiebaut de Schotten et al. 2015). Moreover, these studies are crucial to test causal predictions of theoretical models at the subject level (cf. Dubois and Adolphs 2016). For relevant examples focused on IR, see Chapter 4. 2.2.2 Single-group designs Questions about generalizability, which loom large in single-case studies, can be partially addressed via single-group designs, in which several participants perform the same experiment(s) and are treated as a unified sample. Within cognitive TIS, these are widely used in psycholinguistic studies (Ruiz et al. 2008), assessments of executive functions (Shlesinger 2003), and, of course, neuroscientific experiments (Lehtonen et al. 2005). Together with between-group designs, they account for the majority of research in the field. Typically, these designs include different tasks (e.g., BT and L2 reading), or different stimulus types (e.g., BT of cognates and non-cognates), or different conditions (e.g., verbal recall with and without articulatory suppression), so that patterns specific to one of them can be revealed through statistical comparisons. Moreover, additional tests can be performed to assess whether two variables are correlated or whether performance in one task or condition predicts performance in another one. If the sample is large enough (i.e., if it has sufficient statistical power) and if it has been properly selected, then results may be presumed to be representative of a larger population. Abundant brain-based studies have relied on single-group designs to assess translation and interpreting. In particular, they have proven instrumental to examine various questions related to directionality (see Chapter 5), the mechanisms recruited to process diverse translation units (see Chapter 6), and the key substrates

44 The Neurocognition of Translation and Interpreting

of specific IR modalities (see Chapter 7). Note, in passing, that single-group designs in the study of IR have been combined with the most varied neuroscientific techniques, including fMRI, PET, EEG, and fNIRS (García 2013a). These designs are sometimes undermined by poor sampling and analysis procedures – an issue that has prompted ardent debates in the literature on neuroplasticity in SIs (Hervais-Adelman, Moser-Mercer, and Golestani 2018) and in neuroscience at large (Munafò et al. 2017). Moreover, they are often unable to fully rule out alternative explanations of a particular hypothesis, given the impossibility of controlling for all potential lurking variables in the construction of stimulus sets (see Section 2.4). Notwithstanding, when these issues are properly addressed, single-group designs offer a powerful framework to trace robust links between brain and behavior, beyond subject-level observations. In particular, they can offer hints on particular contributions of different brain regions to several co-occurring processes during IR (e.g., Hervais-Adelman et al. 2014). Moreover, they allow assessing how variable certain neurocognitive patterns are within the sample, thus accounting for inter-individual heterogeneity and so-called individual factors. 2.2.3 Between-group designs The main limitation of single-group designs is that they are moot on whether the observed patterns are specific to the target population or general across various ones. If an experiment on interpreting students reveals differences between BT and FT, can one assume the same to be true of professional SIs and lay bilinguals? This is where between-group designs make their greatest contribution, as they can identify differences between two or more samples with specific profiles. Ideally, the groups should differ in one (and only one) critical variable, so that observed differences can be attributed to it. This approach has been used, for example, to detect patterns specific to professional translators (e.g., García et al. 2014) and interpreters (e.g., Bajo, Padilla, and Padilla 2000) relative to subjects with less or no experience in IR. As regards neuroscientific studies, between-group designs have been mainly used to detect biological signatures of expertise in IR via structural and functional neuroimaging, ERPs, and EEG-derived functional connectivity (see Chapter 7). A key problem facing these designs in cognitive TIS is the lack of standardized criteria on how to match groups so that key confounds can be ruled out. Moreover, additional concerns related to sampling have been raised in the literature (Koshkin and Ossadtchi 2017). Despite such issues, between-group designs are critical to link specific neurocognitive patterns to particular levels of IR competence, expertise, and performance. Their growth in recent years is arguably one of the main developments in the field.



Chapter 2.  The toolkit 45

2.2.4 Pre/post designs For all their strengths, between-group designs cannot reveal whether differences between samples are causally attributable to the key variable separating them. For example, if a single assessment reveals increased gray matter volume in professional SIs than in non-interpreter bilinguals, it is possible that such a difference was present even before the former started practicing interpreting. So, is that particularity a result of sustained practice of interpreting, or is it a precondition for successful insertion in the field? This question and its ramifications can be addressed through pre/post designs. Pre/post schemes comprise three parts. First, in a pre-treatment phase, one or more groups are assessed to obtain a baseline measure. Second, the group or groups are subjected to a particular treatment (e.g., a period of training in simultaneous interpreting). Third, in the post-treatment phase, the group or groups are administered the same measures applied at the beginning. If the outcomes between the first and third phases differ, the effect can be causally attributed to the intervention – and if that effect is absent in a control group which did not undergo the intervention, the evidence becomes even stronger (e.g., Trevisan et al. 2017). Despite their overall paucity, pre/post experiments have been growing in the study of effects related to interpreting experience. They have been used, for instance, to examine its impact on lexico-semantic (Bajo, Padilla, and Padilla 2000) and WM (Dong and Lin 2013) skills, and on relevant neuroanatomical (Hervais-Adelman et al. 2017) and neurofunctional (Hervais-Adelman, Moser-Mercer, and Golestani 2015) mechanisms. Moreover, a pre/post design has shown that tDCS over motor brain regions can modulate the translation of specific word types (Liuzzi et al. 2010). Pre/post studies are demanding as they require the sustained commitment of several participants, who must remain enrolled in the study for several days, weeks, or months. Also, depending on the type and length of the treatment phase, uncontrolled situations experienced by the subjects outside the research setting may give rise to important confounds in the results. Be that as it may, the combination of carefully matched tasks or conditions, comparable treated and non-treated samples, and a theoretically relevant intervention represents one of the strongest frameworks to support causal claims and postulate mechanistic relations in cognitive neuroscience.

46 The Neurocognition of Translation and Interpreting

2.3

Mind games: A sampler of experimental paradigms

Depending on its objectives, a study may comprise one or more cognitive tasks. Oftentimes, these are variations of well-established experimental paradigms, that is, data-collection set-ups in which pre-selected stimuli are presented in a particular way so that subjects perform predefined operations on them. Given their processing demands, a number of experimental paradigms are agreed to tap into specific cognitive functions, whose efficacy, efficiency, and level of engagement can be measured by reference to behavioral outcomes (see Section 2.5) or neural correlates (see Section 2.7). Many of them have become fundamental sources of insight into the mechanisms supporting IR, either as the main targets of inquiry or as control conditions for comparison purposes. Broadly speaking, experimental set-ups in the study of IR can be classified into naturalistic and atomistic paradigms, with the latter comprising a host of verbal and non-verbal tasks. Importantly, the results obtained therefrom can certainly be analyzed in conjunction with neuroscientific data, but they can also be profitably interpreted on their own – as done, for example, in psycholinguistic research. 2.3.1 Keeping it real A number of studies in the field have used naturalistic paradigms. These are based on tasks that closely resemble the actual conditions in which certain cognitive processes are deployed in daily settings. For example, real-life stimuli have been used by Borius et al. (2012), who explored neural correlates of sight translation of newspaper excerpts; and by Kurz (1995), who examined functional brain connectivity during simultaneous interpretation of speeches delivered at international conferences. Naturalistic interpreting tasks have also been used in behavioral studies assessing the relationship between WM and TT quality in oral (e.g., Injoque-Ricle et al. 2015) and signed (Macnamara and Conway 2016) modalities. The main attribute of naturalistic tasks is that they reduce artificial manipulations to the minimum, thus meeting the requisite of ecological validity. However, precisely because of the absence of those manipulations, they are blind to the specific contributions of fine-grained mechanisms. For example, they cannot conclusively reveal whether behavioral or neural differences between two conditions were driven by verbal or executive demands or by a specific level of processing. Also, unconstrained responses from the participants can create considerable motor artifacts and thus introduce noise in ongoing brain signals. These problems can be partially overcome via atomistic paradigms.



Chapter 2.  The toolkit 47

2.3.2 Piece by piece Though certainly useful, naturalistic paradigms have not been widely used in the field, arguably because cognitive neuroscience, for better or worse, mainly focuses on the role of particular mechanisms subserving fine-grained sub-functions (e.g., semantic access, cross-linguistic activation, STM capacity). Therefore, the bulk of the field relies on atomistic paradigms: highly controlled tasks involving decontextualized, token-level stimuli, such as single words, individual sentences, or series of shapes. In these settings, irrelevant variables can be matched and thus factored out between two or more contrastive conditions, so that these differ only in one variable of interest (e.g., translation directionality, cognate status, or WM load). Ensuing differences in the behavioral and neurocognitive correlates of each condition can then be reasonably attributed to that manipulated variable (see Section 2.4). In particular, atomistic paradigms shed light on particular verbal and non-verbal operations involved in IR, as illustrated below. 2.3.2.1 Verbal paradigms Verbal paradigms aim to assess specific linguistic mechanisms. In a typical experiment, individual words or sentences are successively presented on a computer screen and participants must execute particular actions in response to them (such as reading them out loud or pressing buttons to classify them based on predefined criteria). While these computerized tasks will be the focus of the present overview, note that others can be performed with paper and pencil (e.g., Yudes et al. 2013). The most straightforward paradigm to study IR consists in word or sentence translation. In either case, participants view the corresponding stimuli in one language and they must utter their equivalents in the pre-established target language. Typically, oral responses are recorded and timed via a voice-activated device (e.g., Christoffels, De Groot, and Waldorp 2003), although some studies have relied on keyboard-button presses, either synchronized with overt production (e.g., García et al. 2014; Santilli et al. 2018) or followed by typing of the target unit (Grabner et al. 2007) – Figure 2.1. The assumption is that this paradigm taps directly into deliberate cross-linguistic operations. Moreover, specific manipulations of the stimuli warrant inferences regarding which cognitive mechanisms (e.g., form-level or conceptually-mediated routes) were mainly recruited during the task (García et al. 2014). Numerous variations of this basic configuration have been used in behavioral (e.g., Kroll and Stewart 1994), neuropsychological (e.g., Aglioti et al. 1996), electrophysiological (e.g., Christoffels, Ganushchak, and Koester 2013), and neuroimaging (e.g., Klein et al. 1995) studies. The data thus gleaned has informed different models and hypotheses (see Chapters 4 through 7).

48 The Neurocognition of Translation and Interpreting

House beep time (ms)

Fixation line 1000 R

Word 2000

House Haus Response (Space key)

3000 A

confirm

next trial

Inter-Trial-Interval 3000 ms

Figure 2.1  Example of a word translation paradigm. In this English-into-German translation task, used to measure EEG correlates of IR, each trial began with a fixation line followed by an acoustic warning signal at 2000 ms. After 3000 ms, the source stimulus was shown and subjects had to press the spacebar if they knew an adequate translation for it. In such a case, an input box appeared below the fixation line, in which subjects had to type the corresponding German word and confirm their response by pressing Enter. The following trial began 3000 ms afterwards. A: activation interval; R: reference interval. Reproduced with permission from Grabner et al. (2007).

The above paradigm has often been accompanied by single-language reading tasks. These normally follow the same layout and timeline as their complementary translation tasks, but participants must simply name each stimulus in the same language in which it appears. In the context of IR research, word- and sentence-reading tasks usually serve as baseline or control conditions. This is useful, for example, to establish the extent to which potential asymmetries between BT and FT reflect IR-specific modulations as opposed to more general differences between L1 and L2 processing. In neuroimaging (e.g., Klein et al. 1995) and functional connectivity (García, Mikulan, and Ibáñez 2016) studies, reading conditions have also been used to subtract single-language activity patterns from those involved during translation – once again, with the aim of revealing neural modulations proper to IR or one of its possible directions. They have also been combined with translation tasks to test for possible dissociations in neuropsychological (e.g., Weekes and Raman 2008) and brain stimulation (Borius et al. 2012) protocols. Another highly useful paradigm consists in equivalent recognition. In the standard version of this task, participants are presented with two words (one in each language) and they must decide whether the second one is an adequate translation of the first. In some trials, the words are indeed viable equivalents (e.g., casa-house, in a Spanish-English experiment), but in others they are not. Crucially, in the latter case, non-equivalents may still be tacitly linked at the phonological/orthographic level. For example, in the pair casa-face, the tacit translation of the second word (cara) has major phonological/orthographic overlap with the first one (casa). Therefore, differential performance on such non-equivalent pairs can be taken to reflect implicit cross-language activation at a sublexical level (e.g., de Groot 1992). Similarly, non-equivalent pairs may be semantically linked (e.g., casa-yard) and thus reflect conceptually-mediated cross-linguistic effects. This paradigm has been



Chapter 2.  The toolkit 49

used mainly to assess the role of form-level and conceptually-mediated routes linking the L1 and the L2 lexical systems (e.g., de Groot 1992; Ferré, Sánchez-Casas, and Guasch 2006), even in combination with ERP recordings (Moldovan et al. 2016). Also, equivalent recognition tasks involving multiple target options have been used in fMRI research (Mayer et al. 2015). The connections between cross-language equivalents can also be tapped via bi­ lingual word-association tasks (e.g., van Hell and de Groot 1998a). Here, participants view series of isolated words and they must produce a related word in each case. In the intra-linguistic condition, lexical associates must belong to the same language as the stimulus; in the inter-linguistic condition, the associate must be given in the other language. The crux of the paradigm consists in comparing the associates given for the same item in each condition. Imagine that, in an English-Spanish experiment, a participant is shown the word book. If his/her responses in the intra-linguistic and inter-linguistic conditions are pencil and lápiz, the pattern will be registered as an equivalent response. By contrast, if the responses were pencil and página, the pattern will constitute a non-equivalent response. Therefore, by manipulating the types of items presented to the participants (e.g., using both cognate and non-cognate words), researchers can infer whether the relational structure of L1 and L2 equivalents is more or less similar in the context of specific lexical categories. Also relevant are lexical decision paradigms, in which participants are shown strings of letters and must decide whether each sequence constitutes a real word or not.13 This canonical approach for examining lexical access mechanisms (Yap et al. 2006) has given rise to variations which illuminate aspects of IR. In particular, in lexical decision experiments with cross-linguistic priming (Kiran and Lebel 2007), a proportion of the target items is preceded by their translation equivalents,14 so that inferences can be drawn regarding the cognitive connections between them (Figure 2.2). This approach can reveal whether cross-linguistic lexical processing is easier for some categories than others, or even for one direction over another (i.e., from L2 to L1 or from L1 to L2).

13. The strings which do not constitute real words are called pseudowords (when they adhere to phonotactic and/or graphotactic restrictions of the language in question) or non-words (when they do not follow such conventions). A typical approach for their construction is to replace only one letter from a real word (so that wig gives rise to wug), although some researchers prefer to combine real syllables in novel ways (so that the first syllable in pencil and the second one in window give rise to pendow). 14. These are sometimes shown for such a brief period that participants are not aware of their presence, which can shed light on unconscious or automatic processes.

50

The Neurocognition of Translation and Interpreting

Target item

Fixation cross

Prime word

+

colchón

500 ms

300–500 ms

matress

Equivalent target word

mountain

Non-equivalent target word

Real word? Yes

marstain

No

Pseudoword

max 1400 ms

Figure 2.2  Example of a lexical decision paradigm with cross-linguistic priming. In this hypothetical task, applicable to Spanish-English or English-Spanish bilinguals, each trial begins with a fixation cross lasting 500 ms, after which a Spanish prime is shown for a random interval between 300 and 500 ms. This is followed by a target item (shown for a maximum of 1400 ms), which could be an English equivalent of the prime, an unrelated English word or a pseudoword. Participants must press one key if the target item is a real word and another one if it is not. The predicted effect is that both types of real word should elicit faster responses than pseudowords, and that translation equivalents should be processed faster than non-equivalents, due to pre-activation of shared semantic features relative to the cross-linguistic prime.

Just as lexical access can be assessed via lexical decision tasks, so too processing of particular meanings can be tapped through semantic decision tasks. In the simplest form of these tests, participants are shown randomized lists of words from different semantic categories and they must press a specific button depending on which category they belong to – e.g., animate vs. inanimate nouns (Menenti 2006) or action vs. non-action verbs (Dalla Volta et al. 2014). The results can illuminate whether those categories differ in their implied cognitive effort and/or underlying neural mechanisms. Of note, certain semantic decision experiments involve pairs of words presented in the same or in different languages, so that participants must decide whether each dyad is semantically related. This framework has proven quite informative for brain-based research on IR, as it has revealed that bilinguals unconsciously perform inter-linguistic translation even during single-language tasks (Thierry and Wu 2007) and that professional SIs evince distinctive electrophysiological responses for processing word pairs in their most practiced direction – from L2 to L1 (Elmer et al., 2010).



Chapter 2.  The toolkit 51

Finally, dichotic listening tests have been used to infer hemispheric dominance in relevant populations. In these studies, different verbal stimuli are administered simultaneously to each ear, and participants must decide which one they heard or how similar both were. Since initial auditory operations are contralateralized, an advantage for right-ear responses can be interpreted as a left-hemisphere preference for the process in question, and vice versa. This paradigm has been used, for instance, to assess whether professional SIs and interpreting students differ in their hemispheric specializations for syntactic and semantic processing (Fabbro, Gran, and Gran 1991).15 Overall, atomistic verbal paradigms are easy to administer, capable of illuminating narrowly defined mechanisms, and flexible enough to incorporate virtually any type of linguistic stimulus in varied presentation modes. Moreover, whereas most of them require explicit responses, specific configurations can be used to investigate implicit or unconscious processes (e.g., Thierry and Wu 2007). Furthermore, although they fail to capture all the complexities of discourse-level, skopos-driven forms of IR, they afford most useful evidence to understand critical sub-operations of translation and interpreting. Note, first of all, that token-level processes are indispensable steps in rendering translation units of any length and complexity. What is more, situated translation activities often include SL segments (such as list items and titles) that actually coincide with single words and sentences – let alone the fact that spontaneous segmentation of discourse pieces for IR oftentimes results in units comprised of one lexical or sentential item. Also, word-for-word IR has been claimed to play a role in professional interpreting (Christoffels, de Groot, and Waldorp 2003), especially during periods of fatigue or stress (Darò and Fabbro 1994). For these reasons, the verbal paradigms described above, as well as others that will be introduced in upcoming chapters, are a cornerstone to understand highly specific sub-processes crucial to IR. Also, as will be shown in Chapters 4 through 7, they have already given shape to an empirical corpus capable of motivating fundamental constraints for full-fledged neurocognitive models of translation and interpreting. Those prospects appear even more attainable when verbal paradigms are considered alongside non-verbal ones, as described next.

15. Note, however, that strong arguments have been raised against the validity of dichotic listening tests to assess hemispheric dominance for language processing (Paradis 1992, 1995, 2003, 2008).

52

The Neurocognition of Translation and Interpreting

2.3.2.2 Non-verbal paradigms In the study of IR-related topics, non-verbal paradigms have been mainly used to assess executive functions in SIs. While some of them are simple pencil-and-paper tests, others require additional elements (e.g., cards) and still others are fully computerized. Although they may sometimes involve verbal stimuli, they are intended to shed light on domain-general functions which mediate various aspects of cognitive control across processing modalities – hence their designation as ‘non-verbal paradigms’. Of note, each executive function has been related to well-defined regions and networks in the bilingual brain (Abutalebi and Green 2007), so that behavioral performance on them can be assumed to reflect the efficiency of particular biological mechanisms. The most widely used executive tests are STM span tasks. Participants are presented with ever-longer lists of stimuli (first three items, then four, then five, and so on) and asked to repeat them immediately afterwards or after a delay. In most cases, the extent of the longest list repeated without errors is taken as an index of the maximum amount of information that can be kept transiently active in relevant circuits, in the absence of overt or even covert (subvocal) rehearsal (Injoque-Ricle et al. 2015). The STM span paradigm has been widely used to assess particularities of expertise in simultaneous interpreting, with versions based on digits (Bajo, Padilla, and Padilla 2000), letters (Babcock and Vallesi 2017), shapes (Babcock and Vallesi 2017), and words in both L1 and L2 (Christoffels, de Groot, and Kroll 2006). Research on SIs has also relied on WM storage-plus-processing tasks. Here, the ever-longer lists described above are accompanied by concurrent tasks, such as uttering written sentences (e.g., Signorelli, Haarmann, and Obler 2012) or performing mathematical operations (Babcock and Vallesi 2017). Accordingly, they indicate how much information subjects can retain in WM while they are engaged in other cognitive activities. This skill, which is crucial for successful performance of simultaneous interpreting, has been assessed in both professional practitioners and interpreting trainees (e.g., Köpke and Nespoulous 2006; Tzou et al. 2012). Other studies have employed free and cued recall paradigms. In a typical incarnation of these tasks, participants are shown lists of words and asked to memorize them for later retrieval. In so-called free conditions, no specific constraint is given regarding which items must be recalled. By contrast, in cued conditions, participants are instructed to recall only those items holding a specific relation to a cue word shown at the end of each list; for example, they may be required to name the word in the list occurring immediately after the cue word or those belonging to its overarching semantic category. Both tasks have been used to assess professional and prospective SIs relative to other bilingual and monolingual groups (Bajo, Padilla, and Padilla 2000; Köpke and Nespoulous 2006; Signorelli, Haarmann, and Obler 2012). Taken together, they give hints on the capacity to selectively access particular types of information in WM.

Chapter 2.  The toolkit 53



The capacity to handle concurrent processes can be further illuminated through so-called dual tasks. In general, these involve making decisions on two sets of stimuli from different perceptual modalities. For instance, subjects may be presented with series of tones and shapes, and asked to press a button each time they hear a particular sound or view a particular image (Figure 2.3). While control conditions are restricted to only one modality (e.g., just detecting sounds in a sequence of tones), the dual task proper consists in a combination of both modalities and thus reflects the subjects’ efficiency at coordinating disparate cognitive processes. As concerns research on SIs, this paradigm has been employed in strictly behavioral protocols (Strobach et al. 2015) and together with fMRI recordings (Becker et al. 2016). A Task 1 “

Task 2 ”

Low–Middle–High Small–Middle–Big

B Button presses

Next

Task 2 Task 1 Trial start

“

max. 2500 ms

”

SOA 50, 100 or 400 ms 500 ms time

Figure 2.3  Example of a dual-task paradigm. In this dual-task paradigm, used to measure relevant structural and functional brain correlates in simultaneous interpreters and non-interpreter bilinguals, participants had to perform an auditory and a visual task (panel A). In the former, they listened to a tone and had to classify its pitch as low, middle or high. In the latter, they viewed a triangle and had to classify it as small, middle or big. In the single-task conditions, only one of the tests was administered. Instead, in the dual-task condition (panel B), the two tests were presented together, with stimuli from each test appearing in random succession and with the interval between tests varying among 50, 100, and 400 ms. Reproduced with permission from Becker et al. (2016).

54

The Neurocognition of Translation and Interpreting

Further insights into executive functions can be gained through mental-set shifting paradigms, mainly represented by the Wisconsin Card Sorting Test (Nelson 1976). In this task, participants are shown four guide cards and asked to place a new card below one of them. The cards vary in color (red, blue, green, yellow), shape (triangle, circle, star, cross), and number of shapes (one, two, three, four). In each trial, the examiner sets a tacit categorization parameter (e.g., based on color) and subjects have to infer the implicit rule to find the correct position for the cards. After a few trials, the examiner changes the rule (e.g., the categorization parameter switches to shape) without informing the participant, who has to adapt his/her active mental scheme to infer the new rule and respond accordingly. This and other relevant tasks (e.g., Babcock and Vallesi 2017; Becker et al. 2016) have been used to assess mental flexibility in both professional (Santilli et al. 2018; Yudes, Macizo, and Bajo 2011a) and aspiring (Dong and Xie 2014) SIs. Another critical component of executive functioning is the ability to suppress conflicting processes. This can be examined via inhibitory control tasks, the most popular ones corresponding to the Stroop paradigm. In a typical version (Figure 2.4), participants view color terms written in different colors, and they must name the color of the ink. In the congruent condition, the color term matches the ink color (e.g., red is written in red). However, in the incongruent condition there is a discrepancy between both elements (e.g., red is written in blue). Here, successful performance requires inhibiting the automatic tendency to read the word so as to utter ‘blue’ instead of ‘red’. Stroop tasks, as well as other inhibitory control paradigms, have been employed to examine whether SIs outperform non-interpreter bilinguals in their capacity to resolve various types of cognitive conflict (e.g., Köpke and Nespoulous 2006; Yudes, Macizo, and Bajo 2011a). As a final example, consider attentional paradigms, arguably best illustrated by flanker tasks (Eriksen and Eriksen 1974). Participants are faced with several arrows on a screen and they must focus on a central one, which may point leftwards or rightwards. The surrounding arrows can point in a congruent or an incongruent direction, and their distance relative to the central arrow may vary from trial to trial. Subjects are instructed to press a left or a right key, depending on the orientation of the central arrow, and their performance is taken as a proxy of particular components of attention. Variations of these tasks have been employed to evaluate alerting and orienting functions in professional SIs (Babcock and Vallesi 2015; Morales et al. 2015). When jointly considered, non-verbal paradigms offer a composite picture of multiple executive mechanisms which mediate every instance of translation and interpreting. Importantly, their inclusion in protocols assessing verbal functions in relevant populations has afforded valuable insights into the relations between linguistic and non-linguistic mechanisms during IR (e.g., Christoffels, de Groot, and

Chapter 2.  The toolkit 55



Congruent condition

A Yellow

Response +

1000 msec Red

Response

Conflict condition

B Yellow

Response +

1000 msec Red

Response

Figure 2.4  Example of a color Stroop task. In each trial, participants view a fixation cross lasting 1000 ms, followed by a color term that remains on the screen until a response is given. They are instructed to name the color of the ink in which the words appear, as fast as possible. In the congruent condition (panel A), the ink color matches the color term. In the conflict condition (panel B), the word appears in an incongruent color. Reproduced with permission from Oh et al. (2012).

Waldorp 2003; Santilli et al. 2018). Moreover, as anticipated here, they have been instrumental to understand how sustained practice of simultaneous interpreting impacts bilingual neurocognition – a topic that will be thoroughly addressed in Chapter 7. 2.4

The craft of manipulation

In any given study, the relevance and robustness of the above paradigms depends on the scrupulousness with which researchers have controlled the key variables involved. Imagine you want to study differences in the neural circuits subserving the translation of cognates and non-cognates. You choose a random list of words for each category, have your subjects translate them while you record their brain activity, and detect increased modulations in specific brain regions for non-cognates. Does that mean that those regions are differentially related to processing of this word type? Hardly so. In the absence of additional precautions during stimulus

56

The Neurocognition of Translation and Interpreting

selection, the resulting differences could be driven by any other factor: maybe the lists differed in their frequency, length, familiarity or concreteness, and the observed neural pattern is merely reflecting the role of these features. Therefore, in this case, no firm conclusion could be advanced regarding cognate status as a specific phenomenon modulating neurocognitive responses during translation. The key to circumventing such uncertainties lies in proper control of variables. Whereas dependent variables are contingent on the subjects’ performance or characteristics (see Section 2.5), two other types of variable fall under the researcher’s control: independent and controlled variables. Independent variables correspond to factors manipulated by design in the experiments. In the above example, the intended independent variable was cognate status. Similarly, if a study involves a BT condition and an FT condition, the independent variable is directionality; and if a study aims to track differences between professional interpreters and interpreting students, then the independent variable is interpreting experience. The validity of these constructs depends on how well the chosen tasks, stimuli, or subjects operationalize the contrasts at hand – e.g., whether the cognates and non-cognates truly differ in their level of phonological or orthographic overlap, or whether the professional interpreters recruited are demonstrably more experienced than the interpreting students. In this sense, as already suggested, a critical point is to rule out the influence of other factors that might explain observed differences. Such is the function of controlled variables, that is, factors which are established as similar between tasks, conditions, or groups. At least three main types need to be contemplated. First, researchers must account for condition-related variables. For example, if an experiment seeks to explore differences between sentence translation and reading (e.g., Lehtonen et al. 2005), then those two conditions should be identical in their timeline, number of trials, presentation mode, and required response. It is also important for the conditions to be counterbalanced (i.e., sequenced differently) across subjects; otherwise, if one condition were always administered after the other, then its differential outcomes might be reflecting fatigue effects (e.g., if it involved worse performance) or task familiarization effects (e.g., if it involved better performance). Any study which fails to contemplate these issues will be marked by severe methodological flaws. Second, a strict control of stimulus-related variables is also critical. If a study involves two or more tasks or conditions, the items included in each of them should be matched for all variables other than the independent variable. If, as proposed at the outset of this section, an experiment aims to detect neurocognitive mechanisms implicated in the processing of cognates and non-cognates, then the list of items for each category should be statistically similar in terms of key variables known to modulate behavioral outcomes and associated neural patterns. In most



Chapter 2.  The toolkit 57

experiments, these include word length, frequency, familiarity, and concreteness, although the words’ age of acquisition and translation entropy should also be controlled for similar reasons. Relevant data can be obtained through ad hoc questionnaires or, preferably, from normative databases – such as the classical MRC psycholinguistic database (Coltheart 1981), for English, or the EsPal repository (Duchon et al. 2013), for Spanish. In the case of translation experiments, however, matching stimulus lists from different languages may be challenging due to the lack of comparability between their respective databases (e.g., the number of words included in their underlying corpora or the range of values considered for certain variables), not to mention that most languages do not have validated databases for several key factors. However, at least for certain language pairs, appropriate combinations of databases can be found (e.g., García et al. 2014). Whatever the case may be, the main point is that if possible stimulus-related confounds are not properly accounted for, then differences between tasks or conditions cannot be firmly attributed to the (independent) variable of interest. Finally, between-group designs entail the added challenge of controlling for subject-related variables. This is so because demographic and experiential traits have a massive impact on neurocognition. The main sociodemographic factors to consider are sex, age, and education level. In addition, given the impact of L2 proficiency and age of acquisition on translation mechanisms (García 2015a), studies on IR should also ensure that groups are matched for such variables. Moreover, particular research questions may require forming groups with comparable executive skills (Santilli et al. 2018). For all these purposes, while most researchers resort to ad hoc questionnaires, validated tools are available in the literature. For example, L2-related factors can be assessed with the Language History Questionnaire 2.0 (Li et al. 2014), whereas specific aspects of translation and interpreting competence can be tapped via the Translation and Interpreting Competence Questionnaire (Schaeffer et al. forthcoming). The latter, in particular, offers information on several aspects of critical importance for research in cognitive TIS, such as self-reported competence in BT and FT, years of experience, and practical habits. Depending on the research question, these should be matched or explicitly shown to differ between groups. Be that as it may, systematic reporting of subject-related variables is indispensable to establish a study’s internal validity and to enable comparability of results across experiments.

58

The Neurocognition of Translation and Interpreting

2.5

Do it well, do it fast

In any experimental study, if key condition-, stimulus-, and subject-related variables are properly controlled, the participants’ performance can be measured by reference to behavioral outcomes. In brief, these are external indices of the inner processes targeted by the tasks at hand. Two main types of behavioral outcomes are considered in neurocognitive research: accuracy and response time. In fact, instructions for participants in most experiments explicitly require that they “perform the tasks as accurately and fast as possible.” In most paradigms, each trial has a correct and one or more incorrect response options.16 For example, in a lexical decision task, the left and right arrows of a keyboard may be used to indicate ‘word’ and ‘pseudoword’, respectively. If a participant views the stimulus dog and presses the left arrow, the response is registered as correct; otherwise, it is recorded as incorrect. The number or percentage of correct responses from a subject or sample determines the degree of accuracy. Depending on the task, accuracy may be taken as a measure of knowledge (e.g., overall vocabulary in lexical decision tasks) or processing efficacy in a given domain (e.g., inhibitory control in a Stroop task) – see Section 2.3.2. Also, behavioral actions in cognitive paradigms are usually timed, so that researchers can measure the lapse between stimulus presentation and the subject’s response (e.g., key-press or articulation onset). This yields so-called response (or reaction) times. These are reliable measures of the global duration of a process, from its onset until its outward manifestation, but they are blind to its inner time course – which can be measured via high-temporal-resolution techniques, such as ERPs (see Section 2.7.1.2). Response times are indices of processing efficiency, and they can be interpreted as proxies of cognitive effort and/or connection strength. They can actually be considered the cornerstone of experimental cognitive science, as they afford quantitative evidence to establish whether a task engages more cognitive resources than another, which processing level is predominantly involved in each case, and even what type of unconscious operations underlies the subjects’ performance – for examples, see García (2015a). Besides accuracy and response time, other behavioral outcomes are sometimes contemplated in the field. For instance, verbal fluency is calculated as the ‘number of hits’ in an experimental run (e.g., Santilli et al. 2018).17 Task-specific outcomes can 16. In some translation experiments aimed to assess particular lexical categories, such as cognates, only one possible target word is predefined as a valid response. For details and an example, see García et al. (2014). 17. Although ‘number of hits’ is similar to ‘accuracy’ as an outcome measure, the difference lies in that the number of items per subject cannot be pre-established by design. For example, in a



Chapter 2.  The toolkit 59

also be found in the Wisconsin Card Sorting Test, which includes such measures as ‘number of perseverative errors’ and ‘number of same-category errors’ (e.g., Yudes, Macizo, and Bajo 2011a). Also, assessments of translation quality have been used to correlate specific cognitive or neurological results with overall IR performance (e.g., Hervais-Adelman, Moser-Mercer, and Golestani 2015; Injoque-Ricle et al. 2015). When verbal or non-verbal paradigms are administered on their own to healthy participants (whatever their level of IR competence may be), the ensuing behavioral outcomes warrant but limited, indirect inferences on the underlying neurocognitive mechanisms. As stated before, for instance, faster responses for one task than another can be interpreted as greater strength of the active connections (García 2015a), but they do not support any firm conclusion on the actual neural substrates involved. However, behavioral evidence is indispensable to reach well-rounded accounts of neurocognition, as distinct neural patterns deprived of external correlates cannot be strongly related to a specific cognitive function (Krakauer et al. 2017). In particular, behavioral outcomes become maximally informative for brain-based research when combined with lesion models or neuroscientific techniques, to which we turn in the next two sections. 2.6 System breakdown A classical approach to relate behavioral performance to particular neural substrates consists in assessing brain-damaged patients. These subjects provide so-called ‘lesion models’, that is, unplanned scenarios in which cognitive functions can be examined in the presence of more or less circumscribed cerebral injuries. The aim is to establish single or double dissociations. The former stem from individual patients whose focal lesions disturb a given process (e.g., BT) while others (e.g., FT) remain fully or relatively spared. The latter require joint analyses of two or more cases to establish contrastive patterns in which damage to area A compromises function X but not Y, while a lesion to area B yields opposite results (Dunn and Kirsner 2003). Unless combined with neuroscientific techniques, behavioral outcomes from lesion models do not offer real-time data on brain activity. However, the detection of single and (more crucially) double dissociations is a strong indication that functions X and Y rely on partially independent neural circuits, whose critical substrates lexical decision task with thirty trials, if two subjects produce only correct responses, their performance will be identical. However, in a verbal fluency task, both subjects may produce only correct responses, but one of them may have uttered twenty words while the other produced merely twelve. Moreover, the responses given by each of them may or may not be similar, whereas no such variation is possible among the stimuli used in controlled tasks yielding accuracy outcomes.

60 The Neurocognition of Translation and Interpreting

differentially correspond to areas A and B. This approach has proven crucial for the development of neurocognitive theories on multiple domains, including memory (Squire 2009), social cognition (Ibáñez et al. 2018), conceptual knowledge (Capitani et al. 2003), and language (Paradis 2004; Ullman 2001a). More particularly, a considerable body of evidence exists on how various brain lesions compromise IR skills in bilinguals. Such findings have often been produced via ad hoc tests, whereas others have been obtained with standardized instruments. Part C of the Bilingual Aphasia Test (Paradis 1979, 2011) stands out among the latter, as all of its versions (covering dozens of language pairs) include ten equivalent recognition trials, twenty word-translation trials, and twelve sentence-translation trials (half in each translation direction). This allows for systematic comparisons of IR capacities across multiple patients with similar or different patterns of brain damage (e.g., Fabbro and Paradis 1995). Through the use of diverse research protocols, lesion models have led to the discernment of four translation neuropathologies, which will be discussed in Chapter 4. These have informed the relationship between executive and linguistic mechanisms during translation (García-Caballero et al. 2007), their partial independence from those supporting single-language operations (Paradis, Goldblum, and Abidi 1982), and the inner organization of IR-specific systems (Aglioti and Fabbro 1993). Evidence of this kind is crucial to characterize the anatomical organization of relevant systems, as shown by extant neurological models of IR mechanisms (Fabbro 1999; García 2012a) – see Chapter 4. 2.7

The brain, in vivo

The cognitive processes recruited by the tasks above (and by any mental operation, for that matter) are associated with, if not dependent on, the synchronized activity of millions of interconnected neurons. These networks are widely distributed across the brain, but the key contributors to each particular function are mainly concentrated in critical regions evincing marked task-related metabolic changes. Moreover, different cognitive processes entail systematic modulations of electrophysiological signals through time. Multiple aspects of these phenomena can be captured (and even altered) by neuroscientific techniques, which offer a window into the real-time biological workings of IR. Neuroscientific techniques can be separated into two broad types. Non-invasive techniques provide online measures of hemodynamic or electrophysiological brain activity in the absence of deliberate lesions. On the other hand, invasive techniques involve incisions, perforations or electrical perturbations for human clinical assessment, during which researchers can examine the role of a brain region in a given



Chapter 2.  The toolkit 61

function. With their benefits and limitations, all of them can be and have been used to study the biological basis of IR. Let us describe their main principles to understand how they work. 2.7.1 Non-invasive techniques 2.7.1.1 Functional neuroimaging Functional neuroimaging techniques have excellent spatial resolution. They offer accurate information about which part of the brain is implicated in the cognitive operation at hand. However, they lack good temporal resolution, as they do not reveal precisely when that process took place. The most popular ones are fMRI and PET, both of which have been used to examine aspects of IR. 2.7.1.1.1 fMRI The discoveries and developments that led to modern fMRI span the whole twentieth century. The most recent breakthrough was achieved by Ogawa et al. (1990), who found that magnetic resonance imaging (MRI) scanners could measure blood flow and indirectly index neural activity. During cognitive tasks, blood flow increases in relevant brain areas (Huettel, Song, and McCarthy 2008). This increment generates a slight shift in the magnetic resonance signal, known as blood-oxygen-level-dependent (BOLD) effect. Habitually, by comparing hemodynamic concentrations during a target task (e.g., BT) with those of a control or baseline condition (e.g., L2 reading), researchers can infer which areas were distinctively engaged by the former. MRI is based on a complex quantum phenomenon but it can be schematically described in terms of classical physics (Weishaupt, Köchli, and Marincek 2006). A superconductive magnet produces a strong magnetic field that causes hydrogen nuclei’s spins to align parallel or anti-parallel with it. A radio-frequency pulse is then applied at a particular frequency, which changes the orientation of the spins until they are perpendicular to the main field (T1) and synchronously moving (T2). Next, the pulse is turned off and the spins return to the orientation imposed by the main field while emitting a signal that can be detected by an antenna or coil. The time needed for the spins to return to the main field’s orientation and desynchronize (‘relaxation’) depends on the tissue where the nuclei are immersed. Color-coding the different relaxation times gives rise to the differences observed in structural MRI between white matter, gray matter, and cerebrospinal fluid. In fMRI, the magnetic properties of oxygenated and deoxygenated hemoglobin allow measuring blood flow and mapping significant activation differences between conditions. Consider, for example, Figure 2.5, which shows activation increases during reading of naturalistic texts.

62

The Neurocognition of Translation and Interpreting

L

R

p < .01

p < .001

p < .0001

Figure 2.5  Example of fMRI recordings during language processing. The images show activated areas during reading of paragraphs from The Emperor’s New Clothes (by Hans Christian Andersen) and a Nelson-Denny Practice Test. Significant hemodynamic increases were observed bilaterally in regions related to language processing (inferior frontal, middle temporal, superior temporal, and angular gyri), visual processing (e.g., cuneus, lingual gyrus, occipital pole), and eye-movement control (e.g., superior colliculus, thalamus). Images are displayed on inflated brain surfaces. As indicated by the color bar, the smaller the p-value, the greater the statistical significance of the activation. L: left; R: right. Reproduced with permission from Choi, Desai, and Henderson (2014).

The fMRI technique is characterized by low temporal resolution, relatively high cost, and strict infrastructural requirements. In addition, the interpretation of fMRI results is usually challenging due to the complex interactions between neural, vascular, and metabolic factors, as well as the statistical problems arising from multiple-comparison tests (Buxton 2009; Logothetis 2008). Nevertheless, these limitations can be largely overcome by proper study design, data analysis, and interpretation strategies. Under such circumstances, fMRI is a most valuable tool to obtain spatially precise recordings of task-relevant regions in a non-invasive context. Indeed, it is one of the main sources of evidence on the neural bases of IR. Among other things, it has been used to examine the impact of syntactic complexity during FT (Lehtonen et al. 2005), the role of semantic factors in the relation between translation equivalents (Correia et al. 2014), the networks recruited by translation units with different sensorimotor associations (Mayer et al. 2015), and the neurofunctional particularities of SIs (Becker et al. 2016) and interpreting students (Hervais-Adelman, Moser-Mercer, and Golestani 2015) – see Chapters 5, 6, and 7.



Chapter 2.  The toolkit 63

2.7.1.1.2 PET PET imaging was developed in the second half of the twentieth century. It relies on the detection of gamma rays by a brain scanner to measure blood flow, metabolism, and neurotransmitter binding or uptake (Kandel 2013). When neurons in a given location increase their firing rate, regional oxygen consumption and blood flow increases as well. Therefore, different neuronal processes can be monitored by introducing specific tracers in the bloodstream. Subjects are injected with an isotope produced in a cyclotron by accelerating protons into the nuclei of chemical elements, such as oxygen and fluorine. Once unstable, such elements are used to synthesize a tracer that can be detected in the bloodstream. Two widely used tracers are 18F-deoxyglucose and H215O. These are employed to measure glucose metabolism and blood flow, respectively, thus indexing local neuronal activity (Portnow, Vaillancourt, and Okun 2013). The isotopes emit positrons that travel in tissue at the speed of light and eventually collide with an electron, an event that annihilates both particles and emits two gamma rays at 180 degrees from each other. When two diagonally placed sensors make a near-simultaneous detection, the emission’s source can be identified and reconstructed in a 3D model (Herholz, Herscovitch, and Heiss 2004). By way of illustration, Figure 2.6 shows results from an experiment comparing activation patterns during synonym generation and word repetition in L1 and L2. Like fMRI, PET involves elevated costs, specialized equipment, and highly qualified personnel. Moreover, the sluggish nature of hemodynamic responses renders it temporally imprecise. However, this technique possesses very good spatial resolution and offers the chance to measure aspects of brain function that could not be observed otherwise in a non-invasive setting (e.g., serotonin or dopamine metabolism). As regards research on IR, PET has been employed to track brain regions differentially engaged by translation relative to single-language processing, and by BT relative to FT (Klein et al. 1995; Price, Green, and von Studnitz 1999; Rinne et al. 2000) – see Chapter 5. 2.7.1.2 EEG methods Typically, EEG methods are characterized as possessing excellent temporal resolution but poor spatial resolution. This is true of ERPs, which reveal when specific processes take place (within the order of milliseconds) but fail to indicate the exact source of significant modulations. However, the use of ERP with dense-array electrodes (128 or 256 recording units), 3D position monitoring systems, recent source-location algorithms, and fMRI co-recordings reduce spatial resolution limitations and offer a more complete view of brain processes.

64 The Neurocognition of Translation and Interpreting

A L1 synonym generation minus L1 word repetition

t value 6

4.7 B L2 synonym generation minus L2 word repetition

3.7 2.5

Figure 2.6  Example of PET recordings during language processing. Averaged PET subtraction images of cerebrospinal fluid increases during verbal tasks, superimposed upon averaged MRI scans. The images show areas yielding stronger activation for synonym generation than word repetition, both in L1 (panel A) and L2 (panel B). The contrast is aimed to capture circuits differentially implicated in semantic processing in each language. Activity increases can be mainly observed in and around the left inferior frontal gyrus. The color bar indicates activation intensity based on t-values. Reproduced with permission from Klein et al. (1995).

2.7.1.2.1 ERPs The first human scalp EEG was achieved by Hans Berger in 1924, roughly fifty years after Caton’s pioneering recordings of electrical activity from exposed brains of monkeys and rabbits (Freeman and Quian Quiroga 2012). EEG measures the



Chapter 2.  The toolkit 65

oscillations of electrical potentials on the scalp as produced by the cooperative action of neurons. One single sensor (i.e., an electrode) placed on the human scalp provides estimates of activity from between 100 million and one billion averaged neurons (Nunez and Srinivasan 2006), arriving from a surface of approximately 10 cm2, mostly from the outer layers of the cortex (Buzsáki 2006). Interneuronal communication generates two types of electrical signals: action potentials and postsynaptic potentials. Whereas the former travel along the axons of neurons, the latter are confined to the dendrites and cell bodies (they arise when neurotransmitters bind to receptors in the afferent neuron, causing ion channels to open or close). Action potentials last approximately one millisecond and do not easily allow for non-invasive measurement. In contrast, inhibitory and excitatory postsynaptic potentials can last tens or even hundreds of milliseconds and generate a small dipole (a negative and a positive electrical charge separated by a small distance). When several such potentials occur synchronically in a spatially aligned arrangement, their aggregate activity can be recorded from the scalp (Luck and Kappenman 2012). EEG signals can be analyzed in various ways: through visual inspection, by decomposing them into specific frequencies, by estimating relationships among sensors, and, frequently, by averaging evoked responses (Freeman and Quian Quiroga 2012). The latter approach is known as the ERP technique. The electrical activity patterns associated with several presentations of stimuli of the same category are averaged for each condition across subjects, so that random fluctuations of the signal and non-stimulus-related activity tend to be cancelled out, leaving what is common among multiple presentations (Luck and Kappenman 2012). Normally, the result is a waveform featuring negative and positive peaks, whose amplitudes and latencies can be compared between conditions. Many typical waveforms have been identified as reliable indexes of specific neurocognitive processes. For example, modulations of the P2 component (a positive-going potential peaking at about 200 ms post-stimulus onset) have been systematically related to differences in attentional demands. Instead, modulations of the N400 component (negative peaks appearing roughly 400 ms post-stimulus onset) index how difficult it is to integrate lexico-semantic information with the preceding context, among other effects (Luck and Kappenman 2012). Figure 2.7 illustrates the latter ERP associated to the reading of a semantically anomalous word at the end of a sentence. A key limitation of ERP results is that they do not reveal which brain areas triggered the reported modulations. Moreover, it is highly sensitive to ocular and motor artifacts. However, these limitations can be overcome with well-established methodological precautions. Also, the portability, cost-effectiveness, and unparalleled temporal resolution of this technique render it a crucial tool to assess diverse aspects of neurocognition. This, of course, is also true of IR, as shown by studies

66 The Neurocognition of Translation and Interpreting

N400

–5

Cz µV 0

5 0

200

400

ms

600

800

The man wrote the letter with a pen. The man wrote the letter with a mop.

Figure 2.7  Example of an ERP modulation during language processing. This schematic representation illustrates a typical modulation of the N400 component over the Cz electrode, which typically shows greater amplitude for semantically incongruent than congruent final-sentence words – these signals being represented by the red and orange lines, respectively. Such an effect is taken as an index of increased semantic integration efforts. In the context of the present example (The man wrote the letter with a…), this is due to the lower semantic predictability of the incongruous word (mop) compared to the congruous one (pen). The Y axis indicates voltage, with positive values toward the bottom and negative values towards the top. The X axis plots time in milliseconds, starting at target-word onset (0 ms).

revealing that this activity can occur unconsciously during single-language processing (Thierry and Wu 2007), and that particular cognitive mechanisms are engaged by different translation units (Christoffels, Ganushchak, and Koester 2013) and directions (García, Mikulan, and Ibáñez 2016) – see Chapters 5 and 6. 2.7.1.2.2 Oscillatory activity ERPs are time- and phase-locked to stimulus onset. In other words, these responses are generated by the transient synchronization of neural networks, independently of the overall neurophysiological dynamics occurring at the time. Therefore, they do not track delayed or temporally variable neurophysiological events, and their underlying averaging procedure renders them blind to non-phase-locked activity (Mouraux and Iannetti 2008). Such patterns ignored by the ERP technique can be captured by measurements of ongoing dynamics induced by externally or internally generated events. This is achieved through time-frequency analyses, which allow detecting spectral power changes across time points and frequency bands. Shortly, an EEG signal can be divided into frequency bands, each of which is sensitive to distinct cognitive phenomena indexed by modulations within a range of Hz (for examples, see Section 3.3.2.2). Most researchers recognize five such bands, termed delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), and gamma (> 30 Hz) – and some of these are often further subdivided (Burgess 2012). With this technique, power modulations are averaged across trials and normalized

Chapter 2.  The toolkit 67



by reference to baseline activity occurring before stimulus onset. The cognitive processes implicated can thus be related to a power increase (event-related synchronization) or decrease (event-related desynchronization), two forms in which distributed neuronal populations become spatially linked to form transient functional networks (Singer 1993). Figure 2.8 illustrates different patterns of synchronization and desynchronization in the theta and beta bands during processing of correct and incorrect sentences. Semantic

Hz

Control

30

Hz

30

25

25

25

20

20

20

15

15

15

10

10

10

5

5

5

0 Syntactic

Violation

30

0.2

0.4

0.6

0.8

Control

30

0

0.2

0.4

0.6

0.8

Violation

30

0

30

25

25

25

20

20

20

15

15

15

10

10

10

5

5

5

0

0.2

0.4

0.6

0.8

0

0.2 0.4 0.6 Seconds

– 0.25

0 dBV

0.8

Violation-control

0.2

0.4

0.6

0.8

Violation-control

0

0.2

0.4

0.6

0.8

+ 0.25

Figure 2.8  Example of oscillatory activity during language processing. The figure shows time-frequency charts capturing power changes during reading of correct sentences and sentences with semantic or syntactic violations. The final column shows power differences between the correct sentences and those featuring each type of violation. Semantic violations were associated with increased theta power, whereas grammatical violations involved a decrease in beta power. The color bar, ranging from blue to red, indicates the intensity of desynchronization (towards the blue side of the spectrum) or synchronization (towards the red side of the spectrum). Reproduced with permission from Davidson and Indefrey (2007).

68 The Neurocognition of Translation and Interpreting

As regards IR, measurements of oscillatory EEG dynamics have been used to illuminate neurocognitive mechanisms underlying the translation of low- and high-frequency words, and those engaged by successful (as opposed to failed) translations (Grabner et al. 2007). 2.7.1.2.3 Functional connectivity Neuroscientific explorations of a cognitive domain can go beyond the identification of relevant individual sites and their independent temporal dynamics. Recent approaches allow researchers to investigate how different areas interact with one another during mental activity. For instance, in connectivity studies, a distinction is often made between two key forms of neurocognitive interaction: segregation and integration. Two or more cortical areas are segregated if they are specialized for different aspects of a function.18 Instead, functional integration refers to how those specialized areas mutually exchange information during cognitive processing (Friston 2011). Functional networks are created by the coordinated activity of various brain structures, from the microscale of neurons and synapses to the macroscale of cognitive systems (Sporns 2011). These can be explored from three complementary perspectives: structural connectivity (how anatomical connections are organized), effective connectivity (how a neural system influences another one), and functional connectivity (how remote neurophysiological events are linked in terms of statistical dependencies) (Friston 2011). The latter approach, in particular, uses statistical measures such as correlations, coherence or transfer entropy to infer concerted activity patterns among different micro- or macro-structures. If signals from two brain areas show high mutual dependence, they can be said to be interacting (Buzsáki 2006; Varela et al. 2001; Perez Velazquez and Wennberg 2009). So far, brain networks related to IR have been studied through EEG signals, although they can also be examined considering PET, fMRI, and magnetoencephalographic data. Their anatomical and functional properties can be characterized using tools from complex network analysis, a multidisciplinary approach to the study of complex systems with roots in graph theory (Rubinov and Sporns 2010). Task-related patterns of inter-regional connectivity can be visualized as in Figure 2.9, which shows differential forms of inter-hemispheric coupling for two types of relative clauses during sentence comprehension.

18. This notion should be distinguished from functional localization, the incorrect assumption that a function can be confined to a particular area.

Chapter 2.  The toolkit 69



Functional connectivity after relative-clause processing Subject-subject sentences

T6

Br F8

Subject-object sentences

T6

Br F8

O2

Wr

F4

O2

Wr

F4

Fz

Fz

Pz

F3

Pz

F3 Wl

F7 Bl

O1

Bl

T5

p≤

Wl

F7

.01

.02

O1 T5

.05

Figure 2.9  Example of functional connectivity during language processing. The figure shows functional connectivity patterns during post-relative-clause processing for subject-subject and subject-object sentences, relative to mean connectivity before sentence onset in the theta band. Results were derived from EEG signals via the ‘coherence’ metric. Inter-hemispheric connectivity was greater for subject-object than subject-subject sentences, as indicated by the thickness of the lines (the thicker the line, the greater the statistical significance of the connection). Reproduced with permission from Weiss et al. (2005).

In the last decades, connectivity approaches to brain networks have grown exponentially and yielded important results (Fox and Greicius 2010; Greicius 2008). Their use in IR research is incipient but promising, as seen in studies on translation directionality (García, Mikulan, and Ibáñez 2016). 2.7.2 Invasive techniques Invasive techniques are applied only during presurgical assessment in patients who require removal of brain tissue, mainly due to tumors or epilepsy. Therefore, they are seldom used to explore language functions. The evidence they offer is invaluable not just because it is rare, but mainly because it is much more direct than that afforded by neuroimaging or neurophysiological techniques. Though very sparingly, invasive methods have been strategically used to assess IR.

70 The Neurocognition of Translation and Interpreting

2.7.2.1 Direct electrostimulation The practice of direct electrical cortical stimulation began in the nineteenth century, in animals, and was first applied in humans in 1874. Since then, it has been mainly used to treat brain tumors, localize epileptic foci, and infer the function of brain areas during surgery (Borchers et al. 2012). Since cortical organization and post-lesion reorganization vary among patients, crucial functions must be mapped presurgically to avoid resecting a functionally putative region. While the patient is performing a specific task, the cortex is stimulated with a bipolar electrode that induces transient and reproducible changes in behavior. As a rule, if a function F is disrupted by stimulation of area A, it means that A plays a critical role in F. In addition to its clinical relevance, the information gathered during stimulation protocols (including language tasks) provides critical information for neuroscientists. Indeed, mapping of brain functions has grown notably thanks to the data thus obtained (for a review, see Selimbeyoglu and Parvizi 2010). Despite its advantages, this technique is not without shortcomings. Electrical stimulation evokes a complex effect in a large volume of tissue that depends on many physiological and morphological factors; therefore, findings cannot be easily interpreted (Borchers et al. 2012). Moreover, since direct electrostimulation studies are always performed on pathological brains, the generalization of results to models of normal neurocognition must be taken with reserve. Nevertheless, electrostimulation data can fruitfully complement insights derived through other high-spatial-resolution methods. As regards IR, this technique has been used to examine the extent to which single-language and translation processes rely on shared neural substrates (Borius et al. 2012). 2.7.2.2 Intracranial recordings Human intracranial EEG recordings began in the mid-nineteenth century (Jasper and Penfield 1949) and are still commonly used. They provide unique information about foci location in patients with medically intractable epilepsy and are also used in patients with brain atrophy or tumors. When neurologists are unable to locate epileptic foci using scalp EEG and seizure semiology, intracranial electrodes are implanted and left into place for approximately one or two weeks, in expectation of spontaneous seizures. The high temporal and spatial resolution of this technique allows experts to recognize the seizure’s origin and dynamics and to evaluate the chance and feasibility of a resection (Engel Jr. 2005). Two main types of electrodes are used: grids and depth electrodes. The former are placed subdurally, on the surface of the cortex, while the latter are needle-shaped electrodes that penetrate into deeper (even subcortical) structures. While patients remain in the hospital waiting for the seizures to occur, they normally agree to perform cognitive tasks. Researchers can thus obtain intracranial

Chapter 2.  The toolkit 71



EEG recordings to assess the role of very precise brain areas in a given function – for a review of varied domains examined through this technique, see Jacobs and Kahana (2010) and Lachaux, Rudrauf, and Kahane (2003). The ensuing data is characterized by excellent spatial and temporal resolution as well as a relatively high signal-to-noise ratio, since recordings are essentially free of muscular and ocular artifacts (Buzsáki 2006). Moreover, fine-grained neural information can be obtained even in single-case studies (Jacobs and Kahana 2010). Note that electrode sites are chosen exclusively for clinical reasons, resulting in limited brain coverage. Moreover, as in direct stimulation studies, data generalization onto healthy brains must be done with reserve. Despite these limitations, intracranial EEG recordings offer a unique window into neurocognition, and they have afforded preliminary evidence on the particular mechanisms recruited by BT and FT (García, Mikulan, and Ibáñez 2016). 2.8

How (not) to interpret the data

Each neuroscientific technique involves specific assumptions to interpret the data it offers. For example, in high-spatial-resolution studies, greater or broader activation patterns may reflect additional cognitive demands, among other effects. For techniques with high temporal resolution, a waveform’s peak-time and amplitude may indicate varied aspects of the process at hand, including its processing level (e.g., phonological, semantic) and degree of cognitive effort. However, the interpretation of such patterns is typically less straightforward than it may seem. Whether we are reading published works or making sense of our own data, a number of conceptual and epistemological precautions must be taken when interpreting results from neurocognitive research. First, as stated in Section 2.4, an experiment’s internal validity is often tied to how carefully its stimuli have been constructed. To recap a previous example, if a study aims to assess differential mechanisms involved in the translation of cognates and non-cognates, then the stimulus lists from both conditions must be matched in terms of major potential confounds. However, in most studies this is only partially achieved. For example, whereas frequency and length are usually controlled when constructing the stimulus lists, very few studies match the latter in terms of other variables known to modulate neurocognitive activity and behavioral performance, such as imageability or age of acquisition. While this scenario may sometimes be attributed to insufficient care during the research design process, it usually stems from practical constraints: if researchers were to use only those subsets of cognates and non-cognates matched for absolutely all relevant variables (e.g., frequency, length, familiarity, concreteness, imageability, age of acquisition, number

72

The Neurocognition of Translation and Interpreting

of orthographic neighbors, translation entropy), they would likely find themselves with very few items per list. Therefore, their stimuli would be too restrictive and their results would be statistically underpowered. Something similar occurs when selecting groups of participants (e.g., professional and aspiring translators): these are usually matched for age and sex, but not so frequently in terms of education level, L2 proficiency, or age of L2 acquisition. While some of these variables may be more relevant than others, a strict control of too many variables may result in low sample sizes. A compromise solution is often adopted: the most sensitive variables are matched and those presumed to have a less critical effect on the task are left uncontrolled. The bottom line is that no study on IR (or on language in general, for that matter) can perfectly isolate a target variable, so all conclusions must always be taken as tentative or partial. It is up to each reader to judge how successful an experiment has been at ruling out potential confounds and, thus, how firmly its findings can be related to the targeted independent variable. Needless to say, those studies that fully overlook even basic controls of stimulus- and subject-related variables should be read with major reservations. A second point concerns findings from lesion models. If a study shows that a domain X becomes dysfunctional upon damage to an area A, it does not follow that X is anatomically located in A. The damaged area may well be implicated in a more general function, which is involved in (but not reduced to) X. At most, what can be claimed in these cases is that some part of A is critically related to some component of X (Jackson 1878). On the other hand, the absence of a deficit in X following a lesion in area A does not prove that A is not involved in X: it may well be that A plays a role in X but that spared regions sufficed for task completion; and it is also possible that the task employed was not properly designed to tap into the critical contributions of area A. In sum, extreme localizationist interpretations are rarely (if ever) warranted by lesion models: what these offer are hints about which areas play critical or differential roles in one function relative to another one, further suggesting that such functions depend on partially separate mechanisms, whatever their complete putative substrates might be. This, of course, does not support the postulation of one-to-one relationships between structures and functions. A similar observation can be made regarding data from neuroscientific techniques. If a process X, relative to a process Y, involves more activity or greater modulations in an area A, it does not mean that A is the sole substrate of X, nor that A is exclusively devoted to X. Such patterns simply indicate that the neural mechanism in question plays a differential role in the former process. Indeed, if one considers brain activity patterns from virtually any given task without comparing it to activity in a control condition, widespread activity across multiple brain regions will most likely be observed. It must also be noted that activation is not an all-or-nothing phenomenon: different regions may be activated at various intensity levels during



Chapter 2.  The toolkit 73

a given task, but only those surpassing a pre-established statistical threshold will appear as significant contributors in an experiment’s final outcomes. Moreover, neuroscientific results are constrained by the possibilities and limitations of the devices employed. Recordings will vary greatly, for instance, depending on the number of electrodes in an EEG experiment or the number of Teslas of a scanner in an fMRI experiment. In addition, if one technique yields null differences between two conditions, it does not mean that these are neurally equivalent: it may well be that their differential processing is indexed by contrastive patterns on another level of inquiry which could be tapped with other techniques. Note, too, that neuroimages, being snapshots of brain activity, fail to capture the moment-to-moment flow of neurocognition and the multiple interactions it entails among various dimensions (including chemical, electrical, anatomical, and hemodynamic levels). Therefore, neuroscientific findings should always be taken as partial technology-constrained approximations to complex phenomena whose unfolding escapes the possibilities of any single analysis tool. Another important issue concerns the relation between behavioral and neural measures. Note that these are independently captured through different indices, which can be correlated, used for regression analyses, or used to predict one another. However, neither of them can be firmly said to be the cause of the other. Biological, phenomenological, and behavioral dimensions of our dealings with the world occur synergistically and with constant omnidirectional co-determinations. While direct causal relationships do exist among such dimensions, these can rarely be tracked with neuroscientific designs. Only certain approaches (e.g., some brain stimulation studies, certain pre/post designs) allow advancing causal conclusions in the field, but even these are controversial. The lesson is that, except for particular exceptions, most findings in cognitive neuroscience show relations between neurological and behavioral patterns, but these do not often justify mechanistic conclusions in terms of cause and effect. In sum, neuroscientific research does not yield perfect, irrefutable or comprehensive findings, and neither does it reveal exclusive causal mappings between brain mechanisms and associated cognitive functions. What brain-based experiments offer is a rich set of partial peeks at specific aspects of broader, more complex events. Of course, the same could be said of virtually any other field of inquiry within and outside TIS, with perhaps one important difference: the tools of cognitive neuroscience are sufficiently varied so as to provide multidimensional views of its target phenomena, including IR. Understanding the peculiarities, possibilities, and limitations of extant methods is a key step towards integrating fragmentary pieces of knowledge into a panoptic picture of the translating and interpreting brain.

74

The Neurocognition of Translation and Interpreting

2.9 Final remarks Maslow (1966: 15) once wrote that “it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail.” Fortunately, researchers investigating the neurocognitive basis of IR do not have to worry about such a temptation, as a diverse methodological repertoire lies at their disposal. The challenge consists in choosing the right resources for the question at hand – and while this is daunting in itself, the task would be directly unfeasible if they followed MacGyver’s one-tool-for-all philosophy. As we shall see in Chapters 4 through 7, the designs, tasks, measures, and techniques described above have been variously employed to study the neural correlates of translation and interpreting. Yet, before plunging into such contents, newcomers to the field will benefit from general notions about neurology, the neural organization of language, and neurocognitive aspects of bilingualism as a prerequisite for IR. Such prolegomena are offered in Chapter 3.

Chapter 3

Prolegomena to the translating and interpreting brain

3.1

Laying the groundwork

In 2013, the town of Benidorm, in Alicante, earned a good deal of notoriety. Newspapers the world over discussed the irregularities behind the construction of the InTempo building, a 47-story skyscraper heralded as a standard for the future of Spain. Above and beyond the financial and political scandals framing the project, what caught the attention of several media outlets was the news of an apparent architectural blunder. As reported by different sources, the project’s blueprints had made no provisions for elevator shafts. The undertaking, some journalists concluded, was thus doomed due to poor planning. The story was eventually dispelled as false – although the building was never completed –, but it foregrounded a major concern for those engaged in complex creative processes: a massive endeavor could be rendered irreparably flawed if basic elements are not considered from the outset. The point, it must be noted, proves just as relevant for the construction of skyscrapers as it does for the construction of theoretical models. This lesson should resonate loudly in contemporary cognitive TIS. Most models of translation and interpreting processes are set forth without acknowledging basic constraints, such as the organization of the overall system in which those processes occur, the generalities of the language mechanisms they call on, and the particularities of bilingualism as a precondition for IR to occur. In the absence of such foundations, most conceptual edifices in the field are hard to navigate. If broader theoretical stipulations are left unmentioned, claims about translation and interpreting are bound to prove underdetermined and untestable at best, or implausible and uninterpretable at worst. In view of this admonition, the present chapter introduces basic concepts and interpretive constraints for neurocognitive characterizations of IR. Its working premise is that translation and interpreting processes are partially but decisively molded by the general organization of brain structures, the mapping of linguistic functional networks, and the neurocognitive specificities of bilingualism. Through successive overviews of these topics, the following sections aim to endow readers

76

The Neurocognition of Translation and Interpreting

with general knowledge needed to understand, contextualize, and criticize the contents of Chapters 4 through 7. First, insofar as cognitive operations are non-randomly related to brain structures and functions (Kandel 2006, 2013), general principles of neurology must be taken as a key source of constraints for any postulation regarding the biological basis of IR. Such is the crux of Section 3.2. Importantly, too, the identification of diverse brain regions and processes in that passage provides indispensable reference points for newcomers to make sense of the multiple findings that will be mentioned in the following chapters. Second, with the exception of particular intersemiotic modalities, all instances of translation and interpreting are necessarily rooted in linguistic mechanisms. Indeed, from a neural standpoint, IR systems are embedded in more general language systems (García, Mikulan, and Ibáñez 2016). Considering these points, Section 3.3 describes the broad organization of linguistic networks as a critical basis to grasp the particularities of IR systems and processes. Third, due to its very nature, IR is shaped by the distinctive properties of the bilingual brain. Given that “[b]ilingualism is the substrate for considering all manifestations of translation ability” (Shreve 2012: 2), understanding how the brain handles two co-existing languages is crucial for apprehending the neurocognitive particularities of that skill. Accordingly, Section 3.4 presents core findings from bilingualism research, thus meeting recent pleas for its rapprochement with cognitive TIS (Schwieter and Ferreira 2017). In short, what follows is a series of prolegomena mainly directed at readers unfamiliar with neuroscience and neurolinguistics. Those who are well-versed in such disciplines could simply skip this chapter and move on to the more specific issues treated in the remainder of the book. Be that as it may, the following sections offer but very general overviews of their respective topics. For in-depth treatments, the reader is welcome to consult the works of Kandel (2013) on neural science, Hickok and Small (2015) on the neurobiology of language, and Hernandez (2013) on the bilingual brain. 3.2

A primer on neurology

The brain is part of the central nervous system. It is located in the skull, weighs roughly three pounds, and contains over 100 billion neurons. A traditional view of its structures distinguishes among the myelencephalon (medulla oblongata), the metencephalon (pons and cerebellum), the mesencephalon (midbrain), the diencephalon (thalamus and hypothalamus), and the telencephalon (neocortex and cerebral hemispheres). These structures can be discerned in a midsagittal section, as shown in Figure 3.1.



Chapter 3.  Prolegomena to the translating and interpreting brain 77

Cerebral hemisphere Corpus callosum Diencephalon Midbrain Cerebellum Pons Medulla Spinal cord

Figure 3.1  Midsagittal section of the brain. From top to bottom, the figure shows a schematic depiction of a cerebral hemisphere, the corpus callosum, the diencephalon, the midbrain, the cerebellum, the pons, the medulla, and the spinal cord.

The outer surface of the cerebral hemispheres is known as the neocortex. The regions included in the diencephalon and midbrain are called subcortical structures. The next two sections address them in turn. 3.2.1 The neocortex The neocortex contains the most important areas for language processing. It is divided into two roughly symmetrical halves called cerebral hemispheres. Separated by the longitudinal fissure, the hemispheres are bidirectionally connected through the corpus callosum, a bundle of long-distance fibers. Although both hemispheres work in a coordinate fashion during most human activities, each of them is differentially related to specific functions. Language processes, in particular, are lateralized to the left hemisphere (LH) in approximately 97% of the population (Springer et al. 1999) – although the right hemisphere (RH) also plays important roles during verbal communication (see Section 3.3). Also, as shown in Figure 3.2, each hemisphere is divided into anatomically defined lobes, which are further subdivided into smaller portions called gyri. Figure 3.2 shows two main grooves: the Sylvian fissure and the central sulcus. Both constitute anatomical landmarks to delimit broad areas within each hemisphere. The Sylvian fissure separates the temporal lobe from the frontal and the parietal lobes, while the central sulcus marks the boundary between the latter two. The frontal lobe lies anterior to the central sulcus and superior to the Sylvian fissure. It is mainly implicated in motor action and functions such as planning behavior and coordinating information coming from the rest of the brain. The temporal lobe

78

The Neurocognition of Translation and Interpreting

Central sulcus

Frontal lobe

Parietal lobe

Occipital lobe Temporal lobe Sylvian fissure

Figure 3.2  Lateral section of the left hemisphere. The figure depicts four lobes as well as the Sylvian fissure and the central sulcus.

lies posterior and inferior to the Sylvian fissure and is specialized for auditory processing, in addition to more specific functions. Superior to the temporal lobe and posterior to the central sulcus is the parietal lobe, which processes somatosensory information coming from all over the body. The hindmost part of each hemisphere comprises the occipital lobes, which are critical for processing visual information. The left frontal and temporal lobes include the most critical regions for language processes, including key functions such as phonological recognition, lexico-semantic access, morphosyntactic processing, and phonological production, among many others (Ardila, Bernal, and Rosselli 2015). As these regions surround the Sylvian fissure, they are known as perisylvian areas. Of note, the Sylvian fissure is longer and more horizontal in the (language-dominant) LH than in the RH, and some language-critical perisylvian areas (e.g., the planum temporale and the pars triangularis) typically present more gray matter density in the former. Each lobe has its own internal divisions. In particular, the frontal and temporal lobes consist of three gyri (inferior, middle, and superior). For example, the superior temporal gyrus is the part of the temporal lobe lying immediately below the Sylvian fissure. The middle and inferior temporal gyri can be found in succession. Each gyrus subserves partially specific functions within each lobe. All brain regions are widely connected with one another. No region is selfsufficient for processing higher-order functions, such as thought, visuospatial orientation, memory or language. These require the concerted action of multiple neural structures, which may be quite distant from one another. However, certain areas are critical to or indispensable for the proper functioning of specific cognitive domains. The brain’s astounding connectivity is enabled by neurons. In the neocortex, the cell bodies of neurons make up the outer surface of the hemispheres – more

Chapter 3.  Prolegomena to the translating and interpreting brain 79



precisely, a layer ranging from 3 to 6 mm in thickness called gray matter. Neurons may be interconnected either locally (to other neighboring neurons) or remotely (to other distant neurons). Long-distance connections depend on white matter, that is, long myelinated axons beneath the gray matter which link cells within or across lobes, or even across the hemispheres. White matter also affords connections between neighboring cells. All neurons possess similar features (e.g., they all consist of a cell body, an axon branching into output terminals, and multiple dendrites with input spines). However, there are different types of neurons, which can be classified according to their form and function. A detailed account of the physiological and molecular aspects of neurons falls outside the scope of this chapter, but it is worth mentioning that there are subtle differences in the distributions of cell types from one cortical area to another. The first scientist to draw a map of such differential distributions was Korbinian Brodmann (1909), who identified 52 cytoarchitectonic areas within each hemisphere. Such areas have become known as Brodmann areas (BAs). Figure 3.3 depicts several of them. 8

4

9

6

5 1,2,3

7

46 39

40

10 11

47

44

45

43

41 21

38

19

42 22

18 37

17

20

Figure 3.3  A depiction of some Brodmann areas. The figure shows a number of cytoarchitectonic regions over the left hemisphere.

As seen in Figure 3.3, BAs can be mapped throughout neocortical (and also subcortical) regions. They enable us to conceive of anatomical structures in terms of groupings of cell types. As they are typically more fine-grained in their reference than gross anatomical landmarks, they are useful to identify precisely which regions are crucial for a specific cognitive system or function. (Note, however, that there is no one-to-one relation between BAs and cognitive domains.) Table 3.1 lists some BAs which prove important for different linguistic processes.

80 The Neurocognition of Translation and Interpreting

Table 3.1  Brodmann areas and rough neuroanatomical correspondences. Brodmann area

Approximate neuroanatomical location

BA 44, BA 45 BA 4 BA 41 BA 42, BA 22 BA 1, BA 3 BA 17

Inferior frontal gyrus (Broca’s area) Frontal lobe: precentral gyrus (primary motor cortex) Superior temporal gyrus (primary auditory cortex) Posterior portion of the superior temporal gyrus Parietal lobe: postcentral gyrus (primary somatosensory cortex) Posterior occipital lobe

3.2.2 Some language-related subcortical structures Subcortical structures are varied and complex. While a full treatment of them escapes the scope of this chapter, it is worth mentioning three of them (the basal ganglia, the hippocampus, and the cerebellum) because of their role in some functions discussed elsewhere in this book. Such structures are illustrated in Figure 3.4. Basal ganglia

Hippocampus Cerebellum

Figure 3.4  Some subcortical structures involved in language processing. The figure illustrates the location of the basal ganglia, the hippocampus, and the cerebellum.

The basal ganglia are composed of various parts, such as the neostriatum (caudate nucleus and putamen), the globus pallidus, and the substantia nigra. While each of these structures subserves different functions, the basal ganglia, as a whole, are implicated in body posture, planning, motor coordination, and learning of sequential and hierarchical information patterns. As regards language, they are critically involved in syntax (Birba et al. 2017), pragmatics (Holtgraves and McNamara 2010; Monetta and Pell 2007), verbal fluency (Raskin, Sliwinski, and Borod 1992), and action semantics (Bak 2013; Birba et al. 2017). The hippocampus is included in the vast medial temporal lobe network. It is profusely connected to varied locations within the neocortex. One of its main functions is to regulate the exchange of signals between cognitive and emotional mechanisms distributed throughout cortical and subcortical regions. Crucially, the hippocampus is involved in long-term memory, as it proves indispensable for the encoding and mnesic consolidation of new information. In the language



Chapter 3.  Prolegomena to the translating and interpreting brain 81

domain, the hippocampus and its connections with the temporal lobe are crucial for lexico-semantic processing (Ullman 2001a, 2004). Posterior to the pons lies the cerebellum. This very complex structure represents just 10% of the brain’s overall volume, but it contains more than half of its neurons. Its main function is to integrate sensory and motor signals for regulating perceptual input and coordinating bodily action. Thus, the cerebellum is critical for controlling and sequencing fine motor movements, such as those needed to play the piano or shuffle cards. More generally, the cerebellum has been implicated in searching and retrieving conceptual information for subsequent processing in other brain areas. With respect to language, it appears to contribute to the mapping of lexico-semantic patterns onto syntactic structures (Ullman 2001a, 2004), as well as verbal fluency and sequencing skills (Ardila, Bernal, and Rosselli 2015). 3.2.3 Two key language-related networks Beyond the contributions of individual regions, several language processes rely on two broad cortico-subcortical networks. These involve frontostriatal pathways and their projections, on the one hand, and temporo-parietal regions and their connections with diverse areas, on the other. Both systems are illustrated in Figure 3.5. Frontostriatal networks originate in prefrontal regions and project to basal ganglia structures. These include the caudate nucleus and the putamen, followed by the globus pallidus, the substantia nigra, and then the thalamus, which includes feedback loops projecting back to the prefrontal cortex. In particular, this circuit includes connections from Broca’s area to the putamen, which continue towards the globus pallidus and the thalamus, finally leading back to Broca’s area (Ford et al. 2013). This system possesses serotoninergic, noradrenergic, and cholinergic inputs that modulate motor functions and higher-level processes (Tekin and Cummings 2002). Although frontostriatal networks participate in a myriad of linguistic operations, they are particularly critical for syntactic, phono-articulatory, and semantic processes (see Section 3.3.1.2). For their part, the temporal lobes possess two major sulci, parallel to the Sylvian fissure, which represent the anatomical boundary between the superior, medial, and inferior temporal gyri. The medial temporal gyrus, in particular, is part of a cortico-subcortical system implicated in declarative memory functions, comprising the hippocampus, the dentate gyrus, the subicular complex, and the perirhinal, entorhinal, and parahippocampal cortices. These regions, in turn, connect directly and indirectly with the angular gyrus (located between the intraparietal sulcus and the horizontal branch of the Sylvian fissure) and the supramarginal gyrus (in the inferior parietal lobe). All such areas are profusely connected with one another and

82

The Neurocognition of Translation and Interpreting

A. Frontostriatal pathways and linguistically important connections

B. Temporo-parietal pathways and linguistically important connections

Figure 3.5  Main frontostriatal and temporo-parietal pathways implicated in verbal processing. The figure illustrates some of the key linguistically relevant circuits comprised by frontostriatal (panel A) and temporo-parietal (panel B) pathways, alongside their connections with other regions.

with frontal regions, mainly via two pathways. The ventral pathway, mediated by the uncinated fasciculus, links the primary auditory cortex with anterior temporal sites. The dorsal pathway connects temporal areas with the inferior parietal lobe. Although these circuits participate to different degrees in basically every linguistic process, their contributions are particularly critical for phonological and semantic processes (see Section 3.3.1.3). 3.2.4 Neurons and synapses Neurocognitive activity results from complex processes occurring both within and between neurons. Each neuron is composed of three main parts: a soma (or cell body), multiple dendrites, and a single axon typically possessing many branches (Figure 3.6).



Chapter 3.  Prolegomena to the translating and interpreting brain 83

Axon collaterals

Axon

Nucleus Dendrite

Soma (cell body) Apical dendrite

Figure 3.6  Structural components of a neuron. The figure illustrates the soma and nucleus of a pyramidal cell, as well as its axon, some axon collaterals, and some dendrites.

The soma, which contains the neuron’s nucleus, stores genetic material and produces proteins and other molecules needed for the cell’s survival. The dendrites and the axon are nerve fibers enabling inter-neuronal communication. Dendrites are afferent (input) pathways, meaning that they receive signals from other neurons. The axon is an efferent (output) pathway, as it sends signals to other neurons via its multiple collaterals. Axons can reach both neighboring and distant cells; some of them, in fact, are several centimeters long. Due to their structure, neurons possess two key properties: convergence (each neuron may receive input from multiple efferent neurons) and divergence (each neuron may send output to multiple afferent neurons). The point of contact between any two neurons is called a synapse. Each neuron in the brain establishes synapses with thousands of neurons in various brain locations. According to DeFelipe and Farinas (1992), a typical pyramidal neuron has roughly fifty thousand efferent synapses and fifty thousand afferent synapses. This profusion of connections is largely responsible for the complexity of human cognition. In general terms, communication between these cells occurs when the axon of an efferent neuron releases specific molecules of a neurotransmitter (at the presynaptic membrane), which flow across the synaptic cleft (typically about 20 nm) and through the postsynaptic membrane on a dendrite or cell body of the afferent neuron. This process is illustrated in Figure 3.7. Typically, receiving neurotransmitters from a single presynaptic terminal is not enough stimulation for the input neuron to respond. However, when neurotransmitters arrive from a sufficient number of efferent connections at the same time (or within the lapse of a few milliseconds), the input neuron generates an

84

The Neurocognition of Translation and Interpreting

Neurotransmitter Axon terminal (end of axon collateral) Synaptic vesicle

Presynaptic membrane Dendrite

Synaptic cleft Postsynaptic membrane

Figure 3.7  A rough illustration of a synapse. The figure depicts the process whereby neurotransmitters are released by the synaptic vesicle through the presynaptic membrane, across the synaptic cleft, and into the receptors in the postsynaptic membrane.

action potential, that is, an electrical signal which propagates along the axon in order to promote activity in other receiving neurons. Thus, the input neuron now acts as an output neuron. In sum, inter-neuronal communication follows an electrical-chemical-electrical sequence: electrical signals running along axons result in the release of neurotransmitters (chemical substances), which, if certain conditions apply, will enter the afferent neuron and trigger a new electrical signal. Connections between neurons can be either excitatory or inhibitory, with each connection type involving different neurotransmitters. The former contribute to the activation of the receiving neuron and can be established between neighboring or distant cells. Conversely, inhibitory connections tend to mitigate activity within the receiving neuron. The latter are typically established between neighboring cells. The main neuron types featuring excitatory connections are pyramidal and spiny stellate cells. Neurons with inhibitory properties come in various forms. In the neocortex, some of the most typical ones are large basket cells, double bouquet cells, chandelier cells, and smooth stellate cells. There are two types of inhibitory connections. Axoaxonic inhibitory connections are established on the initial segment of a neighboring axon, whereas axosomatic inhibitory connections are established on the soma of a neighboring cell. At the point where the soma meets the axon, every neuron features a structure known as the axon hillock. Signals entering the receiving neuron are here summated, and if their aggregated intensity increases rapidly or if they exceed the neuron’s activation threshold, an action potential is triggered (that is, the neuron fires).



Chapter 3.  Prolegomena to the translating and interpreting brain 85

By contrast, if the overall voltage falls below the threshold, the neuron remains inactive. These processes reflect the main difference between excitatory and inhibitory connections: while the former increase the voltage of the receiving neuron, thus favoring its activation, the latter reduce the voltage and tend to deactivate it. Action potentials are all-or-none responses, that is, either they are generated or they are not. There is no such thing as a partial action potential. Also, the action potential does not decrease as it propagates from the axon to the multiple synapses located next to each of the axon’s collaterals. Conversely, postsynaptic potentials (generated in the receiving neuron’s postsynaptic membrane and capable of increasing or reducing overall voltage) are analog signals which become gradually attenuated as a function of the time elapsed since their release and the distance covered across the corresponding dendrite. The strength of a synaptic connection may change through time. Specifically, inter-neuronal links are strengthened if the synapse is frequently used. This follows from the so-called Hebbian principle, which states that “any two cells or systems of cells that are repeatedly active at the same time will tend to become ‘associated’, so that activity in one facilitates activity in the other” (Hebb 1949: 70). At the same time, neuronal connectivity depends on the opposite process: independent (antiphasic) activation of any pair of cells or cell systems will weaken their reciprocal links (Tsumoto 1992). Such modifications in connectivity strength result from varied biochemical and even structural changes in the neurons involved (Kandel 2013), such as the growth of dendritic spines (Braitenberg and Schüz 1998). In sum, any connection between two neurons is reinforced by their joint activity and debilitated by their independent firing. Another important feature of the neocortex is that connections between areas tend to be reciprocal. If there is a group of neurons sending signals from region A to region B, more often than not there will be another group of neurons sending signals from region B to region A. However, note that the groups of neurons involved in signal transmission in each direction are not the same, so that they are subject to different processes of activation, inhibition, strengthening, and weakening over time (Pandya and Yeterian 1985; Young, Scannell, and Burns 1995). 3.2.5 Cognitive processing as neuronal teamwork No neuron can support a cognitive operation on its own. A single cell is insufficiently efficient to process complex information by itself, as its inner processes may be rendered noisy by other signals around it (Pulvermüller 2002). Moreover, cellular deterioration and death may occur at such rapid rates that if each knowledge pattern depended on a single neuron, it would be practically impossible to retain information for long periods.

86 The Neurocognition of Translation and Interpreting

It has been proposed that the minimal processing module in the neocortex is the cortical minicolumn (Mountcastle 1998). A minicolumn is a group of roughly 100 neurons aligned vertically throughout the six layers of gray matter and operating together as a functional unit (Arbib, Érdi, and Szentágothai 1998; Mountcastle 1998). Each minicolumn is approximately 4 mm in length and 35 µm (microns or micrometers) in diameter. Some 70% of neurons in a typical minicolumn are pyramidal, meaning that they possess excitatory connections. In comparison, a substantial portion of the remaining neurons possess inhibitory connections. Hence, a minicolumn is capable of sending both excitatory and inhibitory signals to other minicolumns. The evidence about the structure and function of minicolumns comes mainly from studies on cats, monkeys, and rats. Hubel and Wiesel (1962, 1977) showed that visual cortex nodes in these species are implemented as hierarchically organized cortical columns: each successive level integrates features processed by the immediately lower level so as to then send activation to higher layers. Therefore, higher hierarchical levels in a cortical structure subserve more abstract cognitive relationships. For his own part, Mountcastle (1998) reached similar conclusions by examining the primary somatosensory and auditory cortices of cats and monkeys. Such brain areas, and presumably other regions, are also organized in terms of columns. Since the cortices of cats and monkeys are similar to those in our own skulls, neuroscientists sometimes extrapolate findings from the former and apply them to their models of the human brain. The topological and cellular similarities between the cortices of these species and our own are such that it is not unusual for neuroscientists to work on the assumption that people’s linguistic, perceptual, and conceptual systems are also organized in hierarchical networks of cortical minicolumns. It requires an additional extrapolation, which might perhaps seem extreme, to derive hypotheses about language processing from research on the brains of cats or monkeys. Indeed, the specific complexities and functional properties of language are simply absent in other species. However, just as cats and monkeys possess cortical columns specialized for processes so specific that they fire upon stimulation of a single finger, so too the human cortex possesses minicolumns and even individual neurons which fire selectively or preferentially in the presence of certain visual stimuli. This has been demonstrated by Quian Quiroga et al. (2005), whose experiments with epileptic patients showed that a single neuron may distinctly respond to stimuli denoting entities as precise as the American actress Jennifer Aniston, NBA legend Michael Jordan, or the Leaning Tower of Pisa. There are other types of neuronal systems which also work as integrated processing units but are more widely distributed throughout the brain. Functional webs are a case in point. These structures support complex composites of cognitive

Chapter 3.  Prolegomena to the translating and interpreting brain 87



information processed by remote systems – consider, for example, the concept dog, which subsumes unimodal information such as four-legged (a visual percept, mainly subserved by occipital networks), bark (an auditory percept, mainly subserved by temporal networks), and soft fur (a somatosensory percept, mainly subserved by parietal networks). According to Pulvermüller (2002), a functional web is a widely distributed functional microsystem of nerve cells which features a well-defined topography and whose dynamics can be characterized in terms of four activity states: rest, ignition, priming, and reverberation. Rest refers to the lack of (significant) electrochemical activity in a web; ignition is the state attained by a web receiving sufficient activation to make it fire; priming is the effect whereby a web receives a small amount of activation as a result of its connections to other fully active webs; and reverberation refers to the state in which a web retains its internal activity for a few seconds. That cognitive information is supported by cortical minicolumns and functional webs implies that not every neuron participates in every process. This fact has an important consequence: since the number of neurons in the brain at any point in life is finite, each cognitive process necessarily depends on a finite number of neurons, which needs not always be the same. Note, too, that a neuron or a column may participate in different neural networks and, consequently, in different cognitive processes. Nevertheless, the particular population of neurons subserving a given operation is partially specific, despite possible overlap among different neurocognitive networks. 3.3

The verbal brain

Verbal processing involves several interrelated but identifiable functions with relatively well-defined neural correlates. These include specific networks and electrophysiological mechanisms implicated in diverse linguistic and pragmatic domains, as described below. 3.3.1 Tell me where: The functional neuroanatomy of language 3.3.1.1 A tale of two hemispheres One of the most robust findings in neuroscience is that each cerebral hemisphere plays differential roles in language processing (Josse and Tzourio-Mazoyer 2004). While the LH is dominant for basic language functions (e.g., phonological and morphosyntactic processing), the RH is characterized by less critical contributions to those functions and includes key substrates of pragmatic domains.

88

The Neurocognition of Translation and Interpreting

Classical language areas are located in the vicinity of the Sylvian fissure. These regions are mainly comprised by Broca’s area (encompassing BAs 44 and 45, as well as portions of BAs 6, 47, 46, 43, 8, and 9) and Wernicke’s area (covering BAs 42, 22, and part of 21) in the LH (Amunts 2008). This has been corroborated via multiple approaches. Pioneering reports by Paul Broca (1824–1880) and Carl Wernicke (1848–1905) showed that lesions to left inferior frontal and superior temporal areas were linked to differential impairments in language production and comprehension, respectively (Broca 1861a; Wernicke 1874) – Figure 3.8. Compatibly, intraoperative studies have long demonstrated that direct cortical stimulation of left perisylvian regions disturbs linguistic processing in roughly 95% of cases, but that such difficulties occur only rarely upon stimulation of the RH (Penfield and Roberts 2014). Similarly, basic linguistic processes become transiently compromised when the LH is selectively anesthetized, whereas this is seldom the case after RH anesthetization (Stemmer and Whitaker 2008). These functional asymmetries have been confirmed by multiple neuroimaging studies showing that word reading and repetition, synonym and rhyme generation, and sentence processing mainly engage left perisylvian networks, alongside less pronounced activity increases in extrasylvian and RH regions (Vigneau et al. 2006). Taken together, these findings indicate that basic linguistic functions are critically (though not exclusively) related to LH regions. Note, however, that this pattern of organization is estimated to hold for 90% of right-handed and 60% of left-handed individuals, whereas the remainder present right or bilateral dominance for language (Mazoyer et al. 2016). 0

10

20

30

Figure 3.8  MRI images of patient Leborgne’s brain. Leborgne was one of two patients first reported by Broca in support of a critical relation between the left inferior frontal gyrus and language production mechanisms. The damage produced by the stroke is indicated in red. Reproduced with permission from Thiebaut de Schotten et al. (2015).



Chapter 3.  Prolegomena to the translating and interpreting brain 89

That being said, the RH also supports verbal abilities in left-dominant subjects. In fact, rudimentary linguistic skills remain spared even after full left hemispherectomy, and verbal deficits following LH damage typically worsen if a second lesion affects the RH (Caplan 1987). More particularly, the RH includes key hubs of vast networks subserving verbal communication at large (Stemmer and Whitaker 2008). For example, it participates critically in prosodic processes, including the production and comprehension of affective intonation, among other suprasegmental domains (Lindell 2006). Moreover, it has also been implicated in inferential processes (Stemmer, Giroux, and Joanette 1994) and the understanding of figurative language and indirect speech (Kaplan et al. 1990; Prat, Mason, and Just 2012) – in cooperation with left perisylvian regions (Bohrn, Altmann, and Jacobs 2012). Also, the RH plays major roles in integrating and synthesizing information distributed throughout unfolding pieces of discourse (Benowitz, Moya, and Levine 1990; Joanette et al. 2008). 3.3.1.2 Functions of frontostriatal networks and their connections As stated in Section 3.2.3, frontostriatal pathways and their connections are key substrates of several linguistic functions, crucially including syntactic, phono-articulatory, and semantic processes. Frontostriatal circuits are critical for the acquisition and implementation of structured cognitive routines, such as those supporting the combination of morphemes into words and of words into sentences (Ullman 2001a). In fact, their damage is consistently associated with disproportionate deficits in such functions (Birba et al. 2017; Grodzinsky 2000), and the activity levels of some of their hubs become differentially greater during processing of simple and complex sentences (Grodzinsky and Friederici 2006). In particular, BA 44, in the posterior part of Broca’s area, contributes critically to the processing of hierarchical syntactic patterns (Nuñez et al. 2011). Importantly, frontostriatal activity during syntactic operations is supported by posterior regions. Indeed, the anterior temporal lobe (including BAs 21 and 22 as well as the posterior part of BA 38) supports grammatical processing during sentence comprehension and is coactivated with frontal circuits in the face of syntactically anomalous sentences (Humphries et al. 2005) – Figure 3.9. As regards speech, Broca’s area is a core substrate for phonological information to reach motor circuits, allowing for its articulatory implementation (Indefrey and Levelt 2004; Vigneau et al. 2006). In fact, the disruption of frontostriatal mechanisms is linked to production deficits in various neurodegenerative disorders, such as Parkinson’s and Huntington’s disease (Hartelius et al. 2003; Ho, Iansek, and Bradshaw 1999). Also, particular portions of these networks participate in higher-level verbal processes. Broca’s area, for example, is significantly more engaged

90 The Neurocognition of Translation and Interpreting

Syntax > Rest

8.0 R

L

2.3

Figure 3.9  Activation patterns associated with syntactic processing. In the present task, subjects had to decide whether two consecutive sentences with different constructions (e.g., active vs. passive voice) had the same meaning. Hemodynamic increases were observed in the three left frontal gyri, the left temporo-parietal cortex, the bilateral primary auditory cortex, and the right cerebellum. L: left; R: right. Reproduced with permission from Nuñez et al. (2011).

by word repetition than passive listening, as is the case with posterior regions, such as the left supramarginal gyrus (Hautzel et al. 2002). More specifically, the dorsal part of the left pars triangularis supports verbal WM processes – e.g., during tasks requiring the retention of word lists (Hautzel et al. 2002). In addition, frontostriatal pathways are variously implicated in semantic processes, which actually involve widely distributed networks across the whole brain. Some of their hubs (e.g., BAs 45 and 47) participate in multimodal conceptual processes – i.e., those that prove general to diverse semantic categories (Amunts et al. 2004; Bookheimer 2007; Newman, Ikuta, and Burns 2010). In combination with temporal structures, these regions integrate a global semantic processing system (Patterson, Nestor, and Rogers 2007). Also, frontostriatal networks are critical for grounding specific semantic categories (Pulvermüller 2013). In particular, they are distinctively associated with processing of verbs and, more specifically, action verbs – words denoting bodily movement. Indeed, frontostriatal motor circuits are



Chapter 3.  Prolegomena to the translating and interpreting brain 91

significantly activated in the presence of these verbs (Pulvermüller 2013). Moreover, their stimulation selectively modulates processing of such a category (Shebani and Pulvermüller 2013) and their atrophy leads to differential impairments of action (as opposed to non-action) words (Birba et al. 2017). 3.3.1.3 Functions of temporo-parietal regions and their connections Temporo-parietal regions, and the vast networks they are a part of, are engaged to a greater or lesser extent by virtually any linguistic task. However, their role is distinctively decisive for phonological and semantic operations. In healthy subjects, phonological stimuli elicit activation peaks in the posterior portion of the left superior temporal gyrus and in the vicinity of Wernicke’s area. In fact, the more complex the stimulus, the more extended the activation pattern over anterior and ventral temporal regions (Indefrey and Levelt 2004; Vigneau et al. 2006). The left superior temporal gyrus, in fact, has been proposed to afford interfaces between phonological sequences and corresponding lexical and articulatory patterns (Hickok and Poeppel 2007). The superior temporal gyrus would be critical to mediate links between a word’s auditory form and relevant semantic information. No less important are the contributions of the angular gyrus, which is involved in verbal repetition, object naming, reading, and writing (Vigneau et al. 2006). Also, the inferior parietal lobe, in coordination with Broca’s area and the premotor cortex, subserves the transient maintenance of verbal information, prior to speech planning and articulation (Vigneau et al. 2006). Thus, it represents a key substrate of phonological WM (Ravizza et al. 2004). Temporal regions also constitute the main neural basis of semantic memory, the system that organizes knowledge about words, their meanings, their referents, and their interrelations (Tulving 1986). In particular, the anterior temporal lobe acts as a central conceptual hub for the integration of information from different sensorimotor modalities (Patterson, Nestor, and Rogers 2007). This form of multimodal synergy is crucial, among other things, to establish conceptual generalizations (Ralph et al. 2017). Moreover, diverse temporal, parietal, and hippocampal sites are differentially engaged in lexico-semantic tasks (Newman et al. 2001), and their lesions impair the comprehension and production of words related to various sensory modalities (Ardila, Bernal, and Rosselli 2015). Taken together, this evidence supports the view that the learning and use of words and their meanings would mainly depend on the integrity of the so-called medial temporal lobe system (dentate gyrus, subicular complex, hippocampus), with distinct contributions from other temporal and temporo-parietal regions (Ullman 2001a) – Figure 3.10.

92

The Neurocognition of Translation and Interpreting

Figure 3.10  Convergent evidence for the distributed neural network underpinning semantic cognition, including findings from multiple studies. The figure shows (i) lesions from patients with semantic dementia (blue) and semantic aphasia (pink); (ii) activation peaks in the anterior temporal lobe identified via distortion-corrected fMRI (purple circle); (iii) fMRI peaks related to semantic manipulations (green circles); and (iv) sites yielding significant semantic effects upon transcranial magnetic stimulation (yellow circles). ATL: anterior temporal lobe; BA: Brodmann area; IFS: interior frontal sulcus; IPL: inferior parietal lobe; IPS: intraparietal sulcus; pMTG: posterior middle temporal gyrus. Reproduced with permission from Jefferies (2013).

3.3.2 Electrified words: The neurophysiology of language The linguistic processes mentioned above, just as every other cognitive event, rely on the diverse neurophysiological operations that are modulated in fractions of a second. These can be tapped through high-temporal-resolution techniques, such as those derived from EEG recordings (Chapter 2, Section 2.7.1.2). 3.3.2.1 ERP signatures ERP evidence has revealed the time course of numerous linguistic mechanisms. For example, relative to infrequent words, frequent items modulate the P1 component over occipital regions and the N1 component over distributed scalp locations (Scott et al. 2009). In addition, the frequency of phonological, as opposed to orthographic, syllables is reflected in the amplitude of the attention-sensitive P200 component (Kwon, Lee, and Nam 2011). Phonemic discrimination is also indexed by the so-called mismatch negativity, a negative deflection peaking at 100–250 ms after stimulus onset, mainly over fronto-central regions (Cheour et al. 1998).



Chapter 3.  Prolegomena to the translating and interpreting brain 93

This ERP also exhibits greater amplitude for real words than pseudowords, and for syntactically incongruent than congruent word pairs, suggesting a role in the integration of phonological and semantic information and in early grammatical processes (Pulvermüller and Shtyrov 2006). The latter operations are also indexed by the early left anterior negativity, observed between 100 and 300 ms over left frontal and anterior temporal topographies (Friederici 2004). Moreover, efforts to integrate incoming lexico-semantic information with the preceding textual context are captured by the N400, a negative waveform distributed over temporal and prefrontal scalp sites (Friederici 2004; Kutas and Federmeier 2011) – see Chapter 2, Section 2.7.1.2.1 and Figure 2.7. This component also discriminates between several other lexical dimensions, such as concreteness, predictability, and semantic proximity, even when stimuli are unattended (Kutas and Federmeier 2011). Finally, structural re-analysis processes are signaled by modulations of the P600, a component with a mostly centro-parietal and frontal distribution whose amplitude is proportional to syntactic integration difficulty (Friederici 2004; Kaan et al. 2000). Taken together, these findings constitute a firm reference point to interpret ERP modulations in IR research. 3.3.2.2 Oscillatory signatures Additional evidence has begun to illuminate the oscillatory correlates of linguistic operations. Although basically all frequency ranges are sensitive to diverse processes within the language domain, the most systematic findings to date correspond to the theta (4–8 Hz), alpha (8–13 Hz), and beta (13–30 Hz) bands – see Chapter 2, Section 2.7.1.2.2 and Figure 2.8. Theta power is augmented during word recognition and retrieval. Also, during reading of correct sentences, theta power increases word by word, which would reflect successively greater demands on WM (Bastiaansen, van Berkum, and Hagoort 2002). Furthermore, while correct sentence reading is also related to cumulative increases of alpha power, this band exhibits reduced power in the face of syntactically or semantically incorrect sentences – an effect that is mirrored by beta-band modulations (Kielar et al. 2014). Once again, the effects detected in these frequency bands can constrain interpretations of oscillatory dynamics in IR studies. 3.3.2.3 Functional connectivity signatures Also, linguistic processes involve joint activity patterns across different brain regions, as revealed by functional connectivity research. Studies on sentence processing show that inter-hemispheric interaction in anterior and posterior regions increases in proportion to stimulus complexity – see Chapter 2, Section 2.7.1.2.3, and Figure 2.9. Furthermore, the comprehension of variously complex sentences is associated with the lineal synchronization of activity across wide brain networks

94 The Neurocognition of Translation and Interpreting

in the theta, beta, and gamma bands (Weiss et al. 2005). Theta modulations would be related to memory processes, whereas those in beta and gamma frequencies would reflect semantic-pragmatic integration and attentional effort, respectively. The simultaneity of those effects further suggests that sentence comprehension depends on different but parallel cognitive operations. Sentence processing also entails non-lineal connectivity dynamics. For instance, semantic incongruences are related to global theta desynchronization at roughly 285 ms, whereas physical incompatibilities (e.g., manipulations of the color of written words relative to their content) increase global alpha synchrony at roughly 200 ms. Although the interpretation of these results is not clear, both precede corresponding modulations observed in ERP studies. Moreover, the lag between synchronization effects and associated ERP patterns correlates with the complexity of the underlying process. Therefore, connectivity metrics reveal temporal dynamics which complement and extend the possibilities of ERPs as neurophysiological markers of language processing, in general (Allefeld, Frisch, and Schlesewsky 2005), and IR mechanisms, in particular (García, Mikulan, and Ibáñez 2016). 3.4

It takes two to tango: The prerequisite of bilingualism

Of necessity, IR processes take place in a ‘bilingual brain’. In contemporary neurolinguistic jargon, such a label applies to roughly one half (Grosjean 1994) of the world’s population, given that bilingualism is acknowledged as an attribute of any person who frequently uses two languages or dialects and who can choose to communicate in one or the other according to the circumstances, irrespective of proficiency or age of acquisition, among other variables (García, Manoiloff, and Wagner 2016). Across this vast population, the bilingual experience is associated with particular neurocognitive phenomena that differ from those encountered in monolingual systems, as seen in both linguistic and executive domains. 3.4.1 Linguistic mechanisms in the bilingual brain Broadly speaking, linguistic processes in bilinguals engage the same gross regions and mechanisms detailed in Section 3.3. However, a number of specifications are in order. Whereas verbal functions in both languages mainly recruit anterior and posterior perisylvian sites, each language involves partially segregated networks within temporal, parietal, and frontal regions (e.g., Chee, Soon, and Lee 2003; Klein et al. 2006). Indeed, brain lesions across such areas may lead to selective or differential deficits in L1 and L2 (Paradis 2004), and the same can be transiently observed following direct cortical stimulation (Lucas, McKhann, and Ojemann 2004; Ojemann



Chapter 3.  Prolegomena to the translating and interpreting brain 95

and Whitaker 1978). Also, activation patterns for L1 tasks are more consistent across individuals than they are for L2 tasks (Dehaene et al. 1997), arguably reflecting the greater variability in learning modes and processing strategies proper to L2s. Of note, some linguistic mechanisms seem to differ systematically between bilinguals and monolinguals. For instance, the former have been reported to possess more gray matter density in the inferior parietal lobule (Richardson and Price 2009) and the left inferior temporal gyrus (García-Pentón et al. 2016) – this difference being positively correlated with overall vocabulary. Bilinguals also exhibit greater fronto-posterior connectivity as well as more white matter integrity in tracts of the corpus callosum, the inferior fronto-occipital fasciculus, and the superior longitudinal fasciculus (García-Pentón et al. 2016). Interestingly, however, bilinguals tend to possess less developed receptive vocabularies in L1 and L2 than monolingual users of the same languages, and they evince reduced verbal fluency across many age ranges (Bialystok 2009; Bialystok, Craik, and Luk 2012). Moreover, use-related factors entail neurocognitive changes in relevant systems. For instance, relative to bilinguals with low levels of proficiency or L2 exposure, those with higher levels manifest activation changes in key linguistic hubs, such as the middle temporal gyrus (Grant, Fang, and Li 2015). Also, they are characterized by greater neural overlap between both languages (Sebastian, Laird, and Kiran 2011) as well as increased recruitment of right prefrontal regions for word production (Videsott et al. 2010) and frontostriatal circuits for morphosyntactic processing (Paradis 2009; Ullman 2001b). At the same time, various linguistic processes are significantly more efficient in high- than low-proficiency bilinguals (e.g., Guasch et al. 2008; Sunderman and Kroll 2006), with each group recruiting different cognitive routes to perform identical L1 and L2 tasks (Guasch et al. 2008; Talamas, Kroll, and Dufour 1999). These particularities represent part of the neurocognitive foundations on which IR processes take place. 3.4.2 Executive mechanisms in the bilingual brain Linguistic operations in bilinguals are in constant interplay with executive mechanisms (Miyake and Friedman 2012; Miyake et al. 2000; Zillmer and Spiers 2001), such as those described in Chapter 1 (Section 1.2.5.4). These are indispensable to control when and how each language is to be used depending on contextual cues. The neural networks supporting language selection, inhibition, and switching are widely distributed across the brain, with key nodes in the basal ganglia and the prefrontal, anterior cingulate, and inferior parietal cortices (Abutalebi and Green 2007, 2008; Luk et al. 2011) – Figure 3.11. Lesions to some of these areas (in particular, the frontal and cingulate cortices) disturb language control in bilinguals, as seen in cases of pathological switching (Fabbro, Skrap, and Aglioti 2000).

96 The Neurocognition of Translation and Interpreting

Prefrontal cortex • Executive functions • Decision-making • Response selection • Response inhibition • Working memory

Anterior cingulate cortex • Attention • Conflict monitoring • Error detection

Basal ganglia • Language selection • Set switching • Language planning • Lexical selection

Inferior parietal lobule • Maintenance of representations • Working memory

Figure 3.11  Schematic depiction (on a single axial slice) of the main neural hubs subserving executive control and language production in bilinguals. Cognitive control emerges from the integration of separable neural systems, including the anterior cingulate cortex, the basal ganglia, the inferior parietal lobule, and, most prominently, the prefrontal cortex. Each of these systems is responsible for distinct aspects of cognitive control as outlined in the callout boxes of the figure. Reproduced with permission from Abutalebi and Green (2007).

Arguably due to higher control demands during verbal communication, bilinguals have been reported to feature neurocognitive advantages in various executive domains. Although the evidence is not fully systematic (Dunabeitia and Carreiras 2015; Paap, Johnson, and Sawi 2015), relevant hubs in cortical and subcortical regions exhibit structural differences between bilinguals and monolinguals (García-Pentón et al. 2016; Richardson and Price 2009). Compatibly, numerous dimensions of WM, inhibitory mechanisms, mental-set shifting, and updating skills appear to be enhanced by the bilingual experience (Bialystok 2009; Bialystok, Craik, and Luk 2012). Moreover, some of the above effects increase in proportion to the subjects’ bilingual skills, as diverse executive processes appear to be more efficient in high- than low-proficiency subjects (Linck et al. 2014; Linck and Weiss 2015). In fact, as shown by meta-analytic evidence, WM skills correlate positively with L2 proficiency and

Chapter 3.  Prolegomena to the translating and interpreting brain 97



L2 processing outcomes (Linck et al. 2014). The age of L2 acquisition and the degree of L2 exposure also play a role in this regard. For example, relative to monolinguals and simultaneous bilinguals, individuals exposed to an L2 since roughly age 4 exhibit greater recruitment of prefrontal executive networks during linguistic tasks (Jasinska and Petitto 2013). Also, executive performance is better for bilinguals with higher levels of L2 exposure (Bosma et al. 2017). In sum, the specific configurations traceable across bilingual brains constitute the neurocognitive framework in which IR processes can be deployed. 3.5

In a nutshell

The notions covered above constitute indispensable anchoring points for those who began reading this book without prior knowledge of neurology and neurolinguistics. Moreover, they are fundamental to prevent constructional scandals like the one described at the outset: though not directly related to IR mechanisms per se, they capture essential constraints for erecting a workable conceptual edifice of such phenomena. A neurocognitive model of IR must be compatible with the basics of neurology and it must contemplate the existence of specialized multidimensional mechanisms for different levels of linguistic processing (e.g., phonological, lexical, semantic, syntactic), partially independent networks for L1 and L2, profuse interactions between verbal and executive mechanisms, and various particularities proper to bilingualism. These and other tenets described throughout the present chapter must be assumed to lie at the core of the contents that will be addressed next. So, without further ado, let us finally begin our in-depth exploration of the neurocognition of translation and interpreting.

Chapter 4

Building up from breakdown

4.1

Lessons from lesions

If every now and then you find yourself falling down the YouTube rabbit hole, you may come across a channel titled ‘What’s Inside?’. The oddly captivating initiative consists of videos in which a father and his child break random objects to examine their contents. One of the clips focuses on an ‘Apple Helicopter’, a fruit-shaped flying toy equipped with a fast-spinning plastic rotor, a motion-detection device, and a music player. Hatchet in hand, the father cuts it open and reveals a series of components (a motor, a speaker, a chip, a sensor, a battery) interconnected by several cables. During the exploration, he presses a switch that turns on the motor and the music. Then, when the boy rips two white cables linked to the speaker, the music stops, but the motor keeps running. The result is quite fulfilling, as the tune was decidedly annoying. At least two conclusions can be derived from this household experiment. First, each of the components, though variously interrelated, plays a particular role in different functions. Second, some of those mechanisms (e.g., the motor) can remain operative even when others (e.g., the white cables and the speaker) are disrupted. As a corollary, we may deduce, the toy’s flying and musical capabilities are autonomous from each other and dependent on distinct circuits, although both are jointly activated whenever the ‘on’ button is pressed in a normally assembled unit. A somewhat similar procedure has long become the backbone of neurocognitive research. Of course, the human brain is infinitely more complex than a nine-dollar gadget (among other things, because there is no one-to-one relation between its structures and functions), and contemporary scientists would hardly condone smashing a patient’s head on purpose. Yet, the fact is that much of our current knowledge about neurocognition comes from the investigation of dysfunctional systems – in particular, lesion models (see Chapter 2, Section 2.6). To recap, these are “natural experiments” in which, for accidental or pathological reasons, a person suffers from partially circumscribed brain damage and manifests specific cognitive alterations. Crucially, the study of such patients allows establishing single dissociations (i.e., observations from individual cases in which damage to area A impairs function X but not Y) and double dissociations (i.e., scenarios in which the previous finding is refined by the additional demonstration

100 The Neurocognition of Translation and Interpreting

that damage to area B compromises function Y but not X). Put succinctly, these patterns would suggest that functions X and Y are subserved by partially independent neural systems and that they are differentially related to areas A and B, even if both of them seamlessly interact in non-dysfunctional cognition (Damasio 1994; Dunn and Kirsner 2003). This is a foundational approach in the biological study of the mind, yielding invaluable reports of anatomo-clinical correlations that helped molding neuroscience as practiced today (Finger 1994). In fact, it is thanks to the study of individual brain-lesioned patients, such as Louis Victor Leborgne, Phineas Gage, and Henry Gustave Molaison, that the world first learned about the crucial role of the inferior frontal gyrus in speech production (Broca 1861a, 1861b), the orbitofrontal cortex in social behavior (Damasio 1994; Harlow 1868), and the medial temporal lobe in declarative memory (Dossani, Missios, and Nanda 2015; Scoville and Milner 1957), respectively. Though originally couched in inadequate localizationist frameworks, these discoveries – and the fine-tunings they have motivated– still constrain modern models of cognition and inspire new cutting-edge research (Thiebaut de Schotten et al. 2015). Above and beyond those seminal cases, the lesion-model approach has afforded insights into the organization of countless neurocognitive systems, including those involved in visual perception, facial recognition, color discrimination, STM, and inhibitory control, to name but a few. More particularly, it has prompted key findings on the bilingual brain. These include insights into the partial autonomy of L1 and L2 subsystems, the differential role of procedural and declarative memory circuits in lexico-semantic and morphosyntactic mechanisms for each language, and the impact of subject-level variables (such as L2 proficiency) on the latter’s neural distribution (Paradis 2004, 2009; Ullman 2001b). Also, and most crucially for our present purposes, this line of research has provided vital data on the neurocognitive basis of IR. As mentioned in Chapter 1 (Section 1.5), translation disorders in brain-damaged bilinguals have been documented for the better part of one century, offering otherwise unattainable insights for cognitive TIS. They can be classified into four types, namely: compulsive translation, inability to translate, paradoxical translation behavior, and translation without comprehension. Each of these neuropathologies entails distinct behavioral manifestations, with partial or complete disruptions of particular IR processes, even when other linguistic domains remain largely or fully unaffected. Evidence of this kind is immensely informative for theory construction in cognitive TIS, as noted over two decades ago by Franco Fabbro and Laura Gran. In a pioneering position paper, titled “Neurolinguistic research in simultaneous interpretation,” the authors emphasized the relevance of the lesion approach for the field, maintaining that

Chapter 4.  Building up from breakdown 101



[…] models hypothesizing subsystems and modules within a general language system are not a product of the fervent imagination of scholars in the field, but rather the interpretation of real phenomena observed in neurological patients showing so-called “dissociations” and “double dissociations”.  (Fabbro and Gran 1997: 14–15; italics in the original)

Unfortunately, despite its prolific and multifarious growth, cognitive TIS has not incorporated this tradition into its mainstream ethos. The primary aim of this chapter is to help reverting such a scenario, distilling the most salient facts across multiple reports and pinpointing their main implications for characterizing the neurocognitive territory in which translation and interpreting take place. Beyond its direct value, the outcome is also fundamental to better interpret the neuroscientific findings detailed in Chapters 5 through 7. 4.2

Disruptions of IR

The evidence presented next comes from twenty-three bilingual/multilingual patients, ranging from 15 to 91 years old. They spoke various combinations of languages. All were either confirmed or presumed to be right-handers. Most of them presented focal lesions resulting in some stable or transient form of aphasia, whereas one suffered from presenile dementia and another one exhibited a sociolinguistic disorder. Different linguistic dysfunctions were reported in each case, including anomia, language mixing, and disfluency. However, other intellectual faculties, such as memory, attention, and visuospatial skills, were virtually intact in all patients, which highlights the partial specificity of their IR deficits. 4.2.1 Compulsive translation Compulsive translation consists in the immediate, involuntary translation of utterances. This behavior is often accompanied by an inability to translate willingly. In some cases, patients have been noted to compulsively translate expressions into a language that was unavailable for spontaneous production in single-language communicative settings. Perecman (1984) has suggested that this disorder may reflect a dysfunction at the conceptual level. However, it may also be caused by an impairment of the mechanisms responsible for inhibiting contextually irrelevant language systems (Green 1986). More generally, it can be seen as an inter-linguistic form of echolalia (the spontaneous repetition of utterances in the same language). The first report, by Kauders (1929), corresponds to patient D.O., a trilingual who spoke German (L1), French (L2), and English (L3). D.O. was highly proficient

102 The Neurocognition of Translation and Interpreting

in the latter two languages, which he had used constantly during the twenty-five years he spent living and working in Paris and London. However, this changed in 1925, when he suffered an apoplexy that focally damaged superior temporal and parietal areas of his LH. Comprehension was disrupted in all three languages, while speech was unintelligible and characterized by language mixing. The patient also exhibited reading difficulties and mild signs of dysgraphia. Yet, his most notable symptoms pertained to the domain of IR. When asked to name objects in L1, D.O. would first translate the target word into his two other languages and only then produce the L1 word. For example, when asked to name the color of a yellow figure, D.O. responded “yellow… jaune… gelb.” Likewise, when presented with a brush, his response was “zum Bresen (with an accompanying combing gesture)… brosse… chaumière… Bürste.”19 The second case of compulsive translation was reported two years later (Veyrac 1931). Patient Ch. was a speaker of English (L1) and French (L2) who had lived in Paris since age 15. Fifty years later she suffered a stroke resulting in Broca’s aphasia. Although voluntary speech and translation were impossible, Ch. would compulsively translate utterances from L1 to L2. For example, when asked “What time is it?,” the patient responded “Quelle heure est-il?”; and later, when the examiner told her “Show me your tongue,” the patient failed to comply with the request but responded “Montrez-moi la langue.” Similar patterns were observed in the frontal patients described by Stengel and Zelmanowitz (1933) and Weisenberg and McBride (1935). Further anecdotal evidence was offered in first person by Roman Jakobson (1964). After a left-sided cranial trauma sustained during a car accident, the famous linguist reports, he spent hours compulsively translating his own utterances into five different languages. Unfortunately, however, no further details are available on his condition. A few years later, Schulze (1968) described the case of a German literature professor diagnosed with motor aphasia after a left parieto-temporal abscess. The patient was a native speaker of Bulgarian with high proficiency in German, Russian, and French. In addition to signs of language mixing at the morphological level, the most conspicuous consequence of his condition was the production of unsolicited translations into non-relevant languages. For instance, when asked to repeat the German word “Jacke” (meaning ‘jacket’), the patient replied “Jacke… Jackett… Dschakett.”20

19. The pseudo-expression zum Bresen is likely the result of a paraphasia. The last three words mean ‘brush’ (in French), ‘cottage’ (in French), and ‘brush’ (in German). 20. The final word means ‘jacket’ in Bulgarian.



Chapter 4.  Building up from breakdown 103

Then, in the 1980s, Perecman (1984) reported the case of H.B., a man who was fluent in German (L1), French (L2), and English (L3), the latter being his dominant language. At age 75, the patient was involved in a traffic accident that resulted in bilateral temporal damage. In addition to showing symptoms of anomia, alexia, and repetition deficits, the patient would mix languages and compulsively translate sentences from L1 into L3. When the examiner told him “no, das ist blau” (‘no, that’s blue’, in German), the patient responded “blau oder gelb… blue or yellow” (both sentences meaning the same). This compulsive inter-linguistic behavior was even more striking for the patient’s own utterances. For example, he was reported to have said: “verstehen sie Deutsch… do you know German […] verstehen sie Deutsch… aber nur ein bischen… but only a little.”21 Yet, H.B.’s voluntary IR skills were severely affected in every language combination. The same disorder was documented in a case of dementia (De Vreese, Motta, and Toschi 1988). N.T. was a multilingual with Alzheimer’s disease, featuring severe temporal damage. The patient was a frequent speaker of Italian (L1), French (L2), and English (L3), and he sometimes used German (L4), too. His verbal production was good, but he exhibited anomia and severe impairments of reading, writing, and comprehension. More generally, his spontaneous conversation skills were preserved in L1 but greatly compromised in L2. Among these dysfunctions, the patient’s most notable alteration consisted in the compulsive translation of short phrases and full sentences produced by himself, his interlocutors, or even third parties. Such a behavior was most frequent from L1 into L2, but it was also observed from L1 into L3 and L4 – two languages in which the patient could not express himself and into which he could not translate willingly. For example, when the examiner asked him for a pipe in Italian (“Mi dia la pipa”), the patient replied “Questa è una pipa, this is a pipe.” Spontaneous translations were also recorded during repetition tasks: when the examiner uttered the word “Mattino” (Italian for ‘morning’), N.T. responded “Mattino, c’est matin ça.”22 Moreover, similar patterns were observed in the face of written stimuli. Upon reading the word “Commissario” (Italian for ‘police captain’), the patient said “Commissaire, c’est un mot important ça,” as he correctly copied it in Italian.23 Of note, the patient’s linguistic compulsions only manifested inter-linguistically, as there were no signs of echolalia. An arguably atypical case was documented by Lebrun (1991). Following RH damage, the patient would compulsively translate written words and full texts from 21. The patient’s utterance could be glossed as ‘Do you understand German’ (in German)… ‘do you know German’ (in English)… ‘do you understand German’ (in German)… ‘but only a little’ (in German)… ‘but only a little’ (in English). 22. The response could be glossed as ‘morning’ (in Italian), ‘that is morning’ (in French). 23. The utterance could be glossed as ‘police captain, that’s an important word’ (in French).

104 The Neurocognition of Translation and Interpreting

L2 (French) into L1 (Flemish), in the absence of oral aphasia. This tendency was interpreted as an inability to follow sociolinguistic conventions, although the actual underlying dysfunctions remain unknown. Further evidence comes from patient R.K. (Eviatar, Leikin, and Ibrahim 1999), a speaker of Russian (L1) and Hebrew (L2) who suffered a stroke damaging the basal ganglia and the corona radiata of her LH. She was diagnosed with fluent aphasia, dysgraphia, and dyslexia in both her languages. Although on-demand translation was completely abolished, she engaged in compulsive translation during L2 lexical-association and antonym-generation tasks. Nevertheless, no instances of compulsive translation were observed from L2 into L1. Finally, García-Caballero et al. (2007) described the case of an elderly crossed aphasic showing deficits in both her L1 (Galician) and her L2 (Spanish). After a cerebral infarction compromising her right basal ganglia, the patient could no longer produce L1 utterances voluntarily and her comprehension was impaired in both languages. Compulsive translation into L2 was observed during L1 lexeme-repetition tasks. However, voluntary translation in this direction was virtually impossible. The case stands out as the patient belongs within the small population of individuals who present RH dominance for language despite being right-handed. In sum, compulsive translation results mainly from lesions to the LH (or, more precisely, the language-dominant hemisphere), as observed in ten out of eleven cases. Damaged areas include frontal and posterior sites involved in linguistic and executive functions, which aligns with the view that this disorder may represent an impairment of language control mechanisms (see Chapter 2, Section 3.4.2 and Figure 3.11). Finally, note that compulsive translation can occur in both directions, regardless of whether the TL is available for spontaneous production. Table 4.1 summarizes all cases and includes additional information for each patient. 4.2.2 Inability to translate The disorder known as inability to translate involves a severe or complete incapacity to voluntarily engage in IR. This type of impairment can affect one or both directions, even when production skills are preserved for single-language tasks. Pertinent evidence is not only anecdotal, but also quantitative. As seen in the previous section, some patients with compulsive translation found it difficult or impossible to produce adequate TL expressions on demand in any direction. When patient H.B. (Perecman 1984) was asked to translate the word “essen,”24 he responded “English.” Then he translated the phrase “the wall” as 24. The word means ‘eat’, in German.

Chapter 4.  Building up from breakdown 105 Table 4.1 Compulsive translation. Summary of evidence and additional patient data. Case

Age

Sex

Handedness

Languages

Spontaneous Voluntary speech translation

Compulsive translation

Etiology

Lesion site

D.O.

62

M

R*

L1: Ger L2: Eng L3: Fr

L1: X L2: X L3: X

?

L2 and/or L3 into L1

Str

LH: STG, PL

Ch.

65

F

R*

L1: Eng L2: Fr

L1: X L2: X

L1 L2: X* L2 L1: X*

L1 into L2

Str

LH*

Case in Stengel and Zelmanowitz (1933)

57

M

R*

L1: Cz L2: Ger

L1: X* L2: X*

?

L1 into L2 L2 into L1

TCH

LH: FL

Case in Weisenberg and McBride (1935)

49

M

R

L1: Eng L2: Sp L3: Fr 4 more

?

L1 L2:— L1 L3:—

L1 into L2 & L3

?

LH*: FL*

Jakobson (1964)

?

M

R

L1: Rus L2: Fr, Ger, others

?

?

L1 into L2, L3, CC L4 & L5

LH

Case in Schulze (1968)

?

M

R

L1: Bul L2: Ger others

L1:— L2:—

?

L2 into L1

Abs

LH: PTR

H.B.

80

M

R*

L1: Ger L2: Fr L3: Eng

L1: ? L2: ? L3: X

erratic in every direction

L1 into L3

Hem

Bil: TL

(continued)

106 The Neurocognition of Translation and Interpreting Table 4.1 (continued) Case

Age

Sex

Handedness

Languages

Spontaneous Voluntary speech translation

Compulsive translation

Etiology

Lesion site

N.T.

65

M

R

L1: It L2: Fr L3: Eng L4: Ger

L1:— L2: X L3: X L4: X

L1 L2: ✓ L1 L3: X L2 L1: X L3 L1: X

into every language, but specially L2

Alz

LH: TL

Case in Lebrun (1991)

?

M

R*

L1: Fl L2: Fr

L1: ✓ L2: ✓

?

L2 into L1, written mode

?

RH

R.K.

68

F

R

L1: Rus L2: Heb

L1:— L2:—

L1 L2: X L2 L1: X

L2 into L1

Str

LH: BG, PPCR

Case in G. Caballero, G Lado et al. (2007)

91

F

R

L1: Gal L2: Sp

L1: X L2: ✓

L1 L2: ? L2 L1:— (words)

L1 into L2

CI

RH: BG (crossed aphasia)

Symbols: asterisk (*): presumably; question mark (?): information not provided; checkmark (✓): from very good to acceptably good; lines (—): considerably impaired; cross (X): severely impaired. Sex: F: female; M: male;. Handedness: R: right-handed. Languages: Bul: Bulgarian; Cz: Czech; Eng: English; Fl: Flemish; Fr: French; Gal: Galician; Ger: German; Heb: Hebrew; It: Italian; Rus: Russian; Sp: Spanish. Etiology: Abs: abscess; Alz: Alzheimer’s disease; CC: cerebral contusion; CI: cerebral infarction; Hem: hematoma; Str: stroke; TCH: traumatic cerebral hemorrhage. Lesion site: bil: bilateral; LH: left hemisphere; RH: right hemisphere; BG: basal ganglia; FL: frontal lobe; PL: parietal lobe; PPCR: posterior part of the corona radiata; PTR: parieto-temporal region; STG: superior temporal gyrus; TL: temporal lobe.



Chapter 4.  Building up from breakdown 107

“la val est langue française.”25 Faced with the noun “Seife,” he uttered “Französiche Auskunft.”26 Something similar was observed in the case of R.K. (Eviatar, Leikin, and Ibrahim 1999). Although the authors provide no specific data, they state that “[when] required to translate Russian words to Hebrew and vice versa […], R.K. was unable to perform this task” (Eviatar, Leikin, and Ibrahim 1999: 129). Another case was described by Gastaldi (1951). His patient, a speaker of German (L1) and Italian (L2), suffered from chronic inflammation in the LH, which resulted in aphasia. Object naming was partially spared in both languages, and he could comprehend simple instructions. However, BT and FT of words proved impossible. Byng et al. (1984) reported the case of B., a dyslexic child with severe focal damage to his left parietal and temporal lobes. His out-loud reading skills were normal in L2 (English) but very poor in L1 (Nepalese). More interestingly, a sight-translation test revealed a significant dissociation between FT (72%) and BT (0%). Some of the mistakes during such tasks seem to have been semantic in nature (e.g., he translated the Nepalese word for horse as duck, and the English word horse as the Nepalese equivalent of dog). Patient A.S. (Nilipour and Ashayeri 1989) spoke Farsi (L1), English (L2), and German (L3). He was involved in an explosion causing severe frontotemporal trauma. For over a month, he presented alternating antagonism between L1 and L3 (i.e., when one language was available for spontaneous production, the other one was not). Accuracy in word translation was 10% in L1-L2, 20% in L2-L1, 0% in L1-L3, and 50% in L3-L1. Sentence translation skills were completely lost, although SL comprehension was intact. Some of the patient’s mistakes in BT were semantic paraphasias – e.g., Fahrrad (German for ‘bicycle’) was translated as the Farsi word meaning car. Aglioti and Fabbro (1993) reported the case of E.M., a speaker of Venetian (L1) and Italian (L2). After a stroke, E.M. presented an ischemic lesion to her left basal ganglia. L1 verbal production was abolished, but her L2 remained functional. Comprehension was well preserved in both languages. E.M. performed three translation tasks, all revealing significant differences between FT and BT. Results for these directions were 69% vs. 41% in oral word translation, 95% vs. 5% in written word translation, and 72% vs. 35% in oral sentence translation, respectively. No such dissociations were observed in the patient’s husband, who served as a healthy control subject. This case shows that inability to translate can manifest in a directionally differential fashion, and that it can be more severe for 25. An approximate translation would be ‘the little valley is French language’. 26. The stimulus meant ‘soap’, in German, and the response was ‘French information’, in German.

108 The Neurocognition of Translation and Interpreting

some translation units (e.g., sentences) than others (e.g., words). A follow-up word translation task conducted three years later confirmed the former pattern: whereas E.M. had an accuracy of 65.7% in FT, her performance reached only 39.2% in BT (Aglioti et al. 1996). Fabbro and Paradis (1995) also reported the case of E.M. as part of a chapter that documented translation deficits in three other patients with left basal ganglia damage, called C.B., El.M., and O.R. Like E.M., they all had greater impairments in their L1s than in their non-native languages. Their translation skills were assessed with Part C of the BAT (Paradis 1979, 2011). C.B., a 71-year-old trilingual (L1: Friulian, L2: Italian, L3: English), suffered an ischemic stroke damaging parts of her left basal ganglia (the caudate nucleus, a small portion of the putamen, and the internal capsule). She was unable to translate words from L3 into L1, and neither could she translate sentences in either direction between L1 and L3 and between L2 and L3. However, SL comprehension was spared. Her performance attests to the functional independence of BT relative to FT, since her added scores for word translation from L2 and L3 into L1 (25%) were significantly lower than those for word translation from L1 into L2 and L3 (52%). Collapsing all language pairs and directions, the patient performed better on word (43%) than on sentence (11%) translation. Moreover, if only the two most dominant languages of the patient (L1 and L2) are considered, an even greater difference can be observed between the translation of words (70%) and sentences (33.3%). El.M. was fluent in Friulian (L1) and Italian (L2). A hemorrhage caused a vast subcortical lesion compromising left basal ganglia structures (the caudate nucleus, the putamen, the internal capsule, part of the globus pallidus) and producing non-fluent aphasia in both languages. His inability to translate sentences was virtually total in BT and FT. However, word translation skills were largely spared for both directions. Collapsing BT and FT, mean scores for word and sentence translation were 65% and 1%, respectively, corroborating that these abilities are dissociable. O.R., also a speaker of Friulian (L1) and Italian (L2), sustained severe damage to the insula and the left basal ganglia (caudate nucleus, putamen) subsequent to an ischemic stroke. While he remained fluent in both languages, he was unable to translate sentences (his accuracy was 16% in each direction). Word translation was better from L1 into L2 (90%) than in the opposite direction (40%). Also, whereas overall performance on this unit type reached 65%, deficits were much more severe for sentence translation (with an accuracy of 17%). In discussing the last four cases, Fabbro and Paradis (1995) observed that the left basal ganglia may play an important role in translation and that the routes



Chapter 4.  Building up from breakdown 109

supporting BT and FT may be independently damaged. However, they appear to have overlooked other interesting patterns in the data from these patients. First, BT was more significantly impaired than FT for both word and sentence translation. Second, collapsing both directions, sentence translation (46%) was more significantly compromised than word translation (61%). Third, the advantage of word translation over sentence translation was greater in BT than in FT (mean differences being 19% and 10%, respectively). Further evidence comes from patient S.M. (Detry, Pillon, and de Partz 2005), a speaker of French (L1) and English (L2) who sustained damage to left perisylvian regions. She exhibited signs of agrammatism and anomia, but her comprehension was well preserved in both languages. However, her translation performance in both directions was considerably compromised, with means of 60.4% in BT and 50% in FT. The latter result is particularly noteworthy, given that picture-naming skills in L2 were almost fully abolished. Finally, Weekes and Raman (2008) reported the case of B.R.B., a bilingual (L1: Turkish, L2: English) who suffered a stroke producing severe left parieto-occipital damage. He exhibited deep dysphasia, with fluent but semantically empty spontaneous speech in both languages. Translation tests were administered only in BT. A first examination revealed a marked inability to translate verbs and verbal nouns, with greater deficits in oral than in sight translation. A second test corroborated the dissociation, yielding mean outcomes of 5% and 93% for each IR modality, respectively. These deficits cannot be attributed to auditory or production deficits, given the high scores on L2 auditory lexical decision (100%), repetition (L1 = 82%, L2 = 65%), word reading (L1 = 100%, L2 = 85%), and picture naming (L1 = 72%, L2 = 82%). In conclusion, all these cases involve LH damage, and they show that a single lesion may selectively compromise either BT or FT of both words and sentences, even if single-language comprehension and production are spared. Moreover, translation impairments can manifest differentially for words relative to sentences and for oral vs. written stimuli, suggesting that different circuits can be engaged depending on the translation unit and modality. Additional patient data and a summary of the evidence can be found in Table 4.2.

110 The Neurocognition of Translation and Interpreting Table 4.2 Inability to translate. Summary of evidence and additional patient data. Case

Age

Sex

Handedness

Languages

Spontaneous speech

Word translation (oral mode)

Sentence Etiology Lesion site translation (written mode)

Case in Gastaldi (1951)

42

M

R*

L1: Ger L2: It

L1: — L2: —

L1 L2: X L2 L1: X

L1 L2: ? L2 L1: ?

CIL

LH

H.B.

80

M

R*

L1: Ger L2: Fr L3: Eng

L1: ? L2: ? L3: X

L1 L2: X* L1 L3: X* L3 L2: X*

?

Hem

Bil: TL

B.

15

M

R

L1: Nep L2: Eng

L1: X L1 L2: ✓ L2: ✓ L2 L1: X (out loud reading) (sight translation)

L1 L2: ? L2 L1: ?

LwSO

LH: PL, TL

A.S.

49

M

R

L1: Far L2: Eng L3: Ger

L1: aa L2: X L3: aa

L1 L2: X L2 L1: X L1 L3: X L3 L1: —

L1 L2: X L2 L1: X L1 L3: ? L3 L1: ?

CC

LH: FTR

E.M.

70

F

R

L1: Ven L2: It

L1: X L2:—

L1 L2: ✓ L2 L1:

L1 L2: ✓ L2 L1: X

Inf

LH: BG

C.B.

71

F

R

L1: Fri L2: It L3: Eng

L1: X L2: X L3: X

L1 L2: ✓ L2 L1: — L1 L3: — L3 L1: X L2 L3: X L3 L2: X

L1 L2: X L2 L1: X L1 L3: X L3 L1: X L2 L3: X L3 L2: X

II

LH: BG

Chapter 4.  Building up from breakdown 111 Table 4.2 (continued) Case

Age

Sex

Handedness

Languages

Spontaneous speech

Word translation (oral mode)

Sentence Etiology Lesion site translation (written mode)

El.M.

56

M

R

L1: Fri L2: It

L1: X L2: X

L1 L2: ✓ L2 L1: ✓

L1 L2: X L2 L1: X

CH

LH: BG

O.R.

63

M

R*

L1: Fri L2: It

L1: ✓ L2: ✓

L1 L2: ✓ L2 L1: X

L1 L2: X L2 L1: X

II

LH: BG, NPI

R.K.

68

F

R

L1: Rus L2: Heb

L1: — L2: —

L1 L2: X L2 L1: X

?

Str

LH: BG, PPCR

S.M.

40

F

R

L1: Fr L2: Eng

L1: — L2: —

L1 L2: — L2 L1: —

?

Str

LH: PSR

B.R.B.

67

M

R

L1: Tur L2: Eng

L1: — L2: —

L1 L2: ? L2 L1: sight trans.: ✓ oral trans.: X

L1 L2: ? L2 L1: ?

Str

LH: POR

Symbols: asterisk (*): presumably; question mark (?): information not provided; checkmark (✓): from very good to acceptably good; lines (—): considerably impaired; cross (X): severely impaired. Sex: F: female; M: male. Handedness: R: right-handed. Languages: Heb: Hebrew; Eng: English; Far: Farsi; Fr: French; Fri: Friulian; Ger: German; It: Italian; Nep: Nepalese; Rus: Russian; Tur: Turkish; Ven: Venetian. Etiology: CC: cerebral contusion; CH: cerebral hemorrhage; CIL: chronic inflammatory lesion; Hem: hematoma; II: ischemic infarction; Inf: infarction; LwSO: lesion with sharp object; Str: stroke. Lesion site: bil: bilateral; LH: left hemisphere; BG: basal ganglia; FTR: fronto-temporal region; NPI: neocortical portions of the insula; PL: parietal lobe; POR: parieto-occipital region; PPCR: posterior part of the corona radiata; PSR: perisylvian regions; TL: temporal lobe.

112 The Neurocognition of Translation and Interpreting

4.2.3 Paradoxical translation behavior Paradoxical translation behavior is a rare pathology. Patients with this disorder are capable of translating into a language unavailable for spontaneous production (e.g., L1) but incapable of translating into a language that is actually available for spontaneous production (e.g., L2). Notably, the capacity to engage in either BT or FT can alternate during the acute phase of the condition. The first two cases were described by Paradis, Goldblum, and Abidi (1982). Patient A.D. was a nun who spoke French (L1) and Arabic (L2). She was hit by a car and suffered a left temporo-occipito-parietal contusion. Like patient A.S., she exhibited signs of alternating antagonism. For over a month, during the days when only her L2 was available for spontaneous production, she was able to translate into L1 but not into L2. Conversely, when speech was possible in L1 but not in L2, only translation into L2 was possible. In a first evaluation, she was unable to describe a picture in L1. However, her BT skills were intact when administered single words (100%) and sentences (100%, with minor replacements of definite by indefinite articles). The following day, when her production skills improved in L1 and diminished in L2, she could translate only two out of the six Arabic sentences she had translated correctly the day before. The second case reported by Paradis, Goldblum, and Abidi (1982) concerns a young Canadian patient (henceforth, C.P.) who spoke French (L1) and English (L2) fluently. He underwent surgery for removal of a venous malformation in his left temporo-parietal region. Until the edema disappeared, C.P. exhibited alternating antagonism. During a period when his L2 was more impaired than his L1, he was asked to translate six sentences in each direction. FT was flawless, as revealed by his performance on sentences like “Mon voisin travaille à Toronto depuis 2 ans” and “Son frère a traversé la rivière à la nage.”27 On the contrary, BT was virtually impossible despite perfect comprehension of all stimuli: two sentences were left untranslated; another two were reformulated word by word, resulting in incomprehensible texts; and the remaining ones were rendered with great effort after several false starts. Paradoxical translation behavior was also observed in N.T. (De Vreese, Motta, and Toschi 1988), who could communicate only in L1. His cumulative mean scores in sentence translation tests were 70% from L1 into L2 and 20% from L2 into L1. IR was also virtually impossible between L1 and L3. In sum, although N.T could converse in L1, he was unable to translate into that language. However, he was quite 27. A possible rendition of these sentences would be ‘My neighbor has been working in Toronto for two years’ and ‘Her brother swam across the river’.

Chapter 4.  Building up from breakdown 113



capable of translating into L2, despite severe disruptions of spontaneous production in this language. In short, paradoxical translation behavior has been observed to occur following left-sided injuries. This disorder shows that the ability to engage in BT or FT does not depend on the integrity of the routes supporting L1 or L2 production, respectively; and that BT may be possible when FT is severely impaired, and vice versa. Table 4.3 offers a summary of findings and additional patient data.

C.P.

N.T.

23

65

M

M

R

R

L1: ✓ L2: X

L1 L2: ✓ L2 L1: X

L1 L2: ✓ L2 L1: X

L1: X L2: ✓

L1 L2: X L2 L1: ✓

L1 L2: X L2 L1: ✓

L1: ✓ L2: X

L1 L2: ? L2 L1: ?

L1 L2: ✓ L2 L1: X

L1: X L2: ✓

L1 L2: ? L2 L1: ?

L1 L2: ? L2 L1: ?

L1: — L2: X

L1 L2: ? L2 L1: ?

L1 L2: ✓ L2 L1: X

L1: Fr L2: Eng

L1: It L2: Fr and others

Lesion site

L1: Fr L2: Ar

Etiology

R

Sentence translation (written mode)

Handedness

F

Word translation (oral mode)

Sex

48

Spontaneous speech

Age

A.D.

Languages

Case

Table 4.3  Paradoxical translation behavior. Summary of evidence and additional patient data.

CC

LH: TOPR

VM

LH: TPR

Alz

LH: TL

Symbols: question mark (?): information not provided; checkmark (✓): from very good to acceptably good; lines (—): considerably impaired; cross (X): severely impaired. Sex: F: female; M: male. Handedness: R: right-handed. Languages: Ar: Arabic; Eng: English; Fr: French; It: Italian. Etiology: Alz: Alzheimer’s disease; CC: cerebral contusion; VM: venous malformation. Lesion site: LH: left hemisphere; TL: temporal lobe; TOPR: temporo-occipito-parietal region; TPR: temporo-parietal region.

114 The Neurocognition of Translation and Interpreting

4.2.4 Translation without comprehension The fourth neurological condition affecting IR is translation without comprehension. Patients with this disorder are able to translate utterances correctly although they are unaware of the meaning of the SL expressions. Three cases have been reported so far.28 Interestingly, each of these patients also manifested one of the three dysfunctions described above, so that no “pure cases” of this disorder seem to exist in the literature. Patient Ch. (Veyrac 1931) provided the first examples of this dysfunction. After compulsively (and correctly) translating utterances like “What time is it?” and “Show me your tongue” into French (see Section 4.2.1), Ch. never gave signs of comprehending them. She did not attempt to check her watch and made no effort to stick out her tongue. Therefore, IR of stock phrases was possible even though source-unit understanding seemed impaired. Patient C.P. (Paradis, Goldblum, and Abidi 1982) had similar symptoms. On a day in which he was more fluent in French than in English, he was asked to translate concrete nouns, such as “plafond,” “porte,” “fenêtre,” and “table.”29 He accurately provided the English equivalents but was unable to identify and point to those objects. Curiously enough, the patient said he was sure that those objects existed in the room, but he could not tell which one was which. In some cases, after effortlessly providing a correct translation, he pointed to incorrect objects, such as the sink instead of the window or the bed instead of the table. Instances of translation without comprehension were also observed in patient N.T. (De Vreese, Motta, and Toschi 1988). Although no specific details are offered, the authors noted that [the patient] displayed a similar behaviour as the second case reported by Paradis, Goldblum et al. (1982). When asked to point to pictured objects given their Italian names, he correctly translated the words in French although he often remained unable to point to their respective pictures.  (De Vreese, Motta, and Toschi 1988: 253)

28. Surprisingly, Fabbro (2001) cites the work by Fabbro and Paradis (1995) in the context of a brief discussion of this disorder. However, the latter report does not describe a single instance of translation without comprehension in any of the four patients evaluated. Although they exhibited comprehension deficits during single-language tasks, no mention is made of such impairments during IR. In fact, in discussing patient C.B., Fabbro and Paradis (1995: 143–144) affirm that “C.B. could translate words from her L1 into her L3, but not vice-versa, although she perfectly understood the meaning of the words to be translated, since she correctly indicated the related objects (e.g., head, wall).” 29. French words for ‘ceiling’, ‘door’, ‘window’, and ‘table’, respectively.

Chapter 4.  Building up from breakdown 115



All in all, this pathology occurs subsequent to posterior LH lesions and suggests that perceptual and semantic information associated to SL words need not be active during translation – in other words, successful IR does not necessarily hinge on conceptually-mediated routes. Further details on these cases are provided in Table 4.4.

Case

Age

Sex

Handedness

Languages

Spontaneous speech

Word translation (oral mode)

Sentence translation (written mode)

Etiology

Lesion site

Table 4.4  Translation without comprehension. Summary of evidence and additional patient data.

Ch.

65

F

R*

L1: Eng L2: Fr

L1: X L2: X

L1 into L2

?

Str

LH*

C.P.

23

M

R

L1: Fr L2: Eng

L1: ✓ L2: X

L1 into L2

?

VM

LH: TPR

L1: X L2: ✓

?

?

L1: — L2: X

L1 into L2

?

Alz

LH: TL

N.T.

65

M

R

L1: It L2: Fr and others

Symbols: question mark (?): information not provided; checkmark (✓): from very good to acceptably good; lines (—): considerably impaired; cross (X): severely impaired. Sex: F: female; M: male. Handedness: R: right-handed. Languages: Eng: English; Fr: French; It: Italian. Etiology: Alz: Alzheimer’s disease; Str: stroke; VM: venous malformation. Lesion site: LH: left hemisphere; TL: temporal lobe; TPR: temporo-parietal region.

4.3

Charting the territory

Beyond its potential clinical relevance, the evidence above supports preliminary (though non-trivial) conclusions about the neurocognitive organization of IR systems. In particular, empirically grounded inferences can be made about their lateralization, their relative autonomy vis-à-vis other linguistic mechanisms, and the functional individuation of the circuits engaged by different translation directions, processing levels, and types of SL unit. Let us address these issues in turn.

116 The Neurocognition of Translation and Interpreting

4.3.1 Taking sides As seen in Chapter 3 (Section 3.3.1.1), while the two cerebral hemispheres constantly exchange information during verbal communication, a number of basic language mechanisms are critically rooted in left-sided circuits – at least in the vast majority of individuals (Mazoyer et al. 2016). Evidence from the four translation neuropathologies suggests that the same is true for the networks subserving key IR processes. In twenty out of twenty-three cases reviewed, translation disorders resulted from lesions damaging the LH only (see Tables 4.1 through 4.4). Each of those patients was either confirmed or presumed to be right-handed, indicating that injuries were probably confined to the language-dominant hemisphere. In fact, in most of them, basic linguistic functions were variously affected, as revealed through single-language tasks. These observations support the view that LH structures may be putatively related to core IR mechanisms, as further evinced by electrostimulation evidence showing that inhibition of left frontal regions directly disturbs IR processes (Borius et al. 2012). From a broader perspective, these findings indicate that putative IR networks are embedded in more general systems subserving linguistic processes as a whole (García, Mikulan, and Ibáñez 2016). Indeed, the patients’ lesions were mainly located in frontostriatal and temporo-parietal areas implicated in multiple verbal operations during single-language tasks (Chapter 3, Sections 3.3.1.2 and 3.3.1.3). Results from the three remaining patients are not inconsistent with this conclusion. In the case of H.B., damage was bilateral, meaning that it affected the LH and the RH (Perecman 1984). In the remaining two cases, lesions were circumscribed to the RH, but this does not necessarily contradict the hypothesis. First, the patient tested by García-Caballero et al. (2007) was a crossed aphasic, meaning that her RH was subserving the language functions typically supported by the LH. Second, Lebrun (1991) described his patient’s disorder as a “sociolinguistic deficit.” Although sociolinguistic abilities do play a role in translation, they are not part of putative IR routes per se. RH participation likely reflects the use of attentional, pragmatic, and otherwise extralinguistic strategies (Paradis 2003, 2009). Such processes, arguably related to what Lebrun (1991) termed sociolinguistic skills, are distinct and separate from those subserved by IR routes proper. Moreover, as will be seen in the following chapters, whole-brain recordings in neuroimaging studies show that, compared to other language functions, IR yields significant activation increases predominantly or even exclusively in the LH (Klein et al. 1995; Lehtonen et al. 2005; Rinne et al. 2000). Admittedly, this claim may be partially biased by the underrepresentation of patients with RH lesions in the literature – either because they did not exhibit language symptoms or because IR was not assessed in them. However, a recent



Chapter 4.  Building up from breakdown 117

systematic analysis of linguistic domains in a right-handed bilingual with extensive right-sided damage showed full preservation of equivalent recognition, word translation, and sentence translation skills, all in both directions (Calvo et al. 2019). This further reinforces the notion that left-sided regions play an asymmetrically critical role in IR. The above claim, however, does not mean that the RH is irrelevant for this domain. Consider the following findings, all of which will be detailed in later chapters. First, PET evidence shows that predominant LH activation increases during IR (relative to single-language reading) may be accompanied by bilateral engagement of the anterior cingulate and the basal ganglia (Price, Green, and von Studnitz 1999). Second, during both BT and FT, professional translators (García, Mikulan, and Ibáñez 2016) and interpreters (Kurz 1994, 1995) evince significant functional connectivity between both hemispheres – for details, see Chapters 5 and 6. Third, behavioral experiments (Fabbro et al. 1990; Fabbro, Gran, and Gran 1991; Proverbio and Adorni 2011) and electrophysiological research (Proverbio, Leoni, and Zani 2004) on hemispheric specialization suggest that linguistic processes are less left-lateralized in SIs than in other populations. Finally, neuroplastic adaptations related to sustained practice of simultaneous interpreting have been observed in cortical and subcortical regions across the LH and the RH (Hervais-Adelman, Moser-Mercer, and Golestani 2015; Hervais-Adelman et al. 2017) – for details, see Chapter 7. In sum, putative IR mechanisms appear to be distributed mainly over LH regions. However, the processes subserved by those structures also rely on contributions from RH hubs. Such patterns have been reported in subjects with and without professional training in translation and interpreting. Interestingly, too, sustained practice of at least one IR modality (simultaneous interpreting) seems to involve neural changes across both hemispheres. Therefore, despite the dominance of left-sided areas, cross-linguistic operations seem to rely on complex interactions from both halves of the brain. 4.3.2 A thing unto itself As argued in the previous section, IR mechanisms seem to be couched in the general brain regions mediating basic verbal processes across multiple tasks. However, this correspondence in terms of gross neural areas does not imply that IR mechanisms lack neurocognitive specificity. If that were the case, damage to such sites should not cause dissociations between translation and single-language abilities. However, the cases of inability to translate and paradoxical translation behavior suggest otherwise.

118 The Neurocognition of Translation and Interpreting

Four empirical patterns derived from those corpora attest to the partial independence of BT- and FT-preferential routes relative to pathways engaged by L1 or L2 production. First, the circuits supporting BT can remain functional when those involved in L1 production are severely or partially impaired (see patients A.D. and C.P.). By the same token, the pathways supporting FT can be spared even when those subserving L2 production are compromised (see patients A.D., N.T., E.M., C.B., and Ch.). Moreover, the subsystems implicated in BT may become dysfunctional in the context of preserved L1-production skills (see patients A.D., C.P., O.R., N.T., and R.K.). Likewise, FT may be markedly disturbed despite spared performance in L2-only tasks (see patients A.D., O.R., R.K., and the case reported by Gastaldi 1951). Therefore, IR-preferential routes seem to be at least partially autonomous from those involved in single-language processes. Neither does the integrity of receptive single-language processes guarantee that IR can be performed. Consider, in this sense, a potential double dissociation between the cases of translation without comprehension and certain deficits exhibited by patients A.S. and C.B. Whereas patients Ch., C.P., and N.T were capable of translating source units without grasping their meaning, A.S. and C.B. did understand the meaning of input items but were unable to translate them. For example, in the tasks involving his L1 and L3, A.S. showed preserved comprehension of all stimuli, but he could translate only half of them in BT and none in FT. It would thus seem that preserved receptive skills in the SL are not sufficient for IR. This conclusion aligns with previous claims advanced by Paradis (1984) upon consideration of only two cases – namely, those reported in Paradis, Goldblum, and Abidi (1982). It also fits well with Fabbro’s (2001) proposal that translation subsystems are neurofunctionally independent from those engaged during L1 and L2 tasks. What is more, it is consistent with the finding that compulsive translation does not necessarily imply compulsive repetition in the same language (echolalia), as seen in patient N.T. (De Vreese, Motta, and Toschi 1988). Beyond neuropsychological findings, additional confirmatory evidence has been obtained via brain stimulation. In a study with bilinguals possessing various degrees of translation and/or interpreting experience, Borius et al. (2012) assessed how electrical currents applied on LH regions affected BT and single-language tasks (reading, picture naming). IR was disrupted upon stimulation of Broca’s area and the superior frontal gyrus, although the same was true for other tasks. However, IR was the only skill that remained undisturbed when other perisylvian regions were stimulated. Therefore, at least some of the mechanisms implicated in single-language processing are not critical for (if at all related to) translation operations. All in all, it seems that a subset of the neural pathways grounding IR processes are partly independent from other language-preferential networks. Although extant data are insufficient to advance specific claims about the location of those networks,



Chapter 4.  Building up from breakdown 119

their proposed functional individuation would not be exceptional. Indeed, language regions in the brain include several specialized sub-mechanisms responsible for distinct, fine-grained operations (see Chapter 3, Sections 3.2.2, 3.2.3, 3.3.1.2, 3.3.1.3, 3.4.1). In this sense, the relative functional autonomy of IR pathways may be a natural manifestation of the inner organization of the language system at large. 4.3.3 Coming and going In addition to their overall semi-independence from other language pathways, IR systems also encompass partially autonomous routes mediating specific sub-operations. In particular, different networks seem to be involved in BT and FT across modalities. Once again, the evidence comes from cases of inability to translate and paradoxical translation behavior. These conditions show that the neural networks engaged by BT can remain functional when those supporting FT are compromised, as seen for oral translation of words (see patients A.S. and A.D.) and sentences (see patient A.D.). In the same vein, they indicate that the circuits implicated in FT may operate properly even when those recruited during BT are severely affected. This has been observed in tasks requiring sight translation of words (see patient B.) and oral translation of words (see patients E.M., C.B., O.R., A.D.) and sentences (see patients C.P. and N.T.). No such dissociations should be observed if both directions depended on identical substrates. In fact, the functional individuation of BT and FT routes has been postulated in earlier works. On the one hand, it has been acknowledged in previous neuropsychological studies (Fabbro and Paradis 1995; Paradis 1984). On the other, it has been captured by psycholinguistic models based on response-time data (e.g., Kroll and Stewart 1994; Kroll et al. 2010). Notwithstanding, the specific hubs differentially related to each direction remain anatomically underdetermined. Data from the four cases reported by Fabbro and Paradis (1995) suggests that the basal ganglia may play a greater role in BT than in FT, irrespective of the translation unit. Moreover, neuroimaging evidence further indicates that cortical portions of frontostriatal pathways may play differential roles in each direction (see Chapter 5). Yet, at present, any firm conclusion beyond these rather broad preliminary patterns should be dismissed as speculative. In sum, IR appears to recruit partly independent mechanisms depending on the direction in which it is performed. Therefore, IR-preferential systems seem to possess a particular internal organization, such that specific sub-components are unevenly engaged depending on directionality. In fact, this conclusion is variously corroborated and refined by numerous neuroscientific and behavioral studies to be detailed in Chapter 5.

120 The Neurocognition of Translation and Interpreting

4.3.4 Of words and concepts The review also hints at another neurocognitive distinction within IR systems. It would seem that translation and interpreting can be performed through either a conceptually-mediated route (involving access to semantic information) or a form-level route (which allows for the establishment of cross-linguistic associations in the absence of conceptual activations). While the most compelling evidence comes from cases of translation without comprehension, relevant data are available in other conditions as well. The role of conceptually-mediated mechanisms in IR is uncontroversial. In fact, many of the deficits manifested by patients with inability to translate imply disruptions of semantic processes. For example, some translation errors produced by patient A.S. seem to constitute cross-linguistic semantic paraphasias – e.g., he translated bicycle as the L1 word meaning car (Nilipour and Ashayeri 1989). The same was observed in patient B. (Byng et al. 1984), who, for instance, translated horse as the Nepalese word for dog. Also, considering that many compulsive translations produced by patient N.T. were paraphrastic rather than verbatim, De Vreese, Motta, and Toschi (1988: 253) maintain that “we may assume that automatic translation as unintentional accessing of other languages was performed in this patient at the prelinguistic conceptual level.” More notably, evidence from translation without comprehension attests to the relative autonomy of form-level routes. As seen in patients Ch., C.P., and N.T., certain brain lesions can transiently compromise conceptual access without impeding successful retrieval of cross-language equivalents. Indeed, these patients’ inability to designate the referents of the nouns they had just translated was not due to perceptual, attentional, or awareness-related impairments. As proposed by Paradis, Goldblum, and Abidi (1982), their performance indicates a lack of comprehension of SL expressions, which underscores the role of spared cross-linguistic connections operating on non-semantic information. Also informative is a specific pattern observed in patient S.M. (Detry, Pillon, and de Partz 2005). Despite preserved comprehension, she was almost fully incapable of naming pictures in L2, but showed considerably milder deficits to engage in FT. This further speaks to the relative autonomy of form-level connections between languages. As the authors maintain, [t]hat S.M. retrieved more accurately L2 output word forms in translation than in naming strongly suggests the existence of a direct processing route linking L1 input word forms to L2 output word forms […] This route would be relatively spared in SM in comparison with the semantic route for translating from L1 to L2, which was severely impaired. (Detry, Pillon, and de Partz 2005: 41)



Chapter 4.  Building up from breakdown 121

Also, the postulation of separate routes for conceptually-mediated and form-level operations during IR aligns with works from different approaches. This distinction is one of the core components of the Revised Hierarchical Model, a leading account of the organization of bilingual memory (e.g., Kroll and Stewart 1994; Kroll et al. 2010). Mutatis mutandis, a somewhat similar dichotomy is proposed in the Théorie du sens, as seen in the opposition between deverbalization and transcoding (Lederer 1994; Seleskovitch 1981, 1978). Furthermore, as noted elsewhere in this volume, some authors claim that each route could be differentially recruited for strategic reasons during simultaneous interpreting (Paradis 2009), as word-to-word mappings might be preferred by professionals during periods of fatigue or stress (Darò and Fabbro 1994). The distinction between form-level and conceptually-mediated routes is additionally supported by behavioral experiments, which show that differential reliance on them depends on L2 proficiency. For highly proficient bilinguals, conceptual mediation seems to be crucially involved in both BT and FT, as shown by Stroop-type experiments (Heij et al. 1996) and other lexical tasks (e.g., de Groot, Dannenburg, and van Hell 1994; Duyck and Brysbaert 2004). For example, in both BT and FT, these subjects are faster to translate words denoting low (e.g., one, two) than large (e.g., eight, nine) numbers (Duyck and Brysbaert 2004). Insofar as numerical magnitude is an aspect of meaning rather than form, this result would reflect the involvement of conceptually-mediated mechanisms. Although these results do not provide any direct evidence for a role of form-level connections, their differential contributions are clearly observed in low-proficiency bilinguals. Unlike their highly proficient counterparts, these individuals do not experience delays during reading and translation when items are blocked by semantic category (e.g., lists of stimuli belonging to the category fruit) (Kroll and Curley 1988). This indicates that, in this population, cross-linguistic processing may not be crucially mediated by conceptual routes, suggesting that form-level connections were recruited in their stead. Compatible evidence has been obtained through translation recognition tasks in which distractors (wrong translations) could operate at either the semantic (man-mujer) or the word-form (man-hambre) level30 – for details on this task, see Chapter 2, Section 2.3.2.1. Whereas highly proficient bilinguals make more errors in the former condition, low-proficiency ones have poorer performance in the latter (de Groot 1992; Ferré, Sánchez-Casas, and Guasch 2006; Talamas, Kroll, and Dufour 1999).31 Taken together, these data 30. The word hambre has significant orthographic/phonological overlap with the translation equivalent of man, namely, hombre. 31. Additional evidence by Guasch et al. (2008) suggests that word-form manipulations modulate performance across all proficiency levels, whereas semantic relations do so only at intermediate and higher levels, and only when semantic associations are very strong.

122 The Neurocognition of Translation and Interpreting

c­ onfirm that both levels of processing are operative during IR, additionally indicating that their relative contributions depend on L2 competence. In short, the evidence motivates one further constraint for neurocognitive models of translation and interpreting. Although extant findings are limited and somewhat indirect, one could at least hypothesize that cross-linguistic processes can operate at two functional levels, each placing differential demands on conceptually-mediated and form-level routes. A plausible neurobiological account of IR should explicitly capture such a distinction, even if its underlying evidence remains incipient at present. 4.3.5 The unit determines the network The preliminary conclusions in the previous sections were advanced without considering any particular translation unit. However, the evidence indicates that different brain networks are recruited depending on linguistic properties of the SL input. Specifically, frontostriatal and temporo-parietal circuits seem to be dissimilarly engaged by sentence and word translation. Compared to word translation, sentence translation seems to rely more critically on frontostriatal regions. The strongest evidence comes from the four cases described by Fabbro and Paradis (1995). All these patients, presenting basal ganglia lesions, were more significantly impaired in sentence than in word translation. The same was true of the crossed aphasic assessed by García-Caballero et al. (2007), who also presented frontostriatal damage. Furthermore, neuroimaging studies on sentence translation have revealed distinctive activation increases in frontal (but not in posterior) regions (Lehtonen et al. 2005) – for details, see Chapter 6. On the other hand, word translation seems distinctly associated with posterior sites. First, compulsive translation in patients with temporo-parietal damage involves mainly words and lexemes. Spontaneous translation of non-rotely learned sentences was observed only subsequent to bilateral lesions (see patient H.B.).32 Second, in five out of seven reports of compulsive translation including neuroanatomical references, damage was confined to left temporal/temporo-parietal areas. The other two patients (García-Caballero et al. 2007; Weisenberg and McBride 1935), with frontal lesions, were better than these five on word translation. Third, patients with frontostriatal lesions were either largely or completely unimpaired in word translation in specific directions (E.M., C.B. El.M., O.R.). There are no reports in which word translation became impossible subsequent to frontostriatal lesions. 32. Further research is necessary to account for this discrepant pattern, which may involve multiple factors, such as etiology (hematoma), lesion site and extension, and lateralization (indeed, the patient’s manual preference was not reported).



Chapter 4.  Building up from breakdown 123

By contrast, word translation did prove impossible for some patients without posterior brain injuries. It follows that, if both sets of networks were compromised, IR skills should be markedly impaired for sentences and words alike. That is precisely what can be observed in patient A.S. (Nilipour and Ashayeri 1989). Following a brain lesion affecting frontal and temporal areas, his overall translation scores (integrating different language combinations) were 20% for words and 0% for sentences. Less robustly, the evidence also shows that partially different circuits may be recruited depending on specific features of lexical stimuli. For example, in patient S.M. (Detry, Pillon, and de Partz 2005), BT and FT deficits were greater for non-cognates than for cognates. Similarly, in patient El.M. (Fabbro and Paradis 1995), translation in both directions was better for concrete than abstract words. Moreover, although available clinical data is not enough to specify which particular mechanisms are engaged by different word types, abundant neuroscientific and behavioral evidence sheds light on the issue (see Chapter 6). In short, IR does not rely on all-purpose systems which are equally involved in all relevant processes. The complexities and particularities of the linguistic input seem to be key determinants of which neural mechanisms will be recruited for the task. Here lies another demonstration of the cognitive intricacies immanent in translation and interpreting. 4.4 Piecing it all together Some of the general patterns identified above have been integrated into two neuroanatomical models (Fabbro 1999; García 2012a). With their strong and weak points, such formulations provide a composite picture of the overall organization of IR systems, leading to explicit hypotheses that can be empirically tested in new studies. 4.4.1 A neuroarchitectural model of translation routes Building on clinical and neuroscientific findings, I have previously postulated the Neuroarchitectural Translation Model (NTM), a neurocognitive account of the linguistic systems mediating IR (García 2012a) – Figure 4.1. Specifically, the NTM aims to characterize the organization of IR-specific routes within an overall model of linguistic systems. Mainly driven by lesion studies, and in line with previous accounts of the bilingual brain (Paradis 2004, 2009; Ullman 2001b), the model identifies partially independent systems for broad processing levels in L1 and L2 (e.g., phonological production, phonological recognition, morphosyntactic integration, lexico-semantic operations), alongside more general non-verbal systems related to

124 The Neurocognition of Translation and Interpreting

each language. Moreover, the NTM makes explicit predictions regarding the differential relation between those components and two well-established neurocognitive systems, one mediating procedural memory (with key hubs along frontostriatal circuits) and the other mediating declarative memory (spanning various structures within and around temporal, parietal, and hippocampal sites). Amid that overarching architecture, the model includes four relatively independent routes specific to IR. Semantic connections between languages would depend on conceptually-mediated forward and backward routes. These would comprise directionally-specific links between constellations of semantic features in each language, as also proposed in psycholinguistic models of bilingual memory (de Groot 1992, 1993; de Groot, Dannenburg, and van Hell 1994; Kroll and Stewart 1994; Kroll et al. 2010). IR is also proposed to hinge on direct lexical connections, afforded by form-level forward and backward routes. Such sets of connections, which are also acknowledged in the psycholinguistic accounts just cited, would complement the other two routes and, as shown by cases of translation without comprehension, they may even predominate during specific instances of IR – although more research is needed to test the robustness of this hypothesis (see Section 4.3.4). Except for specific pathological cases, the NTM posits that conceptually-mediated and form-level routes are jointly involved in virtually all instances of IR, and that their relative contributions depend on subject- and stimulus-related variables – for specifications on the role of these factors, see Chapters 5 and 6. Note, too, that each direction is proposed to rely on separate routes, in line with the evidence summarized in Section 4.3.3 and convergent findings from neuroscientific experiments (see Chapter 5). Moreover, the model agrees with the conclusions advanced in Section 4.3.2 in that it acknowledges the partial neurofunctional independence of those four routes for IR relative to others supporting single-language processes. In neuroanatomical terms, the NTM posits that the key substrates of IR routes mainly comprise perisylvian and frontostriatal regions of the LH (see Section 4.3.1). Also, in agreement with Section 4.3.5, the model postulates that word and sentence translation are differentially related to posterior networks implicated in declarative memory and to frontostriatal pathways subserving procedural memory, respectively. Still, it is recognized that no translation unit is exclusively reliant on any such circuits. Finally, the NTM proposes that these broad organizational features would hold for any and all bilinguals, irrespective of their level of translation or interpreting training. However, the relative contributions of procedural and declarative memory systems during ST processing and TT production would depend on subject-related factors, such as age of L2 acquisition and L2 competence. On this point, note that the version of the model presented here (Figure 4.1) relates L2 morphosyntactic and phonological production operations with declarative memory systems. Therefore, in line with previous neurolinguistic frameworks (Paradis 2004, 2009; Ullman 2001b), this incarnation of the NTM captures the main organizational

Chapter 4.  Building up from breakdown 125



NEUROCOGNITIVE

SOMATOSENSORY

SYSTEM

SYSTEM

SEMOLOGICAL SYSTEM

LING. SYST. L1

SEMANTIC SYSTEM 1

SYST.

L2

SEMANTIC SYSTEM 2

CMFR

SYST.

A

MORPH. SYNT.

PROD.

RECOG.

PROD.

GRAPH. SYST. 2 RECOG.

PHON.

SYST. 2

PROD.

RECOG.

RECOG.

PROD.

GRAPH. SYST. 1

SYSTEM

SYSTEM

PHON.

LEXICON

FLBR

MORPH. SYNT.

SYST. 1

LEX-GRAM. 2

FLFR

LEXICON

ISUAL

UDITORY

LEX-GRAM. 1

V

CMBR

MOTOR SYSTEM

DECLARATIVE MEMORY (mainly temporo-parietal regions) PROCEDURAL MEMORY (mainly fronto-striatal regions)

Note: Graph. prod. includes parietal and fronto-striatal motor regions, whereas Graph. recog. includes occipital regions.

Figure 4.1  The Neuroarchitectural Translation Model. The figure shows the components included in the model, with each box representing a partially autonomous subsystem – as revealed mainly by lesion studies. The middle part of the diagram captures the four functionally autonomous systems specifically implicated in IR. The brown boxes would be mainly associated with declarative memory systems, whereas those painted in light brown would be critically related to procedural memory systems. Note that this version of the model refers to the functional architecture of a late and/or low-proficiency bilingual. CMBR: conceptually-mediated backward route; CMFR: conceptually-mediated forward route; FLBR: form-level backward route; FLFR: form-level forward route; Graph. syst.: graphemic system; L1 syst.: native-language system; L2 syst.: second-language system; Ling. syst.: linguistic system; Morph.: morphology; Phon. syst.: phonological system; Prod.: production; Recog.: recognition; Synt.: syntax. Reproduced with permission from García (2012a).

126 The Neurocognition of Translation and Interpreting

features of late and/or low-proficiency bilinguals. In fact, another version of the model, aimed to describe early and/or high-proficiency bilinguals, posits that morphosyntactic and phonological production processes in L2 hinge distinctively on procedural memory systems – for details, see García (2012a). Overall, the NTM is the result of a step-by-step construction process, informed by findings on the general organization of language systems, the particular features of bilingual memory, and the specificities of IR routes. It is functionally and anatomically explicit (which paves the way for direct testing and refinement), and it proves consistent with evidence from lesion models and neuroscientific experiments. However, it overlooks the role of relevant non-verbal mechanisms and dynamic aspects of online processing. Neither does it consider the functions afforded by RH regions, which, as seen in Section 4.3.1, also contribute to aspects of IR. Some of these shortcomings are partially addressed in the alternative model presented next. 4.4.2 A neural model of the systems subserving simultaneous interpreting While the above model is proposed to capture organizational aspects of IR systems across translation and interpreting modalities, the one advanced by Fabbro (1999) focuses on simultaneous interpreting (Figure 4.2). This formulation assumes that both hemispheres have a balanced contribution for acoustic analysis of SL input and motor control during TL production – with critical roles of the auditory and sensory-motor cortices, respectively. Yet, each hemisphere is claimed to predominate in different higher-level operations. The LH would be critically involved in strictly linguistic processes, ranging from phonological decoding (linked to temporo-parietal and subcortical cortices) to cross-linguistic processing (ascribed to inferior anterior frontal regions) to articulatory functions (associated with cortical and subcortical motor networks as well as the cingulate gyrus). Instead, the RH is presented as a key contributor to relevant complementary processes, including prosodic, pragmatic, emotional, and attentional mechanisms (attributed to temporo-parietal and premotor areas, in addition to the cingulate gyrus). However, the model is less explicit regarding the inner organization of IR systems. Separate routes are recognized for BT and FT, both of which are tacitly assumed to be partially independent from other general language mechanisms. However, the model does not capture the role of conceptually-mediated and formlevel connections, and neither does it identify specific regions distinctively engaged by different translation units. Fabbro’s model constitutes the first attempt to account for the neuroanatomical organization of verbal and non-verbal systems involved in simultaneous interpreting. Although it fails to incorporate some of the key patterns identified throughout

Chapter 4.  Building up from breakdown 127



Target language (L1 or L2) Motor output Bilateral primary sensory-motor cortex

Left hemisphere (L.H.) Premotor areas supplementary motor area Broca’s area L.H. subcortical structures

Articulatory processing L2

Semantic associations L1 L2 L.H. inferior anterior frontal cortex

Phonological decoding L1

Premotor area R.H. Cingulate gyrus

Cingulate gyrus

Prosodic and emotional processing

R.H. emotional, attentional, and pragmatic functions

L2 → L1

L1 → L2

Articulatory processing L1

Right hemisphere (R.H.)

Prosodic and emotional decoding

Phonological decoding L2

R.H. temporo-parietal areas

L.H. subcortical structures and temporo-parietal cortex

Acoustic analysis Bilateral primary and secondary auditory cortex

Source language (L2 or L1)

Figure 4.2  Main functional components involved in the process of simultaneous interpretation (per Fabbro 1999: 205). The figure shows the components included in the model, with each box representing a partially autonomous subsystem. The left side of the diagram represents systems mainly subserved by the left hemisphere, whereas the right side depicts systems mainly associated with right hemisphere regions. L1: native language; L2: second language; L.H.: left hemisphere; R.H.: right hemisphere. Reproduced with permission from Fabbro (1999).

Section 4.3, it does provide a set of interrelated hypotheses that can be empirically tested via various brain-based methods. In this sense, it represents a potentially useful platform to guide further research in the field.

128 The Neurocognition of Translation and Interpreting

4.5

Interpretive remarks

In the ‘Apple Helicopter’, each function can be fully attributed to an isolable component that is exclusively devoted to it. For instance, provided it is attached to the battery, the toy’s speaker is self-sufficient for, and completely devoted to, its sound-production capacity. The same is true of the relationship between the sensor and the toy’s motion-detection features, or between the motor and its rotor-spinning capabilities. In this sense, and in many others, the gadget represents a poor metaphor of the human brain, where links between structures and functions are hardly so straightforward. Therefore, a few clarifications are in order. A sketch made of boxes and arrows could offer a fair representation of the toy’s functional organization. However, these visual devices need to be taken with caution when used to diagram neurocognitive systems, as done in Figures 4.1 and 4.2. Unlike what those graphics might seem to imply, the brain mechanisms implicated in IR (or in any other domain) do not consist of discrete, self-contained modules. Specific cerebral regions and networks may well be involved in processes captured by different boxes, with various degrees of overlap. In consequence, the boundaries set by the boxes’ edges in these models must not be interpreted isomorphically at a neural level. Furthermore, the areas associated to a particular process need not be activated in their entirety when the process is deployed, and those which do become engaged may do so with uneven intensities. Moreover, different substrates may be recruited in successive iterations of the same task (e.g., translating a specific noun or sentence). An all-or-nothing interpretation of the anatomo-functional links postulated above would thus be incorrect. Also, an arrow linking two boxes does not represent a single neural pathway. The links between two subsystems may actually be rooted in multiple sets of intraand inter-regional connections, some of which may actually participate in different functions, even simultaneously. Moreover, a myriad of connections exist within the regions associated with the boxes themselves. Hence, the relations captured by the above diagrams should not be interpreted as if one were looking at the cables linking isolated components in a mechanical toy. In addition, a localizationist interpretation of anatomo-clinical correlations across case studies would be an oversimplification – and an outdated one at that (Catani et al. 2012). On the one hand, such a view typically implies an incorrect representationalist viewpoint, whereby imaginary symbolic units are assumed to be contained within damaged regions – for a full treatment of this misconception, see Lamb (1999). On the other hand, the symptoms evinced by a patient following focal damage to a given area may actually reflect the impact of the lesion on distant, more critical areas. Though spared in themselves, these key hubs may be



Chapter 4.  Building up from breakdown 129

secondarily affected due to their disrupted connections with other supporting regions. Thus, observable deficits need not always be directly related to the locus of injury (Lichtheim 1885; Wernicke 1874). Moreover, neither the models nor the anatomo-clinical correlations behind them indicate that the region in question is solely responsible for its associated function. Although particular higher-order cognitive processes rely more critically on some areas than others, every mental operation involves distributed patterns of activity over neighboring and remote regions, which may be structurally and/ or functionally connected. In fact, even in cases of severe and profuse focal lesions, patients often exhibit only partial loss of function, suggesting that some domain-relevant circuits were preserved beyond the damaged tissue. Therefore, a putative relation between a brain region and a cognitive function should not be mistaken for one of self-sufficiency. Likewise, no region is exclusively related to any given process. Both gross and fine-grained areas across the brain participate in multiple operations, and these may pertain to highly specific or highly general domains. As a matter of fact, even in cases of selective dysfunctions, patients manifest alterations in several cognitive skills, though typically with varying degrees of severity. Once again, the models above should not be taken to capture any sort of one-to-one mapping. Once these specifications are made, the conclusions set forth in Section 4.3 and the models introduced in Section 4.4 can be appraised fairly. First, such propositions are empirically grounded, which sets them apart from well-known approaches in cognitive TIS (see Chapter 1, Section 1.2). Second, they are synthetically informative, as they summarily capture the main patterns emerging across multiple individual studies. Third, they are explicit, testable, and thus, correctible: they are a summation of unambiguous propositions that can be experimentally assessed and ultimately corroborated, rejected, rectified, or expanded. Importantly, too, they can be put in dialogue with research on the systems’ dynamic, real-time workings, as revealed via neuroscientific and behavioral approaches (see Chapters 5, 6, and 7). In sum, although these models are neither perfect nor exhaustive, they do offer a highly useful and plausible starting point to understand the neurocognitive organization of IR systems. 4.6 From static maps to dynamic pictures The previous pages have offered a panoramic view of the lesion-based approach as applied to the study of IR. Evidence from numerous studies was first scrutinized to reveal specific neurofunctional patterns and then shown to be jointly formalizable in explicit anatomical models. In particular, putative IR systems appear to be

130 The Neurocognition of Translation and Interpreting

asymmetrically dependent on LH regions and partially independent from more general routes implicated in single-language processes. Moreover, these systems appear to include partly independent routes for BT relative to FT, and for form-level relative to conceptually-mediated operations. Available data further indicate that sentence and word translation are differentially subserved by frontostriatal and temporo-parietal regions, respectively. The result is a considerably detailed map of the neural routes though which IR takes place. Though certainly valuable in itself, however, this initial milestone faces the same limitation of most other maps: just as the blueprint of a water pipeline is blind to the moment-to-moment undulations of the liquids inside it, so these static plots prove uninformative of the physiological modulations underlying IR and its outward temporal signatures. Such specifications are the subject matter of the following chapters, which supplement our preliminary architectural scheme with dynamic pictures of its inner events and their external correlates.

Chapter 5

The dynamics of directionality

5.1

A sense of direction

No sightseeing trip across the state of Minnesota would be complete without a tour of Crystal and Brooklyn Park. These picturesque, tree-laden cities have no shortage of trails and recreational activities for children and grown-ups alike. Basically, you have two options to organize the outing. You could first visit Crystal and then drive to Brooklyn Park, in which case you would take the northbound road of Bottineau Boulevard; or you could go the opposite way, riding along the southbound road. Now, before you make your decision, bear in mind that each option entails different conditions. Whereas the northbound road has a speed limit of 45 mph, the southbound road allows for a maximum of 55 mph. Moreover, while the west side of Bottineau runs parallel to the railroad tracks, the east side is typified by a frontage road and several highway crossings. The moral for prospective tourists is that going from A to B is hardly the same as going from B to A. Should such travelers also be interested in cognitive TIS, they could readily extrapolate the lesson to the construct of directionality. This notion, which captures the contrast between BT and FT across all IR modalities (Pokorn 2011), has recently resurfaced as a major research topic for the field. Nowadays, there is no question that translation and interpreting are routinely practiced both into and from L1s and L2s (Donovan 2004; Monti et al. 2005; Pavlović 2007; Whyatt and Kościuczuk 2013), and that the two directions differ in many respects (Ferreira and Schwieter 2017). However, a number of spurious dictums and misconceptions had to be overcome before these facts were acknowledged as such. For decades, numerous theorists and instructors have disseminated the a priori view that brokering into a non-native language is necessarily suboptimal and should thus be avoided whenever possible (Pokorn 2005). This idea looms large in the context of translation, where FT has long been subject to conscious and unconscious rejection (Ferreira and Schwieter 2017; Pokorn 2005). Before the turn of the century, indeed, a common axiom in popular textbooks was that “[translating] into your language of habitual use […] is the only way you can translate naturally, accurately and with maximum effectiveness” (Newmark 1988: 3). From this perspective, translation was basically reduced to BT.

132 The Neurocognition of Translation and Interpreting

Similar ideas have found their way into the interpreting world – or, more precisely, the Western half of it. Tacitly and otherwise, the International Association of Conference Interpreters has repeatedly deprecated the use of retour – i.e., interpreting into a B language, typically, the L2 (Gile 2005). As written in one of its official publications, “most interpreters, and especially teachers of interpreting, insist on the fact that true interpretation […] can occur only into one’s ‘A’ language” (Bros-Brann 1976: 17). Even in the twenty-first century, SIs at the European Parliament conceive of retour as “a necessary evil in a given market situation” (Wooding 2001, quoted in Donovan 2004: 206). This may account for its relative absence in interpreter education and its neglect in some mainstream interpreting models. In fact, an enduring theoretico-pedagogical tenet of the Théorie du sens has been that discourse production can only be fully spontaneous and idiomatic in one’s mother tongue, which renders it perfectly adequate and superior to an L2 for transmitting information (Seleskovitch 1968; Seleskovitch and Lederer 1989). In short, this IR modality has also been marked by an eschewing of processes in forward direction, and, with it, of directionality as a whole. Notwithstanding, language brokering from L1 into L2 is a widespread phenomenon. FT is routinely practiced in several countries, especially those which are not (primarily) Anglophone (Ferreira and Schwieter 2017). For example, work in this direction constitutes a recurrent demand for 91% of translators in Poland (Whyatt and Kościuczuk 2013) and 84% of translators in Spain (Roiss 2001), and it is actually more prevalent for roughly three out of four professionals in Croatia (Pavlović 2007). The same is true of retour, which became dominant during the Cold War in countries that did not welcome foreign practitioners (Donovan 2004). Even contemporarily, retour is preferred over L2-L1 interpreting in post-Soviet states, on the assumption that the former guarantees optimal SL comprehension and thus offers a better ground for reformulation (Gile 2005). Likewise, this direction is widely employed in parliamentary meetings across Europe (Monti et al. 2005), where users see it as perfectly acceptable provided basic quality standards are met (Donovan 2004). Indeed, evidence from prospective SIs revealed that the amount of transmitted information can be statistically similar for both directions (Tommola and Helvä 1998). Now, besides proving biased and shortsighted for educational and commercial purposes, the neglect of L1-L2 brokering has had direct theoretical consequences for cognitive TIS. Most notoriously, it implicitly led to the postulation of non-directional models – i.e., descriptive accounts that overlook directionality as a cognitively relevant factor and characterize IR operations in terms of SL systems, for processing input, and TL systems, for processing output. To name but a few examples, this is the case of Nida’s three-stage model (diagrammed in Nida

Chapter 5.  The dynamics of directionality 133



and Taber 1969: 33), the ESIT’s interpretive model (as depicted in Hurtado Albir 1990: 71), and Bell’s psycholinguistic model (laid out in Bell 1991: 59). While these formulations differ dramatically from one another, they all share the basic structure captured in Figure 5.1 – for details about each model, see Chapter 1, Section 1.2. TL text

Phase 1

Phase 3

TL processes

SL processes

SL text

Cross-linguistic (or non-linguistic) processes Phase 2

Figure 5.1  Basic architecture of non-directional models in cognitive TIS. In several models of translation and interpreting (e.g., Nida and Taber 1969: 33; Hurtado Albir 1990: 71; Bell 1991: 59), the processes underlying such activities are characterized in terms of SL and TL processes. Thus, these accounts are blind to the existence of different routes and mechanisms for operations leading from an L2 to an L1, on the one hand, and those leading from an L1 to an L2, on the other. L1: native language; L2: foreign (or second) language; SL: source language; TL: target language.

As seen in Figure 5.1, the models in question distinguish three macro-phases, related to (i) SL processing, (ii) cross-linguistic (or non-linguistic) processing, and (iii) TL processing. Each model describes these phases assuming different theoretical premises. Based on generative-transformational grammar, Nida and Taber (1969) described them as ‘analysis’, ‘transfer’, and ‘restructuring’, respectively. In the interpretive model, they are characterized as ‘comprehension’, ‘deverbalization’, and ‘re-expression’ (Hurtado Albir 1990; Lederer 1994; Seleskovitch 1978, 1968). For Bell (1991), who draws on systemic-functional linguistics and pragmatics, they involve ST analysis, executive-semantic processes, and TT synthesis. Beyond theory-specific differences, the point is that all these formulations share a fundamental feature: the assumption that none of the phases would differ between L2-L1 and L1-L2 tasks. As a matter of fact, insofar as the constructs of SL and TL are blind to the native or non-native status of the languages, these models imply that IR relies on a unitary set of mechanisms for both translation directions. The picture thus created is objectively ill-suited as a cognitive account of IR mechanisms. As shown in Chapter 4 (see Sections 4.3.3, 4.4.1, 4.4.2), each direction relies on partially autonomous networks for processing varied translation units, and

134 The Neurocognition of Translation and Interpreting

any of them can be selectively or differentially affected by brain lesions. Moreover, the existence of separate form-level and conceptually-mediated connections for BT and FT has been formalized in experimentally-based models of bilingual memory (e.g., Kroll and Stewart 1994; Kroll et al. 2010). In failing to capture these systemic properties, non-directional models prove architecturally underspecified; but, beyond such structural omissions, can a cognitive model of IR be presumed valid, robust, and comprehensive if it fully overlooks dynamic distinctions between both directions? Do BT and FT involve identical activation levels across key mechanisms? Do they entail similar information-sharing patterns between relevant cognitive systems? Do their critical sub-operations present the same time course as the overall process unfolds? Neuroscientific and psycholinguistic evidence indicates that the answer to all these questions is a categorical no. The reasons and details behind this laconic statement are laid out in the following sections. 5.2

Multidimensional signatures of directionality

Findings on the neurocognitive signatures of directionality come from three sources. Data on which specific brain regions are engaged during BT and FT have been obtained through functional neuroimaging experiments. Insights into the cross-regional connectivity and temporal dynamics of each direction are afforded by electrophysiological studies. Further specifications on the global efficiency of relevant processes can be derived from behavioral research. Let us address these strands of evidence in turn. 5.2.1 Functional neuroimaging evidence As seen in Chapter 4, BT and FT do not rely on the same set of brain regions. Yet, which are the key areas showing contrastive activation levels for each direction? The very first study to tackle this question was conducted by Klein et al. (1995), who employed PET to examine the neural correlates of various linguistic processes. They recruited right-handed English-French bilinguals with high L2 competence and asked them to perform eight tasks. Six of them involved single-language conditions, namely: word repetition, synonym generation, and rhyme generation, all in L1 and L2. The remaining two were translation tasks involving nouns, adjectives, and verbs. In the first one, participants were asked to audibly translate orally-presented words in forward direction; in the second one, they had to do the same with another set of words, in the opposite direction.



Chapter 5.  The dynamics of directionality 135

Participants were similarly fast and highly accurate in both translation conditions, with nearly 90% of correct responses in each direction. Moreover, BT and FT coincided in some hemodynamic patterns. First, neither direction yielded differential activation in the RH. Also, they involved moderate-to-weak activity increases in inferior temporal (BAs 37/20), superior parietal (BA 7), and cerebellar regions. More notably, significant increases for both conditions were observed in inferior and dorsolateral regions of the left frontal and prefrontal lobes, particularly in BAs 47, 46, 45, 8, and 9. Given that these areas were also engaged during synonym and rhyme generation, the authors concluded that they play a key role in phonological and semantic processing irrespective of task and language. Despite such similarities, a key finding of the study was that, compared to BT, FT was characterized by more significant engagement of the putamen (Figure 5.2). This structure is a key hub of frontostriatal circuits, which play critical roles in linguistic and executive domains. In particular, these include semantic and phonological processes (see Chapter 2, Section 3.3.1.2) and bilingual control functions, such as language selection, lexical selection, and mental-set shifting (see Chapter 2, Section 3.4.2). Moreover, note that the putamen was consistently damaged in several patients exhibiting dissociations between their translation skills in BT and FT (Fabbro and Paradis 1995) – for details, see Chapter 4. In sum, this pioneering study suggested that whereas both translation directions would mainly implicate left frontal regions (alongside complementary contributions of temporo-parietal hubs), FT would be typified by the recruitment of additional subcortical mechanisms subserving linguistic and executive functions. The next study on the issue also relied on PET recordings (Price, Green, and von Studnitz 1999). Six right-handed native German speakers with high proficiency in English read and translated written words presented in their two languages. Their accuracy was similar for the two translation directions, both of which involved greater bilateral activation than the reading conditions in the anterior cingulate, the putamen, and the head of the caudate, as well as left-sided increases in the anterior insula, the cerebellum, and the supplementary motor cortex. Also, they entailed reduced activity in the left medial temporal (BA 39), superior frontal (BA 10), and posterior cingulate (BA 39) cortices, alongside additional decreases in the right middle and inferior temporal cortices (BAs 20 and 21). However, none of these patterns differed significantly between BT and FT. While this null result clashes with the significant modulations reported by Klein et al. (1995), its validity is severely undermined by the study’s extremely low sample size. Indeed, low statistical power objectively reduces the chance of detecting true effects in neuroscientific studies (Button et al. 2013). Less crucially, the two experiments also differed in the conditions used as a baseline for comparison

136 The Neurocognition of Translation and Interpreting

A BT minus L2 repetition

t value 6

4.7 B FT minus L1 repetition 3.7 2.5

Figure 5.2  Neural signatures of word translation in forward and backward direction. Averaged PET subtractions of cerebrospinal fluid increases in the left inferior frontal cortex for twelve subjects, superimposed upon their averaged MRI scans. The image shows significant activity increases upon subtraction of L2 repetition from BT (panel A) and L1 repetition from FT (panel B). The color bar indicates activation intensity based on t-values. BT: backward translation; FT: forward translation; L1: native language (English); L2: foreign language (French). Reproduced with permission from Klein et al. (1995).

and in their stimulus presentation modalities.33 Therefore, the results obtained by Klein et al. (1995) are not necessarily challenged by this second study. In fact, the finding that each translation direction involves different activity patterns has been replicated and further specified in several other experiments. 33. For a fuller treatment of these discrepancies, see García (2013a).

Chapter 5.  The dynamics of directionality 137



This was shown, for instance, in the very first functional imaging investigation of simultaneous interpreting (Rinne et al. 2000; Tommola et al. 2001). A sample of conference interpreters, with five to twenty years of professional experience, were administered shadowing and interpreting tasks in their L1 (Finnish) and their L2 (English). Stimuli consisted in eight speeches lasting between 3.5 and 4 minutes at a rate of 98 words per minute. All tasks were performed twice, each time with a different ST. Qualitative assessments of interpreting performance revealed greater propositional accuracy in the L1-L2 than in the L2-L1 direction. This difference, the authors maintain, may reflect better comprehension of the STs in L1 than in L2. The study’s most informative results, however, concerned underlying brain activation patterns. While neither interpreting condition involved distinctive RH activations, they differed in their engagement of left sided-regions. Relative to L2 shadowing, interpreting into L1 yielded higher activity levels in the left supplementary motor cortex (BA 6) and a site anterior to Broca’s area (BA 46) – Figure 5.3, panel A. On the other A. L2–L1 interpreting minus L2 shadowing

B. L1–L2 interpreting minus L1 shadowing

1

2

3

Figure 5.3  Neural signatures of simultaneous interpreting into L1 and L2. Statistically significant increases in regional cerebral blood flow, overlaid on a brain template, when simultaneous interpreting from L2 into L1 was compared with L2 shadowing (panel A) and when simultaneous interpreting from L1 into L2 was compared with L1 shadowing (panel B). From left to right, the rows show (1) front and back views, (2) right and left lateral views, and (3) bottom and top views. L1: native language (Finnish); L2: foreign language (English). Reproduced with permission from Tommola et al. (2001).

138 The Neurocognition of Translation and Interpreting

hand, a subtraction between L1 shadowing and interpreting into L2 showed more extended left frontal activations in the same areas, alongside additional activations in the left inferior temporal lobe (BAs 20 and 28) – Figure 5.3, panel B. Finally, a direct comparison between those two contrasts showed that L1-L2 interpreting entailed greater activity in Broca’s area than its opposite condition. This difference might reflect higher demands on WM and morphosyntactic processing mechanisms (García 2012a). Additional insights have been obtained through fNIRS, a non-invasive technique that allows targeting a preselected brain region to measure task-related changes in the concentration of oxyhaemoglobin (O2Hb) and deoxyhaemoglobin (HHb) (Quaresima, Bisconti, and Ferrari 2012). In particular, O2Hb changes are considered highly sensitive indicators of cerebral blood flow during cognitive activity. Given its design, functioning, and portability, this method can track brain modulations while subjects are sitting in upright position, which offers a more realistic scenario to assess translation and interpreting processes.34 Available comparisons between BT and FT through fNIRS have yielded congruent results. In a first experiment, Quaresima et al. (2002) monitored hemodynamic activity in and around Broca’s area (Figure 5.4, panel A) as early Dutch-English bilinguals engaged in BT and FT of simple sentences (e.g., I’m eating fish and chips, She writes with a pencil). Relative to sentence reading, both translation directions involved similar activity patterns, with more pronounced increases in different parts of the inferior frontal cortex, including Broca’s area. However, sites adjacent to this region evinced differential modulations for BT and FT (Figure 5.4, panel B). Such a finding reinforces the notion that while both directions share broad processing mechanisms, they might also depend on partially distinct microanatomic circuits (García 2012a, 2013a). Further fNIRS evidence on directionality effects has been provided by He et al. (2017). In their study, highly proficient Chinese-English bilinguals majoring in translation studies were asked to read and translate naturalistic texts in both languages. The texts were matched in several variables, including their translatability. During the tasks, hemodynamic changes were recorded over 14 left-hemisphere sites, covering portions of Broca’s area, the dorsolateral prefrontal cortex, and other frontal structures (e.g., premotor and supplementary motor cortices). Average translation times for BT and FT were statistically similar.35 However, O2Hb concentrations revealed underlying processing differences. Specifically, 34. For further details on the advantages and disadvantages of fNIRS, see Quaresima, Bisconti, and Ferrari (2012). 35. Curiously, the authors state that translation times differed significantly between directions. However, they only report a significant main effect of task including all four conditions (BT, FT,

Chapter 5.  The dynamics of directionality 139



B. Activity modulations during translation

Front

B1. Backward translation

1.5 1.0 0.5 0.0 –0.5 –1.0 ∆ Conc (µM)

A. Optical probe position

1.5 1.0 0.5 0.0 –0.5 –1.0 0 10 20 30 40 50 60

0 10 20 30 40 50 60 1.5 1.0 0.5 0.0 –0.5 –1.0

Time (s)

1.5 1.0 0.5 0.0 –0.5 –1.0 0 10 20 30 40 50 60

0 10 20 30 40 50 60

0 10 20 30 40 50 60

O2Hb HHb

1.5 1.0 0.5 0.0 –0.5 –1.0 0 10 20 30 40 50 60

1.5 1.0 0.5 0.0 –0.5 –1.0 0 10 20 30 40 50 60

Front

B2. Forward translation

∆ Conc (µM)

1.5 1.0 0.5 0.0 –0.5 –1.0

1.5 1.0 0.5 0.0 –0.5 –1.0

0 10 20 30 40 50 60

0 10 20 30 40 50 60

Time (s)

1.5 1.0 0.5 0.0 –0.5 –1.0

1.5 1.0 0.5 0.0 –0.5 –1.0 0 10 20 30 40 50 60

0 10 20 30 40 50 60

0 10 20 30 40 50 60

O2Hb HHb

1.5 1.0 0.5 0.0 –0.5 –1.0 0 10 20 30 40 50 60

1.5 1.0 0.5 0.0 –0.5 –1.0 0 10 20 30 40 50 60

Figure 5.4  Differential modulations of activity in and around Broca’s area during backward and forward translation. A. Schematic representation of the optical probe and the twelve measurement sites over the left lateral frontal cortex, centered around Broca’s area as localized by the 10–20 system. Sources and detectors are indicated with white and black circles, respectively. B. Topographic presentation of time courses of Δ[O2Hb] (continuous line) and Δ[HHb] during backward translation (panel B1) and forward translation (panel B2). The figures show the grand average of results from selected locations over eight subjects. The vertical lines indicate the translation period. HHb: deoxyhaemoglobin; O2Hb: oxyhaemoglobin. Reproduced with permission from Quaresima et al. (2002).

L1 reading, L2 reading), without presenting results of post-hoc tests. When post-hoc comparisons are conducted (based on reported means and standard deviations), response-time differences between BT and FT turn out to be non-significant (p > .10).

140 The Neurocognition of Translation and Interpreting

greater modulations of neural activity were observed for FT than BT exclusively over Broca’s area. In sum, this study suggests that directionality differences involving this crucial linguistic hub hold across various translation units. However, given that fNIRS offers limited brain coverage, dissimilar contributions of other areas for each direction cannot be ruled out based on this evidence. In sum, functional neuroimaging results convergently show that BT and FT involve different activation levels in regions related to linguistic and executive processing. In general, FT seems characterized by increased engagement of frontostriatal circuits, cutting across subcortical (e.g., the putamen) and cortical (e.g., Broca’s area) hubs. This pattern has been observed for different language pairs, units, and modalities, highlighting the neurocognitive ubiquity of directionality effects. Yet, given the analytical approaches employed in the above studies and intrinsic limitations of hemodynamic techniques, this evidence is moot on the cross-regional connectivity and temporal dynamics of IR processes. Fortunately, these gaps have been bridged in a series of EEG studies, shedding further light on the dynamic differences between BT and FT. 5.2.2 Electrophysiological evidence Beyond the distinct contributions of individual brain regions, the mechanisms underlying translation and interpreting (just as those supporting any other cognitive process) depend on coordinated modulations across distant circuits which manifest distinct dynamics through time. Such biological signatures can be tapped via measures of functional connectivity and ERPs, as detailed in Chapter 2 (Section 2.7.1.2). Their application in IR studies allows us to refine our understanding of the neurocognitive particularities of BT and FT. Preliminary data were first reported by Kurz (1994, 1995). Based on the socalled coherence method, this study examined functional connectivity changes in professional conference interpreters as they rendered speeches in both directions. To circumvent motor artifacts, all participants were asked to perform the task silently – i.e., without overt articulation. Each condition involved alternations between four-minute periods of simultaneous interpreting and one-minute resting periods. Compared to resting activity, interpreting in both directions was characterized by increased coupling between frontal and temporal recording sites over the LH. Also, compared to L2-L1 interpreting, L1-L2 interpreting was associated with greater RH involvement. Such a pattern could reflect differential reliance on attentional or pragmatic functions (Muñoz, Calvo, and García 2018; Paradis 2009), but no firm conclusions can be derived from this study on account of its small sample size and additional methodological limitations.

Chapter 5.  The dynamics of directionality 141



Contrastive connectivity patterns for BT and FT were also reported by García, Mikulan, and Ibáñez (2016). In this study, Spanish-speaking translators with high competence in English were required to read and translate series of words in both languages, including concrete, abstract, cognate, and non-cognate nouns. Cross-regional coupling and decoupling was estimated via the weighted phase lag index (Vinck et al. 2011). Subtraction of L2 reading from BT and L1 reading from FT revealed differential patterns for each translation direction. Notably, beta frequency results showed greater information exchange in right-sided temporo-occipital networks for BT and in bilateral fronto-temporal networks for FT (Figure 5.5). The authors surmised that the latter pattern may reflect additional recruitment of cognitive control mechanisms. A. BT minus L2 reading

wPLI 0.05 0.17 β band

B. FT minus L1 reading

wPLI 0.04 0.16 β band

Figure 5.5  EEG connectivity patterns during translation from ten professional translators. Illustration of the 10% stronger connections in the beta band (13–30 Hz) between pairs of electrodes according to the weighted phase lag index (wPLI). The images depict significant connections upon subtraction of L2 reading from BT (panel A) and L1 reading from FT (panel B). BT: backward translation; FT: forward translation; L1: native language (Spanish); L2: foreign language (English). Reproduced with permission from García, Mikulan, and Ibáñez (2016).

These patterns broadly align with those observed in the first-ever intracranial EEG study on IR (García, Mikulan, and Ibáñez 2016), which allowed tracking directionality differences, literally, within the brain. The same protocol described above was administered to a proficient Spanish-English bilingual undergoing presurgical assessment for refractory epilepsy. Despite lesions to the left parietal and superior temporal cortices, the patient exhibited full preservation of her language functions, including verbal concept formation, reasoning, comprehension, and expression. Real-time recordings during translation tasks were acquired over 105 channels

142 The Neurocognition of Translation and Interpreting

implanted in left frontal, temporal, and parietal regions – but note that the subject showed RH dominance for language processing. Connectivity analyses, based on the weighted symbolic mutual information metric (King et al. 2013), showed that intra- and inter-lobe activity differed markedly between BT and FT. Inter-regional connections for BT were more widely and less densely distributed across frontal, temporal, and parietal regions, notably converging in a posterior node located in the superior portion of the precuneus (Figure 5.6, panel A). Conversely, FT was characterized by more intense information sharing among anterior temporal, frontal, and prefrontal regions (including the medial frontal and orbitofrontal cortices), with significant connectivity patterns being mostly concentrated in those frontal areas (Figure 5.6, panel B). Such discrepant patterns might reflect greater executive control demands for FT than BT, which, in turn, would differentially rely on more specific perceptual and lexical mechanisms (García, Mikulan, and Ibáñez 2016). A. Backward translation

B. Forward translation .010 × 10–3 .009 .007 .006 .005 .004 .002 .001 × 10–3 p value

Figure 5.6  Intracranial EEG connectivity during translation by a proficient bilingual. Illustration of the 5% most significant differences in connectivity values in the beta band using weighted symbolic mutual information (wSMI) analysis. The images depict significant connections for backward translation (panel A) and forward translation (panel B). Recordings were obtained from the language-non-dominant hemisphere. Reproduced with permission from García, Mikulan, and Ibáñez (2016).

In addition to inter-regional connectivity, the two translation directions also differ in their temporal dynamics. As an instance of IR unfolds, several processes occur in relative succession. The following two studies have aimed to track their time course comparing BT and FT. Christoffels, Ganushchak, and Koester (2013) recruited proficient Dutch-English bilinguals and had them perform overt production tasks, including BT and FT of single words. Although no behavioral differences emerged between these conditions, directionality effects were indexed by two ERPs, both evincing maximal modulations over centro-parietal sites.

Chapter 5.  The dynamics of directionality 143



On the one hand, the N400 component was more significantly modulated during BT than FT (Figure 5.7). This effect could reflect more effortful semantic access to input words in BT, as shown in previous studies (García 2015a; Moreno, Rodríguez-Fornells, and Laine 2008). However, considering the absence of latency effects in this component, the authors further surmised that both directions could involve access to the conceptual system, in line with results from a translation recognition study (Guo, Misra, Tam, and Kroll 2012). On the other hand, as compared to BT, FT involved greater amplitude of the P2 component over central and parietal sites (Figure 5.7). Given the relevance of this component as an index of early processing demands, the observed pattern would indicate greater lexical retrieval effort and longer-lasting word selection processes for FT. Moreover, on account of the latency of P2 modulations, this finding suggests that differential neurocognitive dynamics for each direction begin to be deployed as early as 200 ms after source-segment viewing. –10

Frontal

–10

(μV) –6

(μV) –6

–2

–2

2

200 400 600 8001000

(ms)

2

Central

–10 (μV) –6

200 400 600 8001000

(ms)

–2 2

6

6

6

10

10

10

14

14

14

IH BT

Parietal

Ctrl BT

IH FT

200 400 600 8001000

Ctrl FT

Figure 5.7  ERP waveforms underlying translation of two word classes in both directions. In the 150–270 ms window, FT yielded more positive amplitude than BT over central and parietal sites. In the 300–600 ms window, BT yielded more negative amplitude than FT, also over central and parietal sites. The same was true for IHs relative to control words over central sites. For presentation purposes only, the signal was low-pass filtered (10 Hz) and the waveforms were averaged across electrodes of three regions: frontal, central, and parietal. BT: backward translation (solid line); FT: forward translation (dotted line); Ctrl: control words (black); IHs: interlingual homographs (gray). Reproduced with permission from Christoffels, Ganushchak, and Koester (2013).

Similar findings were obtained in the intracranial study by García, Mikulan, and Ibáñez (2016). Significant differences were observed in two of the patient’s implanted areas: the posterior fusiform gyrus and the anterior middle temporal gyrus (Figure 5.8). While these recordings were obtained from the language-non-dominant hemisphere, note that language-related activity in such regions spreads broadly across the two hemispheres (Cohen et al. 2003; Kennepohl et al. 2007). Moreover, engagement of the language-non-dominant hemisphere during verbal processing is greater for L2s than L1s (Perani et al. 2003).

144 The Neurocognition of Translation and Interpreting

BT involved more negative amplitude than FT in the two regions in question, with maximal modulations between 500 and 580 ms (Figure 5.8, panels A and B). Such deflections may be interpreted as part of the N400 family, especially since this component peaks considerably later in bilinguals than in monolinguals during verbal processing (Moreno and Kutas 2005). Accordingly, and given the language-related roles of their triggering regions, such patterns mirror those reported by Christoffels, Ganushchak, and Koester (2013), further suggesting that SL comprehension is more effortful in BT than in FT. An additional effect recorded within the posterior fusiform gyrus (Figure 5.8, panel B) also aligned with a previous finding by Christoffels, Ganushchak, and Koester (2013). Specifically, FT evinced larger amplitude than BT between 220 and 250 ms after word presentation. On account of its latency and polarity, this effect likely represents a modulation of the P2 component, suggesting increased cognitive effort for FT in the initial stages of IR. Furthermore, these data corroborate that distinct neurocognitive dynamics for each translation direction become significant in less than one fourth of a second after process onset. Additional data have been contributed by Jost et al. (2018). This study involved native speakers of French who had learned English after the age of seven and attained intermediate to high proficiency levels. In addition to within-language tasks, participants were asked to overtly perform BT and FT of single words. Behavioral outcomes revealed similar accuracy for both conditions, with slower responses for FT. Crucially, this effect was mirrored by neural differences between the two tasks. As shown by source estimation analyses, FT involved stronger activation in the posterior cingulate gyrus and the thalamus in a late window (ranging from 589 to 680 ms). As noted in the paper, activity changes in the cingulate gyrus could reflect increased attentional demands and higher arousal for FT, whereas the pattern traced to the thalamus might reflect greater motoric and semantic demands. In sum, EEG evidence reveals further specificities of directionality as a key modulator of neural activity during IR. First, relative to BT, FT seems to imply more widespread functional coupling of linguistic and executive mechanisms distributed across lobes and hemispheres. Second, converging ERP evidence from scalp-level and intracranial recordings suggests that FT would involve more effortful lexical retrieval and selection operations as well as higher attentional load, whereas BT would be typified by greater source-segment processing demands. Finally, discriminatory dynamics for each direction become apparent around 200 ms after source-segment visualization, revealing fast cognitive differentiations between directions that could hardly be tapped through non-neuroscientific methods.

Chapter 5.  The dynamics of directionality 145



A. Anterior middle temporal gyrus 40

BT FT p < .01

30 20 10 μV 0 –10 –20 –30 –100

0

100

200

300

400 500 Time (ms)

600

700

800

900

1000

B. Posterior fusiform gyrus 80 60

BT FT p < .01

40 20 μV

0 –20 –40 –60 –80 –100

0

100

200

300

400 500 Time (ms)

600

700

800

900

1000

Figure 5.8  Intracranial ERP recordings from a proficient bilingual during BT and FT. Significant differences between directions occurred between 550 and 650 ms in the anterior middle temporal gyrus (panel A), and at 220–250 and 500–750 ms in the posterior fusiform gyrus (panel B). BT: backward translation; FT: forward translation. Reproduced with permission from García, Mikulan, and Ibáñez (2016).

146 The Neurocognition of Translation and Interpreting

5.2.3 Psycholinguistic evidence Neurological phenomena, on their own, do not reveal whether directionality effects have significant outward manifestations. In fact, hemodynamic and electrophysiological differences between BT and FT may show that each direction relies more or less heavily on specific pathways or cognitive mechanisms, but those patterns need not be mirrored by behavioral effects. The latter can be (and have repeatedly been) assessed via psycholinguistic translation experiments. As stated elsewhere in this volume, the evidence thus obtained can inform neurocognitive research in three major ways. First, it offers indirect insights into the strength of the underlying connections, on the assumption that slower responses reflect greater cognitive effort. Second, thanks to hypothesis-driven methodological manipulations, they can reveal which processing levels (e.g., form-level, semantic) prove more crucial for BT and FT. Third, they also offer clues on how relevant systems change their functional organization depending on subject-level variables, such as L2 proficiency and translation competence. The ensuing findings represent an important complement to the neuroscientific results reviewed above. Building on a well-established trend, Kroll and Stewart (1990) and SánchezCasas, García-Albea, and Davis (1992) designed bilingual single-word experiments which included BT and FT tasks. A novel finding of their studies was that, at least for specific word classes, BT was performed significantly faster than FT, but that such an asymmetry was attenuated at (relatively) high levels of L2 proficiency. Complementary results reported shortly afterwards replicated and extended these findings. In the study by Kroll and Stewart (1994), participants translated words presented in two different groupings. In one condition, successive words belonged to the same semantic category (e.g., weapons); in the other, they were sequenced in a random fashion. Two main differences were observed between translation directions. FT was performed more slowly than BT and, unlike the latter, it evidenced a categorical interference effect – i.e., slower performance for the semantically-grouped than the randomly-grouped condition. This motivated the preliminary conclusion that semantic connections played more critical roles in FT than BT. More notably, such results led to the postulation of an integrative theoretical proposal known as the Revised Hierarchical Model – Figure 5.9. This model aims to account for the relations among three systems in bilingual memory: an L1 lexicon, an L2 lexicon, and a shared conceptual system. Note that, in this diagram, the model focuses on low-competence bilinguals whose dominant language is the L1. In this context, the L1 lexicon is proposed to be wider than its L2 counterpart, hence the larger size of its corresponding box. The two systems are linked directly by reciprocal form-level connections, and indirectly by separate sets of conceptually-mediated connections. A non-trivial aspect of this model is that it

Chapter 5.  The dynamics of directionality 147



CONCEPTS

L1

L2

Figure 5.9  The Revised Hierarchical Model. The model sets forth an account of the organization of bilingual memory, based on three components: an L1 lexicon, an L2 lexicon, and a shared conceptual system. These are linked by reciprocal routes, including separate form-level and conceptually-mediated connections for each translation direction. Solid lines represent strong(er) connections, whereas dotted lines indicate weak(er) connections. L1: native language; L2: foreign language. Adapted from Kroll and Stewart (1994).

aligns well with the neuroanatomical models introduced in Chapter 4 (Section 4.4). In particular, it explicitly captures the claims that (i) (at least some of) the routes involved in IR are independent from those supporting single-language processes, (ii) BT and FT rely on partially separate connections, and (iii) IR in either direction can recruit both form-level and conceptually-mediated pathways (see Chapter 4, Sections 4.3.2, 4.3.3, and 4.3.4). However, the model introduces a number of specifications not present in the neuroanatomical proposals previously introduced. First, it posits that different connections have different weights. In particular, the form-level route for BT would be stronger than the one for FT (as depicted by the solid and dotted lines, respectively). Also, it assumes that links to the conceptual system are stronger for the L1 than the L2 lexicon. However, the model recognizes that the relative strength of the different routes tends to become more balanced as L2 proficiency increases. On these assumptions, faster performance for BT than FT can be explained by two factors. First, form-level connections are stronger for the former direction. Second, because of the entrenchment of such connections, BT would be able to largely bypass semantic access and predominantly rely on direct lexical links. By contrast, FT would necessarily depend on semantic mediation, hinging on the strong connections between the conceptual system and L1 word forms. The model further posits that the conceptual route would prove slower than its form-level counterpart because it involves more, longer links. As regards the presence of categorical interference for FT but not for BT, the model posits that presentation of multiple words belonging to the same semantic

148 The Neurocognition of Translation and Interpreting

field would promote the activation of multiple exemplars of the category at hand, which would complicate lexical selection and reduce processing speed. Given that connections between conceptual and lexical information would be stronger for L1 than for L2, only FT would be affected by categorical interference. However, the Revised Hierarchical Model further postulates that increasing levels of L2 proficiency lead to stronger connections between concepts and words in that language, so that these would be able to swiftly access relevant meanings. Thus, in highly proficient bilinguals, semantic factors would play critical roles in both directions, diminishing or abolishing differences between BT and FT. This, in fact, has been confirmed in several studies (e.g., Christoffels, de Groot, and Kroll 2006; García et al. 2014; McElree, Jia, and Litvak 2000). In this sense, subsequent experiments with high-proficiency bilinguals indicated that conceptual mediation plays significant roles in both directions (de Groot, Dannenburg, and van Hell 1994; Duyck and Brysbaert 2004; Heij et al. 1996). For example, Duyck and Brysbaert (2004) showed that the magnitude of number words affected performance in BT and FT. Specifically, numbers pointing to low quantities (e.g., one, two) are translated faster than those pointing to larger quantities (e.g., eight, nine), even when matched for sublexical features. Insofar as number magnitude is a semantic feature, this evidence indicates that, at least for certain word classes, conceptually-mediated routes are similarly recruited irrespective of directionality – although they may still play a more dominant role for FT at low proficiency levels. Furthermore, behavioral performance in both translation directions is susceptible to translation competence, but up to a certain level. In a study by García et al. (2014), beginner translation students, advanced translation students, and professional translators performed a series of word-level tasks, including BT and FT. Importantly, all groups differed in their translation competence ratings, but successive groups were matched for L2 proficiency. Results showed that the advanced students outperformed beginners, but their response times did not differ from those of professionals. This suggests that, after a certain degree of strengthening related to field-specific training, translation routes for both directions may reach ceiling levels. Moreover, complementary analyses showed that, whenever directionality effects emerged for specific lexical categories (e.g., concrete, abstract, cognate or non-cognate nouns) in any of the groups, they consistently revealed advantages for BT over FT. Therefore, even though directionality effects may cease to emerge cross-categorically at high levels of competence, they may still be observed for specific lexical domains. Note that the Revised Hierarchical Model features a very simple architecture and that dozens of relevant papers have been published since its formulation. Although this proposal has been criticized in several aspects (Brysbaert and Duyck 2010),

Chapter 5.  The dynamics of directionality 149



the core tenets outlined above still seem broadly plausible to this day. In fact, a very recent computational model has integrated the Revised Hierarchical Model with additional theoretical principles and managed to simulate several well-established effects in the bilingual lexicon, including directionality and competence-related effects (Dijkstra et al. 2018). Although the Revised Hierarchical Model, as depicted in Figure 5.9, certainly underspecifies the collection of mechanisms needed to account for all patterns in the literature, it offers a basic approximation to the minimal functional distinctions underlying asymmetric translation effects. In sum, neurocognitive differences between BT and FT are systematically mirrored by behavioral effects in low-proficiency bilinguals, who are typically faster in the former direction. However, this is not the case in high-proficiency L2 users, including professional translators. Still, even the latter may manifest faster processing for BT than FT in circumscribed lexical fields. More generally, form-level and conceptually-mediated processes seem operative in both directions, although their relative contributions also vary according to subject- and task-related factors. 5.3

Back and forth

The findings described above should make it abundantly clear: from a cognitive standpoint, directionality matters. The resources recruited during IR differ greatly depending on whether SL materials are presented in the L2 or the L1. Specifically, BT and FT can be distinguished in terms of the regions they recruit, the relative engagement of shared substrates, the interactions among distant hubs, the time course of ongoing processes, and their behavioral manifestations. First, as seen in Chapter 4, each direction seems to depend on relatively autonomous pathways. Irrespective of the translation unit and modality, it is rare for a single lesion to affect BT and FT in an equal manner. Therefore, a partial anatomical differentiation can be postulated between both operations. Second, BT and FT further differ in the intensity with which they engage critical neural substrates across units and modalities. In particular, FT seems typified by hyper-activation of frontostriatal hubs (e.g., the putamen, Broca’s area) implicated in linguistic and executive functions – although additional distinctions could be expected in other hitherto underexplored areas. This is the case even when behavioral performance proves similar between directions, stressing the value of neuroscientific approaches to tap otherwise undetectable patterns. Accordingly, BT and FT place discrepant demands on various cognitive systems. Third, the integration of information across distant brain areas is also dissimilar between directions. BT appears to involve more intense coupling among posterior hubs implicated in lexico-semantic processes. Instead, FT features comparatively

150 The Neurocognition of Translation and Interpreting

sparser connectivity patterns distinctively involving frontal sites, which may reflect higher implicit control demands for its underlying operations. Fourth, the temporal unfolding of key sub-processes also changes as a function of directionality. The electrophysiological dynamics of BT and FT become significantly different before the first quarter of a second since source-segment viewing. Also, stimulus-locked modulations indicate greater demands for FT in early processing windows (≈ 200 ms), and for BT in later windows (≈ 400 ms). Thus, the inner time course of IR is also sensitive to this factor. Finally, the behavioral manifestation of directionality effects seems broadly driven by subject-related variables. Specifically, BT typically proves faster than FT at low levels of L2 proficiency, but such a difference tends to disappear in highly competent individuals, including aspiring and professional translators. Not­ withstanding, even in those cases, faster performance for BT than FT can be observed in circumscribed lexico-semantic domains. In short, the neurocognitive impact of directionality can be traced even in external conduct. All things considered, the direction in which IR is performed constitutes a multidimensional determinant of the neurocognitive operations involved. BT and FT can be linked to partially distinct neural structures, regional activity levels, cross-areal interactions, temporal dynamics, and outward manifestations. As one can reckon by traversing a certain boulevard in Minnesota, the road from A to B is not nearly the same as that from B to A. These observations can be directly incorporated in ongoing discussions of directionality within cognitive TIS. On the one hand, they align with previous results from non-neural approaches. For example, indicators of greater cognitive effort for FT than BT have been previously identified in eye-tracking (Pavlović and Jensen 2009) and keylogging (Ferreira 2012) studies, including pupil dilation, average fixation duration, overall processing time, and pause duration. Likewise, the finding that gaze-time values are higher for BT than FT in both student and professional translators (Pavlović and Jensen 2009) is compatible with the finding that N400 modulations are greater for the former direction (Christoffels, Ganushchak, and Koester 2013), since both could point to greater demands for processing L2 input. On the other hand, some neuroscientific results challenge previous claims stemming from non-neural trends. For instance, they clash with the contention that ST processing is equally demanding for both directions, as proposed by Pavlović and Jensen (2009) based on eye-tracking data. In fact, all EEG results surveyed above capture task-related neurophysiological modulations which are time-locked to the stimulus, meaning that they index discrepant processes related to source-segment processing in each direction. Insofar as different methods track different dimensions of the cognitive processes at hand, the lesson here is that a null effect in one dependent variable (e.g., gaze time, pupil dilation) must not be taken to reflect a



Chapter 5.  The dynamics of directionality 151

null effect of the independent variable (here, directionality) as a whole. Specific differences between BT and FT may be absent in ocular dynamics but still prove significant and pervasive at a neurocognitive level, as seems to be the case. Moreover, whereas directionality effects in non-neural approaches are typically discussed in functionally unspecific terms (e.g., as “greater cognitive effort”), the particularities of neuroscientific methods allow for more precise interpretations. For example, the detection of directionality differences in a given brain area allow researchers to raise hypotheses regarding which specific sub-processes could be at play in the effect, provided putative functions of that area are well characterized in the literature. Similarly, given that most ERPs can be associated to very specific cognitive operations, differences in specific components (e.g., the P2 or the N400, as seen above) indicate that the differences between direction are driven by specific executive (e.g., attentional) or linguistic (e.g., lexico-semantic) dimensions. In this sense, brain-based research allows decomposing the coarse-grained notion of “cognitive effort” into more precise factors, thus paving the way for a refined understanding of directionality effects and other phenomena involved in IR. From a broader theoretical perspective, neuroscientific findings cast serious doubts on the validity and generalizability of non-directional models. If those accounts were correct, any process in which SL input is reformulated in a TL should involve the same cognitive systems. However, this is not the case: succinctly, while both BT and FT imply rendering SL input as TL output, each of them relies on variedly different mechanisms. Therefore, on the basis of neurocognitive evidence, non-directional models can be deemed exceedingly restrictive or underspecified, if not altogether wrong. At the very least, they are misleading, in that they nurture the fallacy that all instances of IR can be identically characterized irrespective of which language (L1 or L2) is used as input and output. Even if one were to defend them on the grounds that they are actually aimed to describe BT only, they would have to be recognized as extremely shortsighted, for they would be oblivious to the widespread use of both directions across nations, settings, and modalities. Even worse, they would also be culturally biased, as they would not be applicable to those contexts in which FT is the rule. More generally, note that non-directional models such as those mentioned at the outset are uninformed by experimental evidence. Insofar as present claims for the relevance of directionality are correct, the inadequacy of those models can be largely attributed to their reliance on purely analytical, observational, anecdotal or introspective approaches. It would thus appear that solid cognitive accounts of IR can hardly be reached by neglecting objective, direct, and quantifiable proxies of mental activity.

152 The Neurocognition of Translation and Interpreting

In addition, the data reviewed throughout the chapter afford valuable interpretive constraints for the neuroanatomical models described in Chapter 4 (Section 4.4). First, although the models by García (2012a) and Fabbro (1999) depict each direction via fully separate arrows, these must be assumed to represent broadly overlapping circuits. This follows directly from the observation that only a few isolated regions yield hemodynamic differences between BT and FT, which means that broader areas which fail to exhibit direction-specific activity patterns may be similarly engaged in both cases. Therefore, those graphic links, like most pictorial devices in boxes-and-arrows models, should not be understood in discrete, isomorphic terms. Neither should directionality effects be reduced to discrepant activation patterns for specific isolated regions. In fact, as seen in functional connectivity analyses, BT and FT manifest differing levels of cross-regional interaction, which, in turn, likely reflect the interaction among varied cognitive domains in each case. Accordingly, the direction-specific arrows in the preceding neuroanatomical models might be better conceived as a short-hand for discrepant patterns of coupling and decoupling dynamics spanning multiple neurocognitive hubs. Likewise, the systems mediating each translation direction must not be assumed to operate isochronously. In fact, the temporal windows in which specific functions are recruited, and the overall processing time of the task, are also dissimilar between the distributed and partially overlapping mechanisms subserving BT and FT. In this sense, although the models in question fail to explicitly capture the point in their visual designs, access to verbal and non-verbal information must be acknowledged as following unequal time trajectories in each directional route. Specifications like this showcase the constructive benefits of erecting models with multimethodological (and hence, multidimensional) inputs. 5.4

In the right direction

Brain-based research on directionality proves pivotal for uncovering the myriad cognitive particularities of BT and FT, evincing the limitations of popular models and establishing fine-grained empirical constraints to inform superior accounts of IR. Yet, directionality is not the only variable molding the inner workings of translation and interpreting. In fact, the translation unit further determines which systems are engaged and how and when they operate. As seen in the following chapter, neuroscience also affords valuable knowledge on this issue.

Chapter 6

Process is as unit requires

6.1

The process’s raw material

Should your reading habits be anything like mine, chances are that, as you progressed through this book, you kept eating and drinking several palatable foods. Your tongue, perhaps, has been covered with bits of salty crackers, surrounded with sour chocolate, washed with bitter coffee, and refreshed with sweet juice. If you look carefully, in fact, traces of your choices may actually be found in some of the preceding pages – or on the corner of your tablet’s screen, for that matter. Now, if you were able to recognize such different flavors it is because each tastant engages specific buds and chemical pathways. Discernible sets of receptor cells specialize in the detection and transduction of saltiness, sourness, bitterness, and sweetness, in connection with taste-specific nerve fibers that trigger distinct brain processes for each flavor. Thus, although the tongue is the general playground where all taste experiences first occur, different sub-mechanisms are engaged depending on the qualities of the input it receives. The principle of “all for one and one for all,” as it were, is hardly applicable to the workings of our gustatory system. By the same token, the Musketeerean motto would fail to describe how the mind processes different translation units. Indeed, if you will indulge my metaphor, translation units come in several flavors. Some expressions are short and simple, but others are lengthy and complex; some evoke clearly defined images, whereas others point to rather diffuse ideas; some look and sound very similar to their potential target versions, but others have no such resemblances; some can be worked out in a flash with full confidence, while others take longer and leave you questioning the result… Considering such diversity, how plausible are those accounts of IR positing a single, undifferentiated set of operations for any and all translation units? Would it not be more likely that, as is the case with taste processing, each translation unit recruits specific mechanisms depending on its particular properties? And should this be the case, is it possible to identify which particular cognitive systems are implicated as a function of key variables, such as the complexity, frequency, semantic associations, and form-level particularities of the unit at hand? As we shall see below, neuroscientific and psycholinguistic studies can valuably inform these issues.

154 The Neurocognition of Translation and Interpreting

6.2 Conceiving translation units From a process-oriented stance, a translation unit can be conceived as the portion of text being attended to in a given instance of IR (Alves and Vale 2009; Bennett 1994; Carl and Kay 2011). In other words, it is the textual segment acting as cognitive input in any of the successive cross-linguistic processes whereby SL material is rendered into a TL throughout a translation or interpreting task – see Malmkjær (1998).36 Now, while this definition may be broadly shared by most scholars in cognitive TIS, divergent proposals have been advanced on whether this construct can be systematically mapped to a particular type of linguistic unit. A number of product-oriented scholars have characterized ST-TT pairings with a focus on lexical items, proposing that translation units typically corresponded to single words or lexemes – for a historical overview, see Hurtado Albir (2001). In corpus studies, for example, it has been claimed that “[l]exical unit and relevant context together form the translational unit” (Teubert 1996: 256). Early work in cognitive TIS partially supported such a view, but only for non-experts. Results from TAPs, for example, suggested that this was the case for language learners, but not for professional translators (Lörscher 1991, 1993). In the latter population, Lörscher (1991, 1993) reported, dominant units would be semantically defined and realized by multiword constructions (e.g., phrases, clauses, sentences). This coincides with the views of Matthiessen (2001: 116), who maintained that “the clause (complex) is a likely candidate as the ‘unit of translation,’ and Malmkjær (1998: 286), who adopted a similar position arguing that “the clause is a manageable unit of attentional focus [and that] it tends to be at clause level that language represents events.” Moreover, a compatible conclusion was advanced in a corpus study and a complementary survey of professional translators and translation editors (Huang and Wu 2009). Irrespective of their approach and conclusions, accounts like those above prove somewhat reductionist in that they seek to identify the unit of translation. However, the fact is that source segments vary greatly as one progresses through a text. Whether IR is performed on the basis of a single word, a word group, a clause, a sentence, or anything in between depends on multiple factors, including the translation modality (e.g., written vs. oral), the text type (e.g., a brochure listing product names vs. a scientific paper), and organizational features of specific ST 36. Note that this conception differs from others advanced in product-oriented accounts, such as those based on comparative stylistics (Vinay and Darbelnet 1995 [1958]) or functional models (Nord 1997). From these perspectives, translation units are defined by considering ST segments and their non-reducible correspondences in the final translation product, so that they may be better dubbed ‘alignment units’ (Carl and Kay 2011).



Chapter 6.  Process is as unit requires 155

passages (e.g., lists of items sequenced with bullets vs. multi-sentential paragraphs), alongside individual tendencies, ad hoc strategies, and ready-made segmentations provided by computer-assisted tools. A cognitively plausible view of the construct thus calls for flexible, non-deterministic conceptions. In this line, Hurtado Albir and Alves (in Munday 2009: 238) refer to translation units as “[any] dynamic segment of the ST, independent of specific size or form, to which, at a given moment, the translator’s focus of attention is directed.” Their account resonates with prior work on translation strategies, indicating that clauseand sentence-level segmentation typically co-occurs with word-for-word renderings in both non-experts and professionals (Tirkkonen-Condit 2005). Likewise, both sentence-by-sentence (Kalina 1998) and word-for-word (Christoffels, de Groot, and Waldorp 2003) reformulation have been recognized as viable practices for interpreters – the latter being arguably preferred under exacting conditions (Darò and Fabbro 1994). Real-time processing data support the relevance of these positions. Combined measures of gazing and typing activities during translation (Carl and Kay 2011) indicate that the segments acting as input can shift depending on complex interactions between ST- and TT-oriented attentional patterns, which in turn differ as a function of translation experience. Besides, although translation units may only partially coincide with linguistically defined elements, such mappings repeatedly occur at the single-word and single-sentence level (Carl and Kay 2011). Furthermore, as shown in dozens of behavioral experiments, different cognitive operations are associated to particular psycholinguistic properties of source words and sentences (García 2015a; Starreveld et al. 2014; van Hell and de Groot 2008). Notwithstanding, several models in cognitive TIS either disregard the role of translation units or address the matter via coarse-grained postulations, often devoid of concrete evidence. Some accounts of IR systems, rooted in rationalist formulations (e.g., Nida and Taber 1969) or limited introspective data (e.g., Kiraly 1995), directly overlook the issue by proposing unit-blind cognitive architectures for their modeling domain. Others prove aprioristically reductionist in characterizing translation mechanisms as a whole by reference to a single unit type. This is the case, for instance, of Bell’s (1991) psycholinguistic account, which posits a general description of IR systems assuming that all translation processes start from, and result in, simple clauses. Still others set forth purely conjectural associations between certain units and particular cognitive operations. Consider, for example, the Théorie du sens, which assumes that, although IR typically hinges on deverbalization (see Chapter 1, Section 1.2.2), specific units (e.g., proper names, numbers, and technical terms) would be translated as correspondences via direct form-level links (Lederer 1994; Seleskovitch 1975). What is more, even the neuroanatomical models introduced in Chapter 4 (Section 4.4) are underdetermined in

156 The Neurocognition of Translation and Interpreting

this sense. Fabbro’s (1999) model of simultaneous interpreting makes no provision for differential processes related to specific translation units, and García’s (2012a) neuroarchitectural account of IR only acknowledges a broad distinction between the general brain systems engaged by lexical and sentential units. Although all of these models have clear merits in many respects, neurocognitive and psycholinguistic findings show that they are either too general or imprecise, if not misleading. In fact, tacitly or otherwise, they may prompt readers to believe that multiple types of translation units are processed via the exact same cognitive resources, with identical demands, and following a similar time course. Yet, as specified in the following sections, none of these conclusions would be correct. The human brain recruits distinct mechanisms depending on the specific “tastes” of to-be-translated segments. In particular, available data reveal fine-grained particularities of the mechanisms engaged by various types of word- and sentence-level translation units. Though far from exhaustive, research conducted on them is decidedly informative, not because they are the units of translation, but rather because they correspond to a considerable subset of actual processing segments. Also, focusing on words and sentences is methodologically convenient, as it allows controlling for several relevant factors (see Chapter 2, Section 2.4). Moreover, as these two types of stimuli have been widely studied in neurolinguistics and psycholinguistics, ensuing results on IR can be interpreted in terms of well-established constraints. Finally, it is worth noting that several neurocognitive patterns related to certain word and sentence types, such as action language, were first detected through atomistic tasks (e.g., García and Ibáñez 2016c; Pulvermüller 2005, 2013) and then proven to be valid even for naturalistic textual materials (e.g., Desai et al. 2016; García et al. 2016a, 2018; Trevisan et al. 2017). Thus, it is not capricious to believe that broad effects detected for specific unit types via fragmentary translation experiments could remain operative in more ecological settings. The question pursued below, then, is not so much ontological as it is descriptive. At this juncture, and especially from a cognitive standpoint, it seems pointless to try to justify a single linguistic unit that all of IR can be traced back to. Rather, once we accept that translation and interpreting are based on myriad types of source segments, the truly informative mission is pinpointing which specific mechanisms are engaged as a function of their distinctive properties. This approximation is no longer a definitional, prescriptive quest for a monolithic and all-valid construct, but a scientifically approachable program with multiple ramifications – many of which can be aptly explored with the tools of cognitive neuroscience.

Chapter 6.  Process is as unit requires 157



6.3

Spatiotemporal correlates of lexical and sentential translation units

Current knowledge on which neurocognitive mechanisms are engaged depending on the translation unit comes from three strands of evidence. Findings from fMRI and tDCS shed light on how crucially particular neural regions are recruited as a function of that variable. EEG studies reveal distinct temporal dynamics associated with semantic and form-level properties of source segments. Finally, psycholinguistic studies illuminate how such factors impact on the overall efficiency of IR in untrained bilinguals, translation students, and professional translators. 6.3.1 Functional neuroimaging evidence The most obvious fact that unit-blind models overlook is that processing demands during IR differ greatly for lexical and sentential segments. First of all, as seen in Chapter 4 (Section 4.3.5), word- and sentence-level processes rely contrastively on declarative and procedural memory systems, mainly subserved by frontostriatal and temporo-parietal circuits, respectively. In fact, when frontostriatal regions are focally damaged, sentence translation becomes significantly more impaired than word translation. Consider, for example, the performance of patient El.M., one of several subjects exhibiting inability to translate upon sustaining left basal ganglia damage (Fabbro and Paradis 1995). As shown in Figure 6.1, such a lesion entails greater deficits in sentence than in word translation, irrespective of directionality – arguably because of the putative role of frontostriatal networks for syntactic processing in L1 and in well-consolidated L2 systems (Birba et al. 2017; Paradis 2009; Ullman 2001a, 2001b). Neuroimaging studies align with this broad distinction. Evidence from hemodynamic techniques indicates that left perisylvian networks are implicated in IR of both words and sentences (Hervais-Adelman et al. 2014; Lehtonen et al. 2005; Quaresima et al. 2002). However, the engagement of frontostriatal hubs is markedly greater for the latter type of units, whereas temporo-parietal activity is more distinctively linked to word translation (Klein et al. 1995) – for a review, see García (2013a). Note, in passing, that these anatomo-functional relations for such units can also be observed in single-language tasks (Paradis 2009; Ullman 2001a, 2001b), suggesting that at least broad IR operations are embedded in general language-processing mechanisms (García, Mikulan, and Ibáñez 2016). Needless to say, however, neither of the above gross regions is exclusively devoted to a particular unit type. Indeed, specific substructures of frontostriatal networks may be distinctively recruited for translating words and sentences. For example, the PET study by Price, Green, and von Studnitz (1999) revealed that

158 The Neurocognition of Translation and Interpreting

Mean accuracy

80

BT FT

60 40 20 0

Words

Sentences

Figure 6.1  Dissociation between word and sentence translation following frontostriatal damage. The figure shows the performance of patient El.M. (see Chapter 4, Table 4.2) on tasks requiring word and sentence translation. Results were significantly lower for the latter unit in both directions. BT: backward translation; FT: forward translation. Data from Fabbro and Paradis (1995).

increased engagement of the basal ganglia during word translation was confined to the putamen and the head of the caudate nucleus. However, an fMRI study on sentence translation showed that hyperactivity in the basal ganglia for such a condition, relative to reading, mainly occurred in the globus pallidus (Lehtonen et al. 2005). These discrepancies likely reflect the dissimilar semantic, syntactic, and executive demands entailed by each unit type, although available evidence does not warrant more precise functional interpretations (García 2013a). Presumably, different sentence types also trigger distinct processes during IR. Indeed, this is the case during single-language tasks (Carreiras and Clifton 2004; Constable et al. 2004; Feng et al. 2015; Mashal et al. 2009). However, the only neuroimaging report on translation assessing contrastive sentence structures failed to reveal differential patterns. In the above-cited fMRI study, Lehtonen et al. (2005) recruited a group of native Finnish-Norwegian speakers and asked them to perform two tasks. In the critical one, subjects viewed 54 sentences in L1 and translated them silently into their L2. After each trial, they were shown another L2 sentence and asked to decide whether it was an acceptable rendition of the original stimulus. In the control task, subjects were asked to read L1 sentences, keep them in memory, and then judge whether they were identical to the sentence shown immediately afterwards. A key manipulation in both tasks was that half the sentences required a word-order change for translation, whereas the other half included more elaborate noun phrases but required no such change. The latter sentence type yielded longer reaction times, which the authors attributed to greater WM demands due to the presence of double adjectives before the head noun. However, hemodynamic



Chapter 6.  Process is as unit requires 159

patterns revealed no congruent differences. Relative to the control condition, translation of both sentence types was associated with activation increases in the ventrolateral prefrontal cortex (BA 47) and the globus pallidus. Given that contrastive syntactic constructions have been previously related to dissociable neural patterns (Birba et al. 2017; Constable et al. 2004; Feng et al. 2015; Mashal et al. 2009), it may be that the imaging protocol implemented was not sufficiently precise to capture condition-specific modulations. While the above topic remains unresolved and calls for further research, extant evidence has successfully revealed many neurocognitive peculiarities of specific word-level units. In particular, two separate studies converge to show that the establishment of cross-language equivalence involves markedly different neural systems depending on the lexical category (nouns vs. verbs) and dominant semantic associations (perceptual vs. action-related) of the unit at hand. In an fMRI protocol, Correia et al. (2014) aimed to identify which brain mechanisms subserve cross-linguistic mappings for words denoting animate entities. Once inside the scanner, native speakers of Dutch, with high proficiency in English, listened to L1 and L2 nouns referring to animals (e.g., paard, eend, horse, duck) and everyday objects (e.g., fiets, jurk, bike, dress). Their task was to press a key each time a non-animal word was heard.37 Of note, three different voices were used to present the stimuli, so that results could be interpreted in a speaker-independent fashion. Taken together, animal nouns elicited broad activation patterns across the bilateral superior temporal lobe, the right inferior frontal cortex, and the bilateral anterior insula. However, the most informative results stemmed from across-language generalization analyses. The aim was to detect which regions allowed classifying contrastive words in one language (e.g., paard vs. eend) based on the discriminative patterns of their translation equivalents in the other language (e.g., horse vs. duck). Significant activation clusters affording such a classification were circumscribed to the anterior temporal lobe, the angular gyrus, and superior temporal and postcentral sites, together with portions of the insula and the occipital cortex. Interestingly, the result remained the same irrespective of directionality. These regions, therefore, seem to code for lexico-semantic information shared between translation equivalents within the category of animals and, presumably, for other noun types, too. Interestingly, findings from numerous language- and image-processing experiments show that most of these areas (specially the anterior temporal, superior temporal, and angular cortices) are putatively implicated in

37. This is an elegant strategy, in that it requires sustained attention during the task but keeps target events (i.e., neural activity elicited by animal names) free from response-related motor artifacts.

160 The Neurocognition of Translation and Interpreting

A1 Neuropsychological evaluation tDCS: 20 min (anodal/cathodal/sham) AWL: action word learning (40 min)

NP day 0

AWL day 1

AWL day 2

A2

“apef” T: translations

AWL

970 ms

Do word and action match?

AWL T

day 3

day 4

Re: Reassessments (no stimulation) 200 ms

Re

Re

Re

day 7

day 14

day 28

1400 ms

2000 ms (ISI)

Response window (≤1400 ms) A4

A3 8 60

Correct couplings

tr i

correct

sp

al er ar

le ng

ni

“apef”

ss

se

“apef”

n

io

incorrect

“apef”

“apef”

“apef”

“apef”

“apef”

Incorrect couplings

correct “apef”

“apef”

“apef”

“apef” B Percentage of correct translations (%)

100 *

90 80 70 60 50 40 30 20 10 0

Anodal

Sham

Cathodal

Figure 6.2  Critical role of the primary motor cortex in the translation of action verbs. A. Learning paradigm. A1. The training of novel action-word learning was spread over four single learning sessions (40 minutes each), separated by 24 hours (days 1–4). Prior to each learning session, subjects received tDCS. At days 7, 14, and 28, session 1 was repeated for reassessments without stimulation. 2. Correct and incorrect couplings of spoken pseudowords and photographic illustrations of body-related actions were presented. In the learning sessions, subjects were instructed to decide intuitively whether the actions matched the accompanying pseudowords. Only responses during photo presentation (1.4 s) were accepted. No feedback regarding the correctness of responses was provided. The intertrial interval was fixed at 2 seconds. A3. Each learning session



Chapter 6.  Process is as unit requires 161

processing multimodal semantic information – i.e., conceptual composites which bring together input from various sensory modalities (Binder and Desai 2011; Patterson, Nestor, and Rogers 2007; Ralph et al. 2017; Seghier 2013). It would thus appear that semantic overlaps for translation equivalents denoting concrete entities are mapped on the same overall regions subserving their access in single-language (and even non-linguistic) tasks. However, radically different circuits are implicated in the translation of other lexical classes, such as action verbs. In a tDCS experiment, Liuzzi et al. (2010) recruited right-handed monolingual German speakers, split them into three groups, and had them complete a three-step neurolinguistic protocol (Figure 6.2, panel A1). First, each group underwent a specific type of brain stimulation, namely: anodal stimulation (which typically increases activity in the target area), cathodal stimulation (which typically inhibits activity in the target area), and sham stimulation (a control condition in which no significant modulation is induced). In all cases, the target brain region was the primary motor cortex, which is known to play critical roles in processing of action verbs (García and Ibáñez 2016c; Pulvermüller 2005, 2013). Second, all participants underwent an intensive, four-day word-acquisition paradigm, whereby they learned the meanings of novel (pseudo)words (e.g., apef) by associating them with photos showing specific bodily movements (Figure 6.2, panels A2-A4). Finally, participants performed a translation task in which they had to provide the German equivalent of the newly learned word. The primary outcome measure was the mean of action verbs correctly translated by each group.

consisted of 608 trials and was subdivided into two blocks. Over the course of the learning session, correct couplings occurred more frequently than incorrect couplings. A4. Each pseudoword was coupled four times with the correct action and twice with two different incorrect actions (ratio 4:2). Photos of actions were taken from different perspectives and with different actors. Each single photo was coupled once with a correct pseudoword and once with an incorrect pseudoword. The respective action (e.g., to hole) remained constant across four correct couplings. Thus, to learn the correct meanings of the pseudowords, subjects had to extract the correct action-related information from four different photos. From one learning session to the next, the correct couplings were kept identical, but one incorrect coupling (e.g., to eat and apef ) was exchanged. The ratio of 4:2 between correct and incorrect couplings was preserved in each learning session. Subjects were not informed about the underlying statistical principle of the learning paradigm and they received no feedback regarding their performance until completion of the protocol. At the end of day 4, subjects were required to translate the 76 novel action words into German. B. The tDCS protocol had a significant effect on the translation test. The posthoc Scheffe’s test showed that the percentage of action-related words correctly translated into the native language was significantly reduced after cathodal tDCS to the left motor cortex (star above line: p  Listening Interpreting > Shadowing

Left hemisphere

p(FWE-corrected) < .01

Right hemisphere

Figure 7.1  Differential patterns of regional activation during simultaneous interpreting. Significant differences in activation levels in thirty-four multilinguals, rendered on a canonical single-subject brain. Contrasts shown are speech shadowing in L2 versus listening to L2 (blue) and simultaneous interpretation into L1 versus shadowing (red). The latter comparison revealed that simultaneous interpreting involves a differential engagement of the left premotor and ventrolateral prefrontal cortices, together with the presupplementary motor area and the caudate nucleus. Results were obtained at a family-wise-error-corrected significance level of p < .01. Reprinted with permission from Hervais-Adelman, Moser-Mercer, and Golestani (2011).

And how soon do these changes become significant after the onset of field-specific training? As seen next, answers to these and other questions can be gleaned from studies on the particularities of interpreting students. 7.3

En route to expertise

SIs are made rather than born. The skills of the trade are neither innate nor incidentally acquired; rather, they are the result of intense periods of guided practice that impose exceeding demands on several neurocognitive systems. In simultaneous interpreting programs, students face these conditions for several hours each week, typically starting with no experience at all. Therefore, studies on this population allow establishing whether sustained engagement in a particular IR modality can directly trigger neural and behavioral changes, and how soon these become significant after the onset of training. The findings reported so far are as compelling as they are surprising.

182 The Neurocognition of Translation and Interpreting

Hervais-Adelman and colleagues conducted two neuroimaging studies in which simultaneous interpreting students were scanned at the beginning and at the end of their training program. Over the course of fifteen months, these subjects received ten hours of formal interpreting instruction per week, together with additional practice sessions. In both cases, a group of non-interpreter multilinguals (NIMs) was also tested at the same time points as the trainees. One of the reports (Hervais-Adelman et al. 2017) showed that, by the second assessment, only the interpreting students had developed increased cortical thickness in the left planum temporale, superior temporal gyrus, and anterior supramarginal gyrus, as well as in the right parietal, angular, and dorsal premotor cortices (Figure 7.2). As noted by the authors, and as specified in Chapter 3, these regions have been implicated in phonetic, lexico-semantic, and executive domains, many of which are extremely taxed during simultaneous interpreting.

Right

p < p 10 – < 2 p 10 – < 3 p 10 – FD 4 R < .0 5

Left

–5

0 –5

STG*†

5 0 –5

SPL†

5 0 –5

Angular Gyrus

5

IPS

0 –5

Symm. % Change/yr

0

5

Symm. % Change/yr

5

Symm. % Change/yr

–5

Post. Lateral Fissure

Symm. % Change/yr

0

SMG

Symm. % Change/yr

5

Symm. % Change/yr

Symm. % Change/yr

Controls Interpreters

Sup. Frontal Gyrus [BA6]

5 0 –5

Figure 7.2  Training-induced cortical thickness changes in simultaneous interpreting students. Regions showing a significant main effect of group in mean percent annualized cortical thickness change, projected on a canonical inflated white-matter surface. Dark patches represent sulci; light patches represent gyri. For clarity of display, the clusters having reached significance (at p < .0001) are displayed at a threshold of p < .01. Color coding indicates significance level. The asterisk (*) denotes clusters in which the peak reaches whole-brain FDR-corrected significance at p < .05. The cross (†) denotes clusters that reach whole-brain cluster-corrected significance at p < .05, with a cluster-forming threshold of p < .0005. Bar plots show symmetrized percent change for both groups at the peak co-ordinates of the cluster. Error bars represent 95% confidence intervals. IPS: intraparietal sulcus; SMG: supramarginal gyrus; SPL: superior parietal lobule; STG: superior temporal gyrus. Reprinted with permission from Hervais-Adelman et al. (2017).



Chapter 7.  The interpreter’s brain 183

The other study (Hervais-Adelman, Moser-Mercer, and Golestani 2015) used fMRI to examine neural activity changes during simultaneous interpreting and singlelanguage shadowing. After the training period, shadowing outcomes exhibited no enhancements in either group. Instead, interpreting quality did improve, but only for the trainees. Even more crucially, this pattern was accompanied by significant brain adaptations in those subjects. As shown in Figure 7.3, performance of simultaneous interpreting after the program involved reduced engagement of the right caudate nucleus. Of note, this frontostriatal hub has been implicated in several cognitive domains of relevance to this IR modality. These include cognitive control (see Chapter 3, Section 3.4.2) and word translation (see Chapters 4, 5, and 6), two functions that seem boosted by systematic interpreting practice, as will be shown below. Remarkably, at least some neurocognitive changes triggered by simultaneous interpreting training seem specific to this modality. Van de Putte et al. (2018) assessed one group of multilinguals before and after a nine-month program in simultaneous interpreting, and another group at the beginning and end of a nine-month program in translation. During fMRI sessions at both time points, all participants completed tasks tapping inhibitory control, switching, and verbal fluency skills. Whereas behavioral and fMRI results failed to reveal robust differences between groups and between scans, simultaneous interpreting students were the only ones showing structural connectivity changes triggered by training, particularly in frontostriatal, temporo-parietal, and fronto-cerebellar networks implicated in cognitive control and semantic processing (see Chapter 3). These studies show that sustained practice of this modality can significantly change structural and functional aspects of brain regions supporting a diversity of linguistic and executive functions. However, such neural peculiarities do not reveal which specific cognitive domains exhibit outward enhancements after training. Fortunately, insights in this direction have been abundantly reported elsewhere in the literature. A first observation is that simultaneous interpreting training improves TT quality. This was reported in the above-cited longitudinal study by Hervais-Adelman, Moser-Mercer, and Golestani (2015) and supported by previous evidence. For example, Tzou et al. (2012) showed that performance on a 15-minute-long simultaneous interpreting task was better for freshman trainees than NIMs, and even higher in sophomore trainees. Admittedly, these results are rather unsurprising, as they basically confirm that the training programs are effective in fostering better renditions of the STs. Also, they prove somewhat unspecific for our present concerns, since they fail to reveal which particular sub-functions lie at the heart of those overall enhancements. Yet, this issue is illuminated by numerous studies employing reliable atomistic tasks, as defined in Chapter 2.

184 The Neurocognition of Translation and Interpreting

LIFG [BA47]

L Postcentral gyrus [BA3] 0.03

0.03

0

0.01

–0.01 T2 T1

T1

T1

T2

Int Ctrl L Precentral gyrus

x = –40

0.012 0.008 0.004

0.012

T2 T1

T2

Int Ctrl L Paracentral lobule

x = –30

R Caudate

T2 Ctrl

R Cingulate cortex

0.012

0.02 0.015 0.01

0.008

0.005

0.004

0

0 T1

T2 T1 Int

0.016

0

Interpreting

0.01

0.02

0

Shadowing

0.02

T1

T2 T1 Int

T1

T2 T1 Int

T2 Ctrl

T2 Ctrl

0.008 0.004 0 –0.004 T1

T2 T1 Int

T2

x = –12

x = 15

Z = 13

Ctrl

Figure 7.3  Training-induced changes in brain activation during simultaneous interpreting for interpreter trainees relative to untrained multilinguals. Three-way interaction of group, time, and condition on BOLD response, showing regions within the search volume thresholded at p < .01, with a cluster extent threshold of five voxels, projected on a canonical single-subject brain. Interaction plots represent the mean contrast estimate for each condition vs. the silent baseline (in arbitrary units) at the peak voxel of the indicated clusters at each time point. Error bars represent standard errors of the mean (corrected to be appropriate for repeated-measures comparisons). The asterisk indicates a region in which there is a significant effect of training (the three-way interaction is driven by the pre- vs. post-training difference in the BOLD response during simultaneous interpretation in the trained participants). BA: Brodmann area; Ctrl: untrained multilingual controls; Int: simultaneous interpreting trainees; L: left; LIFG: left inferior frontal gyrus; R: right; T1: pre-training results; T2: post-training results. Reprinted with permission from Hervais-Adelman, Moser-Mercer, and Golestani (2015).

In a pioneering study with conservative sample sizes, Bajo, Padilla, and Padilla (2000) compared the performance of freshman simultaneous interpreting students and NIMs in different linguistic tasks. These included assessments of sentence comprehension and semantic categorization (indicating whether the second item in a word pair denoted an exemplar of the category denoted by the first). Although both groups had similar outcomes in an initial test, only the former showed significant improvements in the two tasks during a second evaluation roughly one year later, suggesting a direct impact of training on conceptual mechanisms.



Chapter 7.  The interpreter’s brain 185

Nevertheless, other language domains seem unaffected by early interpreting experience. Indeed, null differences between trainees and proficient NIMs have been reported in reading (Chincotta and Underwood 1998). Tentatively, then, it would seem that shallow-level (or highly automatized) processes do not benefit from initial training in this IR modality – still, as will be seen below, some of these functions do actually seem more robust in full-blown professionals. As regards non-verbal domains, Dong and Xie (2014) used the Wisconsin Card Sorting Test (detailed in Chapter 2, Section 2.3.2.2) to evaluate mental-set shifting skills in simultaneous interpreting students and NIMs. Crucially, they found better performance in second- than first-year trainees, even when controlling for L2 proficiency. This pattern seems to reflect a direct relationship between cumulative practice of simultaneous interpreting and the ability to alternate among cognitive schemas. Note that similar findings have been reported in professional practitioners – for a discussion, see Section 7.4. Other studies have focused on memory skills. Köpke and Nespoulous (2006) tested beginner trainees and NIMs via two free recall tasks. In the first one, subjects were allowed to subvocally rehearse the testing material; instead, in the second one, they were deprived of this possibility, as they were asked to continuously repeat a syllable while they encoded new information. Interestingly, the trainees were advantaged only in the latter task – which more closely replicates key processing conditions of simultaneous interpreting, as it measures the capacity to retain incoming input during phonological production. One would thus be tempted to propose that enhanced mnemonic performance for interpreting students is observed only or mostly in the face of elevated processing demands proper to their specialization. Such a conjecture is actually supported by additional works tapping other memory skills. For example, when asked to memorize new patterns of information in the absence of intervening cognitive processes, this population does not seem to possess any systematic advantage. Longitudinal studies with such tasks have yielded mixed results, showing that simultaneous interpreting training can boost STM for letters but not for digits or spatial locations (Antonova Ünlü and Sağın Şimşek 2018; Babcock et al. 2017). Similarly, the performance of NIMs on several STM measures is similar to that of freshman (Köpke and Nespoulous 2006; Tzou et al. 2012) and advanced (Antonova Ünlü and Sağın Şimşek 2018; Chincotta and Underwood 1998) simultaneous interpreting students.46 In sum, when cognitive resources can be fully devoted to transient retention of information, prospective SIs do not manifest any clear advantages. 46. The only discrepant pattern in the literature can be found in Tzou et al. (2012), who reported differences between sophomore simultaneous interpreting students and NIMs in STM for digits.

186 The Neurocognition of Translation and Interpreting

By contrast, when memory mechanisms are taxed in combination with other cognitive processes, as is the case in simultaneous interpreting, trainees do exhibit consistently better outcomes. This has been shown through complex span tasks, in which subjects must encode increasingly longer lists of items while they engage in concurrent mental operations, such as producing sentences of varying length. In these tasks, NIMs are outperformed by simultaneous interpreting students at the end (Tzou et al. 2012) and even at early stages (Köpke and Nespoulous 2006) of their programs, irrespective of the language used for testing. In fact, longitudinal experiments with these paradigms show significant enhancements in students after five (Antonova Ünlü and Sağın Şimşek 2018) and four (Chmiel 2018) semesters of practice, there being no improvements in NIMs evaluated at the same time-points.47 Collectively, these findings indicate that advantages in pre-professional stages are mainly manifested in tasks which require allocating diverse cognitive resources to co-occurring processes – i.e., one of the hallmarks of simultaneous interpreting. The selective nature of training-induced enhancements in interpreting students is further corroborated by evidence on inhibitory control (Dong and Xie 2014; Köpke and Nespoulous 2006) and switching skills (Babcock et al. 2017; Van de Putte et al. 2018). Neither of these domains shows coherent behavioral advantages for prospective SIs after one or even two years of training. As will be argued later on, the absence of effects in such dimensions might reflect their marginal relevance for thriving in the trade. 7.3.1 So different, so fast As seen throughout the preceding chapters, all bilinguals are capable of IR, albeit in a rudimentary fashion (Malakoff 1992). However, those who specialize in this skill (or at least in the particular modality under discussion) seem characterized by specific psychobiological adaptations. Granted, claiming that people with different abilities have different brains probably amounts to little more than a truism at this juncture, but the evidence above goes well beyond that generality. Its richness lies in the very specific insights it affords concerning the scope, nature, and swiftness of the neurocognitive changes related to sustained interpreting practice. First, separate studies using different techniques show that taking classes in simultaneous interpreting is directly linked to anatomical and functional brain changes (Hervais-Adelman et al. 2017; Van de Putte et al. 2018). Although each of

47. Note, however, that no such improvements appear when intervening cognitive tasks involve non-verbal materials (Babcock et al. 2017), which might suggest a specific effect on WM for language-mediated operations, at least in pre-professional stages.



Chapter 7.  The interpreter’s brain 187

them measured particular neural dimensions (cortical thickness, hemodynamic increases, structural connectivity), they all related interpreting training with distinct patterns in perisylvian, extrasylvian, and subcortical regions previously associated with task-relevant verbal and non-verbal domains. Of course, those results cannot be interpreted as reflecting cognitive improvements in any specific subfunction. Yet, it is at least plausible to conceive them as partially driven by the demands that this IR modality places on circumscribed cognitive skills. Importantly, all these experiments have employed acceptable sample sizes – ranging from eighteen trainees in Van de Putte et al. (2018) to thirty-four in Hervais-Adelman et al. (2017)– and controlled for important confounds, such as L2 proficiency, age of acquisition, and more basic sociodemographic factors. By circumventing some of the main limitations of other studies in the field, this research constitutes firm evidence of training-induced neuroplasticity. The preceding sentence denotes causality, and this is not accidental. Whereas most evidence in neuroscience is correlational, the studies just mentioned have used longitudinal designs, indicating that the reported effects could be directly triggered by accruing experience in the trade. The same is true of some of the behavioral studies reviewed, which have revealed improved outcomes in linguistic (Bajo, Padilla, and Padilla 2000) and executive (Antonova Ünlü and Sağın Şimşek 2018; Chmiel 2018) domains after periods of intensive practice. The absence of comparable enhancements in the control samples rules out the possibility that the reported gains are due to task familiarization or other unspecific factors related to general bilingual skills. These results are of particular theoretical relevance, since they allow taking a stance on the so-called “self-selection issue” (Christoffels and de Groot 2005). Let us put it this way: mere single-point comparisons between a group of subjects who already possess specialized training and others who do not are insufficient to claim that differences between them were triggered by practice. It could well be the case that subjects who enter the field are those who acknowledged pre-existing aptitudes for the trade, so that their advantages over NIMs would be related to unspecified task-irrelevant factors. However, convergent findings argue against this possibility. In a recent study, Rosiers et al. (2019) compared NIMs with aspiring SIs about to start their training and found similar performance in a set of executive tasks. Likewise, the interpreting students tested in extant longitudinal experiments did not differ from NIMs in the initial assessment, but they did differ in the second one, even when ruling out the influence of other variables, such as L2 competence. Also, to anticipate some of the evidence that will be reviewed below, some neurocognitive differences and behavioral advantages in professional SIs correlate with their hours of practice (Elmer, Hänggi, and Jäncke 2014) and with their years of professional experience (Santilli et al. 2018). In sum, the distinguishing traits of

188 The Neurocognition of Translation and Interpreting

our target populations seem to be absent before the onset of training; rather, they emerge in the course of specialized programs, regardless of other factors. Moreover, at least in some cases, they increase in proportion to the time devoted to simultaneous interpreting. It follows that accruing experience in this IR modality constitutes a causal (albeit partial) determinant of the particular neurocognitive profile of SIs. What is more, these training-induced changes can occur rather fast. In most cases, significant neural and behavioral effects are detected before the first eighteen months of training, and often before the first year. Once again, pre/post designs are most compelling in this sense, but additional compatible evidence abounds in the studies offering single-point comparisons between groups. Therefore, it would not be completely accurate to ascribe these effects to ‘expertise’ in a technical sense. Though sometimes misused and confounded with other constructs (Muñoz Martín 2014), this term refers to consolidated mastery of a specific skill-set after several years of deliberate, goal-oriented practice (Ericsson et al. 2006). It is hardly the case that the trainees recruited in these studies had already attained masterful standards; indeed, it is usually in the course of actual professional life that these levels can be achieved. Rather, the changes and boosts reported above seem to reflect accruing levels of simultaneous interpreting competence, namely, the acquisition of skills needed to accomplish this task (Tiselius and Hild 2017), at levels of performance that pave the way for, but are not tantamount to, interpreting expertise. Succinctly, then, the neurocognitive impact of simultaneous interpreting training becomes significant way before the attainment of ultimate performance standards. Sustained exposure to this particular socio-cognitive experience results in swift adaptations, even in adult proficient bilinguals. Finally, such fast changes do not seem to manifest in every cognitive domain. In fact, only a handful of functions have been found to be systematically better in interpreting trainees than in NIMs. If, as proposed above, those functions are the ones that are distinctively taxed during simultaneous interpreting, training-induced effects would be characterized by limited (or null) generalizability. In other words, they would be considerably selective and specific, which would further illuminate their partial functional independence – see Chapter 4. Yet, before jumping to conclusions on this point, let us consider an even larger body of evidence from professional SIs.

Chapter 7.  The interpreter’s brain 189



7.4

Keep the change (and make it broader)

At some point, after several months of intense training, interpreting students will earn their diplomas. Most of them will then get certified and start partaking regularly in the profession. A few years later, with several conferences on their backs, they will have mastered several areas of the trade. With the expertise thus acquired, which of the prior neurocognitive effects will remain the same and which ones will change? Do any further effects appear de novo in the course of professional life? As shown next, there is no shortage of evidence to begin answering these questions. A number of experiments indicate that the interpreter’s brain may possess particular features, some of which resemble the ones detected in trainees. Elmer, Hänggi, and Jäncke (2014) performed volumetric comparisons between the brains of professional SIs and NIMs. The main finding was that the former subjects exhibited reduced volumes in diverse brain regions, such as the bilateral middle anterior cingulate gyrus, the middle anterior insula, the superior middle gyrus, and the pars triangularis (Figure 7.4). As noted by the authors, these areas have been implicated in numerous functions which lie at the heart of simultaneous interpreting, including error detection, sensory-to-motor coupling, and other verbal and non-verbal functions, such as WM and phonetic processing. Moreover, gray matter density in these and other relevant regions (e.g., the bilateral caudate nucleus) correlated negatively with the interpreters’ hours of practice, suggesting a link between cumulative interpreting training and cortical pruning for more efficient language control (Elmer, Hänggi, and Jäncke 2014). Yet, another study suggests that experience in simultaneous interpreting may also involve volumetric increases. Becker et al. (2016) found that professional SIs had more gray matter density in the left frontal pole, a region implicated in cognitive control. Moreover, the volume of this region in the interpreters was negatively associated with reduced mixing costs in switching paradigms, further suggesting that enhanced skills to meet extreme processing demands imply morphometric changes in critical substrates. The differences between these results and those reported by Elmer, Hänggi, and Jäncke (2014) may reflect discrepancies in subject-level variables (such as the participants’ years of training) and methodological factors (including image acquisition parameters). Certainly, further research is necessary to clarify which type of volumetric difference is more characteristic of this population. However, these data do converge in showing that interpreting expertise could entail anatomical specificities in regions subserving linguistic and executive functions. Interpreting experience has also been related to functional adaptations during relevant tasks. In a dichotic listening test, only professional SIs, as opposed to

190 The Neurocognition of Translation and Interpreting

Figure 7.4  Gray matter differences between professional simultaneous interpreters and multilingual control subjects. The figure shows reduced regional volume for interpreters relative to controls (blue) and significant correlations between gray matter volume and the cumulative number of practice hours within the interpreters group (red). ACC: anterior cingulate cortex; AIC: anterior insular cortex; IFG: inferior frontal gyrus; MCC: middle cingulate cortex; MIC: middle insular cortex; NC: nucleus caudatus; Operc: pars opercularis; Tri: pars triangularis. Reprinted with permission from Elmer, Hänggi, and Jäncke (2014).

simultaneous interpreting students, exhibited hemispheric asymmetries when judging syntactic and semantic errors in translated sentence pairs (Fabbro, Gran, and Gran 1991). According to the authors, these results could suggest a rearrangement of hemispheric specializations driven by modality-specific processing strategies. However, this interpretation must be taken with reservation due to the limitations of dichotic listening to assess language lateralization (Paradis 1992, 1995, 2003, 2008). More direct evidence comes from semantic decision tasks. Elmer and colleagues recruited NIMs and professional SIs (specialized in the L2-L1 direction), presented them with noun pairs, and had them decide whether each dyad was congruent or incongruent. The stimuli were presented in four language combinations (L1-L1, L2-L2, L1-L2, L2-L1), which allowed assessing both intra- and inter-linguistic effects. A first report (Elmer, Meyer, and Jäncke 2010) showed that the interpreters exhibited enhanced N400 responses in all conditions but L2-L1 – i.e., the one corresponding to their professionally trained translation direction (Figure 7.5). This could be indicative of “a training-induced altered sensitivity to semantic processing within and across L1 and L2” (Elmer et al. 2010: 152).

Chapter 7.  The interpreter’s brain 191



N400: Incongruent nouns, L1–L1 4

Voltage

2

ANOVA 8 TW × 2 G

Time window p < .001 Time window × Group p < .05

0

1

2

8 TWs 300–700 ms 3 4 5 6 7

8

2

8 TWs 300–700 ms 3 4 5 6 7

8

8 TWs 300–700 ms 3 4 5 6 7

8

–0.5

0

µV

–2

SI C

–1 –1.5

*

–2

–4 –100 0

ROI: CPz, Cz, Pz

–2.5

100 200 300 400 500 600 700 Time N400: Incongruent nouns, L2–L2

4

Voltage

2

ANOVA 8 TW × 2 G

Time window p < .001 Time Group p < .05

0.5

1

0

SI C

–0.5 0

µV

–1 –1.5

–2

*

–2

–4 –100 0

ROI: CPz, Cz, Pz

–2.5

100 200 300 400 500 600 700 Time N400: Incongruent nouns, L1–L2

4

Voltage

2

ANOVA 8 TW × 2 G

Time window p < .001 Time window × Group p < .05

1

2

0

0

µV

–2 –4 –100 0

0.5 –0.5 –1 –1.5

ROI: CPz, Cz, Pz 100 200 300 400 500 600 700 Time –4.0 µV

SI C

*

–2

0 µV

4.0 µV

Figure 7.5  N400 differences during semantic decision for simultaneous interpreters and multilingual controls. The figure shows responses to incongruent nouns for both groups (blue: simultaneous interpreters; red: multilingual controls). The data refer to voltage values within a region of interest. The x-axis shows time course in milliseconds (ms), the y-axis depicts mean voltage values in μV. The figure also reports mean amplitudes in eight consecutive time windows for the L1-L1, L2-L2, and L1-L2 combinations (the L1-L2 condition yielded no differences between groups). 1 = 300–350 ms, 2 = 350–400 ms, 3 = 400–450 ms, 4 = 450–500 ms, 5 = 500–550 ms, 6 = 550–600 ms, 7 = 600– 650 ms, 8 = 650–700 ms. The asterisk (*) indicates significant differences at p < .05. Scalp maps show modulations between 300 and 700 ms. The upper map refers corresponds to the interpreters and the lower one to controls. C: multilingual controls; G: group; L1: native-language (German); L2: foreign language (English); SI: simultaneous interpreters; ROI: region of interest; TW: time window. Reprinted with permission from Elmer, Meyer, and Jäncke (2010).

192 The Neurocognition of Translation and Interpreting

A second report (Elmer and Kühnis 2016) focused on functional connectivity between two left-hemisphere regions implicated in sensory-to-articulation mappings: the auditory-related cortex (BAs 41 and 42) and Broca’s area (BAs 44 and 45). The key finding was that, during semantic decision, theta-band coupling between these areas was greater for SIs than for NIMs. Notably, too, this connectivity pattern in the interpreters correlated positively with their amount of training, suggesting that distributed linguistic networks may adapt in proportion to the demands placed on them. Also, relative to NIMs, professional SIs seem typified by distinct interactive dynamics between frontotemporal hubs during executive tasks. In an fMRI study assessing switching processes and dual-task performance, Becker et al. (2016) found that the interpreters’ left frontal pole featured higher global network efficiency and was more functionally connected to the left inferior and middle temporal gyri. These results, the authors propose, might reflect neural adaptations underlying enhanced cognitive control. In sum, professional SIs also seem to present distinguishing neuroanatomical and neurofunctional traits. As observed in trainees, these involve regions and networks related to verbal and non-verbal processes which are fundamental for successful task fulfilment. Interestingly, some such patterns are proportional to the amount of practice, further emphasizing their experience-related nature. However, as previously stated for trainees, it does not follow that all functions related to those regions must manifest behavioral enhancements. Indeed, significant gains are observed in only some of them. First, as compared to professional musicians and non-expert controls, SIs demonstrate enhanced auditory perception for both verbal and non-verbal sounds (Elmer et al. 2014), possibly reflecting increased capacities for extracting and recognizing relevant input from complex auditory signals. Also, relative to NIMs, expert SIs seem advantaged when it comes to comprehending sentences (Bajo, Padilla, and Padilla 2000) as well as understanding (Yudes et al. 2013) and recalling (Dillinger 1990) longer pieces of discourse. Moreover, they are better at recognizing semantic inaccuracies, but not lexical or syntactic errors (Fabbro, Gran, and Gran 1991; Yudes et al. 2013). Considering that simultaneous interpreting requires grasping salient semantic patterns in unfolding texts, this pattern may constitute a cognitive gain developed through accumulated experience in conference settings. Also, professional SIs outperform NIMs in word knowledge (Christoffels, de Groot, and Kroll 2006), verbal fluency (Santilli et al. 2018), and word translation (Santilli et al. 2018) tests, suggesting wider vocabulary and enhanced intra- and inter-linguistic lexical search skills (Figure 7.6). While such advantages may be driven by the continual need to retrieve highly specific words under time pressure,

Chapter 7.  The interpreter’s brain 193



Number of words

20 15 10 5

NIBs

Number of words

25 20 15 10 NIBs

10

L2 semantic fluency

15 10 NIBs

C2. 1200

800

800

200

PSIs

20

1000

400

NIBs

25

1000

600

PSIs

NIBs

30

5

RT (ms)

RT (ms)

C1. Overall word translation * 1200

15

B2. L1 semantic fluency 35

PSIs

L2 phonological fluency

20

5

PSIs

B1. Overall semantic fluency 35 * 30

5

A2. L1 phonological fluency 25

Number of words

Number of words

A1. Overall phonological fluency # 25

PSIs

NIBs

BT

PSIs FT

600 400

NIBs

PSIs

200

NIBs

PSIs

NIBs

PSIs

Figure 7.6  Verbal fluency and word translation speed in simultaneous interpreters and bilingual controls. Considering performance in both languages, interpreters produced more words than bilingual controls in phonological fluency (panels A1-A2) and semantic fluency (panels B1-B2) tasks. Also, results from translation tasks in both directions revealed faster processing for interpreters than bilingual controls (panel C1-C2). The asterisk (*) indicates significant differences at p < .05 for mixed ANOVAs only. The hash (#) identifies significant differences at p < .05 for both the mixed ANOVAs and the ANCOVAs (with interpreting expertise as a covariable). BT: backward translation; FT: forward translation; L1: native language (Spanish); L2: foreign language (English); NIBs: non-interpreter bilinguals; PSIs: professional simultaneous interpreters; RT: response time. Reprinted with permission from Santilli et al. (2018).

194 The Neurocognition of Translation and Interpreting

they may also be triggered by other recurrent demands placed on bilingual systems. Indeed, the same effects have been reported in L2 teachers (Christoffels, de Groot, and Kroll 2006; Stavrakaki et al. 2012). Still, it would be wrong to ascribe these patterns to unspecific L2-related variables, as they emerge even when SIs and controls are matched for L2 competence, age of acquisition, years of study, and weekly exposure (Santilli et al. 2018). Therefore, in the case of these professionals, and arguably in the case of L2 teachers as well, signs of enhanced lexico-semantic processing can be reasonably attributed to field-specific expertise. In addition, professional SIs are more efficient at processing unfamiliar materials. In fact, they outperform NIMs or even L2 teachers when asked to recognize (Bajo, Padilla, and Padilla 2000), repeat (Signorelli, Haarmann, and Obler 2012) or recall (Stavrakaki et al. 2012) non-words – i.e., items that do not form part of the lexicon. In a similar vein, they show advantages in categorizing non-typical exemplars of particular semantic categories (Bajo, Padilla, and Padilla 2000). Tentatively, this may reflect better implicit strategies to deal with unexpected, non-rotely-learned verbal information, which is typically encountered in actual interpreting scenarios. On the other hand, several linguistic processes seem unaffected by experience in the trade. No significant differences have been observed between professional interpreters and NIMs or L2 teachers in picture naming (Christoffels, de Groot, and Kroll 2006; Santilli et al. 2018) – Figure 7.7, panel A. Also, experiments tapping semantic congruency judgments have yielded mixed results (Elmer and Kühnis 2016; Elmer, Meyer, and Jäncke 2010). Finally, null effects of interpreting experience have been reported in over-practiced or shallow-processing tasks, including lexical decision on real words (Bajo, Padilla, and Padilla 2000), number counting (Signorelli, Haarmann, and Obler 2012), and word reading in both languages (Hiltunen et al. 2016; Santilli et al. 2018) – Figure 7.7, panel B. Note that these lexical tasks rely on highly automatized mechanisms which hardly face greater demands during simultaneous interpreting than in other daily linguistic activities. Therefore, they may fail to benefit from the systematic entrenchment that other functions seem to manifest in this population. It appears that the impact of simultaneous interpreting experience on linguistic processing is limited to those subdomains that are distinctly taxed during situated performance. In line with the characterization offered in Section 7.2, expert interpreters exhibit advantages in auditory perception, conceptual processing (including text comprehension and semantic error detection), vocabulary search, inter-linguistic operations, and handling of unfamiliar materials. However, other processes of marginal relevance to interpreting seem impervious to the accumulation of experience in the profession. This broad pattern represents a first confirmation of the conjecture announced at the end of Section 7.3: it seems that

Chapter 7.  The interpreter’s brain 195



A1. Overall picture naming 1400

A2. L1 picture naming 1400 1200 RT (ms)

RT (ms)

1200 1000 800 600 400

1000 800 600

NIBs

400

PSIs

B1. Overall word reading 800 700

700

600

600

500 400

PSIs

PSIs

L2 word reading

400 300

200

200

PSIs

NIBs

500

300 NIBs

NIBs

B2. L1 word reading 800

RT (ms)

RT (ms)

L2 picture naming

NIBs

PSIs

NIBs

PSIs

Figure 7.7  Picture-naming and word-reading speed in simultaneous interpreters and bilingual controls. Considering performance in both languages, no differences emerged between groups in picture-naming (panels A1-A2) or word-reading (panels B1-B2) tasks. L1: native language (Spanish); L2: foreign language (English); NIBs: non-interpreter bilinguals; PSIs: professional simultaneous interpreters; RT: response time. Reprinted with permission from Santilli et al. (2018).

linguistic boosts related to interpreting practice become significant only for restricted subdomains. Suggestively, research on executive processes also reveals a clear divide between enhanced and unaffected functions. We have seen that recall capacity in trainees proves better only if subvocal rehearsal is impeded. An identical pattern characterizes professional SIs, whose performance is similar to that of other groups (NIMs, L2 teachers, consecutive interpreters) when rehearsal is unhindered (Bajo, Padilla, and Padilla 2000; Hiltunen et al. 2016; Hiltunen and Vik 2017; Köpke and Nespoulous 2006; Signorelli, Haarmann, and Obler 2012), but significantly better when the task involves articulatory suppression (Bajo, Padilla, and Padilla 2000; Köpke and Nespoulous 2006; Yudes, Macizo, and Bajo 2011b). Once again, this evidence points to selective boosts confined to processes that replicate the defining constraints of simultaneous interpreting. The same is true of WM outcomes, which, as was the case in trainees, show better performance in the presence of concurrent cognitive operations (Figure 7.8).

196 The Neurocognition of Translation and Interpreting

# recalled series

6

A. Word span

38

Dutch English

5 4 3

34 32 30

2 1 Students

B. Reading span Dutch English

36 # recalled words

7

Teachers

28

Interpreters

34

Students

Teachers Interpreters

C. Speaking span Dutch English

# recalled words

32 30 28 26 24

Students

Teachers

Interpreters

Figure 7.8  Working memory outcomes for simultaneous interpreters, foreign-language teachers, and bilingual students. The figure shows that interpreters outperformed bilingual students and foreign-language teachers in word-span (panel A), reading-span (panel B), and speaking-span (panel C) tasks. In the latter two tasks, participants were required to encode information while engaging in concurrent verbal processes. These differences emerged in both of the participants’ languages (L1: Dutch; L2: English). Reprinted with permission from Christoffels, de Groot, and Kroll (2006).



Chapter 7.  The interpreter’s brain 197

Indeed, professional SIs have larger memory spans than NIMs when encoding occurs as subjects are required to pronounce written (Signorelli, Haarmann, and Obler 2012; Yudes, Macizo, and Bajo 2011a; Yudes et al. 2013) or spontaneous (Christoffels, de Groot, and Kroll 2006) sentences, judge the sensibility of verbal materials (Chmiel 2018) or even perform mathematical calculations (Babcock and Vallesi 2017).48 It follows that these experts are also characterized by superior capacity for storing transient information while performing additional cognitive tasks, as crucially required to thrive in their profession. What is more, unlike trainees, professional SIs possess better retention skills than NIMs even in the absence of concurrent cognitive operations. This superiority has been reported for letters (Babcock and Vallesi 2017), shapes (Babcock and Vallesi 2017), non-words (Stavrakaki et al. 2012), words in L1 and L2 (Christoffels, de Groot, and Kroll 2006), and digits (Bajo, Padilla, and Padilla 2000; Stavrakaki et al. 2012) – although results for the latter type of materials are inconsistent (Köpke and Nespoulous 2006; Santilli et al. 2018). In fact, memory for letters is better in SIs than in other IR experts, such as translators (Henrard and Van Daele 2017). Therefore, it seems that expertise in this interpreting modality also entails advantages in STM, even when no coherent effects are observed in students developing their interpreting competence. Professional SIs are also better than NIMs and other IR experts in specific aspects of cognitive control. In particular, they are faster at dual tasks, which require allocating specific resources to two parallel but unrelated processes, such as identifying specific tones and shapes in concomitant sequences of visual and auditory targets (Becker et al. 2016; Morales et al. 2015; Strobach et al. 2015) – Figure 7.9. This speaks to superior mental coordination skills, arguably forged through the need to optimally manage joint input and output demands during simultaneous interpreting. Also, professional SIs seem faster and more accurate than NIMs at mental-set shifting paradigms, requiring successive updates of the active cognitive scheme (Yudes, Macizo, and Bajo 2011a) – although not all outcome measures capture this effect (Santilli et al. 2018). By the same token, when SIs are asked to swap between two different tasks (e.g., color and shape discrimination), their switching costs are similar to those of NIMs, translators, and consecutive interpreters, but their mixing

48. Note, however, that specific WM tasks have yielded similar results for professional SIs and L2 teachers (Stavrakaki et al. 2012), indicating that this domain may also benefit from other specialized bilingual activities.

198 The Neurocognition of Translation and Interpreting

A. Experimental paradigm ISI = 2500 ms 500 ms

Visual distractor

Visual target

Visual distractor

+

+

+

+

MA

PA

TA

PA

Auditory distractor Auditory distractor

Simultaneous presentation

Auditory target

B. Results Block 1

4.0 3.0 2.0 1.0 0.0

One

Two Load

Block 2

5.0 Accuracy (d’)

Accuracy (d’)

5.0

Three Single interpreter Single control

4.0 3.0 2.0 1.0 0.0

One

Two Load

Three

Dual interpreter Dual control

Figure 7.9  Performance on single and dual n-back tasks for simultaneous interpreters and bilingual controls. The figure shows a schematic representation of a stream of trials in the dual n-back task (panel A) and the mean accuracy (d′) on single and dual n-back tasks by block and group (panel B), indicating superior performance for interpreters. ISI: inter-stimulus-interval. Reprinted with permission from Morales et al. (2015).

Chapter 7.  The interpreter’s brain 199



210

Controls Sls

190

Reaction time (in ms)

170 150 130 110 90 70 50

Switching costs (RT)

Mixing costs (RT)

Figure 7.10  Switching and mixing costs in simultaneous interpreters and bilingual controls. The figure shows costs for both conditions considering performance in a switching paradigm (depicted in Chapter 2, Figure 2.3). Error bars represent standard errors. Covariates appearing in the model are evaluated at the following values: age = 41.83, gender = 1.75. RT: reaction times; SIs: simultaneous interpreters. Reprinted with permission from Becker et al. (2016).

costs are significantly reduced (Babcock and Vallesi 2017; Becker et al. 2016)49 – Figure 7.10. Of note, the latter pattern correlates negatively with gray matter density in the left frontal pole, an important hub of the cognitive control network (Becker et al. 2016). Such a dissociation also seems closely related to the demands of simultaneous interpreting, which requires practitioners to prevent mixing the L1 and L2 while both languages are jointly activated (rather than alternately inhibited), there being no need to switch between discrepant (e.g., visual and auditory) processing modalities. Finally, as previously shown for trainees, professional SIs show no advantages in basic attentional functions (including alerting and orienting capacity) (Babcock and Vallesi 2017; Morales et al. 2015) or in inhibitory skills for minimizing the impact of linguistic and spatial information on perceptual processes (Babcock and 49. In classical dual tasks, mixing costs are calculated as the difference between repetition trials in mixed-task (color and shape) blocks and all trials in the single-task (color-only, shape-only) blocks. Instead, switching costs are indexed by the difference between non-switch and switch trials within the mixed-task condition (Becker et al. 2016).

200 The Neurocognition of Translation and Interpreting

Vallesi 2017; Köpke and Nespoulous 2006; Yudes, Macizo, and Bajo 2011a) – in fact, the only study showing faster responses for SIs in a conflict resolution experiment found that this pattern was accompanied by less accurate performance (Aparicio, Heidlmayr, and Isel 2017).50 Such results also align with the hypothesis that interpreting-related boosts are essentially demand-driven. The focalized types of attention tapped by the paradigms reported are mostly irrelevant for expert interpreting performance, which turns out to be a highly automatic skill (Cowan 2000). Likewise, the ability to suppress interference from different processing modalities (linguistic, spatial, perceptual) is arguably of little importance to comply with the requisites of simultaneous interpreting. Briefly put, experience in this profession seems to confer no attentional or inhibitory enhancements either before or after the attainment of expert performance standards. Taken together, then, the evidence corroborates an incipient pattern observed in interpreting trainees: sustained practice of this IR modality is associated with boosts in only some cognitive functions. Both in the linguistic and in the executive domain, processing advantages seem robust only for those functions which are directly and specifically taxed during simultaneous interpreting. The theoretical implications of this observation are laid out below, together with several other postulates on the adaptive properties of the neurocognitive systems at play. 7.4.1 Brains interpreting interpreting brains What many have long intuited is now scientifically proven: there is something special about the brains of SIs. The changes that began to emerge in the course of their training remain present once they become professionals, and some adaptations seem to emerge only in the latter stage of their careers. With the vast collection of findings we have summarized, the ground is set to specify the ways in which these subjects’ minds differ from those of other multilinguals. First, SIs possess specific neuroanatomical and neurofunctional traits. Although the evidence is not abundant, it may seem surprising that the specific neural structures exhibiting differential patterns do not precisely coincide with those observed in trainees. Nevertheless, these discrepancies may be driven by accumulated experience and ensuing expertise. In fact, as noted before, some of the neuroscientific particularities of SIs are significantly associated with their amount of practice. Yet, such inconsistencies may also reflect the very impact of age (given that the professional SIs considered are, on average, much older than the trainees examined in the first set of studies) or even methodological inconsistencies among the metrics and analytical parameters used in each report. 50. However, see Henrard and Van Daele (2017) for different results.



Chapter 7.  The interpreter’s brain 201

Be that as it may, what seems to remain constant in both populations is that their distinctive neural patterns correspond to perisylvian, extrasylvian, and subcortical regions and networks mediating myriad verbal and non-verbal operations of direct relevance to simultaneous interpreting. Indeed, key functions associated with the particular areas showing these effects (as detailed in Chapter 3) coincide with those enumerated as crucial for successful performance at the outset of this chapter. The major biological changes associated with simultaneous interpreting practice do not seem to emerge randomly in the brain; rather, they appear to be confined to the systems critically involved in this modality. That general finding calls for a qualification regarding the neural systems which subserve IR. The neuroanatomical models described in Chapter 4, as well as the functional specifications introduced in Chapters 5 and 6, emerged from synchronic comparisons between tasks in a given individual or in a group of subjects with similar linguistic profiles. Now, after interpolating the variables of training and expertise, and by detecting diachronic changes in the systems involved, we can refine our account of those systems and acknowledge them as inherently plastic. The anatomical, hemodynamic, and electrophysiological correlates of IR functions are not only recruited during task performance, but also molded by the very processes they enable. Just as a strong biceps allows for repeated weight-lifting, leading to stronger biceps, so too the neurocognitive systems in question allow for continual practice of this IR modality, leading to significant changes in them. Of course, neural differences alone are insufficient to ascertain which of the many functions associated with each region or mechanism actually manifest performance gains. This is where behavioral evidence makes its crucial contribution, as it reveals the extent and nature of the effects related to interpreting experience. We have already seen that at least many of the reported advantages in trainees were causally driven by task-specific practice and visible shortly after its initiation. We can now expand on these conclusions and state that the verbal and non-verbal enhancements observed in these populations are non-generalizable beyond directly taxed functions, mutually independent, subject to distinctive adaptive time courses, and uninfluenced by other bilingual-experience-related factors. The behavioral outcomes of professional SIs corroborate a claim that had been preliminarily identified in the trainees: significant improvements are confined to selected domains that are specifically taxed during simultaneous interpreting. For example, some of the most robust linguistic advantages have been revealed through assessments of discourse comprehension, word translation, and processing of unfamiliar verbal materials. All of these skills must be optimized to succeed in the trade. Indeed, successful interpreting is contingent on the capacity to grasp key concepts across unfolding discourse, while quickly activating adequate TL equivalents of SL words and efficiently identifying and handling unexpected

202 The Neurocognition of Translation and Interpreting

pieces of information (Christoffels and de Groot 2005; Padilla, Bajo, and Macizo 2005). Likewise, as regards executive functions, the most systematic gains for SIs emerge in tasks tapping recall under articulatory suppression, WM, and dual-task processing. Elevated efficiency in those domains proves fundamental to properly engage in simultaneous interpreting, which requires storing, manipulating, and retrieving information in memory as phonological systems are unavailable for internal rehearsal and limited resources are being allocated to co-occurring (input and output) operations (Chernov 2004; Christoffels, de Groot, and Kroll 2006; Gerver 1975; Pöchhacker 2004). Now, what proves even more revealing is the set of cognitive functions that seem impervious to the impact of interpreting practice. For instance, SIs present no advantages when it comes to naming pictures or reading words out loud in either language, or when task performance hinges on the efficiency of inhibitory or switching mechanisms. Unlike the functions discussed in the previous paragraph, these ones seem marked by a common denominator: arguably, none of them is more taxed during simultaneous interpreting than during other forms of bilingual processing. Indeed, the ability to quickly name static objects is mostly irrelevant for simultaneous interpreting, and so is the capacity to quickly utter written words (Santilli et al. 2018). Similarly, SI does not hinge so much on the inhibition of competing processes as it does on their joint management (Dong and Xie 2014; Ibáñez, Macizo, and Bajo 2010; Yudes, Macizo, and Bajo 2011a), and neither does it require sudden shifts between processing modalities. The same goes for the tasks tapping aspects of focal attention, all of which have yielded null effects of interpreting expertise: unlike divided attention skills, which are certainly crucial for the profession (Pöchhacker 2004), the dimensions tested so far are arguably immaterial for interpreting performance – which, if anything, turns out to be highly automatic in trained practitioners (Cowan 2000; Gile 1995; Gran 1989). Given the tasks whence they stem, significant and null results motivate a nontrivial conclusion: effects related to interpreting expertise seem characterized by demand-based domain specificity. The boosts consolidated in a particular verbal or non-verbal function do not percolate to other functions of the same general (linguistic or executive) system, no matter how similar or overlapping these might be. Far from an oddity of SIs, this appears to be a constant of expertise-related effects. For example, highly skilled mnemonists are characterized by superior WM and verbal memory, with no accompanying enhancements of visual memory (Maguire et al. 2002). Moreover, the demands of bilingual processing across the lifespan have been associated with gains in only some executive functions (Bialystok 2009; Bialystok, Craik, and Luk 2012), and even in only some dimensions within specific domains, such as WM (Calvo, Ibáñez, and García 2016). In line with these antecedents, then,



Chapter 7.  The interpreter’s brain 203

the impact of simultaneous interpreting practice on cognitive performance seems typified by limited (and possibly null) cross-domain generalizability. Another observation compatible with the above claim is that linguistic and executive gains in SIs are independent from each other. Needless to say, verbal and non-verbal functions are in constant interplay during simultaneous interpreting (Christoffels, de Groot, and Kroll 2006; Christoffels, de Groot, and Waldorp 2003). However, effects within each of these broad domains are not explained by those emerging in the other. For example, advantages in lexical tasks are uncorrelated with enhancements in STM (Christoffels, de Groot, and Kroll 2006), and they have actually been reported in samples with low and high STM spans (Yudes et al. 2013). What is more, SIs exhibit better translation performance than NIMs even when these have comparable executive skills (Santilli et al. 2018). In fact, using graphical modeling, Christoffels, de Groot, and Waldorp (2003) found that translation efficiency and STM capacity represent statistically independent determinants of interpreting outcomes. The bottom line is that, in all probability, verbal and non-verbal adaptations in SIs are not co-dependent. Also, it would seem that not all cognitive systems follow the same developmental trajectories since the onset of systematic interpreting practice. We have seen that several effects observed in professionals can also be tracked in trainees. However, this is not always the case. Enhanced vocabulary search skills, as tapped through verbal fluency tests, appear robust in experts (Santilli et al. 2018) but not in students (Van de Putte et al. 2018). The same is true of STM effects, which have been repeatedly replicated in professionals but prove inconsistent at best in trainees. This raises the conjecture that some processing advantages associated with sustained experience in simultaneous interpreting emerge only after the accumulation of several years of practice, exceeding those necessary to complete training programs. Alternatively, it may be that systematic reliance on STM is only discerned as a useful practical strategy in the course of professional life, so that stringent demands on relevant mechanisms only begin to be exerted in post-graduation stages of the subjects’ careers. Although the reasons remain unclear, the detection of both early- and late-onset boosts associated with this IR modality further attests to how complex and varied such changes can be. Finally, note that the distinctive cognitive traits identified in trainees and professional SIs seem specifically related to the recurring practice of this IR modality. Most of the studies have effectively matched their samples in terms of general bilingual-experience variables, such as age of L2 acquisition and/or proficiency. Given that these factors have been shown to modulate cognitive outcomes in different populations (Bosma et al. 2017; Ferré, Sánchez-Casas, and Guasch 2006; Grant, Fang, and Li 2015; Jasinska and Petitto 2013; Linck et al. 2014; Matusevych,

204 The Neurocognition of Translation and Interpreting

Alishahi, and Backus 2015), ruling them out is an important milestone for the observed effects to be reliably attributed to interpreting practice. Furthermore, comparisons between subjects specializing in simultaneous interpreting and other IR modalities (translation or consecutive interpreting) suggest that some of the adaptations and enhancements identified so far may be distinctive of the former activity (Becker et al. 2016; Henrard and Van Daele 2017; Van de Putte et al. 2018). However, this is hardly true of all the domains examined so far. In fact, comparisons between professional SIs and consecutive interpreters (or even other language professionals, such as L2 teachers) have yielded non-significant differences in aspects as varied as fluency, listening span (Stavrakaki et al. 2012), and information recall (Hiltunen et al. 2016; Hiltunen and Vik 2017). Establishing which cognitive modulations are exclusive of experience in simultaneous interpreting, vis-à-vis other IR modalities, is one of the outstanding challenges for neurocognitive approaches within TIS. 7.5

The plastic nature of IR systems

The basic configuration of the bilingual brain suffices to perform IR, and several neurocognitive particularities of IR systems are equally present in amateurs, trainees, and professionals. However, in many respects, the human brain is not indifferent to the frequency, intensity, and quality with which one engages in translation or interpreting. As illustrated here by reference to simultaneous interpreting, recurring practice of a specific IR modality is associated with neuroplastic changes and increased efficiency in several sub-functions of crucial importance for the trade. Plasticity, then, must be factored in as another key property of the neurocognitive mechanisms described throughout this book and as a major research topic for neural approaches in cognitive TIS. By acknowledging this point, we should be able to better understand not only how the human brain reshapes messages across languages, but also how that very process reshapes the human brain.

Chapter 8

A story in the making

8.1

The tale of the attic

A large house was once built, with chambers aplenty and dwellers galore. Each resident spent most of the time in a particular room, but they all knew what lay behind every other door and seamlessly wandered from one area to the next. There was, however, one exception: no matter how many new tenants entered the house, virtually no one ever visited the attic. Why would they, in any case? Life was productive and invigorating on the main floor, whereas the attic, rumor had it, was obscure, inscrutable, and even off-limits. Yet, one day the hatch was opened and the attic could be seen for what it really was. Some of the corners were shady and many parts were incomplete, yes, but how ample and absorbing it turned out to be! Certainly, it was far from obscure, as light had found its way in through multiple windows. Neither was it inscrutable; a number of people, mainly settled in other houses, had accessed it before, carefully chronicled their visits, and left their accounts scattered over the floor. The attic, it soon became apparent, was anything but inaccessible, and thanks to this realization all inhabitants could now benefit from an even larger building. So goes the tale of the neurocognitive attic in the house of TIS – or, at least, so one hopes it will go. This book has documented an exhaustive tour of that neglected space, during which we have collected the testimonies from previous visitors and organized them to reveal a coherent, interpretable story. The full narrative is yet to be written, and it may actually never be completed, but the passages penned so far are eloquent in themselves. Thousands of years after its inception, IR can finally be understood in relation to its most fundamental precondition: the human brain. With gaps and slopes, with twists and turns, this story already spans almost a full century. Through descriptions of cerebral lesions, blood flow patterns, voltage changes, vocal productions, and manual actions, we have gained a new, biologically grounded view of IR and its intervening processes. The emerging picture, though preliminary, offers explicit insights into several topics, including the lateralization of putative networks; their relation with, and partial independence from, other linguistic mechanisms; the functional organization and temporal dynamics of the circuits engaged by different translation directions, processing levels, and types of

206 The Neurocognition of Translation and Interpreting

SL unit; the system’s susceptibility to training-induced plasticity; and the behavioral correlates of its main operations. Not too bad for a first journey around the attic. Now, having perused the first few episodes of that evolving draft, we can come down the hatch and pinpoint their highlights, reminisce over their strong and weak points, and ascertain what the story needs to grow longer, firmer, and richer. Before signing off, then, let us distill what has been learned and indulge in some constructive criticism. 8.2

Q&A

In Chapter 1, we identified several central questions for cognitive TIS that proved underexplored or downright unapproachable in the light of non-neural trends. Seven chapters later, we are in a position to recapitulate them and assemble plausible, empirically-driven answers. Though provisional, these claims attest to the distinct contributions that brain-based research can bring to our overarching field. Question A.  Which functional systems can be identified as independent from one another within the overall architecture of the translating and interpreting mind? Previous models in cognitive TIS have either overlooked this question or tacitly implied that the same set of mechanisms was equally operative in any instance of IR. However, brain-based research shows that a robust account of the system’s internal organization should incorporate a number of specifications. The networks supporting IR overlap greatly with gross regions implicated in general language functions. These regions include perisylvian areas (e.g., Broca’s area, superior temporal lobe) as well as extrasylvian hubs (e.g., motor cortex) and subcortical pathways (e.g., basal ganglia). Basic linguistic functions involved in the task, such as retrieval of TL equivalents for SL words, mainly rely on LH regions, with important but less critical contributions from the RH. However, at least some of the circuits subserving IR are partially autonomous, as they do not fully depend on the system’s capacity to engage in other linguistic operations, such as single-language production. At the same time, the system seems to possess a complex internal configuration, including partially dissociable circuits for different directions (BT and FT), processing levels (conceptually-mediated and form-level operations), and types of SL unit (sentences, words with distinctive psycholinguistic features). With varying levels of detail, the above claims have been explicitly captured in two neuroanatomical models (Fabbro 1999; García 2012a), both of which identify key regions underlying each sub-function. Although these models and their



Chapter 8.  A story in the making 207

supporting studies are unspecific and preliminary in several respects, all of their postulations constitute testable, a posteriori hypotheses that should be incorporated or explicitly rebutted in ongoing discussions about the internal configuration of IR mechanisms. Question B.  How critically do those systems participate in different forms of IR? The above systems manifest dynamic, task-sensitive activity changes depending on the particularities of the IR process at hand. As compared to single-language operations, such as word reading, IR involves activation increases in different frontostriatal and temporal regions. However, the level of engagement of these circuits differs between translation directions, with FT typically eliciting greater modulations in cortical and subcortical hubs (e.g., Broca’s area, putamen). Some of the neurocognitive systems involved also exhibit different activity levels depending on the type of unit being processed. For example, frontostriatal and temporo-parietal regions appear to be differentially critical for the translation of sentences and words, respectively. Also, such regions seem to be distinctly specialized for processing different word types during IR, with the former playing key roles in the translation of action verbs and the latter proving more crucial for the translation of concrete nouns. Notably, some of these neural distinctions have been reported in the absence of behavioral effects (e.g., in terms of accuracy or response time). This shows that, in the cognitive study of IR, neuroscientific methods can capture major functional distinctions that escape the possibilities of non-neural approaches. Question C.  What are the inner temporal particularities of IR? Measures based on response speed, such as those yielded by keylogging or classical response-time measures, allow examining aspects of IR in their overall temporal extent, subsuming all intervening operations (from process onset to behavioral response) under a single result. However, that overarching outcome encompasses several discernible processes, which can be examined with EEG, intracranial recordings, and other neuroscientific methods. Available results indicate that differential activity patterns for BT and FT emerge before the first quarter of a second after source-segment visualization. Also, each translation direction seems to entail distinct dynamics in successive time windows (around 200 and 400 ms) known to index attentional and lexico-semantic demands. Moreover, both before and shortly after the first 500 ms, significant modulations emerge depending on the cognate status, concreteness, and frequency of the source segment. Therefore, both taskand unit-related variables seem to influence the inner time course of IR, yielding informative neurocognitive effects that cannot be tapped considering outward behaviors alone.

208 The Neurocognition of Translation and Interpreting

Question D.  What types of interaction take place between (potentially discernible) cognitive systems during IR? Beyond region-specific modulations and topographically imprecise temporal effects, neuroscientific experiments have offered the first glimpses into how different cognitive systems become coupled during IR. Though still incipient, the evidence points to different functional connectivity patterns for BT and FT (with greater and more distributed cross-regional interactions for the latter) and even structural connectivity changes across cortical and subcortical regions associated with repeated practice of IR. This is one of the least explored topics in the field, but the preliminary findings just summarized suggest that major insights could be gained through more systematic investigation. Question E.  How does sustained practice of certain modalities modulate fine-grained cognitive domains? The neurocognitive systems involved in IR are susceptible to significant changes when constantly put to use in formal training and professional settings. This has been amply demonstrated by multidimensional evidence from prospective and expert SIs. After only one or one and a half years of practice, and then in professional career stages, these populations show structural and functional adaptations in circuits that subserve task-relevant domains. Moreover, these neuroplastic effects occur alongside outwardly measurable boosts in verbal and non-verbal functions, although these seem to be confined to those dimensions that are distinctly taxed during the activity. In addition, such verbal and non-verbal enhancements appear to be independent from each other and from other modulators of cognitive outcomes in bilinguals (e.g., L2 competence and age of acquisition), and at least some of them seem to require longer periods of practice to manifest consistently. In sum, then, these patterns show that, and specify how, the continual practice of IR can reshape key neurocognitive systems involved in the task. Question F.  And, more generally, can cognitive TIS enter in fruitful, reciprocally informative dialogue with the natural sciences? The answer to this question should be self-evident by now. Neuroscience and other relevant fields in the natural sciences represent highly valid and valuable interlocutors for cognitive TIS. This dialogue not only provides new approaches to explore long-standing questions, but also gives rise to novel queries of interest to translation and interpreting scholars from all orientations. Also, as seen throughout the book, brain-based research affords empirical arguments to formulate new models and hypotheses that directly oppose widely disseminated accounts of the translating mind. Yet, perhaps the most noteworthy aspect of this approach is that it is not revolutionary. Whereas the last decades have witnessed an overabundance of

Chapter 8.  A story in the making 209



self-proclaimed “paradigm shifts” in TIS, this long-neglected avenue of interdisciplinary rapprochement harmoniously aligns with previous trends, creating opportunities for virtuous synergies rather than radical reinventions. With the lines of inquiry introduced here, ongoing discussions about the mental basis of translation and interpreting can be enriched and widened, as neural and non-neural trends converge in a co-constructive give-and-take. 8.3

The good…

The answers above constitute the final output of abundant work from a wide community of researchers distributed across time and space. For all its flaws, the extant research program presents several virtues that should be retained by those operating inside it and emulated by those who favor other approaches. Whatever degree of progress can be attributed to brain-based explorations of IR stems from the following methodological and conceptual merits. First, unlike other trends in cognitive TIS, this empirical corpus rests on a considerable variety of research designs. Most of the conclusions reached so far are jointly informed and constrained by single-case reports, single-group studies, between-group comparisons, and pre/post experiments. Moreover, available reports include data from subjects with varying degrees of competence and expertise, ranging from bilinguals with no formal training to students and professionals in the fields of translation and interpreting. The juxtaposition of results from those populations, either within or between studies, is pivotal for establishing which findings are general to any and all bilinguals, and which ones are contingent on field-specific practice. Second, the main patterns identified stem from a wide repertoire of tasks and materials. Of course, the most direct insights come from translation paradigms proper, but major specifications have been obtained through other verbal (e.g., single-language reading, equivalent recognition, word association, lexical decision, semantic decision) and non-verbal (STM, WM, recall, dual-task, mental-set shifting, inhibitory control, attentional) paradigms. Moreover, all of these tools have been used to assess the processing of multiple types of units, varying in their grammatical complexity, form-level features, and semantic properties. Through strategic comparisons of different tasks and conditions, this trend offers a quite comprehensive (though non-exhaustive) coverage of multiple functions involved in its target phenomena. Third, brain-based research affords a multidimensional view of IR. Whereas other approaches are limited to assessments of accuracy and speed, it complements such classic indexes with insights from anatomo-clinical dissociations, structural brain measures, hemodynamic changes, ERPs, oscillatory modulations, functional

210 The Neurocognition of Translation and Interpreting

connectivity patterns, and even brain stimulation techniques. This results in a complex (though still subspecified) conception of the systems and processes in question, which adds new layers of insight to those supporting mainstream work in the field. That toolkit is not merely another way of describing what has already been studied by previous trends. On the contrary, the incorporation of new methods unveils aspects of IR that could hardly be inspected otherwise. Consider, for example, our current possibilities to investigate the time course of the processes involved during translation. Strictly behavioral studies, such as those based on response-time or keylogging outcomes, are blind to the dynamic changes occurring before the subject’s external responses. Instead, high-temporal-resolution techniques, such as ERPs, offer time-locked measures of ongoing processes prior to or even in the absence of a behavioral response. As seen elsewhere in this book, this has led to novel empirical observations, such as the notion that differential cognitive dynamics for BT and FT emerge after only 200 ms once the source segment has been viewed, or the finding that each direction could involve specific attentional and lexico-semantic demands prior to target-segment production (Christoffels, Ganushchak, and Koester 2013; García, Mikulan, and Ibáñez 2016). Moreover, measures of neural activity have revealed significant processing differences between conditions and groups even when behavioral measures failed to do so (Christoffels, Ganushchak, and Koester 2013; Janyan, Popivanov, and Andonova 2009; Klein et al. 1995; Tommola et al. 2001). This shows that different IR processes may involve distinct spatiotemporal dynamics that have no evident correlates in articulation or physical action, which may prevent misinterpreting a null effect in one dependent variable as a null effect of a given factor in its entirety. With new tools come new insights into hitherto unexplored dimensions of IR. Thanks to this and other assets, brain-based research contributes important specifications regarding central topics for cognitive TIS, such as how mental processes vary during IR as a function of directionality, unit-specific variables, or competence and expertise. In addition, neural evidence has paved the way for questions that simply proved unexaminable through other approaches, such as the inner architecture of IR-relevant systems, their degree of activation during particular sub-processes, or the latter’s inner time course, as stated above. Once again, this represents an extension to build on previous achievements, and not a “new turn” from which to recast the entire field. Of course, this approach not only elaborates on previous notions, but it is also capable of challenging them. In particular, findings on the functional organization of IR-relevant routes, the dynamic differences between BT and FT, and the recruitment of specific mechanisms depending on the translation unit have motivated a direct rejection of models which neutralize or underspecify the architecture of



Chapter 8.  A story in the making 211

the intervening systems. As seen in the preceding chapters, we now know that structure-free, non-directional, and unit-blind models misrepresent the translating and interpreting mind. In this sense, the contributions of brain-based research to TIS may be not just additive, but also corrective. Finally, this research niche is exemplary regarding how to forge fruitful collaborations across disciplines. The phenomena that can inform how IR occurs in the human mind are just too complex and heterogeneous to be addressed from a single perspective. Showing awareness of this challenge, most teams involved in brain-based research include experts in diverse areas, such as neuroscience, neurology, neuropsychology, cognitive science, bilingualism, linguistics, and, of course, TIS. If we are ever to understand the inner workings of translation and interpreting at an appliable level, this pluralist policy is not just desirable, but actually indispensable. There lies another lesson to be learned from neural approaches to the topic. 8.4 … and the bad In most scientific endeavors, each achievement is accompanied by a caveat. Brainbased research on IR does not escape this tendency. The work surveyed in this book features a number of limitations that need to be acknowledged and actively addressed as further investigations are produced. Some of the problems concern terminological and conceptual inconsistencies across studies. On the one hand, one and the same term has often been used to refer to different phenomena. This is perhaps most notorious in the use of the word ‘translation’, which can be found to denote varying modalities (written translation, sight translation, silent translation, simultaneous interpreting, consecutive interpreting) in relation to different unit types (single words, sentences, multi-sentential texts). On the other hand, sometimes the same construct is presented under different labels. In the literature on simultaneous interpreting and executive functions, for instance, the terms STM and WM are employed inconsistently (and sometimes even interchangeably) to describe different tasks and processes. Both scenarios can lead to imprecise statements and errors in the quest for systematic patterns across individual reports. Of course, this problem is not exclusive to neural approaches; indeed, the same scenario has been identified in historical overviews of general notions within TIS, such as those of translation methods and translation techniques (Hurtado Albir 2001). Actually, similar inconsistencies can be found in virtually any scholarly arena. Although it is probably impossible to fully circumvent these obstacles in theoretically diverse spaces, it would be desirable for a consensus group to emerge and set forth conventional terminological criteria with a view to augmenting the field’s internal cohesion.

212 The Neurocognition of Translation and Interpreting

Another shortcoming stems from sampling procedures. Some of the multiparticipant experiments reported feature low sample sizes, which undermines the reliability of their results (Button et al. 2013). Indeed, some neuroscientific studies on interpreting have been ardently criticized because of this and other methodological flaws – see Hervais-Adelman, Moser-Mercer, and Golestani (2018).51 This represents a sensitive issue for the progress of the field, because underpowered studies can easily lead to the detection of false positives and fail to capture true effects (Button et al. 2013). Even though the recruitment of large samples often proves costly, dilatory, and logistically challenging, no efforts should be spared to ensure the robustness of reported results. In this sense, the systematic inclusion of power estimation measures at the beginning of a study would offer an objective way to determine how many subjects are needed to obtain a reliable effect. Moreover, important subject-related variables are poorly described and controlled in several studies. In many cases, little to no information is offered about the participants’ general language history, L2 competence, dominant IR modality, translation or interpreting skills in each direction, hours of practice, or years of professional experience – when appropriate. Such missing details may mask informative distinctions among groups and subjects. Indeed, individual traits have been proposed to constitute key modulators of performance in BT and FT (Ferreira 2014), both directions being susceptible to differences in even informal levels of translation competence (García et al. 2014). In this sense, an instructive lesson can be learned from the field of bilingualism, in which standardized questionnaires have been generated and made publicly available to systematically assess numerous variables of interest – for an example, see Li et al. (2014). Additional insufficiencies must be recognized in terms of the materials and tasks employed in some studies. The number of stimuli in certain experiments is objectively low. Most notably, in clinical assessments based on the Bilingual Aphasia Test (Paradis 1979, 2011), each translation condition ranges from five to ten trials, which is enough to demonstrate marked dissociations but may prove insufficient to reveal subtler contrastive patterns between tasks, directions or unit types. Also, not all experiments considered have been equally successful at ruling out the impact of stimulus-related confounds. Whereas most studies involving two or more conditions have effectively controlled for certain factors (e.g., word frequency and length), less systematic attention has been paid to other variables (e.g., familiarity, concreteness, cognate status, age of acquisition). A more strict control of these features will be necessary for future research to yield more focalized conclusions.

51. Note that the study targeted by this criticism (Elmer and Kühnis 2016) has been omitted from Chapter 7 on account of its objective methodological problems.



Chapter 8.  A story in the making 213

Moreover, most investigations have concentrated on the processing of single words and sentences, partially neglecting naturalistic discourse-level materials. Although, as stated in Chapter 6, several effects first discovered via atomistic paradigms were then confirmed to emerge even in ecological textual tasks, this gap has not yet been systematically bridged in extant research. In fact, the effects observed for a given variable in one unit type may not necessarily remain unaltered in more complex ones. For example, although the cognate effect, as first reported for single words, has been corroborated with sentential units, its impact is reduced in proportion to semantic and grammatical constraints (Starreveld et al. 2014; van Hell and de Groot 2008). Ideally, any effect documented in one unit type should then be subject to further testing in progressively more complex units to ascertain its susceptibility or imperviousness to concomitant linguistic demands. In particular, further work focused on supra-sentential texts, be they real or constructed specially for a given research aim, will be crucial to understand how traces of fine-grained sub-processes (i.e., those targeted by atomistic tasks) relate to key markers of the more holistic operations in which they occur during IR. Also, as is often the case in many other areas of neuroscience and cognitive science, statistical criteria vary greatly from one report to the next. Alpha levels and outlier removal thresholds are varyingly established in an unprincipled manner, and the same happens with method-specific parameters set to determine that two conditions reliably differ – such as the minimal number of contiguous voxels or consecutive time-points in fMRI and ERP studies, respectively. Also, target conditions (e.g., activity levels during word translation) are compared with dissimilar baseline conditions in different studies (e.g., reading-related modulations or resting-state activity). Moreover, effect sizes are not always reported. Whereas identical inconsistencies have not prevented major progress in the study of other domains, such as memory, perception, attention, and social cognition, further brain-based studies on IR would benefit from the consolidation of more consensual analysis settings. In sum, the work conducted so far is not flawless. Truth be told, however, neither could it be so, for science is a human – and thus, necessarily imperfect– endeavor. In any case, these problems do not invalidate the core achievements noted above; rather, their explicit identification paves the way for improving upon previous milestones as new research is conducted. In the face of a problem, the worst policy is to ignore it. By adopting a self-critical spirit, neural approaches will be able to maximize their contributions to cognitive TIS.

214 The Neurocognition of Translation and Interpreting

8.5

Needs assessment

Such a virtuous development, if it is going to occur, requires not just improving on what has been done, but also forging new lines of action. Expansive initiatives need to be pursued on a strictly scientific level and, no less importantly, in institutional spheres, too. The future of this trend depends on how well these two dimensions are articulated from now on. 8.5.1 More, better science Looking forward, the field needs to broaden its thematic, methodological, and theoretical horizons. Future endeavors should expand on and beyond the topics covered in this book. First, the main studies conducted so far should be replicated by separate teams in order to distinguish between consistent and inconsistent results. Ideally, these replications should be accompanied by complementary experiments assessing additional factors and their potential influence on previously reported effects. For example, it would be informative to explore the interaction among various psycholinguistic variables (frequency, length, cognate status, concreteness, age of acquisition) and how these impinge on IR as practiced by brain-lesioned patients and healthy subjects with different levels of competence and expertise. It would be highly useful, too, to investigate how neurocognitive processes are modulated as successive segments are processed during a translation or interpreting task. Studies on single-language reading have shown that theta-band power increases as new words are processed in an unfolding sentence, suggesting an incremental WM load (Bastiaansen, van Berkum, and Hagoort 2002). By investigating real-time neurocognitive dynamics as a ST progresses during IR, we could move beyond the cumulative, condition-specific ethos governing the field so far and engage in fruitful explorations of segment-after-segment modulations. Moreover, beyond the advances presented in Chapter 6, the cognitive effects triggered by translation units are not restricted to the latter’s lexico-semantic or morphosyntactic properties. In fact, much of the information evoked by a given segment during actual IR assignments depends on its co-textual and contextual relations. Even though neuroscientific research has not yet ventured into this territory, important progress could be thus achieved. Despite the major challenges involved, one could eventually envisage theoretical formulations that link findings from atomistic paradigms to broader text-level frameworks. This could result in novel accounts of how mental models, schemata, or other situational factors influence the meanings derived from single units and micro-propositions, leading to the construal of integrated concepts or complex propositions – for relevant insights, see Kintsch (1998) and Kintsch and Mangalath (2011).



Chapter 8.  A story in the making 215

Also, whereas several non-linguistic mechanisms have been characterized in relation to interpreting experience, not much is known about their specific role during IR proper. To bridge this gap, future studies could include different translation or interpreting tasks in which linguistic factors are held constant but executive demands are strategically manipulated. This could be achieved, for example, by creating distractive and non-distractive conditions (to assess inhibitory and attentional mechanisms) or by including pre-IR conditions which markedly tax fundamental domains (STM, WM, mental-set shifting) to assess how the depletion of specific resources impacts on IR dynamics and outcomes. Moreover, foundational efforts should be made to investigate key topics which have not yet been approached from a neurocognitive perspective. For instance, specific designs could be devised or adapted to identify neural modulations associated with ST and TT errors, the adoption of particular processing strategies or the development of topic-specific expertise. In addition, although increasing attention is being paid to neuroplastic effects in SIs, there is very little evidence on potential psychobiological adaptations in people who repeatedly practice other IR modalities, including written translation and consecutive interpreting. Given that some cognitive domains (e.g., cross-linguistic search) may be presumed to be similarly boosted across different forms of IR, whereas others (e.g., WM storage-plus-processing) seem to be exceedingly taxed in only some of them, longitudinal neuroscientific comparisons of translation and interpreting trainees relative to controls could discriminate between modality-neutral and modality-specific changes in the development of IR experience. Though certainly challenging, these issues could be effectively tackled by cross-disciplinary teams actively involving brain scientists and TIS scholars. Relatedly, abundant opportunities exist to explore how the neural mechanisms underlying ST and TT processing change depending on input and output modes. As shown by cases of pure agraphia and pure alexia, as well as multiple neuroimaging studies, reading and writing depend on widely segregated neural networks, which also differ from those that sustain listening and speaking (Dehaene 2009; Luzzatti 2008). Therefore, the specific substrates engaged during SL comprehension may be largely determined by the specific sensorimotor demands of each IR modality (e.g., reading in translation vs. listening in interpreting), and the same holds true for TL production (e.g., writing in translation vs. speaking in interpreting). In this sense, carefully designed experiments could compare activation levels for the same texts across IR modalities and begin to disentangle which mechanisms, if any, play more dominant transmodal roles in each of the three macro-phases (see Chapter 5, Figure 5.1). In addition, multicentric studies would be instrumental to ascertain which of the effects reported so far hold universally across language pairs and which ones

216 The Neurocognition of Translation and Interpreting

are driven by typological or idiosyncratic properties of the languages involved. As it happens, the widely varying structural features of languages across the world are often accompanied by differential processing demands (Evans and Levinson 2009; Han et al. 2013). It follows that at least some of the results reported so far may not be fully generalizable, which creates fertile spaces for further investigation. It would also be useful to integrate different measures of particular processes via multimethodological approaches. With a few exceptions (Becker et al. 2016; Van de Putte et al. 2018), most studies so far have explored their target phenomena relying on behavioral tests or structural imaging or functional imaging or ERPs or functional connectivity metrics. In addition to these single-method strategies, novel projects could favor combinations of different techniques, as previously done in healthy subjects and patient populations to reach multidimensional characterizations of particular cognitive functions (Dimigen et al. 2011; García-Cordero et al. 2016; Kliegl et al. 2012; Melloni et al. 2016). From a translational perspective, neuroscientific tools offer several alternatives for establishing predictors of performance in different domains supporting IR. Previous studies on children and adults show that different brain properties correlate with outcomes in L1 and L2 tasks, such as language reading and imitation (Vaquero, Rodriguez-Fornells, and Reiterer 2017; Yeatman et al. 2012). By the same token, the regions and networks identified in previous chapters emerge as obvious candidates to predict cognitive strengths and weaknesses in aspiring translators and interpreters. Research in this direction could bridge the gap between empirico-theoretical and applied interests in the field. Finally, extant and future advances should be given full consideration in the formulation of new theoretical proposals. First and foremost, unless convincingly refuted or fine-tuned, the findings enumerated in the present book should be incorporated as key constraints for models produced within cognitive TIS. Importantly, this concerns both neural and non-neural trends. Irrespective of one’s theoretical and epistemological stance, the architectural and functional principles identified here must be presumed operative behind any instance of IR. Neglecting such specifications could likely lead to simplistic or even incorrect postulations. Just as neural approaches have benefited from progress made in other arenas, so too non-neural ones should profit from the contributions of brain-based research. Embracing this spirit is vital for cognitive TIS to evolve from being a collective to a co-constructive enterprise. Currently, the theoretical landscape in TIS is vast but disconnected. The give-and-take between different approaches towards conceptual formalizations has been partial at best. Yet, the specialties and compartmentalizations of academia are in no way mirrored by the functioning of human neurocognition, which is essentially integrative and transdimensional in its nature and operation. In this sense, the most outstanding challenge is for different models



Chapter 8.  A story in the making 217

to assimilate – rather than merely acknowledge– robust findings captured beyond their encompassing trend. In all likelihood, this is the only way for us to reach a comprehensive account of IR that satisfies descriptive, explanatory, and appliable imperatives. 8.5.2 An institutional architecture At this point, the guidelines above are merely wishful enunciations. Their value, if any, will remain aspirational unless they can be turned into concrete actions. That transition calls for a number of educational and institutional initiatives. First, academic programs in TIS should open their curricula to include courses in, or related to, brain-based research. In addition to introductory classes presenting the rationale, contributions, and implications of extant work, students should be offered workshops on relevant methodological skills to actively engage in novel investigations. Practical knowledge about the possibilities and limitations of neuroscientific methods would be essential to this end, and so would training in experimental design, task administration, statistics, and data interpretation. Most of this knowledge, in fact, would also be of general usefulness to those operating in non-neural cognitive approaches. The aim of these educational innovations should not be just to promote enthusiasm for brain-based research, but also to raise awareness of its problems and shortcomings. In particular, newcomers should be alerted not to overestimate the weight of neuroscientific results. As it happens, the mere presence of brain-related jargon, even when irrelevant to the point at hand, leads people to feel more satisfied with explanations of cognitive phenomena (Fernandez-Duque et al. 2015; Weisberg et al. 2008). This scenario is all the more worrisome given that neuroscience, much like any other experimental approach to human cognition, faces a reproducibility crisis largely triggered by widespread theoretical, epistemic, and methodological flaws (Munafò et al. 2017). Accordingly, from the outset, an explicit distinction should be made between an empirical result and a factual demonstration, so that students become empowered to know the actual scope of the findings they read about and eventually produce. Encyclopedic knowledge, however, will not suffice. Students should have direct access to laboratories where they can learn how to use relevant tools. To this end, financial investments on the part of translation and interpreting schools or departments should be accompanied by a firm determination to overcome the “two cultures” divide. Whatever amount of time and money is spent on neuroscientific research in TIS should not be seen as a sacrilege given the discipline’s humanistic origins, but rather as a trans-ontological extension of such roots. The laboratories listed in Chapter 1 (Section 1.6) stand as living proof of the gains that can be

218 The Neurocognition of Translation and Interpreting

reached therefrom, and one can only hope that similar initiatives will multiply to favor a truly holistic understanding of translation and interpreting. Of course, this does not necessarily mean that translation and interpreting departments should found their own laboratories. Active agreements with psychology, neuroscience, or even health science programs could be forged so that TIS scholars can exchange knowledge and hands-on experience with experts from relevant fields. Frequent collaboration should be encouraged in all steps of a project, ranging from the definition of hypotheses to task design, stimulus construction, data collection and analysis, interpretation of findings, and manuscript writing. As in many other areas of interdisciplinary inquiry, outcomes are likely to be maximized by integrating existing forms of expert knowledge rather than developing all of them from scratch. At the same time, ensuing advances should be disseminated not only in neuroscience venues, but also in outlets within TIS. Fortunately, the latter is already happening. Little by little, neurocognitive approaches are beginning to figure more prominently in journals like Perspectives, Meta, Target, Translation Spaces, and Translation, Cognition, and Behavior, be it in the form of bibliographical input, full articles, thematic sections, and even special issues. These editorial developments should be accompanied by greater presence of brain-based research in international conferences and symposia, as explicitly advocated by the international congress series “Translation, Interpreting, and Cognition,” launched in Argentina in 2017 and continued in Germany in 2019. No less importantly, the circulation of breakthroughs will greatly benefit from increased visibility of wide-ranging multidisciplinary spaces, like those offered by the Translation, Research, Empiricism, Cognition (TREC) network. Institutional resolutions along these lines will fuel a virtuous, recursive cycle of actions favoring the generation and appropriation of relevant knowledge. 8.6 Parting words Translation and interpreting are multilayered activities. In its quest to holistically approach them, TIS has embraced contributions from philosophical, linguistic, textual, pragmatic, sociocultural, literary, and psychological approaches, among others. For decades, however, neurobiological considerations remained outside its central concerns, on the assumption that the brain could not be fruitfully investigated. That reservation may have been true in the early stages of the discipline and maybe even in its moments of consolidation, but now the tides have changed. If the mission of TIS, in general, and cognitive TIS, in particular, is to provide a



Chapter 8.  A story in the making 219

comprehensive account of their objects of study, neurocognitive dimensions must be placed on a par with other levels of inquiry. This demand is not arbitrary. Neural activity unavoidably underlies every instance of IR, in every culture and society, across genres and text types, at any level of competence, irrespective of the modality employed, and no matter the goal of the task. The human brain is arguably the only constant factor cutting across the history of interlingual brokering. Making it part of the field’s inquiries is not a contemporary eccentricity, but a way of settling a long-standing debt. The conclusions reached so far are mostly partial and preliminary, and they vary in their degree of empirical support. Therefore, all of them could be proven wrong. While this may be perceived as a weakness, it is actually an invaluable asset, for a scientific arena can only progress on the basis of falsifiable theses. Unlike the axiomatic dictums that once dominated our conception of the translating mind, the claims presented here can be subjected to further experimental testing, so that they can be corroborated, rejected, or refined as needed. In this sense, too, the narrative they are part of represents a work in progress. Now that the attic is open and its story has been assembled, anyone can peruse the extant passages and contribute to their prolongation. In all likelihood, the most exciting episodes are yet to be produced. It will be our fortune to witness their development and our responsibility to make them solid, brilliant, and useful. These final lines should then not be read as a “The end,” but as a “To be continued.” If the present book inspires any type of extension, or even its own rebuttal, the toil behind its creation will be well justified.

About the author

Dr. Adolfo M. García specializes in the neuroscience of language. Having graduated with honors as Technical-Scientific Translator (MdPCC, Argentina) and Teacher of English as a Foreign Language (UNMdP, Argentina), he obtained his Ph.D. in Language Sciences (UNCuyo, Argentina), with funding from the National Scientific and Technical Research Council (CONICET, Argentina). His award-winning thesis described neurocognitive aspects of the systems subserving interlingual reformulation in bilinguals. Then, with the support of a Postdoctoral Fellowship from CONICET, he extended his scientific education at the Institute of Cognitive Neurology (INECO). His research training on neurolinguistics also included stays at New York University and Rice University (USA). He is now the Scientific Director of the Laboratory of Experimental Psychology and Neuro­ science, at the Institute of Cognitive and Translational Neuroscience (Argentina). He is also Assistant Researcher at CONICET (Argentina), Adjunct Professor of Neurolinguistics at the Faculty of Education of UNCuyo (Argentina), member of the Translation, Research, Empiricism, Cog­ nition (TREC) Network, and honorary member of the Center of Cognitive Neuroscience at La Laguna University (Spain). In addition, he has served as Associate Editor for the Journal of World Languages, the Journal of Alzheimer’s Disease, and Perspectives: Studies in Translation Theory and Practice; as Review Editor for Frontiers in Human Neuroscience, Frontiers in Aging Neuro­science, and Frontiers in Psychiatry; and as ad hoc reviewer for dozens of leading journals in neuroscience, neurolinguistics, cognitive science, linguistics, and translation studies. He has also guest-edited special issues for Cortex and Perspectives: Studies in Translation Theory and Practice. As a professor, he has taught undergraduate, graduate, and professional development courses in Argentina, Chile, China, Colombia, Germany, and the USA. From 2011 through 2014, he worked as Adjunct Professor of Translation Studies at the National University of Córdoba (Argentina). He has also been a Visiting Professor at Macquarie University (Australia), Universidad del Valle (Colombia), Universidad de Antioquia (Colombia), Pontificia Universidad Católica de Valparaíso (Chile), Johannes Gutenberg University (Germany), Durham University (UK), and Macau University (China). He has supervised numerous research fellows and acted as thesis advisor to undergraduate, masters, and Ph.D. students in Argentina and Europe. Dr. García has collaborated in research projects at Diego Portales University (Chile), New York University (USA), Antioquia University (Colombia), and La Laguna University (Spain). He has also led or partaken in academic projects funded by the Argentine Ministry of Education (2009), the Montevideo Group (2012), the Argentine Chancellery (2014), the Australian Systemic-Functional Linguistics Association (2014), CONICET (2015), COLCIENCIAS (2016), the Argentine Agency for the Promotion of Science and Technology (2018), and the Latin-American-Swiss Center at the University of St. Gallen (2019). Furthermore, he has served as external consultant for the British Council in the field of bilingualism and higher education. In addition, he regularly reviews research projects for national agencies in several countries.

222 The Neurocognition of Translation and Interpreting

He has presented over 150 works in national and international academic events, including numerous conferences as keynote speaker. In addition, he has organized several scientific meetings with worldwide impact, including the first edition of the international conference series “Translation, Interpreting, and Cognition.” Also, he routinely disseminates research for the general public in newspapers, radio broadcasts, and TV shows, including appearances in major outlets such as BBC, Nature News, Discovery Channel, Popular Science, Daily Mail, Newsweek, and Le Monde. A highlight in this area is his participation in the contents and design team of Cere­breando, a public, nation-wide exhibition promoting the social appropriation of neuroscientific knowledge in Argentina. He has published more than 130 scientific works, including books, chapters, and articles in leading journals within the fields of neuroscience (e.g., Brain, Neuroscience and Biobehavioral Reviews, Philosophical Transactions of the Royal Society, Cortex, Journal of Medical Genetics, Brain and Language, Journal of the International Neuropsychological Society), cognitive science (e.g., Nature Human Behavior, Scientific Reports, Cognition, Bilingualism: Language and Cognition), and translation and interpreting studies (e.g., Perspectives: Studies in Translation Theory and Practice, Target, Meta, Translation and Interpreting Studies). His books include Traductología y neurocognición (UNC, 2012); Qué son las neurociencias, co-authored by Agustín Ibáñez (Paidós, 2015); Mente bilingüe, co-edited with Sonia Suárez Cepeda (Comunicarte, 2016); An Introduction to Relational Network Theory (Equinox, 2017, co-authored by William Sullivan and Sarah Tsiang and prefaced by Michael Halliday); Neuroscience and Social Science: The Missing Link, co-edited by Agustín Ibáñez and Lucas Sedeño (Springer, 2018); and Contextual Cognition, co-authored by Agustín Ibáñez (Springer, 2018). He has also published a collection of short stories titled Incierta forma (Babel, 2012). Dr. García has formulated the Hand-Action-Network Dynamic Language Embodiment (HANDLE) model, a theoretical proposal which accounts for neurocognitive synergies during joint processing of language and manual movements. He has also advanced the Disrupted Motor Grounding Hypothesis, which provides a rationale for several linguistic deficits in neurodegenerative movement disorders. These formulations are now being applied to explain several forms of cross-domain interaction in embodied and situated approaches to healthy and pathological cognition. Moreover, he has consolidated an overarching view of the main systems mediating interlingual reformulation in bilinguals, synthetically captured by the Neuroarchitectural Translation Model. In 2013, he received the Most Outstanding Paper Award from the Linguistic Association of Canada and the United States, and he was distinguished by the Ibero-American Neuroeducation Society. In 2015, he was awarded the Young Investigator Prize, granted by the Argentine Association of Behavioral Science, and a distinction by the House of Representatives of Mendoza (Argentina) for his contributions to the public dissemination of neuroscience. In 2017, he was proclaimed Distinguished Citizen of Balcarce in the field of science, and he obtained an award at the MTC 2017 Neurodegenerative Disease Research Contest, hosted by the Argentina Ataxia Association and the Lorena Scarafiocca Foundation for Huntington’s disease. That same year, his work on language as an early marker of Parkinson’s disease was chosen among the top-10 scientific breakthroughs in Argentina. In 2018, his achievements in this field earned him a recognition from the Legislature of the City of Buenos Aires.

List of figures and tables

Figure 1.1 Milestones in the research of neurocognitive aspects of translation and interpreting 35 Figure 2.1 Example of a word translation paradigm 48 Figure 2.2 Example of a lexical decision paradigm with cross-linguistic priming 50 Figure 2.3 Example of a dual-task paradigm 53 Figure 2.4 Example of a color Stroop task 55 Figure 2.5 Example of fMRI recordings during language processing 62 Figure 2.6 Example of PET recordings during language processing 64 Figure 2.7 Example of an ERP modulation during language processing 66 Figure 2.8 Example of oscillatory activity during language processing 67 Figure 2.9 Example of functional connectivity during language processing 69 Figure 3.1 Midsagittal section of the brain 77 Figure 3.2 Lateral section of the left hemisphere 78 Figure 3.3 A depiction of some Brodmann areas 79 Figure 3.4 Some subcortical structures involved in language processing 80 Figure 3.5 Main frontostriatal and temporo-parietal pathways implicated in verbal processing 82 Figure 3.6 Structural components of a neuron 83 84 Figure 3.7 A rough illustration of a synapse Figure 3.8 MRI images of patient Leborgne’s brain 88 Figure 3.9 Activation patterns associated with syntactic processing 90 Figure 3.10 Convergent evidence for the distributed neural network underpinning 92 semantic cognition, including findings from multiple studies Figure 3.11 Schematic depiction (on a single axial slice) of the main neural hubs subserving executive control and language production in bilinguals 96 Figure 4.1 The Neuroarchitectural Translation Model 125 Figure 4.2 Main functional components involved in the process of simultaneous interpretation (per Fabbro 1999: 205) 127 Figure 5.1 Basic architecture of non-directional models in cognitive TIS 133 Figure 5.2 Neural signatures of word translation in forward and backward direction 136 Figure 5.3 Neural signatures of simultaneous interpreting into L1 and L2 137

224 The Neurocognition of Translation and Interpreting

Figure 5.4 Differential modulations of activity in and around Broca’s area during backward and forward translation Figure 5.5 EEG connectivity patterns during translation from ten professional translators Figure 5.6 Intracranial EEG connectivity during translation by a proficient bilingual Figure 5.7 ERP waveforms underlying translation of two word classes in both directions Figure 5.8 Intracranial ERP recordings from a proficient bilingual during BT and FT Figure 5.9 The Revised Hierarchical Model Figure 6.1 Dissociation between word and sentence translation following frontostriatal damage Figure 6.2 Critical role of the primary motor cortex in the translation of action verbs Figure 6.3 Time course of translation-equivalent processing as a function of semantic and form-level overlap Figure 6.4 Unconscious activation of translation equivalents during single-language tasks Figure 6.5 Oscillatory dynamics during translation of words with different frequency Figure 6.6 A psycholinguistic account of the concreteness effect Figure 6.7 Semantic magnitude effects during number-word translation Figure 6.8 Cognate effects in English-Spanish bilinguals with different levels of informal translation competence Figure 7.1 Differential patterns of regional activation during simultaneous interpreting Figure 7.2 Training-induced cortical thickness changes in simultaneous interpreting students Figure 7.3 Training-induced changes in brain activation during simultaneous interpreting for interpreter trainees relative to untrained multilinguals Figure 7.4 Gray matter differences between professional simultaneous interpreters and multilingual control subjects Figure 7.5 N400 differences during semantic decision for simultaneous interpreters and multilingual controls Figure 7.6 Verbal fluency and word translation speed in simultaneous interpreters and bilingual controls Figure 7.7 Picture-naming and word-reading speed in simultaneous interpreters and bilingual controls

139 141 142 143 145 147 158 160 165 166 167 169 170 172 181 182

184 190 191 193 195



List of figures and tables 225

Figure 7.8 Working memory outcomes for simultaneous interpreters, foreign-language teachers, and bilingual students Figure 7.9 Performance on single and dual n-back tasks for simultaneous interpreters and bilingual controls Figure 7.10 Switching and mixing costs in simultaneous interpreters and bilingual controls Table 3.1 Table 4.1 Table 4.2 Table 4.3 Table 4.4

Brodmann areas and rough neuroanatomical correspondences Compulsive translation Inability to translate Paradoxical translation behavior Translation without comprehension

196 198 199 80 105 110 113 115

List of acronyms and abbreviations

µm BA BOLD BT cm EEG ERP ESIT fMRI fNIRS FT HHb Hz IR L1 L2 LH mm MRI ms NIM NTM O2Hb PET RH SI SL ST STM TAP tDCS TIS TL TT WM

micron(s) or micrometer(s) Brodmann area blood-oxygen-level-dependent backward translation centimeter(s) electroencephalography event-related potential École Supérieure d’ Interprètes et Traducteurs functional magnetic resonance imaging functional near-infrared spectroscopy forward translation deoxyhaemoglobin Hertz interlingual reformulation native (first) language non-native (second or foreign) language left hemisphere millimeter(s) magnetic resonance imaging millisecond(s) non-interpreter multilingual Neuroarchitectural Translation Model oxyhaemoglobin positron emission tomography right hemisphere simultaneous interpreter source language source text short-term memory think-aloud protocol transcranial direct current stimulation translation and interpreting studies target language target text working memory

References

Abrevaya, Sofía, Lucas Sedeño, Sol Fittipaldi, David Pineada, Francisco Lopera, Omar Buriticá, Andrés Villegas, Catalina Bustamante, Diana Gomez, Natalia Trujillo, Ricardo Pautassi, Agustín Ibáñez, and Adolfo M. García. 2017. “The road less traveled: Alternative pathways for action-verb processing in Parkinson’s disease.” Journal of Alzheimer’s Disease 55 (4): 1429–1435.  https://doi.org/10.3233/JAD-160737 Abutalebi, Jubin, and David Green. 2007. “Bilingual language production: The neurocognition of language representation and control.” Journal of Neurolinguistics 20 (3): 242–275. https://doi.org/10.1016/j.jneuroling.2006.10.003 Abutalebi, Jubin, and David W. Green. 2008. “Control mechanisms in bilingual language production: Neural evidence from language switching studies.” Language and Cognitive Processes 23 (4): 557–582.  https://doi.org/10.1080/01690960801920602 Aglioti, Salvatore, Alberto Beltramello, Federico Girardi, and Franco Fabbro. 1996. “Neuro­ linguistic and follow-up study of an unusual pattern of recovery from bilingual subcortical aphasia.” Brain 119 (5): 1551–1564.  https://doi.org/10.1093/brain/119.5.1551 Aglioti, Salvatore, and Franco Fabbro. 1993. “Paradoxical selective recovery in a bilingual aphasic following subcortical lesions.” Neuroreport 4 (12): 1359–1362. https://doi.org/10.1097/00001756-199309150-00019 Ahrens, Barbara. 2017. “Interpretation and Cognition.” In The Handbook of Translation and Cog­ nition, ed. by John W. Schwieter, and Aline Ferreira, 445–460. Hoboken, New Jersey: John Wiley & Sons, Inc.  https://doi.org/10.1002/9781119241485.ch24 Alves, Fabio. 2003. Triangulating Translation. Perspectives in Process Oriented Research. Amster­ dam & Philadelphia: John Benjamins.  https://doi.org/10.1075/btl.45 Alves, Fabio, Adriana Pagano, Stella Neumann, Steiner Erich, and Silvia Hansen-Schirra. 2010. “Translation units and grammatical shifts: towards an integration of product and process-based translation research.” In Translation and Cognition, ed. by Gregory Shreve, and Erik Angelone, 109–142. Amsterdam: John Benjamins.  https://doi.org/10.1075/ata.xv.07alv Alves, Fabio, and Daniel C. Vale. 2009. “Probing the unit of translation in time: Aspects of the design and development of a web application for storing, annotating, and querying translation process data.” Across Languages and Cultures 10 (2): 251–273. https://doi.org/10.1556/Acr.10.2009.2.5 Alves, Fabio, and Daniel C. Vale. 2017. “On drafting and revision in translation: A corpus linguistics oriented analysis of translation process data.” In Annotation, Exploitation and Eval­ uation of Parallel Corpora, ed. by Silvia Hansen-Schirra, Stella Neumann, and Oliver Čulo, 89–110. Berlin: Language Science Press. Alves, Fabio, Karina S. Szpak, and Augusto Buchweitz. 2018. “Translation in the brain: Pre­ liminary thoughts about a brain-imaging study to investigate psychological processes involved in translation. In Researching Cognitive Processes of Translation, ed. by Defeng Li, Victoria Lai Cheng Lei, and Yuanjian He, 121–138. Singapore: Springer.

230 The Neurocognition of Translation and Interpreting

Allefeld, Carsten, Stefan Frisch, and Matthias Schlesewsky. 2005. “Detection of early cognitive processing by event-related phase synchronization analysis.” Neuroreport 16 (1): 13–16. https://doi.org/10.1097/00001756-200501190-00004 Amoruso, Lucía, Agustín Ibáñez, Bruno Fonseca, Sebastián Gadea, Lucas Sedeño, Mariano Sigman, Adolfo M. García, Rircardo Fraiman, and Daniel Fraiman. 2016. “Variability in functional brain networks predicts expertise during action observation.” Neuroimage 146: 690–700.  https://doi.org/10.1016/j.neuroimage.2016.09.041 Amoruso, Lucía, Lucas Sedeño, David Huepe, Ailin Tomio, Juan Kamienkowski, Esteban Hurtado, Juan Felipe Cardona, Miguel Ángel Álvarez Gonzalez, Andrés Rieznik, Mariano Sigman, Facundo Manes, and Agustín Ibáñez. 2014. “Time to Tango: expertise and contextual anticipation during action observation.” Neuroimage 98: 366–385. https://doi.org/10.1016/j.neuroimage.2014.05.005 Amunts, Katrin. 2008. “Architectonic language research.” In Handbook of the Neuroscience of Language, ed. by Brigitte Stemmer, and Harry A. Whitaker, 33–43. London: Elsevier. https://doi.org/10.1016/B978-0-08-045352-1.00004-5 Amunts, Katrin, Peter H. Weiss, Hartmut Mohlberg, Peter Pieperhoff, Simon Eickhoff, Jennifer M. Gurd, John C. Marshall, Nadim J. Shah, Gereon R. Fink, and Karl Zilles. 2004. “Analysis of neural mechanisms underlying verbal fluency in cytoarchitectonically defined stereotaxic space – the roles of Brodmann areas 44 and 45.” Neuroimage 22 (1): 42–56. https://doi.org/10.1016/j.neuroimage.2003.12.031 Anderson, Linda 1994. “Simultaneous interpretation: Contextual and translational aspects.” In Bridging the Gap: Empirical Research in Simultaneous Interpretation, ed. by Sylvie Lambert, and Barbara Moser-Mercer, 101–120. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.3.11and Antonova Ünlü, Elena, and Çiğdem Sağın Şimşek. 2018. “Testing the impact of formal interpreting training on working memory capacity: Evidence from Turkish-English students–interpreters.” Lingua 209: 78–88.  https://doi.org/10.1016/j.lingua.2018.04.003 Aparicio, Xavier, Karin Heidlmayr, and Frédéric Isel. 2017. “Inhibition efficiency in highly proficient bilinguals and simultaneous interpreters: Evidence from language switching and Stroop tasks.” Journal of Psycholinguistic Research 46 (6): 1427–1451. https://doi.org/10.1007/s10936-017-9501-3 Arbib, Michael, Péter Érdi, and John Szentágothai. 1998. Neural Organization: Structure, Function, and Dynamics. Cambridge, MA: MIT Press. Ardila, Alfredo, Byron Bernal, and Mónica Rosselli. 2015. “How localized are language brain areas? A review of Brodmann areas involvement in oral language.” Archives of Clinical Neuro­ psychology 31 (1): 112–122.  https://doi.org/10.1093/arclin/acv081 Babcock, Laura, Mariagrazia Capizzi, Sandra Arbula, and Antonino Vallesi. 2017. “Short-term memory improvement after simultaneous interpretation training.” Journal of Cognitive En­ hancement 1 (3): 254–267.  https://doi.org/10.1007/s41465-017-0011-x Babcock, Laura, and Antonino Vallesi. 2015. “Are simultaneous interpreters expert bilinguals, unique bilinguals, or both?” Bilingualism: Language and Cognition 20 (2): 403–417. https://doi.org/10.1017/S1366728915000735. Baez, Sandra, Eduar Herrera, Adolfo M. García, Facundo Manes, Liane Young, and Agustín Ibáñez. 2017. “Outcome-oriented moral evaluation in terrorists.” Nature Human Behaviour 1: 118.  https://doi.org/10.1038/s41562-017-0118

References 231

Bajo, María Teresa, Francisca Padilla, and Presentación Padilla. 2000. “Comprehension processes in simultaneous interpreting.” In Translation in Context, ed. by Andrew Chesterman, Natividad Gallardo San Salvador and Yves Gambier, 127–142. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.39.15baj Bak, Thomas H. 2013. “The neuroscience of action semantics in neurodegenerative brain diseases.” Current Opinion in Neurology 26 (6): 671–677. https://doi.org/10.1097/WCO.0000000000000039 Barbizet, Jacques. 1964. “Le problème du codage cérébral, son rôle dans les mécanismes de la mé­moire.” Annales médico-psychologiques 1: 1–27. Barbizet, Jacques, and Philippe Duizabo. 1977. Abrégé de neuropsychologie. Paris: Masson. Barik, Henri C. 1975. “Simultaneous interpretation: Qualitative and linguistic data.” Language and Speech 18 (2): 272–298.  https://doi.org/10.1177/002383097501800310 Bastiaansen, Marcel C., Jos J. van Berkum, and Peter Hagoort. 2002. “Event-related theta power increases in the human EEG during online sentence processing.” Neuroscience Letters 323 (1): 13–16.  https://doi.org/10.1016/S0304-3940(01)02535-6 Becker, Maxi, Torsten Schubert, Tilo Strobach, Jürgen Gallinat, and Simone Kühn. 2016. “Simultaneous interpreters vs. professional multilingual controls: Group differences in cognitive control as well as brain structure and function.” NeuroImage 134: 250–260. https://doi.org/10.1016/j.neuroimage.2016.03.079 Bell, Roger. 1991. Translation and Translating. London: Longman. Bender, Heidi A., Adolfo M. García, and William Barr. 2010. “An interdisciplinary approach to neuropsychological test construction: perspectives from translation studies.” Journal of the International Neuropsychological Society 16 (2): 227–232. https://doi.org/10.1017/S1355617709991378. Bennett, Paul. 1994. “The translation unit in human and machine.” Babel 40 (1): 12–20. https://doi.org/10.1075/babel.40.1.03ben Benowitz, Larry I., Kenneth L. Moya, and David N. Levine. 1990. “Impaired verbal reasoning and constructional apraxia in subjects with right hemisphere damage.” Neuropsychologia 28 (3): 231–241.  https://doi.org/10.1016/0028-3932(90)90017-I Bernardini, Silvia. 2001. “Think-aloud protocols in translation research: Achievements, limits, future prospects.” Target 13 (2): 241–263.  https://doi.org/10.1075/target.13.2.03ber Bialystok, Ellen, Fergus I. Craik, and Gigi Luk. 2012. “Bilingualism: consequences for mind and brain.” Trends in Cognitive Science 16 (4): 240–250.  https://doi.org/10.1016/j.tics.2012.03.001. Bialystok, Ellen. 2009. “Bilingualism: The good, the bad, and the indifferent.” Bilingualism: Lan­ guage and Cognition 12 (1): 3–11.  https://doi.org/10.1017/S1366728908003477 Bilalić, Merim. 2017. The Neuroscience of Expertise. Cambridge: Cambridge University Press. https://doi.org/10.1017/9781316026847 Bilalić, Merim, Peter McLeod, and Fernand Gobet. 2008. “Inflexibility of experts – Reality or myth? Quantifying the Einstellung effect in chess masters.” Cognitive Psychology 56 (2): 73–102.  https://doi.org/10.1016/j.cogpsych.2007.02.001 Binder, Jeffrey R., and Rutvik H. Desai. 2011. “The neurobiology of semantic memory.” Trends in Cognitive Science 15 (11): 527–536.  https://doi.org/10.1016/j.tics.2011.10.001 Birba, Agustina, Indira García-Cordero, Giselle Kozono, Agustina Legaz, Agustín Ibáñez, Lucas Sedeño, and Adolfo M. García. 2017. “Losing ground: Frontostriatal atrophy disrupts language embodiment in Parkinson’s and Huntington’s disease.” Neuroscience and Bio­ behaviornal Reviews 80: 673–687.  https://doi.org/10.1016/j.neubiorev.2017.07.011

232 The Neurocognition of Translation and Interpreting

Bohrn, Isabel C., Ulrike Altmann, and Arthur M. Jacobs. 2012. “Looking at the brains behind figurative language – A quantitative meta-analysis of neuroimaging studies on metaphor, idiom, and irony processing.” Neuropsychologia 50 (11): 2669–2683. https://doi.org/10.1016/j.neuropsychologia.2012.07.021 Bookheimer, Susan. 2007. “Pre-surgical language mapping with functional magnetic resonance imaging.” Neuropsychological Review 17 (2): 145–55. https://doi.org/10.1007/s11065-007-9026-x Borchers, Svenja, Marc Himmelbach, Nikos Logothetis, and Hans-Otto Karnath. 2012. “Direct electrical stimulation of human cortex – the gold standard for mapping brain functions?” Nature Reviews Neuroscience 13 (1): 63–70.  https://doi.org/10.1038/nrn3140 Borius, Pierre-Yves, Carlo Giussani, Louisa Draper, and Franck-Emmanuel Roux. 2012. “Sentence translation in proficient bilinguals: A direct electrostimulation brain mapping.” Cortex 48 (5): 614–622.  https://doi.org/10.1016/j.cortex.2011.01.011 Bosma, Evelyn, Eric Hoekstra, Arjen Versloot, and Elma Blom. 2017. “The minimal and shortlived effects of minority language exposure on the executive functions of Frisian-Dutch bilingual children.” Frontiers in Psychology 8: 1453.  https://doi.org/10.3389/fpsyg.2017.01453 Bouton, Charles. 1984. “Le cerveau du traducteur : de quelques propositions sur ce thème.” Meta: Translators’ Journal 29 (1): 44–56.  https://doi.org/10.7202/002129ar Braitenberg, Valentino, and Almut Schüz. 1998. Cortex: Statistics and Geometry of Neuronal Con­ nectivity, second ed. Berlin: Springer.  https://doi.org/10.1007/978-3-662-03733-1 Broca, Paul. 1861a. “Remarques sur le siège de la faculté de la parole articulée, suivies d’une observation d’aphémie (perte de parole).” Bulletin de la Société d’Anatomie 36: 330–357. Broca, Paul. 1861b. “Perte de la parole, ramollissement chronique et destruction partielle du lobe antérieur gauche du cerveau.” Bulletin de la Société Anthropologique 2: 235–238. Brodmann, Korbinian. 1909. Vergleichende Lokalisationslehre der Grosshirnrinde: In ihren Prinzipien dargestellt auf Grund des Zeelenbaues. Leipzig: Barth. Bros-Brann, Eliane. 1976. “Critical comments on H.C. Barik’s article ‘Interpreters talk a lot, among other things’.” AIIC Bulletin 4 (1): 16–18. Brysbaert, Marc, and Wouter Duyck. 2010. “Is it time to leave behind the Revised Hierarchical Model of bilingual language processing after fifteen years of service?” Bilingualism: Lan­ guage and Cognition 13 (3): 359–371.  https://doi.org/10.1017/S1366728909990344. Burgess, Adrian P. 2012. “Towards a unified understanding of event-related changes in the EEG: The firefly model of synchronization through cross-frequency phase modulation.” PloS One 7 (9): e45630.  https://doi.org/10.1371/journal.pone.0045630. Button, Katherine S., John P. A. Ioannidis, Claire Mokrysz, Brian A. Nosek, Jonathan Flint, Emma S. J. Robinson, and Marcus R. Munafo. 2013. “Power failure: Why small sample size undermines the reliability of neuroscience.” Nature Reviews Neuroscience 14 (5): 365–376. https://doi.org/10.1038/nrn3475 Buxton, Richard B. 2009. Introduction to Functional Magnetic Resonance Imaging: Principles and Techniques. second ed. Cambridge, New York: Cambridge University Press. https://doi.org/10.1017/CBO9780511605505 Buzsáki, Gyorgy. 2006. Rhythms of the Brain. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195301069.001.0001 Byng, Sally, Max Coltheart, Jacqueline Masterson, Margot Prior, and Jane Riddoch. 1984. “Bi­ lingual biscriptal deep dyslexia.” The Quarterly Journal of Experimental Psychology Section A 36 (3): 417–433.  https://doi.org/10.1080/14640748408402170

References 233

Calvo, Noelia, Sofía Abrevaya, Macarena Martínez Cuitiño, Brenda Steeb, Dolores Zamora, Lucas Sedeño, Agustín Ibáñez, and Adolfo M. García. 2019. “Rethinking the neural basis of prosody and non-literal language: Spared pragmatics and cognitive compensation in a bilingual with extensive right-hemisphere damage.” Frontiers in Psychology 10: 570.

doi: 10.3389/fpsyg.2019.00570

Calvo, Noelia, Agustín Ibáñez, and Adolfo M. García. 2016. “The impact of bilingualism on working memory: A null effect on the whole may not be so on the parts.” Frontiers in Psychology 7: 265.  https://doi.org/10.3389/fpsyg.2016.00265 Capitani, Erminio, Marcella Laiacona, Bernard Mahon, and Alfonso Caramazza. 2003. “What are the facts of semantic category-specific deficits? A critical review of the clinical evidence.” Cognitive Neuropsychology 20 (3): 213–61.  https://doi.org/10.1080/02643290244000266 Caplan, David. 1987. Neurolinguistics and Linguistic Aphasiology: An Introduction. New York: Cambridge University Press.  https://doi.org/10.1017/CBO9780511620676 Carl, Michael. 2012. “Translog-II: A program for recording user activity data for empirical reading and writing research.” In Proceedings of the Eight International Conference on Language Resources and Evaluation, ed. by Nicoletta Calzolari, Khalid Choukri, Thierry Declerck, Mehmet Uğur Doğan, Bente Maegaard, Joseph Mariani, Asunción Moreno, Jan Odijk and Stelios Piperidis, 4108–4112. Copenhagen: European Language Resources Association. Carl, Michael, Srinivas Bangalore, and Moritz Schaeffer. 2016. New Directions in Empirical Translation Process Research: Exploring the CRITT TPR-DB. Heidelberg: Springer. https://doi.org/10.1007/978-3-319-20358-4 Carl, Michael, and Matthias Buch-Kromann. 2010. “Correlating translation product and translation process data of professional and student translators.” Annual Conference of the European Association for Machine Translation, Saint-Raphaël, France. Carl, Michael, and Martin Kay. 2011. “Gazing and typing activities during translation: A comparative study of translation units of professional and student translators.” Meta: International Translator’s Journal 65 (4): 952–975.  https://doi.org/10.7202/1011262ar Carl, Michael, and Moritz J. Schaeffer. 2017. “Models of the translation process.” In The Hand­ book of Translation and Cognition, ed. by Aline Ferreira and John W. Schwieter, 50–70. Hoboken, New Jersey: John Wiley & Sons, Inc.  https://doi.org/10.1002/9781119241485.ch3 Carreiras, Manuel, and Charles Clifton. 2004. The On-line Study of Sentence Comprehension: Eyetracking, ERPs and Beyond. London: Taylor & Francis. https://doi.org/10.4324/9780203509050 Catani, Marco, Flavio Dell’Acqua, Alberto Bizzi, Stephanie J. Forkel, Steve C. Williams, Andrew Simmons, Declan G. Murphy, and Michael Thiebaut de Schotten. 2012. “Beyond cortical localization in clinico-anatomical correlation.” Cortex 48 (10): 1262–1287. https://doi.org/10.1016/j.cortex.2012.07.001 Cohen, Laurent, Olivier Martinaud, Cathy Lemer, Stéphane Lehéricy, Yves Samson, Michäel Obadia, Andrea Slachevsky, and Stanislas Dehaene. 2003. “Visual word recognition in the left and right hemispheres: Anatomical and functional correlates of peripheral alexias.” Cerebral Cortex 13 (12): 1313–1333.  https://doi.org/10.1093/cercor/bhg079 Coltheart, Max. 1981. “The MRC psycholinguistic database.” The Quarterly Journal of Experimental Psychology Section A 33 (4): 497–505.  https://doi.org/10.1080/14640748108400805 Constable, R. Todd, Kenneth R. Pugh, Ella Berroya, W. Einar Mencl, Michael Westerveld, Weijia Ni, and Donald Shankweiler. 2004. “Sentence complexity and input modality effects in sentence comprehension: an fMRI study.” Neuroimage 22 (1): 11–21. https://doi.org/10.1016/j.neuroimage.2004.01.001

234 The Neurocognition of Translation and Interpreting

Conway, Andrew R. A., Michael J. Kane, Michael F. Bunting, D. Zach Hambrick, Oliver Wilhelm, and Randall W. Engle. 2005. “Working memory span tasks: A methodological review and user’s guide.” Psychonomic Bulletin & Review 12 (5): 769–786. https://doi.org/10.3758/BF03196772 Cooper, Richard P., and Tim Shallice. 2010. “Cognitive neuroscience: The troubled marriage of cognitive science and neuroscience.” Topics in Cognitive Science 2 (3): 398–406. https://doi.org/10.1111/j.1756-8765.2010.01090.x Correia, Joao, Elia Formisano, Giancarlo Valente, Lars Hausfeld, Bernadette Jansma, and Milene Bonte. 2014. “Brain-based translation: fMRI decoding of spoken words in bilinguals reveals language-independent semantic representations in anterior temporal lobe.” Journal of Neuro­science 34 (1): 332–338.  https://doi.org/10.1523/JNEUROSCI.1302-13.2014 Cowan, Nelson. 2000. “Processing limits of selective attention and working memory: Potential implications for interpreting.” Interpreting 5 (2): 117–146. https://doi.org/  https://doi.org/10.1075/intp.5.2.05cow Chee, Michael W., Chun S. Soon, and Hwee L. Lee. 2003. “Common and segregated neuronal networks for different languages revealed using functional magnetic resonance adaptation.” Journal of Cognitive Neuroscience 15 (1): 85–97.  https://doi.org/10.1162/089892903321107846 Cheour, Marie, Rita Ceponiene, Anne Lehtokoski, Aavo Luuk, Jüri Allik, Kimmo Alho, and Risto Näätänen. 1998. “Development of language-specific phoneme representations in the infant brain.” Nature Neuroscience 1 (5): 351–353.  https://doi.org/10.1038/1561 Chernov, Ghelly V. 2004. Inference and Anticipation in Simultaneous Interpreting: A Probabilityprediction Model. Amsterdam: John Benjamins.  https://doi.org/10.1075/btl.57 Chernov, Ghelly V. 1994. “Message redundancy and message anticipation in simultaneous interpretation.” In Bridging the Gap: Empirical Research in Simultaneous Interpretation, ed. by Sylvie Lambert and Barbara Moser-Mercer, 139–153. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.3.13che Chincotta, Dino, and Geoffrey Underwood. 1998. “Simultaneous interpreters and the effect of concurrent articulation on immediate memory.” Interpreting: International Journal of Re­ search and Practice in Interpreting 3 (1): 1–20.  https://doi.org/10.1075/intp.3.1.01chi Chmiel, Agnieszka. 2018. “In search of the working memory advantage in conference interpreting – Training, experience and task effects.” International Journal of Bilingualism 22 (3): 371–384.  https://doi.org/10.1177/1367006916681082 Choi, Wonil, Rutvik H. Desai, and John M. Henderson. 2014. “The neural substrates of natural reading: a comparison of normal and nonword text using eyetracking and fMRI.” Frontiers in Human Neuroscience 8: 1024.  https://doi.org/10.3389/fnhum.2014.01024. Chomsky, Noam. 1959. Syntactic Structures. The Hague: Mouton. Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, Massachusetts: MIT Press. Christoffels, Ingrid. 2004. “Cognitive Studies in Simultaneous Interpreting. Unpublished doctoral dissertation.” Unpublished doctoral dissertation, University of Amsterdam. Christoffels, Ingrid, and Annette M. B. de Groot. 2005. “Simultaneous interpreting: A cognitive perspective.” In Handbook of Bilingualism: Psycholinguistic Approaches, ed. by Judith F. Kroll and Annette M. B. de Groot, 454–479. New York: Oxford University Press. Christoffels, Ingrid K., Annette M. B. de Groot, and Judith F. Kroll. 2006a. “Memory and language skills in simultaneous interpreters: The role of expertise and language proficiency.” Journal of Memory and Language 54 (3): 324–345.  https://doi.org/10.1016/j.jml.2005.12.004

References 235

Christoffels, Ingrid K., Lesya Ganushchak, and Dirk Koester. 2013. “Language conflict in translation: an ERP study of translation production.” Journal of Cognitive Psychology 25 (5): 646–664.  https://doi.org/10.1080/20445911.2013.821127 Christoffels, Ingrid K., Annette M. B. de Groot, and Lourens J. Waldorp. 2003. “Basic skills in a complex task: A graphical model relating memory and lexical retrieval to simultaneous interpreting.” Bilingualism: Language and Cognition 6 (3): 201–211. https://doi.org/10.1017/S1366728903001135 Dalla Volta, R., M. Fabbri-Destro, M. Gentilucci, and P. Avanzini. 2014. “Spatiotemporal dynamics during processing of abstract and concrete verbs: An ERP study.” Neuropsychologia 61: 163–174.  https://doi.org/10.1016/j.neuropsychologia.2014.06.019 Damasio, Antonio. 1994. Descartes’ Error: Emotion, Reason and the Human Brain. London: Penguin. Darò, Valeria, and Franco Fabbro. 1994. “Verbal memory during simultaneous interpretation: Effects of phonological interference.” Applied Linguistics 15: 365–381. https://doi.org/10.1093/applin/15.4.365 Davidson, Donald J., and Peter Indefrey. 2007. “An inverse relation between event-related and time-frequency violation responses in sentence processing.” Brain Research 1158: 81–92. https://doi.org/10.1016/j.brainres.2007.04.082 de Groot, Annette M. B. 1992. “Determinants of word translation.” Journal of Experimental Psy­ chology: Learning, Memory, and Cognition 18: 1001–1018. de Groot, Annette M. B. 1993. “Word-type effects in bilingual processing tasks: Support for a mixed representational system.” In The Bilingual Lexicon, ed. by Robert Schreuder and Bert Weltens, 27–51. Amsterdam: John Benjamins.  https://doi.org/10.1075/sibil.6.04gro de Groot, Annette M. B., and Gerard L. J. Nas. 1991. “Lexical representation of cognates and non­cognates in compound bilinguals.” Journal of Memory and Language 30 (1): 90–123. https://doi.org/10.1016/0749-596X(91)90012-9 de Groot, Annette M. B., Lucia Dannenburg, and Janet G. van Hell. 1994. “Forward and backward word translation by bilinguals.” Journal of Memory and Language 33 (5): 600–629. https://doi.org/10.1006/jmla.1994.1029 de Groot, Annette M., Philip Delmaar, and Stephen J. Lupker. 2000. “The processing of interlexical homographs in translation recognition and lexical decision: support for non-selective access to bilingual memory.” Quarterly Journal of Experimental Psychology A 53 (2): 397–428.  https://doi.org/10.1080/713755891. De Vreese, Luc Pieter, Massimo Motta, and Andrea Toschi. 1988. “Compulsive and paradoxical translation behaviour in a case of presenile dementia of the Alzheimer type.” Journal of Neurolinguistics 3 (2): 233–259.  https://doi.org/10.1016/0911-6044(88)90015-2 DeFelipe, Javier, and Isabel Farinas. 1992. “The pyramidal neuron of the cerebral cortex: morphological and chemical characteristics of the synaptic inputs.” Progress in Neurobiology 39 (6): 563–607.  https://doi.org/10.1016/0301-0082(92)90015-7 Dehaene, Stanislas. 2009. Reading in the Brain: The New Science of How We Read. London: Penguin. Dehaene, Stanislas, Emmanuel Dupoux, Jacques Mehler, Laurent Cohen, Eraldo Paulesu, Daniela Perani, Pierre-Francois van de Moortele, Stéphane Lehéricy, & Denis Le Bihan. 1997. “Anatomical variability in the cortical representation of first and second language.” Neuroreport 8 (17): 3809–3815.  https://doi.org/10.1097/00001756-199712010-00030

236 The Neurocognition of Translation and Interpreting

Delisle, Jean, and Judith Woodsworth. 1995. Translators through History. Amsterdam/ Philadelphia: John Benjamins.  https://doi.org/10.1075/btl.13 Desai, Rutvik H., Woni Choi, Vicky T. Lai, and John M. Henderson. 2016. “Toward semantics in the wild: Activation to manipulable nouns in naturalistic reading.” Journal of Neuroscience 36 (14): 4050–4055.  https://doi.org/10.1523/JNEUROSCI.1480-15.2016 Detry, Caroline, Agnesa Pillon, and Marie-Pierre de Partz. 2005. “A direct processing route to translate words from the first to the second language: Evidence from a case of a bilingual aphasic.” Brain and Language 95 (1): 40–41.  https://doi.org/10.1016/j.bandl.2005.07.014 Diamond, Adele. 2013. “Executive functions.” Annual Review of Psychology 64: 135–168. https://doi.org/10.1146/annurev-psych-113011-143750 Diamond, Bruce J., and Gregory M. Shreve. 2017. “Deliberate practice and neurocognitive optimization of translation expertise.” In The Handbook of Translation and Cognition, ed. by Aline Ferreira and John W. Schwieter, 476–495. Hoboken, New Jersey: John Wiley & Sons, Inc. https://doi.org/10.1002/9781119241485.ch26 Diamond, Bruce, and Gregory Shreve. 2010. “Neural and physiological correlates of translation and interpreting in the bilingual brain: Recent perspectives.” In Translation and Cognition, ed. by Gregory Shreve and Erik Angelone, 289–322. Amsterdam & Philadelphia: John Benjamins.  https://doi.org/10.1075/ata.xv.16dia Dijkstra, Ton, Alexander Wahl, Franka Buytenhuijs, Nino Van Halem, Zina Al-Jibouri, Marcel De Korte, and Steven RekkÉ. 2018. “Multilink: A computational model for bilingual word recognition and word translation.” Bilingualism: Language and Cognition: 1–23. https://doi.org/10.1017/S1366728918000287 Dillinger, Mike. 1990. “Comprehension during interpreting: What do interpreters know that bilinguals don’t?” The Interpreters’ Newsletter 3: 41–58. Dimigen, Olaf, Werner Sommer, Annette Hohlfeld, Arthur M. Jacobs, and Reinhold Kliegl. 2011. “Coregistration of eye movements and EEG in natural reading: analyses and review.” Journal of Experimental Psychology: General 140 (4): 552–572.  https://doi.org/10.1037/a0023885 Dong, Yanping, and Jiexuan Lin. 2013. “Parallel processing of the target language during source language comprehension in interpreting.” Bilingualism: Language and Cognition 16 (3): 682–692.  https://doi.org/10.1017/S1366728913000102 Dong, Yanping, and Zhilong Xie. 2014. “Contributions of second language proficiency and interpreting experience to cognitive control differences among young adult bilinguals.” Journal of Cognitive Psychology 26 (5): 506–519.  https://doi.org/10.1080/20445911.2014.924951 Donovan, Clare. 2004. “European Masters Project Group: Teaching simultaneous interpretation into a B language: Preliminary findings.” Interpreting 6 (2): 205–216. https://doi.org/10.1075/intp.6.2.06don Dossani, Rimal, Symeon Missios, and Anil Nanda. 2015. “The legacy of Henry Molaison (1926– 2008) and the impact of his bilateral mesial temporal lobe surgery on the study of human memory.” World Neurosurgery 84 (4): 1127–1135.  https://doi.org/10.1016/j.wneu.2015.04.031. Dragsted, Barbara. 2005. “Segmentation in translation: Differences across levels of expertise and difficulty.” Target 17 (1): 49–70.  https://doi.org/10.1075/target.17.1.04dra Dragsted, Barbara. 2010. “Coordination of reading and writing processes in translation: An eye on uncharted territory.” In Translation and Cognition, ed. by Gregory Shreve and Erik Angelone, 41–62. Amsterdam: John Benjamins.  https://doi.org/10.1075/ata.xv.04dra Dubois, Julien, and Ralph Adolphs. 2016. “Building a science of individual differences from fMRI.” Trends in Cognitive Science 20 (5): 425–443.  https://doi.org/10.1016/j.tics.2016.03.014

References 237

Duchon, Andrew, Manuel Perea, Nuria Sebastián-Gallés, Antonia Martí, and Manuel Carreiras. 2013. “EsPal: one-stop shopping for Spanish word properties.” Behavioral Research Methods 45 (4): 1246–1258.  https://doi.org/10.3758/s13428-013-0326-1 Duffy, Mary. 2000. “The Internet as a research and dissemination resource.” Health Promotion International 15 (4): 349–353.  https://doi.org/10.1093/heapro/15.4.349 Dunabeitia, Jon A., and Manuel Carreiras. 2015. “The bilingual advantage: Acta est fabula?” Cortex 73: 371–372.  https://doi.org/10.1016/j.cortex.2015.06.009 Dunn, John C., and Kim Kirsner. 2003. “What can we infer from double dissociations?” Cortex 39 (1): 1–7.  https://doi.org/10.1016/S0010-9452(08)70070-4 Durant, Will. 1926. The Story of Philosophy. New York: Simon & Schuster. Duyck, Wouter, and Marc Brysbaert. 2004. “Forward and backward number translation requires conceptual mediation in both balanced and unbalanced bilinguals.” Journal of Experimental Psychology: Human Perception and Performance 30 (5): 889–906. https://doi.org/10.1037/0096-1523.30.5.889 Duyck, Wouter, and Marc Brysbaert. 2008. “Semantic access in number word translation: The role of crosslingual lexical similarity.” Experimental Psychology 55 (2): 102–112. https://doi.org/10.1027/1618-3169.55.2.102 Elmer, Stefan. 2012. “The investigation of simultaneous interpreters as an alternative approach to address the signature of multilingual speech processing.” Zeitschrift für Neuropsychologie 23 (2): 105–116.  https://doi.org/10.1024/1016-264X/a000068 Elmer, Stefan, Carina Klein, Jürg Kühnis, Franziskus Liem, Martin Meyer, and Lutz Jäncke. 2014. “Music and language expertise influence the categorization of speech and musical sounds: behavioral and electrophysiological measurements.” Journal of Cognitive Neuroscience 26 (10): 2356–2369.  https://doi.org/10.1162/jocn_a_00632 Elmer, Stefan, and Jürg Kühnis. 2016. “Functional connectivity in the left dorsal stream facilitates simultaneous language translation: An EEG study.” Frontiers in Human Neuroscience 10: 60. https://doi.org/10.3389/fnhum.2016.00060 Elmer, Stefan, Martin Meyer, and Lutz Jäncke. 2010. “Simultaneous interpreters as a model for neuronal adaptation in the domain of language processing.” Brain Research 1317: 147–156. https://doi.org/10.1016/j.brainres.2009.12.052 Elmer, Stefan, Jürgen Hänggi, and Lutz Jäncke. 2014. “Processing demands upon cognitive, linguistic, and articulatory functions promote gray matter plasticity in the adult multilingual brain: Insights from simultaneous interpreters.” Cortex 54: 179–189. https://doi.org/10.1016/j.cortex.2014.02.014 Engel, Jr., Jerome. 2005. “The emergence of neurosurgical approaches to the treatment of epilepsy.” In From Neuroscience to Neurology: Neuroscience, Molecular Medicine, and the Therapeutic Transformation of Neurology, ed. by Stephen Waxman, 81–105. Amsterdam: Elsevier. https://doi.org/10.1016/B978-012738903-5/50006-0 Ericsson, K. Anders, Neil Charness, Paul J. Feltovich, and Robert R. Hoffman. 2006. The Cambridge Handbook of Expertise and Expert Performance. New York: Cambridge Univer­ sity Press.  https://doi.org/10.1017/CBO9780511816796 Ericsson, K. Anders, and Herbert A. Simon. 1984. Protocol Analysis. Cambridge, Massachussetts: The MIT Press. Eriksen, Barbara A., and Charles W. Eriksen. 1974. “Effects of noise letters upon the identification of a target letter in a nonsearch task.” Perception & Psychophysics 16 (1): 143–149. https://doi.org/10.3758/BF03203267

238 The Neurocognition of Translation and Interpreting

Esposito, Antonella. 2016. Research 2.0 and the Impact of Digital Technologies on Scholarly Inquiry. Philapdelphia: IGI Global. Evans, Nicholas, and Stephen C. Levinson. 2009. “The myth of language universals: Language diversity and its importance for cognitive science.” Behavioral and Brain Sciences 32 (5): 429–448.  https://doi.org/10.1017/S0140525X0999094X Eviatar, Zohar, Mark Leikin, and Raphiq Ibrahim. 1999. “Phonological processing of second language phonemes: A selective deficit in a bilingual aphasic.” Language Learning 49 (1): 121–141.  https://doi.org/10.1111/1467-9922.00072 Fabbro, Franco. 2001. “The bilingual brain: Cerebral representation of languages.” Brain and Lan­guage 79 (2): 211–22.  https://doi.org/10.1006/brln.2001.2481 Fabbro, Franco, Bruno Gran, and Laura Gran. 1991. “Hemispheric specialization for semantic and syntactic components of language in simultaneous interpreters.” Brain and Language 41 (1): 1–42.  https://doi.org/10.1016/0093-934X(91)90108-D Fabbro, Franco, Miran Skrap, and Salvatore Aglioti. 2000. “Pathological switching between languages after frontal lesions in a bilingual patient.” Journal of Neurology, Neurosurgery, and Psychiatry 68 (5): 650–652.  https://doi.org/10.1136/jnnp.68.5.650 Fabbro, Franco. 1999. The Neurolinguistics of Bilingualism: An Introduction. Hove: Psychology Press. Fabbro, Franco, and Laura Gran. 1997. “Neurolinguistic research in simultaneous interpretation.” In Conference Interpreting: Current Trends in Research. Proceedings of the International Conference on Interpreting: What Do We Know and How?, ed. by Yves Gambier, Daniel Gile and Christopher Taylor, 9–27. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.23.02fab Fabbro, Franco, Laura Gran, Gianpaolo Basso, and Antonio Bava. 1990. “Cerebral lateralization in simultaneous interpretation.” Brain and Language 39 (1): 69–89. https://doi.org/10.1016/0093-934X(90)90005-2 Fabbro, Franco, and Michel Paradis. 1995. “Differential impairments in four multilingual patients with subcortical lesions.” In Aspects of Bilingual Aphasia, ed. by Michel Paradis, 139– 176. Oxford: Pergamon. Feng, Shiwen, Jennifer Legault, Long Yang, Junwei Zhu, Keqing Shao, and Yiming Yang. 2015. “Differences in grammatical processing strategies for active and passive sentences: An fMRI study.” Journal of Neurolinguistics 33: 104–117. https://doi.org/10.1016/j.jneuroling.2014.09.002 Fernandez-Duque, Diego, Jessica Evans, Colton Christian, and Sarah D. Hodges. 2015. “Superfluous neuroscience information makes explanations of psychological phenomena more appealing.” Journal of Cognitive Neuroscience 27 (5): 926–944. https://doi.org/10.1162/jocn_a_00750 Ferré, Pilar, Rosa Sánchez-Casas, and Marc Guasch. 2006. “Can a horse be a donkey? Semantic and form interference effects in translation recognition in early and late proficient and nonproficient Spanish-Catalan bilinguals.” Language Learning 56 (4): 571–608. https://doi.org/10.1111/j.1467-9922.2006.00389.x Ferreira, Aline. 2012. “Tempo e segmentação no processamento cognitivo de tradutores experientes em instancias de tradução direta e inversa.” Revista L@el em (Dis-) curso 5 (2): 57–69. Ferreira, Aline. 2014. “Analyzing recursiveness patterns and retrospective protocols of professional translators in L1 and L2 translation tasks.” Translation and Interpreting Studies 9 (1): 109–127.  https://doi.org/10.1075/tis.9.1.06fer

References 239

Ferreira, Aline, and John W. Schwieter. 2017. “Directionality in translation.” In The Handbook of Translation and Cognition, ed. by Aline Ferreira and John W. Schwieter, 90–105. Hoboken, New Jersey: John Wiley & Sons, Inc.  https://doi.org/10.1002/9781119241485.ch5 Ferreira, Aline, and John W. Schwieter. 2015. Psycholinguistic and Cognitive Inquiries into Trans­ lation and Interpreting. Amsterdam: John Benjamins.  https://doi.org/10.1075/btl.115 Finger, Stanley. 1994. Origins of Neuroscience: A History of Explorations into Brain Function. New York: Oxford University Press. Ford, Anastasia, William Triplett, Atchar Sudhyadhom, Joseph M Gullett, Keith McGregor, David FitzGerald, Thomas Mareci, Keith White, and Bruce Crosson. 2013. “Broca’s area and its striatal and thalamic connections: a diffusion-MRI tractography study.” Frontiers in Neuroanatomy 7: 8.  https://doi.org/10.3389/fnana.2013.00008 Fox, Michael D., and Michael Greicius. 2010. “Clinical applications of resting state functional connectivity.” Frontiers in Systems Neuroscience 4: 19.  https://doi.org/10.3389/fnsys.2010.00019. Frauenfelder, Ulrich, and Herbert Schriefers. 1997. “A psycholinguistic perspective on simultaneous interpretation.” Interpreting 2 (1–2): 55–89. https://doi.org/10.1075/intp.2.1-2.03fra Freeman, Walter J., and Rodrigo Quian Quiroga. 2012. Imaging Brain Function with EEG: Advanced Temporal and Spatial Analysis of Electroencephalographic Signals. Berlin: Springer. Friederici, Angela D. 2004. “Event-related brain potential studies in language.” Current Neurology and Neuroscience Reports 4 (6): 466–470.  https://doi.org/10.1007/s11910-004-0070-0 Friston, Karl J. 2011. “Functional and effective connectivity: a review.” Brain Connectivity 1 (1): 13–36.  https://doi.org/10.1089/brain.2011.0008 García-Caballero, Alejandro, Isabel García-Lado, Javier González-Hermida, Ramón Area, María José Recimil, Onésimo Juncos Rabadán, Susana Lamas, Guillermo Ozaita y F. J. Jorge (2007). “Paradoxical recovery in a bilingual patient with aphasia after right capsuloputaminal infarction”. Journal of Neurosurgical Psychiatry 78, 89–91. https://doi.org/10.1136/jnnp.2006.095406 García-Cordero, Indira, Lucas Sedeño, Laura de la Fuente, Andrea Slachevsky, Gonzalo Forno, Francisco Klein, Patricia Lillo, Jesica Ferrari, Clara Rodriguez, Julian Bustin, Teresa Torralva, Sandra Baez, Adrian Yoris, Sol Esteves, Margherita Melloni, Paula Salamone, David Huepe, Facundo Manes, Adolfo M. García, and Agustín Ibañez. 2016. “Feeling, learning from and being aware of inner states: interoceptive dimensions in neurodegeneration and stroke.” Philosophical Transactions of the Royal Society B: Biological Sciences 371: 1708. https://doi.org/10.1098/rstb.2016.0006. García-Pentón, Lorna, Yuriem Fernández García, Brendan Costello, Jon Andoni Duñabeitia, and Manuel Carreiras. 2016. “The neuroanatomy of bilingualism: how to turn a hazy view into the full picture.” Language, Cognition and Neuroscience 31 (3): 303–327. https://doi.org/10.1080/23273798.2015.1068944 García, Adolfo M. 2008. “The circumscribed infinites scheme: A deconstructive approach to trans­lating poetry.” Target 20 (1): 115–134.  https://doi.org/10.1075/target.20.1.07mar García, Adolfo M. 2009. “Semic verbalisation: A systematic procedure for the replication of meaning in translating lexical items.” Linguistica Antverpiensia 7: 75–92. García, Adolfo M. 2010. “Methodological tenets, plausibility and reality in chomskyan biolinguistics.” Linguistics and the Human Sciences 3 (3): 303–324.  https://doi.org/10.1558/lhs.v3i3.303 García, Adolfo M. 2011. “Proceso traductor y equivalencia: cotejo de dos modelos trifásicos e implicaciones para la didáctica de la traductología.” Revista Electrónica de Didáctica de la Interpretación y la Traducción 7: 17–41.

240 The Neurocognition of Translation and Interpreting

García, Adolfo M. 2012a. Traductología y neurocognición: Cómo se organiza el sistema lingüístico del traductor. Córdoba: Facultad de Lenguas de la UNC. García, Adolfo M. 2012b. Aproximaciones teóricas y empíricas a la lingüística cognitiva. Mar del Plata: Editorial Martín. García, Adolfo M. 2012c. “La Teoría de Redes Relacionales: Correlatos neurológicos de un modelo lingüístico conexionista.” Onomázein 26: 221–257. García, Adolfo M. 2013a. “Brain activity during translation: A review of the neuroimaging evidence as a testing ground for clinically-based hypotheses.” Journal of Neurolinguistics 26 (3): 370–383.  https://doi.org/10.1016/j.jneuroling.2012.12.002 García, Adolfo M. 2013b. “Relational network theory as a bridge between linguistics and neuroscience: An interview with Professor Sydney Lamb.” Linguistics and the Human Sciences 8 (1): 3–27.  https://doi.org/10.1558/lhs.v8i1.3 García, Adolfo M. 2014a. “The interpreter advantage hypothesis: Preliminary data patterns and empirically motivated questions.” Translation and Interpreting Studies 9 (2): 219–238. https://doi.org/10.1075/tis.9.2.04gar García, Adolfo M. 2014b. “Neurocognitive determinants of performance variability among world-language users.” Journal of World Languages 1 (1): 60–77. https://doi.org/10.1080/21698252.2014.893671 García, Adolfo M. 2015a. “Psycholinguistic explorations of lexical translation equivalents: Thirty years of research and their implications for cognitive translatology.” Translation Spaces 4 (1): 9–28.  https://doi.org/10.1075/ts.4.1.01gar García, Adolfo. 2015b. “Translating with an injured brain: neurolinguistic aspects of translation as revealed by bilinguals with cerebral lesions.” Meta: Translators’ Journal 60 (1): 112–134. https://doi.org/10.7202/1032402ar García, Adolfo M. 2015c. “A connectionist approach to functional-cognitive linguistics: Spanish pronominal clitics and verb endings in relational-network terms.” Signos: Estudios de Lingüística 48 (88): 197–222.  https://doi.org/10.4067/S0718-09342015000200003 García, Adolfo M. 2016. “El sistema léxico bilingüe: organización y procesamiento.” In Mente bilingüe: abordajes psicolingüísticos y cognitivistas, ed. by Adolfo M. García and Sonia Suárez Cepeda, 69–98. Córdoba: Comunicarte. García, Adolfo M., and María Inés Arrizabalaga. 2013. La traducción bajo la línea de la convergencia. Córdoba: Facultad de Lenguas de la UNC. García, Adolfo M., and Agustín Ibáñez. 2016a. “Processes and verbs of doing, in the brain: Theoretical implications for systemic functional linguistics.” Functions of Language 23 (3): 305–335.  https://doi.org/10.1075/fol.23.3.02gar García, Adolfo M., and Agustín Ibáñez. 2016b. “Hands typing what hands do: Action-semantic integration dynamics throughout written verb production.” Cognition 149: 56–66. https://doi.org/10.1016/j.cognition.2016.01.011 García, Adolfo M., and Agustín Ibáñez. 2016c. “A touch with words: Dynamic synergies between manual actions and language.” Neuroscience & Biobehavioral Reviews 68: 59–95. https://doi.org/10.1016/j.neubiorev.2016.04.022 García, Adolfo M., and Sonia Suárez Cepeda. 2016. Mente bilingüe: Abordajes psicolingüísitcos y cognitivistas. Córdoba: Comunicarte. García, Adolfo M., and Edinson Muñoz. Forthcoming. “Translation, neuroscience, and cognition.” In The Routledge Handbook of Translation and Cognition, ed. by Arnt L. Jakobsen and Fabio Alves. London: Routledge.

References 241

García, Adolfo M., Ezequiel Mikulan, and Agustín Ibáñez. 2016. “A neuroscientific toolkit for translation studies.” In Reembedding Translation Process Research ed. by Ricardo Muñoz Martín, 21–46. Amsterdam: John Benjamins.  https://doi.org/10.1075/btl.128.02gar García, Adolfo M., Laura Manoiloff, and Mónica Wagner. 2016. “Concepciones del bilingüismo y evaluación de la competencia bilingüe.” In Mente bilingüe: abordajes psicolingüísticos y cognitivistas, ed. by Adolfo M. García and Sonia Suárez Cepeda, 17–49. Córdoba: Comunicarte. García, Adolfo M., William Sullivan, and Sarah Tsiang. 2017. An Introduction to Relational Network Theory: History, Principles, and Descriptive Applications. London: Equinox. García, Adolfo M., Agustin Ibáñez, David Huepe, Alexander Houck, Maëva Michon, Carlos Gelormini Lezama, Sumeer Chadha, and Álvaro Rivera-Rei. 2014. “Word reading and translation in bilinguals: The impact of formal and informal translation expertise.” Frontiers in Psychology 5: 1302.  https://doi.org/10.3389/fpsyg.2014.01302 García, Adolfo M., Facundo Carrillo, Juan Rafael Orozco-Arroyave, Natalia Trujillo, Juan Felipe Vargas Bonilla, Sol Fittipaldi, Federico Adolfi, Elmer Noth, Mariano Sigman, Diego Fernandez Slezak, Agustín Ibáñez, and Guillermo A. Cecchi. 2016a. “How language flows when movements don’t: An automated analysis of spontaneous discourse in Parkinson’s disease.” Brain and Language 162: 19–28.  https://doi.org/10.1016/j.bandl.2016.07.008 García, Adolfo M, Sofía Abrevaya, Giselle Kozono, Indira García Cordero, Marta Córdoba, Marcelo Andrés Kauffman, Ricardo Pautassi, Edinson Muñoz, Lucas Sedeño, and Agustín Ibáñez. 2016b. “The cerebellum and embodied semantics: evidence from a case of genetic ataxia due to STUB1 mutations.” Journal of Medical Genetics 54: 114–124. https://doi.org/10.1136/jmedgenet-2016-104148 García, Adolfo M., Lucas Sedeño, Natalia Trujillo, Yamile Bocanegra, Diana Gomez, David Pineda, Villegas Andrés, William Arias, and Agustín Ibáñez. 2017a. “Language deficits as a preclinical window into Parkinson’s disease: Evidence from asymptomatic parkin and dardarin mutation carriers.” Journal of the International Neuropsychological Society 22: 150–158. https://doi.org/10.1017/S1355617716000710 García, Adolfo M., Yamile Bocanegra, Eduar Herrera, Mariana Pino, Edinson Muñoz, Lucas Sedeño, and Agustín Ibáñez. 2017b. “Action-semantic and syntactic deficits in subjects at risk for Huntington’s disease.” Journal of Neuropsychology.  https://doi.org/10.1111/jnp.12120 García, Adolfo M., Lucas Sedeño, Eduar Herrera, Blas Couto, and Agustín Ibáñez. 2017c. “A lesion-proof brain? Multidimensional sensorimotor, cognitive, and socio-affective preservation despite extensive damage in a stroke patient.” Frontiers in Aging Neuroscience 8: 335. https://doi.org/10.3389/fnagi.2016.00335 García, Adolfo M., Yamile Bocanegra, Elena Herrera, Leonardo Moreno, Jairo Carmona, Ana Baena, Francisco Lopera, David Pineda, Margherita Melloni, Agustina Legaz, Edinson Muñoz, Lucas Sedeño, Sandra Baez, and Agustín Ibáñez. 2018. “Parkinson’s disease compromises the appraisal of action meanings evoked by naturalistic texts.” Cortex 100: 111–126. https://doi.org/10.1016/j.cortex.2017.07.003 Gastaldi, Guido. 1951. “Osservazioni su un afasico bilingüe.” Sistema Nervoso 2: 175–180. Gentzler, Edwin. 2001. Contemporary Translation Theories. Clevedon, UK: Multilingual Matters. Gerver, David. 1969. “The effects of source language presentation rate on the performance of simultaneous interpreters.” In Proceedings of the 2nd Louisville Conference on Rate and/or Frequency of Controlled Speech, ed. by Emerson Foulke, 162–184. Louisville: University of Louisville. Gerver, David. 1975. “A psychological approach to simultaneous interpretation.” Meta: Inter­ national Translator’s Journal 20 (2): 119–128.  https://doi.org/10.7202/002885ar

242 The Neurocognition of Translation and Interpreting

Gerver, David. 1976. “Empirical studies of simultaneous interpretation: A review and a model.” In Translation: Applications and Research, ed. by Richard W. Briskin, 165–207. New York: Gardner Press. Gile, Daniel. 1990. “Scientific research vs. personal theories in the investigation of interpretation.” In Aspects of Applied and Experimental Research on Conference Interpretation, ed. by Laura Gran and Christopher Taylor, 28–41. Udine: Campanotto. Gile, Daniel. 1995. Basic Concepts and Models for Interpreter and Translator Training. Amster­ dam: John Benjamins.  https://doi.org/10.1075/btl.8(1st) Gile, Daniel. 2005. “Directionality in conference interpreting: A cognitive view.” In Directionality in Interpreting. The ‘Retour’ or the Native?, ed. by Rita Godijns and Michaël Hindedael, 9–26. Ghent: Communication and Cognition. Gile, Daniel. 2009. “Interpreting studies: A critical view from within.” MonTI: Monografías de Tra­ducción e Interpretación 1: 135–155.  https://doi.org/10.6035/MonTI.2009.1.6 Golestani, Narly, Cathy J. Price, and Sophie K. Scott. 2011. “Born with an ear for dialects? Structural plasticity in the expert phonetician brain.” Journal of Neuroscience 31 (11): 4213–4220. https://doi.org/10.1523/JNEUROSCI.3891-10.2011 Göpferich, Sussane, Arnt Jakobsen, and Inger Mees. 2008. Looking at eyes: Eye-Tracking Studies of Reading and Translation Processing. Copenhagen: Samfundslitteratur. Grabner, Roland H., Clemens Brunner, Robert Leeb, Christa Neuper, and Gert Pfurtscheller. 2007. “Event-related EEG theta and alpha band oscillatory responses during language translation.” Brain Research Bulletin 72 (1): 57–65.  https://doi.org/10.1016/j.brainresbull.2007.01.001 Gran, Laura. 1989. “Interdisciplinary research on cerebral asymmetries: Significance and prospects for the Teaching of Interpretation.” In The Theoretical and Practical Aspects of Teaching Conference Interpretation, ed. by Laura Gran and John Dodds, 93–100. Udine: Camapnotto. Grant, Angela M., Shin-Yi Fang, and Ping Li. 2015. “Second language lexical development and cognitive control: A longitudinal fMRI study.” Brain and Language 144: 35–47. https://doi.org/10.1016/j.bandl.2015.03.010 Green, C. Shawn, and Daphne Bavelier. 2003. “Action video game modifies visual selective attention.” Nature 423 (6939): 534–537.  https://doi.org/10.1038/nature01647 Green, David W. 1986. “Control, activation, and resource: a framework and a model for the control of speech in bilinguals.” Brain and Language 27 (2): 210–23. https://doi.org/10.1016/0093-934X(86)90016-7 Greicius, Michael. 2008. “Resting-state functional connectivity in neuropsychiatric disorders.” Current Opinion in Neurology 21 (4): 424–430. https://doi.org/10.1097/WCO.0b013e328306f2c5 Grodzinsky, Yosef. 2000. “The neurology of syntax: Language use without Broca’s area.” Behavioral and Brain Sciences 23 (1): 1–21.  https://doi.org/10.1017/S0140525X00002399 Grodzinsky, Yosef, and Angela D. Friederici. 2006. “Neuroimaging of syntax and syntactic processing.” Current Opinion in Neurobiology 16 (2): 240–246. https://doi.org/10.1016/j.conb.2006.03.007 Grosjean, François. 1994. “Individual bilingualism.” In The Encyclopaedia of Language and Lin­ guistics ed. by Ronald E. Asher, 1656–1660. Oxford: Pergamon Press. Guasch, Marc, Rosa Sánchez-Casas, Pilar Ferré, and José E. García-Albea. 2008. “Translation performance of beginning, intermediate and proficient Spanish-Catalan bilinguals: Effects of form and semantic relations.” The Mental Lexicon 3 (3): 289–308. https://doi.org/10.1075/ml.3.3.03gua. Gutt, Ernst A. 1991. Translation and Relevance: Cognition and Context. Cambridge: Blackwell.

References 243

Halliday, Michael A. K. 1985. An Introduction to Functional Grammar. London: Arnold. Halliday, Michael A. K. 2007 [2002]. “Applied linguistics as an evolving theme.” In Language and Education. Volume 9 of the Collected Works of M.A.K. Halliday, ed. by Jonathan Webster, 1–19. London & New York: Continuum. Han, Shihui, Georg Northoff, Kai Vogeley, Bruce E. Wexler, Shinobu Kitayama, and Michael E. W. Varnum. 2013. “A cultural neuroscience approach to the biosocial nature of the human brain.” Annual Review of Psychology 64: 335–59. https://doi.org/10.1146/annurev-psych-071112-054629 Hansen‐Schirra, Silvia (2017). “EEG and universal language processing in translation.” In The Handbook of Translation and Cognition, ed. by John W. Schwieter, and Aline Ferreira, 232– 247. Hoboken, New Jersey: John Wiley & Sons, Inc. https://doi.org/10.1002/9781119241485.ch13 Hansen-Schirra, Silvia, Stella Neumann, and Erich Steiner. 2007. “Cohesive explicitness and explicitation in an English-German translation corpus.” Languages in Contrast 7 (2): 241–265. https://doi.org/10.1075/lic.7.2.09han Harlow, John M. 1868. “Recovery from the passage of an iron bar through the head.” Bulletin of the Massachusetts Medical Society (2): 327–347. Harris, Brian, and Bianca Sherwood. 1978. “Translating as an innate skill.” In Language Inter­ pretation and Communication. Proceedings from the NATO Symposium, Venice, Italy, September 26-October 1, 1977, ed. by David Gerver and H. Wallace Sinaiko, 155–170. New York & London: Plenum Press. Hartelius, Lena, Anna Carlstedt, Monica Ytterberg, Malin Lillvik, and Katja Laakso. 2003. “Speech disorders in mild and moderate Huntington disease: Results of dysarthria assessments of 19 individuals.” Journal of Medical Speech-Language Pathology 11 (1): 1–15. Hautzel, Hubertus, Felix M. Mottaghy, Daniela Schmidt, M Zemb, NJ Shah, Hans-Willem Müller-Gärtner, and Bernd J. Krause. 2002. “Topographic segregation and convergence of verbal, object, shape and spatial working memory in humans.” Neuroscience Letters 323 (2): 156–160.  https://doi.org/10.1016/S0304-3940(02)00125-8 He, Yan, Meng-Yun Wang, Defeng Li, and Zhen Yuan. 2017. “Optical mapping of brain activation during the English to Chinese and Chinese to English sight translation.” Biomedical Optics Express 8 (12): 5399–5411.  https://doi.org/10.1364/BOE.8.005399 Hebb, Donald. 1949. The Organization of Behavior: A Neurophysiological Theory. New York: Wiley. Heij, Wido La, Andre Hooglander, Robert Kerling, and Esther van der Velden. 1996. “Nonverbal context effects in forward and backward word translation: Evidence for concept mediation.” Journal of Memory and Language 35 (5): 648–665.  https://doi.org/10.1006/jmla.1996.0034 Henrard, Sébastien, and Agnès Van Daele. 2017. “Different bilingual experiences might modulate executive tasks advantages: Comparative analysis between monolinguals, translators, and interpreters.” Frontiers in Psychology 8: 1870.  https://doi.org/10.3389/fpsyg.2017.01870 Herholz, Karl, Peter Herscovitch, and Ute Heilmann W. D. Heiss. 2004. NeuroPET: Positron Emission Tomography in Neuroscience and Clinical Neurology. Berlin: Springer-Verlag. https://doi.org/10.1007/978-3-642-18766-7 Hernandez, Arturo E. 2013. The Bilingual Brain. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199828111.001.0001 Hervais-Adelman, Alexis, Barbara Moser-Mercer, and Narly Golestani. 2011. “Executive Control of Language in the Bilingual Brain: Integrating the Evidence from Neuroimaging to Neuropsychology.” Frontiers in Psychology 2: 234.  https://doi.org/10.3389/fpsyg.2011.00234

244 The Neurocognition of Translation and Interpreting

Hervais-Adelman, Alexis, Barbara Moser-Mercer, Christoph M. Michel, and Narly Golestani. 2014. “fMRI of simultaneous interpretation reveals the neural basis of extreme language control.” Cerebral Cortex 25 (12): 4727–4739  https://doi.org/10.1093/cercor/bhu158 Hervais-Adelman, Alexis, Barbara Moser-Mercer, and Narly Golestani. 2015. “Brain functional plasticity associated with the emergence of expertise in extreme language control.” NeuroImage 114: 264–274.  https://doi.org/10.1016/j.neuroimage.2015.03.072 Hervais-Adelman, Alexis, Barbara Moser-Mercer, Micah M Murray, and Narly Golestani. 2017. “Cortical thickness increases after simultaneous interpretation training.” Neuropsychologia 98: 212–219.  https://doi.org/10.1016/j.neuropsychologia.2017.01.008 Hervais-Adelman, Alexis, Barbara Moser-Mercer, and Narly Golestani. 2018. “Commentary: Broca pars triangularis constitutes a “Hub” of the language-control network during simultaneous language translation.” Frontiers in Human Neuroscience 12: 22. https://doi.org/10.3389/fnhum.2018.00022. Hickok, Gregory, and David Poeppel. 2007. “The cortical organization of speech processing.” Nature Reviews Neuroscience 8 (5): 393–402.  https://doi.org/10.1038/nrn2113 Hickok, Gregory, and Steven Small. 2015. Neurobiology of Language. San Diego: Academic Press. Hiltunen, Sinikka, Rauni Pääkkönen, Gun-Viol Vik, and Christina M. Krause. 2016. “On interpreters’ working memory and executive control.” International Journal of Bilingualism 20 (3): 297–314.  https://doi.org/10.1177/1367006914554406 Hiltunen, Sinikka, and Gun-Viol Vik. 2017. “Interpreters – experts in careful listening and efficient encoding? Findings of a prose recall test.” International Journal of Bilingualism 21 (2): 194–212.  https://doi.org/10.1177/1367006915610657. Ho, Aileen K., Robert Iansek, and John L. Bradshaw. 1999. “Regulation of Parkinsonian speech volume: The effect of interlocuter distance.” Journal of Neurology, Neurosurgery & Psychiatry 67 (2):199–202.  https://doi.org/10.1136/jnnp.67.2.199 Holmes, James. 2000 [1972]. “The name and nature of translation studies.” In The Translation Studies Reader, ed. by Lawrence Venuti, 180–192. London: Routledge. Holmes, James S. 1988. Translated! Papers on Literary Translation and Translation Studies. Ams­ terdam: Rodopi. Holtgraves, Thomas, and Patrick McNamara. 2010. “Pragmatic comprehension deficit in Parkinson’s disease.” Journal of Clinical and Experimental Neuropsychology 32 (4): 388–397. https://doi.org/10.1080/13803390903130729 Hönig, Hans. 1991. “Holmes’ ‘Mapping Theory’ and the landscape of mental translation processes.” In Translation Studies, The State of the Art. Proceedings of the First James S Holmes Symposium on Translation Studies, ed. by Kitty Van Leuven-Zwart and Ton Naaijkens, 91– 101. Amsterdam: Rodopi. Hönig, Hans. 1992. “Von der erzwungenen Selbstentfrendung des Übersetzers– Ein offener Brief an Justa Holz-Mänttäri.” TexTconTextl 7: 1–14. House, Juliane. 1988. “Talking to oneself or thinking with others.” Fremdsprachen Lehren und Lernen 17: 84–98. Huang, Harry, and Canzhong Wu. 2009. “The unit of translation: Statistics speak.” Meta 54 (1): 110–130.  https://doi.org/10.7202/029796ar Hubel, David H., and Torsten N. Wiesel. 1962. “Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex.” The Journal of Physiology 160 (1): 106–154. https://doi.org/10.1113/jphysiol.1962.sp006837

References 245

Hubel, David H., and Torsten N. Wiesel. 1977. “Ferrier lecture. Functional architecture of macaque monkey visual cortex.” Proceedings of the Royal Society of London B: Biological Science 198 (1130): 1–59.  https://doi.org/10.1098/rspb.1977.0085 Huettel, Scott A., Allen W. Song, and Gregory McCarthy. 2008. Functional Magnetic Resonance Imaging. second ed. Sunderland, Massachussetts: Sinauer Associates. Humphries, Colin, Tracy Love, David Swinney, and Gregory Hickok. 2005. “Response of anterior temporal cortex to syntactic and prosodic manipulations during sentence processing.” Human Brain Mapping 26 (2): 128–138.  https://doi.org/10.1002/hbm.20148 Hurtado Albir, Amparo. 1990. La notion de fidélité en traduction, Col. Traductologie Paris: Didier. Hurtado Albir, Amparo. 2001. Traducción y Traductología: Introducción a la Traductología. Madrid: Cátedra. Hvelplund, Kristian. 2011. Allocation of Cognitive Resources in Translation: An Eye-tracking and Key-logging Study. Ph.D. dissertation Copenhagen: Samfundslitteratur. Ibáñez, Agustín, Adolfo M. García, Sol Esteves, Adrián Yoris, Edinson Muñoz, Lucía Reynaldo, Marcos L. Pietto, Federico Adolfi, and Facundo Manes. 2018. “Social neuroscience: Un­ doing the schism between neurology and psychiatry.” Social Neuroscience 13 (1): 1–39. https://doi.org/10.1080/17470919.2016.1245214 Ibáñez, Agustín, and Adolfo M. García. 2018. Contextual Cognition: The Sensus Communis of a Situated Mind. Heidelberg: Springer.  https://doi.org/10.1007/978-3-319-77285-1 Ibáñez, Agustín, Lucas Sedeño, and Adolfo M. García. 2017. Neuroscience and Social Science: The Missing Link. New York: Springer.  https://doi.org/10.1007/978-3-319-68421-5 Ibáñez, Antonio J., Pedro Macizo, and María Teresa Bajo. 2010. “Language access and language selection in professional translators.” Acta Psychologica 135 (2): 257–266. https://doi.org/10.1016/j.actpsy.2010.07.009 Indefrey, Peter, and Willem J. M. Levelt. 2004. “The spatial and temporal signatures of word production components.” Cognition 92 (1): 101–144. https://doi.org/10.1016/j.cognition.2002.06.001 Injoque-Ricle, Irene, Juan Pablo Barreyro, Jesica Formoso, and Virginia I. Jaichenco. 2015. “Expertise, working memory and articulatory suppression effect: Their relation with simultaneous interpreting performance.” Advances in Cognitive Psychology 11 (2): 56–63. https://doi.org/10.5709/acp-0171-1 Jackson, John. 1878. “On affections of speech from disease of the brain.” Brain 1: 304–330. https://doi.org/10.1093/brain/1.3.304 Jacobs, Joshua, and Michael J. Kahana. 2010. “Direct brain recordings fuel advances in cognitive electrophysiology.” Trends in Cognitive Science 14 (4): 162–71. https://doi.org/10.1016/j.tics.2010.01.005 Jakobsen, Arnt. 1995. “Logging target text production with Translog.” In Probing the Process in Translation. Methods and Results (Copenhagen Studies in Language 24), ed. by Gyde Hansen, 9–20. Copenhagen: Samfundslitteratur. Jakobsen, Arnt. 1998. “Logging time delay in translation.” In LSP Texts and the Process of Trans­lation (Copenhagen Working Papers 1), ed. by Gyde Hansen, 71–101. Copenhagen: Copenhagen Business School. Jakobsen, Arnt. 2014. “The development and current state of translation process research.” In The Known Unknowns of Translation Studies ed. by Elke Brems, Reine Meylaerts and Luc van Doorslaer, 65–88. Amsterdam: John Benjamins.

246 The Neurocognition of Translation and Interpreting

Jakobsen, Arnt L. 2003. “Effects of think aloud on translation speed, revision, and segmentation.” In Triangulating Translation: Perspectives in Process Oriented Research, ed. by Fabio Alves, 69–95. Amsterdam & Philadelphia: John Benjamins.  https://doi.org/10.1075/btl.45.08jak Jakobson, Roman. 1964. “Discussion.” In Disorders of Language, ed. by A. V. S. Dereuck and Maeve O’Conner, 120. Boston: Little, Brown. Jakobson, Roman. 2000 [1959]. “On linguistic aspects of translation.” In The Translation Studies Reader, ed. by Lawrence Venuti, 113–118. London & New York: Routledge. Janyan, Armina, Ivo Popivanov, and Elena Andonova. 2009. “Concreteness effect and word cognate status: ERPs in single word translation.” In Brain Talk: Discourse with and in the Brain, ed. by Kai Alter, Merle Horne, Magnus Lindgren, Mikael Roll and Jane von Koss Torkildsen, 21–30. Lunds: Lunds Universitet. Jasinska, K. K., and L. A. Petitto. 2013. “How age of bilingual exposure can change the neural systems for language in the developing brain: A functional near infrared spectroscopy investigation of syntactic processing in monolingual and bilingual children.” Developmental Cognitive Neuroscience 6 (Supplement C): 87–101.  https://doi.org/10.1016/j.dcn.2013.06.005 Jasper, Herbert, and Wilder Penfield. 1949. “Electrocorticograms in man: Effect of voluntary movement upon the electrical activity of the precentral gyrus.” European Archives of Psy­ chiatry and Clinical Neuroscience 183 (1): 163–174. Jefferies, Elizabeth. 2013. “The neural basis of semantic cognition: Converging evidence from neuropsychology, neuroimaging and TMS.” Cortex 49 (3): 611–625. https://doi.org/10.1016/j.cortex.2012.10.008. Jensen, Astrid, and Arnt Jakobsen. 2000. “Translating under time pressure: an empirical investigation of problem-solving activity and translation strategies by nonprofessional and professional translators.” In Translation in Context. Selected contributions from the EST Congress, Granada 1998, ed. by Andrew Chesterman, Natividad Gallardo San Salvador and Yves Gambier, 105–116. Amsterdam: John Benjamins. Jensen, Cristian. 2008. “Assessing eye-tracking accuracy in translation studies.” Copenhagen Studies in Language 36: 157–174. Joanette, Yves, Ana I Ansaldo, Karima Kahlaoui, Hélène Côté, Valeria Abusamra, Aldo Ferreres, and André Roch-Lecours. 2008. “Impacto de las lesiones del hemisferio derecho sobre las habilidades lingüísticas: perspectivas teórica y clínica.” Revista de Neurología 46 (8): 481–488. Josse, Goulven, and Natalie Tzourio-Mazoyer. 2004. “Hemispheric specialization for language.” Brain Research Reviews 44 (1): 1–12.  https://doi.org/10.1016/j.brainresrev.2003.10.001 Jost, Lea B., Narges Radman, Karin A. Buetler, and Jean-Marie Annoni. 2018. “Behavioral and electrophysiological signatures of word translation processes.” Neuropsychologia 109: 245–254.  https://doi.org/10.1016/j.neuropsychologia.2017.12.034. Just, Marcel A., and Patricia A. Carpenter. 1980. “A theory of reading: from eye fixations to comprehension.” Psychological Review 87 (4): 329–354.  https://doi.org/10.1037/0033-295X.87.4.329 Kaan, Edith, Anthony Harris, Edward Gibson, and Phillip Holcomb. 2000. “The P600 as an index of syntactic integration difficulty.” Language and Cognitive Processes 15 (2): 159–201. https://doi.org/10.1080/016909600386084 Kaindl, Klaus. 2004. Übersetzungswissenschaft im interdisziplinären Dialog. Am Beispiel der Comicübersetzung. Tübingen: Stauffenburg. Kalina, Sylvia. 1998. Strategische Prozesse beim Dolmetschen: Theoretische Grundlagen, empirische Untersuchungen, didaktische Konsequenzen. Tübingen: Gunter Narr.

References 247

Kandel, Eric. 2006. In Search of Memory: The Emergence of a New Science of Mind. New York & London: Norton. Kandel, Eric. 2013. Principles of Neural Science. fifth ed. New York: McGraw-Hill. Kaplan, Joan A, Hiram H Brownell, Janet R Jacobs, and Howard Gardner. 1990. “The effects of right hemisphere damage on the pragmatic interpretation of conversational remarks.” Brain and Language 38 (2): 315–333.  https://doi.org/10.1016/0093-934X(90)90117-Y Kauders, Otto. 1929. “Über polyglotte Reaktionen bei einer sensorischen Aphasie.” Zeitschrift für die gesamte Neurologie und Psychiatrie 149: 291–301. Kennepohl, Stephan, Viviane Sziklas, Krista Garver, David Wagner, and Marilyn Jones-Gotman. 2007. “Memory and the medial temporal lobe: hemispheric specialization reconsidered.” Neuroimage 36 (3): 969–78.  https://doi.org/10.1016/j.neuroimage.2007.03.049 Kielar, Aneta, Jed A. Meltzer, Sylvain Moreno, Claude Alain, and Ellen Bialystok. 2014. “Oscillatory responses to semantic and syntactic violations.” Journal of Cognitive Neuro­science 26 (12): 2840–2862.  https://doi.org/10.1162/jocn_a_00670 King, Jean-Rémi King, Jacobo D. Sitt, Frédéric Faugeras, Benjamin Rohaut, Imen El Karoui, Laurent Cohen, Lionel Naccache, and Stanislas Dehaene. 2013. “Information sharing in the brain indexes consciousness in noncommunicative patients.” Current Biology 23 (19): 1914–1919.  https://doi.org/10.1016/j.cub.2013.07.075 Kintsch, Walter. 1998. Comprehension: A Paradigm for Cognition. New York: Cambridge University Press. Kintsch, Walter, and Praful Mangalath. 2011. “The construction of meaning.” Topics in Cognitive Science 3: 346–370.  https://doi.org/10.1111/j.1756-8765.2010.01107.x Kiraly, Donald C. 1995. Pathways to Translation. Pedagogy and Process. Kent, Ohio: The Kent State University Press. Kiran, Swathi, and Keith R. Lebel. 2007. “Crosslinguistic semantic and translation priming in normal bilingual individuals and bilingual aphasia.” Clinical Linguistics & Phonetics 21 (4): 277–303.  https://doi.org/10.1080/02699200701243873 Klein, Denose, Robert J. Zatorre, Jen-Kai Chen, Brenda Milner, Joelle Crane, Pascal Belin, and Marc Bouffard. 2006. “Bilingual brain organization: a functional magnetic resonance adaptation study.” Neuroimage 31 (1): 366–375.  https://doi.org/10.1016/j.neuroimage.2005.12.012 Klein, Denise, Brenda Milner, Robert J. Zatorre, Ernst Meyer, and Alan C Evans. 1995. “The neural substrates underlying word generation: a bilingual functional-imaging study.” Proceedings of the National Academy of Sciences 92 (7): 2899–2903.  https://doi.org/10.1073/pnas.92.7.2899 Kliegl, Reinhold, Michael Dambacher, Olaf Dimigen, Arthur M. Jacobs, and Werner Sommer. 2012. “Eye movements and brain electric potentials during reading.” Psychological Research 76 (2): 145–58.  https://doi.org/10.1007/s00426-011-0376-x Köpke, Barbara, and Jean-Luc Nespoulous. 2006. “Working memory performance in expert and novice interpreters.” Interpreting 8 (1): 1–23.  https://doi.org/10.1075/intp.8.1.02kop. Koshkin, Roman, and Alex Ossadtchi. 2017. “Commentary: Functional connectivity in the left dorsal stream facilitates simultaneous language translation: An EEG study.” Frontiers in Human Neuroscience 11 (64).  https://doi.org/10.3389/fnhum.2017.00064 Krakauer, John W., Asif A. Ghazanfar, Alex Gomez-Marin, Malcolm A. MacIver, and David Poeppel. 2017. “Neuroscience needs behavior: Correcting a reductionist bias.” Neuron 93 (3): 480–490.  https://doi.org/10.1016/j.neuron.2016.12.041 Kroll, Judith F., and Erika Stewart. 1990. “Concept mediation in bilingual translation.” Bulletin of the Psychonomic Society 28 (6): 510.

248 The Neurocognition of Translation and Interpreting

Kroll, Judith F., and Janet Curley. 1988. “Lexical memory in novice bilinguals: The role of concepts in retrieving second language words.” In Practical Aspects of Memory, ed. by Michael Gruneberg, Peter Morris and Robert Sykes, 389–395. London: John Wiley & Sons. Kroll, Judith F., and Erika Stewart. 1994. “Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations.” Journal of Memory and Language 33: 149–174.  https://doi.org/10.1006/jmla.1994.1008 Kroll, Judith F., Janet G. van Hell, Natasha Tokowicz, and David W. Green. 2010. “The Revised Hierarchical Model: A critical review and assessment.” Bilingualism: Language and Cognition 13: 373–381.  https://doi.org/10.1017/S136672891000009X Kruger, Justin, and David Dunning. 1999. “Unskilled and unaware of it: how difficulties in recognizing one’s own incompetence lead to inflated self-assessments.” Journal of Personaity and Social Psychology 77 (6): 1121–1134.  https://doi.org/10.1037/0022-3514.77.6.1121 Kuhn, Thomas. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Kurz, Ingrid. 1994. “A look into the ‘black box’ – EEG probability mapping during mental simultaneous interpreting.” In Translation Studies. An Interdiscipline, ed. by Mary Snell-Hornby, Franz Pöchhacker and Klaus Kaindl, 199–207. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.2.25kur Kurz, Ingrid. 1995. “Watching the brain at work – An exploratory study of EEG changes during simultaneous interpreting (SI).” The Interpreters’ Newsletter 6: 3–16. Kussmaul, Paul, and Sonja Tirkkonen-Condit. 1995. “Think-aloud protocol analysis in translation studies.” TTR : traduction, terminologie, rédaction 8 (1): 177–199. https://doi.org/10.7202/037201ar Kutas, Marta, and Kara D. Federmeier. 2011. “Thirty years and counting: finding meaning in the N400 component of the event-related brain potential (ERP).” Annual Review of Psychology 62: 621–47.  https://doi.org/10.1146/annurev.psych.093008.131123 Kwon, Youan, Yoonhyoung Lee, and Kichun Nam. 2011. “The different P200 effects of phonological and orthographic syllable frequency in visual word recognition in Korean.” Neuroscience Letters 501 (2): 117–121.  https://doi.org/10.1016/j.neulet.2011.06.060 Lachaux, Jean-Philippe, David Rudrauf, and Philippe Kahane. 2003. “Intracranial EEG and human brain mapping.” Journal of Physiology Paris 97 (4–6): 613–628. https://doi.org/10.1016/j.jphysparis.2004.01.018 Lamb, Sydney. 1999. Pathways of the Brain: The Neurocognitive Basis of Language. Amsterdam: John Benjamins.  https://doi.org/10.1075/cilt.170 Lebrun, Yves. 1991. “Polyglotte Reaktionen.” Neurolinguistik 5: 1–9. Lederer, Marianne. 1994. La traduction aujourd’hui. Le modèle interprétatif. Paris: Hachette. Lederer, Marianne. 2002 [1978]. “Simultaneous interpretation – Units of meaning and other features.” In The Interpreting Studies Reader ed. by Franz Pöchhacker and Miriam Shlesinger, 131–140. London: Routledge. LeDoux, Joseph. 2002. Synaptic Self: How Our Brains Become Who We Are. New York: Penguin. Lehtonen, Minna H., Matti Laine, Jussi Niemi, Tormod Thomsen, Victor A Vorobyev, and Kenneth Hugdahl. 2005. “Brain correlates of sentence translation in Finnish-Norwegian bilinguals.” Neuroreport 16 (6): 607–610.  https://doi.org/10.1097/00001756-200504250-00018 Li, Ping, Fan Zhang, Erlfang Tsai, and Brendan Puls. 2014. “Language history questionnaire (LHQ 2.0): A new dynamic web-based research tool.” Bilingualism: Language and Cognition 17 (3): 673–680.  https://doi.org/10.1017/S1366728913000606 Lichtheim, Ludwig. 1885. “On aphasia.” Brain 7: 433–484.  https://doi.org/10.1093/brain/7.4.433

References 249

Linck, Jared A., Peter Osthus, Joel T. Koeth, and Michael F. Bunting. 2014. “Working memory and second language comprehension and production: a meta-analysis.” Psychonomic Bul­ letin and Review 21 (4): 861–83.  https://doi.org/10.3758/s13423-013-0565-2 Linck, Jared A., and Daniel J. Weiss. 2015. “Can working memory and inhibitory control predict second language learning in the classroom?” SAGE Open 5 (4): 2158244015607352. https://doi.org/10.1177/2158244015607352 Lindell, Annukka K. 2006. “In your right mind: Right hemisphere contributions to language processing and production.” Neuropsychological Reviews 16 (3): 131–148. https://doi.org/10.1007/s11065-006-9011-9 Liuzzi, Gianpiero, Nils Freundlieb, Volker Ridder, Julia Hoppe, Kirstin Heise, Maximo Zimerman, Christian Dobel, Stefanie Enriquez-Geppert, Christian Gerloff, and Pienie Zwitserlood. 2010. “The involvement of the left motor cortex in learning of a novel action word lexicon.” Current Biology 20 (19): 1745–1751.  https://doi.org/10.1016/j.cub.2010.08.034 Logothetis, Nikos K. 2008. “What we can do and what we cannot do with fMRI.” Nature 453 (7197): 869–78.  https://doi.org/10.1038/nature06976 Lörscher, Wolfgang. 1991. Translation Performance, Translation Process, and Translation Strategies: A Psycholinguistic Investigation. Tübingen: Gunter Narr. Lörscher, Wolfgang. 1993. “Translation process analysis.” In Translation and Knowledge: Pro­ ceedings of the Fourth Scandinavian Symposium on Translation Theory, June 4–6, 1992, ed. by Yves Gambier and Jorma Tommola, 195–211. Turku: Centre for Translation and Interpreting. Lörscher, Wolfgang. 2005. “The translation process: Methods and problems of its investigation.” Meta: Translators’ Journal 50 (2):597–608.  https://doi.org/10.7202/011003ar Lu, Fengmei, and Zhen Yuan. 2018. “Explore the brain activity during translation and interpreting using functional near-infrared spectroscopy. In Researching Cognitive Processes of Translation, ed. by Defeng Li, Victoria Lai Cheng Lei, and Yuanjian He, 109–120. Singapore: Springer. Lucas, Timothy H., Guy M. McKhann, and George A. Ojemann. 2004. “Functional separation of languages in the bilingual brain: a comparison of electrical stimulation language mapping in 25 bilingual patients and 117 monolingual control patients.” Journal of Neurosurgery 101 (3): 449–457.  https://doi.org/10.3171/jns.2004.101.3.0449 Luck, Steven J., and Emily S. Kappenman. 2012. Oxford Handbook of Event-related Potential Components. Oxford, New York: Oxford University Press. Luk, Gigi, David W. Green, Jubin Abutalebi, and Cheryl Grady. 2011. “Cognitive control for language switching in bilinguals: A quantitative meta-analysis of functional neuroimaging studies.” Language and Cognitive Processes 27 (10): 1479–1488. https://doi.org/10.1080/01690965.2011.613209 Luzzatti, Claudio. 2008. “Acquired reading and writing disorders.” In Handbook of the Neuro­ science of Language, ed. by Brigitte Stemmer and Harry A. Whitaker, 209–218. London: Elsevier.  https://doi.org/10.1016/B978-0-08-045352-1.00020-3 Macnamara, Brooke N., and Andrew R. A. Conway. 2016. “Working memory capacity as a predictor of simultaneous language interpreting performance.” Journal of Applied Research in Memory and Cognition 5 (4): 434–444.  https://doi.org/10.1016/j.jarmac.2015.12.001. Maguire, Eleanor A., David G. Gadian, Ingrid S. Johnsrude, Catriona D. Good, John Ashburner, Richard S. J. Frackowiak, and Christopher D. Frith. 2000. “Navigation-related structural change in the hippocampi of taxi drivers.” Proceedings of the National Academy of Sciences 97 (8): 4398–4403.  https://doi.org/10.1073/pnas.070039597

250 The Neurocognition of Translation and Interpreting

Maguire, Eleanor A., Elizabeth R. Valentine, John M. Wilding, and Narinder Kapur. 2002. “Routes to remembering: The brains behind superior memory.” Nature Neuroscience 6: 90. https://doi.org/10.1038/nn988 Mahboob, Ahmar, and Naomi Knight (eds.). 2010. Appliable Linguistics. London: Continuum. Malakoff, Marguerite M. 1992. “Translation ability: A natural bilingual and metalinguistic skill.” In Cognitive Processing in Bilinguals, ed. by Richard Jackson Harris, 515–529. Amsterdam: North Holland.  https://doi.org/10.1016/S0166-4115(08)61514-9 Malblanc, Alfred. 1965. Stylistique comparée du français et de l’allemand: essai de représentation linguistique comparée et étude de traduction. Paris: Didier. Malmkjær, Kirsten. 1998. Unit of Translation. In Routledge Encyclopedia of Translation Studies, ed. by Mona Baker, 286–288. London and New York: Routledge. Mary Snell-Hornby, Franz Pöchhacker, and Klaus Kaindl. 1992. Translation Studies: An Inter­ discipline Selected papers from the Translation Studies Congress, Vienna, 1992. Ams­ terdam: John Benjamins. Mashal, Nira, Miriam Faust, Taima Hendler, and Mark Jung-Beeman. 2009. “An fMRI study of processing novel metaphoric sentences.” Laterality 14 (1): 30–54. https://doi.org/10.1080/13576500802049433 Maslow, Abraham H. 1966. The Psychology of Science: A Reconnaissance. New York: Joanna Cotler Books. Matthiessen, Christian. 2001. “The environments of translation.” In Exploring Translation and Multilingual Production: Beyond Context, ed. by Erich Steiner and Colin Yallop, 127–160. Berlin and New York: Mouton de Gruyter.  https://doi.org/10.1515/9783110866193.41 Matusevych, Yevgen, Afra Alishahi, and Ad Backus. 2015. “The impact of first and second language exposure on learning second language constructions.” Bilingualism: Language and Cognition 20 (1): 128–149.  https://doi.org/10.1017/S1366728915000607 Mayer, Katja M., Izzet B. Yildiz, Manuela Macedonia, and Katharina von Kriegstein. 2015. “Visual and motor cortices differentially support the translation of foreign language words.” Current Biology 25 (4): 530–535.  https://doi.org/10.1016/j.cub.2014.11.068 Mazoyer, Bernard, Emmanuel Mellet, Guy Perchey, Laure Zago, Fabrice Crivello, Gaël Jobard, Nicolas Delcroix, Mathieu Vigneau, Gaëlle Leroux, and Laurent Petit. 2016. “BIL&GIN: a neuroimaging, cognitive, behavioral, and genetic database for the study of human brain lateralization.” Neuroimage 124: 1225–1231.  https://doi.org/10.1016/j.neuroimage.2015.02.071 McElree, Brian, Gisela Jia, and Annie Litvak. 2000. “The time-course of conceptual processing in three bilingual populations.” Journal of Memory and Language 42: 229–254. https://doi.org/10.1006/jmla.1999.2677 Melloni, Margherita, Pablo Billeke, Sandra Baez, Eugenia Hesse, Laura de la Fuente, Gonzalo Forno, Agustina Birba, Indira García-Cordero, Cecilia Serrano, Angelo Plastino, Andrea Slachevsky, David Huepe, Mariano Sigman, Facundo Manes, Adolfo M. García, Lucas Sedeño, and Agustín Ibáñez. 2016. “Your perspective and my benefit: multiple lesion models of self-other integration strategies during social bargaining.” Brain 139(11): 3022–3040. https://doi.org/10.1093/brain/aww231 Menenti, Laura. 2006. “L2-L1 word association in bilinguals: Direct evidence.” Nijmegen: Cog­ nitive Neuroscience Journal 1 (1): 17–24. Miyake, Akira, Naomi P. Friedman, Michael J. Emerson, Alexander H. Witzki, Amy Howerter, and Tor D. Wager. 2000. “The unity and diversity of executive functions and their contributions to complex “Frontal lobe” tasks: a latent variable analysis.” Cognitive Psychology 41 (1): 49–100.  https://doi.org/10.1006/cogp.1999.0734

References 251

Miyake, Akira, and Naomi P. Friedman. 2012. “The nature and organization of individual differences in executive functions: Four general conclusions.” Current Directions in Psychological Science 21 (1): 8–14.  https://doi.org/10.1177/0963721411429458 Moldovan, Cornelia D., Josep Demestre, Pilar Ferré, and Rosa Sánchez-Casas. 2016. “The role of meaning and form similarity in translation recognition in highly proficient balanced bilinguals: A behavioral and ERP study.” Journal of Neurolinguistics 37: 1–11. https://doi.org/10.1016/j.jneuroling.2015.07.002 Monetta, Laura, and Marc D. Pell. 2007. “Effects of verbal working memory deficits on metaphor comprehension in patients with Parkinson’s disease.” Brain and Language 101 (1): 80–89. https://doi.org/10.1016/j.bandl.2006.06.007 Monti, Cristina, Claudio Bendazzoli, Annalisa Sandrelli, and Mariachiara Russo. 2005. “Studying directionality in simultaneous interpreting through an electronic corpus: EPIC (European Parliament Interpreting Corpus).” Meta 50 (4).  https://doi.org/10.7202/019850ar Moon, Jihye, and Nan Jiang. 2011. “Non-selective lexical access in different-script bilinguals.” Bilingualism: Language and Cognition 15 (1): 173–180. https://doi.org/10.1017/S1366728911000022 Morales, Julia, Francisca Padilla, Carlos J. Gómez-Ariza, and María Teresa Bajo. 2015. “Simultaneous interpretation selectively influences working memory and attentional networks.” Acta Psychologica 155: 82–91.  https://doi.org/10.1016/j.actpsy.2014.12.004 Moreno, Eva M., and Marta Kutas. 2005. “Processing semantic anomalies in two languages: an electrophysiological exploration in both languages of Spanish-English bilinguals.” Cognitive Brain Research 22 (2): 205–220.  https://doi.org/10.1016/j.cogbrainres.2004.08.010 Moreno, Eva M., Antoni Rodríguez-Fornells, and Matti Laine. 2008. “Event-related potentials (ERPs) in the study of bilingual language processing.” Journal of Neurolinguistics 21 (6): 477–508.  https://doi.org/10.1016/j.jneuroling.2008.01.003 Moser-Mercer, Barbara. 1994. “Paradigms gained or the art of productive disagreement.” In Bridging the Gap: Empirical Research in Simultaneous Interpretation, ed. by Sylvie Lambert and Barbara Moser-Mercer, 17–23. Amsterdam & Philadephia: John Benjamins. https://doi.org/10.1075/btl.3.03mos Moser-Mercer, Barbara. 2010. “The search for neuro-physiological correlates of expertise in interpreting.” In Translation and Cognition, ed. by Gregory Shreve and Erik Angelone, 263–288. Amsterdam & Philadelphia: John Benjamins.  https://doi.org/10.1075/ata.xv.15mos Mountcastle, Vernon. 1998. Perceptual Neuroscience: The Cerebral Cortex. Cambridge, MA: Harvard University Press. Mouraux, André, and Gian Domenico Iannetti. 2008. “Across-trial averaging of event-related EEG responses and beyond.” Magnetic Resonance Imaging 26 (7): 1041–1054. https://doi.org/10.1016/j.mri.2008.01.011 Moya, Virgilio. 2004. La selva de la traducción. Madrid: Cátedra. Munafò, Marcus R., Brian A. Nosek, Dorothy V. M. Bishop, Katherine S. Button, Christopher D. Chambers, Nathalie Percie du Sert, Uri Simonsohn, Eric-Jan Wagenmakers, Jennifer J. Ware, and John P. A. Ioannidis. 2017. “A manifesto for reproducible science.” Nature Human Behaviour 1: 0021.  https://doi.org/10.1038/s41562-016-0021 Munday, Jeremy. 2001. Introducing Translation Studies: Theories and Applications. London: Routledge. Munday, Jeremy. 2009. The Routledge Companion to Translation Studies (revised edition). London & New York: Routledge.

252 The Neurocognition of Translation and Interpreting

Muñoz, Edinson, Noelia Calvo, and Adolfo M. García. 2018. “Grounding translation and interpreting in the brain: What has been, can be, and must be done.” Perspectives: Studies in Trans­lation Theory and Practice.  https://doi.org/10.1080/0907676X.2018.1549575 Muñoz Martín, Ricardo. 2007. “Traductología cognitiva y traductología empírica.” In Quo vadis, Translatologie?, ed. by Gerd Wotjak, 267–278. Berlin: Franck & Timme. Muñoz Martín, Ricardo. 2014. “Situating translation expertise: A review with a sketch of a construct.” In The Development of Translation Competence: Theories and Methodologies from Psycholinguistics and Cognitive Science, ed. by John W. Schwieter and Aline Ferreira, 2–54. Cambridge, England: Cambridge Scholars Publishing. Muñoz Martín, Ricardo. 2016a. Reembedding Translation Process Research. Amsterdam: John Benjamins.  https://doi.org/10.1075/btl.128 Muñoz Martín, Ricardo. 2016b. “Reembedding translation process research: An introduction.” In Reembedding Translation Process Research, ed. by Ricardo Muñoz Martín, 1–19. Amster­ dam: John Benjamins.  https://doi.org/10.1075/btl.128.01mun Nelson, Hazel 1976. “A modified card sorting test sensitive to frontal lobe defects.” Cortex 12 (4): 313–324.  https://doi.org/10.1016/S0010-9452(76)80035-4 Neubert, Albrecht, and Gregory M. Shreve. 1992. Translation as Text. Kent, Ohio: The Kent State University Press. Newman, Aaron J., Roumyana Pancheva, Kaori Ozawa, Helen J. Neville, and Michael T. Ullman. 2001. “An event-related fMRI study of syntactic and semantic violations.” Journal of Psycho­ linguistic Research 30 (3): 339–364.  https://doi.org/10.1023/A:1010499119393 Newman, Sharlene D., Toshikazu Ikuta, and Thomas Burns. 2010. “The effect of semantic relatedness on syntactic analysis: an fMRI study.” Brain and Language 113 (2): 51–58. https://doi.org/10.1016/j.bandl.2010.02.001 Newmark, Peter. 1988. A Textbook of Translation. Hertfordshire, England: Prentice Hall. Nida, Eugene. 1964. Towards a Science of Translating. Leiden: Brill. Nida, Eugene, and Charles Taber. 1969. The Theory and Practice of Translation. Brill: Leiden. Nilipour, Reza, and Hassan Ashayeri. 1989. “Alternating antagonism between two languages with successive recovery of a third in a trilingual aphasic patient.” Brain and Language 36 (1): 23–48.  https://doi.org/10.1016/0093-934X(89)90050-3 Nisbett, Richard E., and Timothy D. Wilson. 1977. “Telling more than we can know: Verbal reports on mental processes.” Psychological Review 84 (3): 231–259. https://doi.org/10.1037/0033-295X.84.3.231 Nord, Christiane. 1997. Translating as a Purposeful Activity: Functionalist Approaches Explained. Manchester: St. Jerome. Nunez, Paul L., and Ramesh Srinivasan. 2006. Electric Fields of the Brain: The Neurophysics of EEG. second ed. Oxford, New York: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195050387.001.0001 Nuñez, S. Christopher, Mirella Dapretto, Tami Katzir, Ariel Starr, Jennifer Bramen, Eric Kan, Susan Bookheimer, and Elizabeth R. Sowell. 2011. “fMRI of syntactic processing in typically developing children: Structural correlates in the inferior frontal gyrus.” Developmental Cognitive Neuroscience 1 (3): 313–323.  https://doi.org/10.1016/j.dcn.2011.02.004 Ogawa, Seiji, Tso Ming Lee, Alan R. Kay, and David W. Tank. 1990. “Brain magnetic resonance imaging with contrast dependent on blood oxygenation.” Proceedings of the National Academy of Science USA 87 (24): 9868–9872.  https://doi.org/10.1073/pnas.87.24.9868

References 253

Oh, Jihoon, Mookyung Han, Bradley S. Peterson, and Jaeseung Jeong. 2012. “Spontaneous eyeblinks are correlated with responses during the Stroop task.” PLoS One 7 (4): e34871. https://doi.org/10.1371/journal.pone.0034871 Ojemann, George A., and Harry A. Whitaker. 1978. “The bilingual brain.” Archives of Neurology 35: 409–412.  https://doi.org/10.1001/archneur.1978.00500310011002 Oléron, Pierre, and Hubert Nanpon. 1964. “Recherches sur la traduction simultanée.” Journal de psychologie normale et pathologique 62: 73–94. Paap, Kenneth R., Hunter A. Johnson, and Oliver Sawi. 2015. “Bilingual advantages in executive functioning either do not exist or are restricted to very specific and undetermined circumstances.” Cortex 69: 265–278.  https://doi.org/10.1016/j.cortex.2015.04.014 Padilla, Francisca, María Teresa Bajo, and Pedro Macizo. 2005. “Articulatory suppression in language interpretation: Working memory capacity, dual tasking and word knowledge.” Bilingualism: Language and Cognition 8 (3): 207–219.  https://doi.org/10.1017/S1366728905002269 Pandya, Deepak N., and Edward H. Yeterian. 1985. “Architecture and connections of cortical association areas.” In Association and Auditory Cortices, ed. by Alan Peters and Edward G. Jones, 3–61. Boston, MA: Springer US.  https://doi.org/10.1007/978-1-4757-9619-3_1 Paradis, Michel 1995. “Another sighting of differential language laterality in multilinguals, this time in Loch Tok Pisin: Comments on Wuillemin, Richardson, and Lynch (1994).” Brain and Language 49 (2): 173–186.  https://doi.org/10.1006/brln.1995.1027 Paradis, Michel 2003. “The bilingual Loch Ness Monster raises its non-asymmetric head again-or, why bother with such cumbersome notions as validity and reliability? Comments on Evans et al. (2000).” Brain and Language 87 (3): 441–448. https://doi.org/10.1016/S0093-934X(03)00136-6 Paradis, Michel 2008. “Bilingual laterality: Unfounded claim of validity: a comment on Hull and Vaid (2007).” Neuropsychologia 46 (5): 1588–1590; author reply 1591–3. https://doi.org/10.1016/j.neuropsychologia.2008.01.029 Paradis, Michel 2011. “Principles underlying the Bilingual Aphasia Test (BAT) and its uses.” Clinical Linguistics & Phonetics 25 (6–7): 427–443. https://doi.org/10.3109/02699206.2011.560326 Paradis, Michel. 1977. “Bilingualism and aphasia.” In Studies in Neurolinguistics, ed. by Harry Whitaker and Haiganoosh Whitaker, 65–121. New York: Academic Press. https://doi.org/10.1016/B978-0-12-746303-2.50008-7 Paradis, Michel. 1979. “L’aphasie chez les bilingues et les polyglottes.” In L’aphasie, ed. by André Roch Lecours, François L’hermitte and et al., 605–616. Paris: Flammarion. Paradis, Michel. 1984. “Aphasie et traduction.” Meta: International Translator’s Journal 29: 57–67. https://doi.org/10.7202/003781ar Paradis, Michel. 1992. “The Loch Ness Monster approach to bilingual language lateralization: A response to Berquier and Ashton.” Brain and Language 43 (3): 534–537. https://doi.org/10.1016/0093-934X(92)90118-X Paradis, Michel. 1994. “Towards a neurolinguistic theory of simultaneous translation: The framework.” International Journal of Psycholinguistics 10: 319–335. Paradis, Michel. 2004. A Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins. https://doi.org/10.1075/sibil.18 Paradis, Michel. 2009. Declarative and Procedural Determinants of Second Languages. Amster­ dam: John Benjamins.  https://doi.org/10.1075/sibil.40

254 The Neurocognition of Translation and Interpreting

Paradis, Michel, Marie-Claire Goldblum, and Raouf Abidi. 1982. “Alternate antagonism with paradoxical translation behavior in two bilingual aphasic patients.” Brain and Language 15 (1): 55–69.  https://doi.org/10.1016/0093-934X(82)90046-3 Parkin, Beth L., Hamed Ekhtiari, and Vincent F. Walsh. 2015. “Non-invasive human brain stimulation in cognitive neuroscience: A primer.” Neuron 87 (5): 932–945. https://doi.org/10.1016/j.neuron.2015.07.032 Patterson, Karalyn, Peter J. Nestor, and Timothy T. Rogers. 2007. “Where do you know what you know? The representation of semantic knowledge in the human brain.” Nature Reviews Neuro­science 8 (12): 976–987.  https://doi.org/10.1038/nrn2277 Pavlović, Nataša. 2007. “Directionality in translation and interpreting practice.” FORUM. International Journal of Interpretation and Translation 5 (2): 79–99. https://doi.org/10.1075/forum.5.2.05pav Pavlović, Nataša, and Kristian Jensen. 2009. “Eye tracking translation directionality.” In Trans­ lation Research Projects 2, ed. by Anthony Pym and Alexander Perekrestenko, 101–119. Tarragona: Universitat Rovira i Virgili. Penfield, Wilder, and Lamar Roberts. 2014. Speech and Brain Mechanisms: Princeton University Press. Perani, Daniela, Jubin Abutalebi, Eraldo Paulesu, Simona Brambati, Paola Scifo, Stefano F. Cappa, and Ferruccio Fazio. 2003. “The role of age of acquisition and language usage in early, high-proficient bilinguals: An fMRI study during verbal fluency.” Human Brain Mapping 19 (3): 170–182.  https://doi.org/10.1002/hbm.10110 Perecman, Ellen. 1984. “Spontaneous translation and language mixing in a polyglot aphasic.” Brain and Language 23 (1): 43–63.  https://doi.org/10.1016/0093-934X(84)90005-1 Perez Velazquez, José Luis, and Richard Wennberg. 2009. Coordinated Activity in the Brain: Measurements and Relevance to Brain Function and Behavior. Dordrecht: Springer. https://doi.org/10.1007/978-0-387-93797-7 Pöchhacker, Franz. 2004. Introducing Interpreting Studies. New York: Routledge. https://doi.org/10.4324/9780203504802 Pokorn, Nike K. 2011. Directionality. In Handbook of Translation Studies ed. by Yves Gambier and Luc van Doorslaer. Amsterdam and Philadelphia John Benjamins. https://doi.org/10.1075/hts.2.dir1 Pokorn, Nike K. 2005. Challenging the Traditional Axioms: Translation into a Non‐mother Tongue. Amsterdam & Philadelphia: John Benjamins.  https://doi.org/10.1075/btl.62 Portnow, Leaj H., David E. Vaillancourt, and Michael S. Okun. 2013. “The history of cerebral PET scanning: From physiology to cutting-edge technology.” Neurology 80 (10): 952–956. https://doi.org/10.1212/WNL.0b013e318285c135 Potter, Mary C., Kwok-Fa So, Barbara von Eckhardt, and Laurie B. Feldman. 1984. “Lexical and conceptual representation in beginning and more proficient bilinguals.” Journal of Verbal Learning and Verbal Behavior 23: 23–38.  https://doi.org/10.1016/S0022-5371(84)90489-4 Prat, Chantel S., Robert A. Mason, and Marcel Adam Just. 2012. “An fMRI investigation of analogical mapping in metaphor comprehension: the influence of context and individual cognitive capacities on processing demands.” Journal of Experimental Psychology: Learning, Memory, and Cognition 38 (2): 282. Price, Cathy J., David W. Green, and Roswitha von Studnitz. 1999. “A functional imaging study of translation and language switching.” Brain 122 (12): 2221–2235. https://doi.org/10.1093/brain/122.12.2221

References 255

Proverbio, Alice Mado, and Roberta Adorni. 2011. “Hemispheric asymmetry for language processing and lateral preference in simultaneous interpreters.” Psychology 2 (1): 12–17. https://doi.org/10.4236/psych.2011.21002 Proverbio, Alice Mado, Giuliana Leoni, and Alberto Zani. 2004. “Language switching mechanisms in simultaneous interpreters: an ERP study.” Neuropsychologia 42 (12): 1636–1656. https://doi.org/10.1016/j.neuropsychologia.2004.04.013 Pulvermüller, Friedemann. 2005. “Brain mechanisms linking language and action.” Nature Re­ views Neuroscience 6 (7): 576–582.  https://doi.org/10.1038/nrn1706 Pulvermüller, Friedemann, and Yury Shtyrov. 2006. “Language outside the focus of attention: the mismatch negativity as a tool for studying higher cognitive processes.” Progress in Neuro­ biology 79 (1): 49–71.  https://doi.org/10.1016/j.pneurobio.2006.04.004 Pulvermüller, Friedemann. 2002. The Neuroscience of Language. Cambridge: Cambridge Univer­ sity Press. Pulvermüller, Friedemann. 2013. “Semantic embodiment, disembodiment or misembodiment? In search of meaning in modules and neuron circuits.” Brain and Language 127 (1): 86–103. https://doi.org/10.1016/j.bandl.2013.05.015 Pym, Anthony. 2003. “Redefining translation competence in an electronic age. In defence of a minimalist approach.” Meta: Translators’ Journal 48 (4): 481–497. https://doi.org/10.7202/008533ar Quaresima, Valentina, Silvia Bisconti, and Marco Ferrari. 2012. “A brief review on the use of functional near-infrared spectroscopy (fNIRS) for language imaging studies in human newborns and adults.” Brain and Language 121 (2): 79–89.  https://doi.org/10.1016/j.bandl.2011.03.009 Quaresima, Valentina, Marco Ferrari, Marco CP van der Sluijs, Jan Menssen, and Willy NJM Colier. 2002. “Lateral frontal cortex oxygenation changes during translation and language switching revealed by non-invasive near-infrared multi-point measurements.” Brain Research Bulletin 59 (3): 235–243.  https://doi.org/10.1016/S0361-9230(02)00871-7 Quian Quiroga, Rodrigo, Leila Reddy, Gabriel Kreiman, Christof Koch, and Itzhak Fried. 2005. “Invariant visual representation by single neurons in the human brain.” Nature 435: 1102. https://doi.org/10.1038/nature03687 Ralph, Matthew A. Lambon, Elizabeth Jefferies, Karalyn Patterson, and Timothy T. Rogers. 2017. “The neural and computational bases of semantic cognition.” Nature Reviews Neuroscience 18 (1): 42–55.  https://doi.org/10.1038/nrn.2016.150 Raskin, Sarah A., Martin Sliwinski, and Joan C. Borod. 1992. “Clustering strategies on tasks of verbal fluency in Parkinson’s disease.” Neuropsychologia 30 (1): 95–99. https://doi.org/10.1016/0028-3932(92)90018-H Ravizza, Susan M., Mauricio R. Delgado, Jason M. Chein, James T. Becker, and Julie A. Fiez. 2004. “Functional dissociations within the inferior parietal cortex in verbal working memory.” Neuroimage 22 (2): 562–573.  https://doi.org/10.1016/j.neuroimage.2004.01.039 Richardson, Fiona M., and Cathy J. Price. 2009. “Structural MRI studies of language function in the undamaged brain.” Brain Structure and Function 213 (6): 511–523. https://doi.org/10.1007/s00429-009-0211-y Rinne, Juha O., Jorma Tommola, Matti Laine, Bernd J. Krause, Daniela Schmidt, Valtteri Kaasinen, Mika Teräs, Hannu Sipilä, and Marianna Sunnari. 2000. “The translating brain: Cerebral activation patterns during simultaneous interpreting.” Neuroscience Letters 294 (2): 85–88.  https://doi.org/10.1016/S0304-3940(00)01540-8 Roiss, Silvia. 2001. “El mercado de la traducción inversa en España: Un estudio estadístico.” Her­ me­neus: Revista de Traducción e Interpretación 3: 379–408.

256 The Neurocognition of Translation and Interpreting

Rosiers, Alexandra, Evy Woumans, Wouter Duyck, and June Eyckmans. 2019. “Investigating the presumed cognitive advantage of aspiring interpreters.” Interpreting 21 (1): 115–134. Rubinov, Mikail, and Olaf Sporns. 2010. “Complex network measures of brain connectivity: Uses and interpretations.” NeuroImage 52: 1059–1069. https://doi.org/10.1016/j.neuroimage.2009.10.003 Ruiz, Carmen, Natalia Paredes, Pedro Macizo, and María Teresa Bajo. 2008. “Activation of lexical and syntactic target language properties in translation.” Acta Psychologica 128 (3): 490–500. https://doi.org/10.1016/j.actpsy.2007.08.004 Sánchez-Casas, Rosa M., José E. García-Albea, and Christopher W. Davis. 1992. “Bilingual lexical processing: Exploring the cognate/non-cognate distinction.” European Journal of Cognitive Psychology 4 (4): 293–310.  https://doi.org/10.1080/09541449208406189 Santamaría-García, Hernando, Sandra Baez, Pablo Reyes, José A. Santamaría-García, José M. Santacruz-Escudero, Diana Matallana, Analía Arévalo, Mariano Sigman, Adolfo M. García, and Agustín Ibáñez. 2017. “A lesion model of envy and Schadenfreude: Legal, deservingness and moral dimensions as revealed by neurodegeneration.” Brain 140 (12): 3357–3377. https://doi.org/10.1093/brain/awx269 Santilli, Micaela, Martina Gonzalez Vilas, Ezequiel Mikulan, Miguel Martorell Caro, Edinson Muñoz, Lucas Sedeño, Agustín Ibáñez, and Adolfo M. García. 2018. “Bilingual memory, to the extreme: Lexical processing in simultaneous interpreters.” Bilingualism: Language and Cog­nition: 1–18.  https://doi.org/10.1017/S1366728918000378 Scarpina, Federica, and Sofia Tagini. 2017. “The Stroop color and word test.” Frontiers in Psy­ chology 8: 557.  https://doi.org/10.3389/fpsyg.2017.00557 Scott, Graham G. Scott, Patrick J. O’Donnell, Hartmut Leuthold, Sara C. Sereno. 2009. “Early emotion word processing: Evidence from event-related potentials.” Biological Psychology 80 (1): 95–104.  https://doi.org/10.1016/j.biopsycho.2008.03.010 Scoville, William B., and Brenda Milner. 1957. “Loss of recent memory after bilateral hippocampal lesions.” Journal of Neurology, Neurosurgery and Psychiatry 20 (1): 11–21. https://doi.org/10.1136/jnnp.20.1.11 Schaeffer, Moritz, David Huepe, Silvia Hansen-Schirra, Sascha Hofmann, Edinson Muñoz, Lucas Sedeño, Agustín Ibáñez, and Adolfo M. García. Forthcoming. “Translation and Interpreting Competence Questionnaire: An online tool for research on translators and interpreters. ” Schulze, Heinz. 1968. “Unterschiedliche Ruckbildung einer sensorischen und einer ideokinetischen motorischen Aphasie bei einem Polyglotten.” Psychiatrie, Neurologie und Medizinische Psychologie 20: 441–445. Schwieter, John W., and Aline Ferreira. 2017. “Bilingualism in cognitive translation and interpreting studies.” The Handbook of Translation and Cognition, ed. by Aline Ferreira and John W. Schwieter, 144–164. Hoboken, New Jersey: John Wiley & Sons, Inc.  https://doi.org/10.1002/9781119241485.ch8 Sebastian, Rajani, Angela R. Laird, and Swathi Kiran. 2011. “Meta-analysis of the neural representation of first language and second language.” Applied Psycholinguistics 32 (4): 799–819. https://doi.org/10.1017/S0142716411000075 Seghier, Mohamed L. 2013. “The angular gyrus: multiple functions and multiple subdivisions.” Neuro­scientist 19 (1): 43–61.  https://doi.org/10.1177/1073858412440596 Seleskovitch, Danica. 1968. L’interprète dans les conférences internationales. Problèmes de langage et de communication. Paris: Minard.

References 257

Seleskovitch, Danica. 1975. Langage, langues et mémoire. Étude de la prise de notes en interprétation consécutive. Paris: Minard. Seleskovitch, Danica. 1978. “Language and cognition.” In Language Interpretation and Com­ munication, ed. by David Gerver and Wallace Sinaiko, 333–341. New York: Plenum Press. https://doi.org/10.1007/978-1-4615-9077-4_29 Seleskovitch, Danica. 1981. “Pourquoi un colloque sur la compréhension du langage ?” In Com­ prendre le langage, ed. by Jacques Barbizet, Maurice Pergnier and Danica Seleskovitch, 9–15. Paris: Didier Érudition. Seleskovitch, Danica, and Marianne Lederer. 1989. Pédagogie raisonnée de l’interprétation. Paris: Didier Erudition. Selimbeyoglu, A., and J. Parvizi. 2010. “Electrical stimulation of the human brain: perceptual and behavioral phenomena reported in the old and new literature.” Frontiers and Human Neuro­science 4: 46.  https://doi.org/10.3389/fnhum.2010.00046 Sharmin, Selina, Oleg Spakov, Kari-Jouko Räihä, and Arnt Jakobsen. 2008. “Where on the screen do translation students look while translating, and for how long?” In Looking at Eyes: EyeTracking Studies of Reading and Translation Processing (Copenhagen Studies in Language 36), ed. by Susanne Göpferich, Arnt Jakobsen and Inger M. Mees, 31–51. Copenhagen: Sam­fundslitteratur. Shebani, Zubaida, and Friedemann Pulvermüller. 2013. “Moving the hands and feet specifically impairs working memory for arm-and leg-related action words.” Cortex 49 (1): 222–231. https://doi.org/10.1016/j.cortex.2011.10.005 Shen, Hui, Zhenfeng Li, Jian Qin, Qiang Liu, Lubin Wang, Ling-Li Zeng, Hong Li, and Dewen Hu. 2016. “Changes in functional connectivity dynamics associated with vigilance network in taxi drivers.” NeuroImage 124: 367–378. https://doi.org/10.1016/j.neuroimage.2015.09.010

Shlesinger, Miriam. 2003. “Effects of presentation rate on working memory in simultaneous interpreting.” The Interpreters’ Newsletter 12: 37–49. Shreve, Gregory. 2009. “Recipient-orientation and metacognition in the translation process.” In Translators and their Readers: In Homage to Euge A. Nida, ed. by Rodica Dimitriu and Miriam Shlesinger, 255–270. Brussels: Les Editions du Hazard. Shreve, Gregory. 2012. “Bilingualism and translation.” In Handbook of Translation, ed. by Yves Gambier and Luc van Doorslaer, 1–6. Amsterdam & Philadelphia: John Benjamins. https://doi.org/10.1075/hts.3.bil1 Shreve, Gregory, and Erik Angelone. 2010. Translation and Cognition. Amsterdam & Philadelphia: John Benjamins.  https://doi.org/10.1075/ata.xv Signorelli, Teresa M., Henk J. Haarmann, and Loraine K. Obler. 2012. “Working memory in simultaneous interpreters: Effects of task and age.” International Journal of Bilingualism 16 (2): 198–212.  https://doi.org/10.1177/1367006911403200 Singer, Wolf 1993. “Synchronization of cortical activity and its putative role in information processing and learning.” Annual Review of Physiology 55: 349–374. https://doi.org/10.1146/annurev.ph.55.030193.002025 Snell-Hornby, Mary. 2006. The Turns of Translation Studies: New Paradigms or Shifting View­ points? Amsterdam: John Benjamins.  https://doi.org/10.1075/btl.66 Sperber, Dan, and Dierdre Wilson. 1986. Relevance: Communication and Cognition. Oxford: Basil Blackwell.

258 The Neurocognition of Translation and Interpreting

Sporns, Olaf. 2011. Networks of the Brain. Cambridge, Mass.: MIT Press. Springer, Jane A., Jeffrey R. Binder Thomas A. Hammeke Sara J. Swanson Julie A. Frost Patrick S. F. Bellgowan Cameron C. Brewer Holly M. Perry George L. Morris Wade M. Mueller. 1999. “Language dominance in neurologically normal and epilepsy subjects: a functional MRI study.” Brain 122 (11): 2033–2046.  https://doi.org/10.1093/brain/122.11.2033 Squire, Larry R. 2009. “The legacy of patient H.M. for neuroscience.” Neuron 61 (1): 6–9. https://doi.org/10.1016/j.neuron.2008.12.023 Starreveld, Peter A., Annette M. B. de Groot, Bart M. M. Rossmark, and Janet G. van Hell. 2014. “Parallel language activation during word processing in bilinguals: Evidence from word production in sentence context.” Bilingualism: Language and Cognition 17: 258–276. https://doi.org/10.1017/S1366728913000308 Stavrakaki, Stavroula, Kalliopi Megari, Mary H. Kosmidis, Maria Apostolidou & Eleni Takou. “Working memory and verbal fluency in simultaneous interpreters.” Journal of Clinical and Experimental Neuropsychology 34 (6): 624–33.  https://doi.org/10.1080/13803395.2012.667068 Steeb, Brenda, Indira García-Cordero, Marjolein C. Huizing, Lucas Collazo, Geraldine Borovinsky, Jesica Ferrari, Macarena M. Cuitiño, Agustín Ibáñez, Lucas Sedeño, and Adolfo M. García. 2018. “Progressive compromise of nouns and action verbs in posterior cortical atrophy.” Frontiers in Psychology 9: 1345.  https://doi.org/10.3389/fpsyg.2018.01345 Stemmer, Brigitte, Francine Giroux, and Yves Joanette. 1994. “Production and evaluation of requests by right hemisphere brain-damaged individuals.” Brain and Language 47 (1): 1–31. https://doi.org/10.1006/brln.1994.1040 Stemmer, Brigitte, and Harry A Whitaker. 2008. Handbook of the Neuroscience of Language: New York: Academic Press. Stengel, Erwin, and Joseph Zelmanowitz. 1933. “Über polyglotte motorische Aphasie.” Zeitschrift für die gesamte Neurologie und Psychiatrie 20: 441–445. Strobach, Tilo, Maxi Becker, Torsten Schubert, and Simone Kühn. 2015. “Better dual-task processing in simultaneous interpreters.” Frontiers in Psychology 6: 1590. https://doi.org/10.3389/fpsyg.2015.01590. Sunderman, Gretchen, and Judith F. Kroll. 2006. “First language activation during second language lexical processing: An investigation of lexical form, meaning, and grammatical class.” Studies in Second Language Acquisition 28 (3): 387–422. https://doi.org/10.1017/S0272263106060177 Talamas, Adrienne, Judith F. Kroll, and Robert Dufour. 1999. “From form to meaning: Stages in the acquisition of second-language vocabulary.” Bilingualism: Language and Cognition 2 (1): 45–58.  https://doi.org/10.1017/S1366728999000140 Tekin, Sibel, and Jeffrey L. Cummings. 2002. “Frontal-subcortical neuronal circuits and clinical neuropsychiatry: an update.” Journal of Psychosomic Research 53 (2): 647–654. https://doi.org/10.1016/S0022-3999(02)00428-2 Teubert, Wolfgang. 1996. “Comparable or parallel corpora?” International Journal of Lexico­ graphy 9 (3): 238–264.  https://doi.org/10.1093/ijl/9.3.238 Thiebaut de Schotten, Michel, Flavio Dell’Acqua, Peter Ratiu, Anoushka Leslie, Henrietta Howells, E. Cabanis, M.-T. Iba-Zizen, Odile Plaisant, Andrew Simmons, Nina F. Dronkers, Suzanne Corkin, Marco Catani. 2015. “From Phineas Gage and Monsieur Leborgne to H.M.: Revisiting disconnection syndromes.” Cerebral Cortex 25 (12): 4812–27. https://doi.org/10.1093/cercor/bhv173

References 259

Thierry, Guillaume, and Yan Jing Wu. 2007. “Brain potentials reveal unconscious translation during foreign-language comprehension.” Proceedings of the National Academy of Sciences of the United States of America 104 (30): 12530–12535.  https://doi.org/10.1073/pnas.0609927104 Tirkkonen-Condit, Sonja. 2005. “The monitor model revisited: Evidence from process research.” Meta 50: 405–413.  https://doi.org/10.7202/010990ar Tiselius, Elisabet, and Adelina Hild. 2017. “Expertise and competence in translation and interpreting.” In The Handbook of Translation and Cognition, ed. by John W. Schwieter and Aline Ferreira, 425–444. Malden, MA/Oxford, England: Wiley-Blackwell. https://doi.org/10.1002/9781119241485.ch23 Tommola, Jorma, and Marketta Helvä. 1998. “Language direction and source text complexity: Effects on trainee performance in simultaneous interpreting.” In Unity in Diversity? Current Trends in Translation Studies, ed. by Lynne Bowker, Michael Cronin, Dorothy Kenny and Jennifer Pearson, 177–186. Manchester: St. Jerome. Tommola, Jorma, Matti Laine, Marianna Sunnari, and Juha O. Rinne. 2001. “Images of shadowing and interpreting.” Interpreting 5 (2): 147–169.  https://doi.org/10.1075/intp.5.2.06tom Toury, Gideon. 1985. “A rationale for descriptive translation studies.” In The Manipulation of Literature: Studies in Literary Translation, ed. by Theo Hermans, 16–41. New York: St Martins Press. Trevisan, Piergiorgio, Lucas Sedeño, Agustina Birba, Agustín Ibáñez, and Adolfo M. García. 2017. “A moving story: Whole-body motor training selectively improves the appraisal of action meanings in naturalistic narratives.” Scientific Reports 7 (1): 12538. https://doi.org/10.1038/s41598-017-12928-w Tsumoto, Tadaharu. 1992. “Long-term potentiation and long-term depression in the neocortex.” Progress in Neurobiology 39 (2): 209–228.  https://doi.org/10.1016/0301-0082(92)90011-3 Tulving, Endel. 1986. “Episodic and semantic memory: Where should we go from here?” Behavioral and Brain Sciences 9 (3): 573–577.  https://doi.org/10.1017/S0140525X00047257 Tymoczko, Maria. 2005. “Trajectories of research in translation studies.” Meta: International Translator’s Journal 50 (4): 1082–1097.  https://doi.org/10.7202/012062ar Tymoczko, Maria. 2012. “The neuroscience of translation.” Target 24 (1): 83–102. https://doi.org/10.1075/target.24.1.06tym Tzou, Yeh-Zu, Zohreh R. Eslami, Hsin-Chin Chen, and Jyotsna Vaid. 2012. “Effect of language proficiency and degree of formal training in simultaneous interpreting on working memory and interpreting performance: Evidence from Mandarin–English speakers.” International Journal of Bilingualism 16 (2): 213–227.  https://doi.org/10.1177/1367006911403197 Ullman, Michael T. 2001a. “A neurocognitive perspective on language: The declarative/procedural model.” Nature Reviews Neuroscience 2 (10): 717–726.  https://doi.org/10.1038/35094573 Ullman, Michael T. 2001b. “The neural basis of lexicon and grammar in first and second language: the declarative/procedural model.” Bilingualism: Language and Cognition 4 (2): 105–122. https://doi.org/10.1017/S1366728901000220 Ullman, Michael T. 2004. “Contributions of memory circuits to language: the declarative/procedural model.” Cognition 92 (1–2): 231–270.  https://doi.org/10.1016/j.cognition.2003.10.008 Van de Putte, Eowyn, Wouter De Baene, Lorna García-Pentón, Evy Woumans, Aster Dijkgraaf, and Wouter Duyck. 2018. “Anatomical and functional changes in the brain after simultaneous interpreting training: A longitudinal study.” Cortex 99: 243–257. https://doi.org/10.1016/j.cortex.2017.11.024 Vandyke Carter, Henry, and Henry Gray. 1918. Anatomy of the Human Body. Philadelphia: Lea & Febiger

260 The Neurocognition of Translation and Interpreting

van Hell, Janet G., and Annette M. B. De Groot. 1998a. “Conceptual representation in bilingual memory: Effects of concreteness and cognate status in word association.” Bilingualism: Language and Cognition 1: 193–211.  https://doi.org/10.1017/S1366728998000352 van Hell, Janet, and Annette M. B. de Groot. 1998b. “Disentangling context availability and concreteness in lexical decision and word translation.” The Quarterly Journal of Experimental Psy­chology Section A 51 (1): 41–63.  https://doi.org/10.1080/713755752 van Hell, Janet G., and Annette M. B. de Groot. 2008. “Sentence context modulates visual word recognition and translation in bilinguals.” Acta Psychologica 128: 431–451. https://doi.org/10.1016/j.actpsy.2008.03.010 Vandermosten, Maaike, Cathy J. Price, and Narly Golestani. 2016. “Plasticity of white matter connectivity in phonetics experts.” Brain Structure and Function 221 (7): 3825–3833. https://doi.org/10.1007/s00429-015-1114-8 Vaquero, Lucia, Antoni Rodríguez-Fornells, and Susanne M. Reiterer. 2017. “The left, the better: White-matter brain integrity predicts foreign language imitation ability.” Cerebral Cortex 27 (8): 3906–3917.  https://doi.org/10.1093/cercor/bhw199 Varela, Francisco, Jean-Phillipe Lachaux, Eugenio Rodríguez, and Jacques Martinerie. 2001. “The brainweb: Phase synchronization and large-scale integration.” Nature Reviews Neuroscience 2 (4): 229–239.  https://doi.org/10.1038/35067550 Veyrac, Guillaume-Julien. 1931. Étude de l’aphasie chez les sujets polyglottes. Paris: L. Arnette. Videsott, Gerda, Bärbel Herrnberger, Klaus Hoenig, Edgar Schilly, Jo Grothe, Werner Wiater, Manfred Spitzer, and Markus Kiefer. 2010. “Speaking in multiple languages: neural correlates of language proficiency in multilingual word production.” Brain and Language 113 (3): 103–112.  https://doi.org/10.1016/j.bandl.2010.01.006 Vigneau, Mathieu, Virginie Beaucousin, Pierre-Yves Herve, Hugues Duffau, Fabrice Crivello, Olivier Houde, Bernard Mazoyer, and Nathalie Tzourio-Mazoyer. 2006. “Meta-analyzing left hemisphere language areas: phonology, semantics, and sentence processing.” Neuroimage 30 (4): 1414–1432.  https://doi.org/10.1016/j.neuroimage.2005.11.002 Vinay, Jean, Paul, and Jean Darbelnet. 1995 [1958]. “A methodology for translation.” In The Trans­ lation Studies Reader, ed. by Lawrence Venuti, 84–93. New York and London: Routledge. Vinck, Martin, Robert Oostenveld, Marijn van Wingerden, Franscesco Battaglia, and Cyriel M. Pennartz. 2011. “An improved index of phase-synchronization for electrophysiological data in the presence of volume-conduction, noise and sample-size bias.” Neuroimage 55 (4): 1548–65.  https://doi.org/10.1016/j.neuroimage.2011.01.055 Von Holzen, Katie, and Nivedita Mani. 2012. “Language nonselective lexical access in bilingual toddlers.” Journal of Experimental Child Psychology 113 (4): 569–586. https://doi.org/10.1016/j.jecp.2012.08.001 Weekes, Brendan S., and Ilhan Raman. 2008. “Bilingual deep dysphasia.” Cognitive Neuro­ psychology 25 (3): 411–436.  https://doi.org/10.1080/02643290802057311 Weisberg, Deena Skolnick, Frank C. Keil, Joshua Goodstein, Elizabeth Rawson, and Jeremy R. Gray. 2008. “The seductive allure of neuroscience explanations.” Journal of Cognitive Neuro­ science 20 (3): 470–477.  https://doi.org/10.1162/jocn.2008.20040 Weisenberg, Theodore, and Katherine McBride. 1935. Aphasia: A Clinical and Psychological Study. New York: Hafner. Weishaupt, Dominik, Victor D. Köchli, and Borut Marincek. 2006. How does MRI Work? An Intro­duction to the Physics and Function of Magnetic Resonance Imaging, second edition. New York: Springer.

References 261

Weiss, Sabine, Horst M. Mueller, Baerbel Schack, Jonathan W. King, Martha Kutas, and Peter Rappelsberger. 2005. “Increased neuronal communication accompanying sentence comprehension.” International Journal of Psychophysiology 57 (2): 129–141. https://doi.org/10.1016/j.ijpsycho.2005.03.013

Wernicke, Carl. 1874. Der Aphasische Symptomencomplex: Eine Psychologische Studie auf Anatomischer Basis. Breslau: Kohn and Weigert. Whyatt, Bogusława, and Tomasz Kościuczuk. 2013. “Translation into a non‐native language: The double life of the native‐speakership axiom.” Minor Translating Major‐Major Translating Minor‐Minor Translating Minor: A Translation Journal 5: 60–79. Yap, Melvin J., David A. Balota, Michael J. Cortese, and Jason M. Watson. 2006. “Single- versus dual-process models of lexical decision performance: insights from response time distributional analysis.” Journal of Experimental Psychology: Human Perception and Performance 32 (6): 1324–1344.  https://doi.org/10.1037/0096-1523.32.6.1324 Yeatman, Jason D., Robert F. Dougherty, Michal Ben-Shachar, and Brian A. Wandell. 2012. “Development of white matter and reading skills.” Proceedings of the National Academy of Sciences 109 (44): E3045–E3053.  https://doi.org/10.1073/pnas.1206792109 Yeung, Andy Wai Kan, Tazuko K. Goto, and W. Keung Leung. 2017. “The changing landscape of neuroscience research, 2006–2015: A bibliometric study.” Frontiers in Neuroscience 11: 120. https://doi.org/10.3389/fnins.2017.00120 Yoris, Adrián, Adolfo M. García, Liliana Traiber, Hernando Santamaría-García, Miguel Martorell, Florencia Alifano, Rafael Kichic, Jason S. Moser, Marcelo Cetkovich, Facundo Manes, Agustín Ibáñez, and Lucas Sedeño. 2017. “The inner world of overactive monitoring: neural markers of interoception in obsessive-compulsive disorder.” Psychological Medicine 47 (11): 1–14.  https://doi.org/10.1017/S0033291717000368 Young, Malcolm P., Jack J. W. Scannell, and Gully Burns. 1995. The Analysis of Cortical Con­ nectivity. Springer: Heidelberg. Yudes, Carolina, Pedro Macizo, and María Teresa Bajo. 2011a. “The influence of expertise in simultaneous interpreting on non-verbal executive processes.” Frontiers in Psychology 2: 309. https://doi.org/10.3389/fpsyg.2011.00309 Yudes, Carolina, Pedro Macizo, and María Teresa Bajo. 2011b. “Coordinating comprehension and production in simultaneous interpreters: Evidence from the Articulatory Suppression Effect.” Bilingualism: Language and Cognition 15 (2): 329–339. https://doi.org/10.1017/S1366728911000150 Yudes, Carolina, Pedro Macizo, Luis Morales, and María Teresa Bajo. 2013. “Comprehension and error monitoring in simultaneous interpreters.” Applied Psycholinguistics 34 (5): 1039–1057. https://doi.org/10.1017/S0142716412000112 Zheng, Binghan, Sandra Báez, Li Su, Xia Xiang, Susanne Weis, Agustín Ibáñez, and Adolfo M. García. Forthcoming. Semantic and attentional networks in bilingual processing: fMRI connectivity signatures of translation directionality. Zhou, Huixia, Baoguo Chen, Meiying Yang, and Susan Dunlap. 2010. “Language nonselective access to phonological representations: evidence from Chinese-English bilinguals.” Quarterly Journal of Experimental Psychology (Hove) 63 (10): 2051–2066. https://doi.org/10.1080/17470211003718705 Zillmer, Eric A., and Mary V. Spiers. 2001. Principles of Neuropsychology. Belmont, CA: Wads­ worth/Thomson Learning.

Index A abstract items  7, 168 accuracy  3, 11, 19, 24, 58, 107–108, 135, 137, 144, 198, 207, 209, 246 activation  10, 23, 29, 47–48, 50, 61–64, 71–72, 84–87, 90–92, 95, 116–117, 122, 134–137, 140, 144, 148–149, 152, 159, 162–163, 165–166, 172–173, 175, 180–181, 184, 207, 210, 215, 223–224, 236, 242–243, 255–256, 258 active paradigms  10 adjectives  134, 158 alpha band  xv, xviii, 167, 242 analytical theories  13 aphasia  37, 60, 92, 101–102, 104, 106–108, 212, 229, 238–239, 247–248, 253, 260 aphasics  36, 43 aphasiological evidence  2 approaches  xi–xii, 2–6, 12–14, 17–19, 23, 25–33, 39–40, 68–69, 73, 88, 121, 129, 140, 149–151, 168, 174, 178, 204, 207–211, 213, 216–218, 222, 234, 237, 252 articulation  8, 29, 58, 91, 140, 192, 210, 234 atomistic paradigms  46–47, 213–214 attention(al)  2, 9, 22, 24–25, 54, 75, 92, 101, 155, 159, 177, 200, 202, 212–213, 215, 234, 242, 255 mechanisms  15, 126, 215 paradigms 54 resources  4, 21 B backward translation (BT)  3, 136, 139, 141–143, 145, 158, 171, 193, 227

basal ganglia  80–81, 95–96, 104, 106–108, 111, 117, 119, 122, 157–158, 206 behavioral  xi–xii, xviii, 2–9, 11, 13–14, 16–18, 26, 31–32, 34, 37–38, 46–47, 52–53, 56, 58–59, 71, 73, 100, 117, 119, 121, 123, 129, 134, 142, 144, 146, 148–150, 155, 164, 168–169, 172–173, 178, 180–181, 183, 186–188, 192, 201, 206–207, 210, 216, 222, 237–238, 242, 246, 250–251, 257, 259 approaches 129 aspects  2, 8, 34 evidence  5, 9, 16, 59, 123, 180, 201 manifestations  3, 100, 149 measures  6, 11, 210 beta band  141–142 Bilingual Aphasia Test  37, 60, 212, 253 bilingual  xvi–xvii, xix, 1, 19, 36–37, 43, 49, 52, 55, 60, 76, 94–97, 100–101, 109, 117, 121, 123–126, 134–135, 141–142, 145–147, 149, 168–169, 172, 175, 178–180, 187, 193–199, 201–204, 212, 224–225, 229, 232–233, 235–239, 243, 246– 251, 253–254, 256, 260–261 bilingualism  xvii, xix, 3, 5–6, 19, 31, 38, 74–76, 94, 97, 175, 211–212, 221–222, 230–236, 238–239, 242, 244, 248, 250– 251, 253, 256–261 bilingual memory  xix, 19, 121, 124, 126, 134, 146–147, 235, 248, 256, 260 biological  2, 8–10, 30, 44, 52, 60–61, 73, 76, 100, 140, 162, 172, 201, 239, 245, 256

black box  xii, 30, 248 brain  xi–xii, xv–xvi, xviii– xx, 1–4, 6–7, 9–12, 25, 27, 29–40, 42–46, 48, 50, 52–53, 55, 59–63, 65, 68–73, 75–78, 81, 83, 86–88, 90, 93–95, 99–100, 116–120, 122–123, 127–129, 134, 137–138, 140–141, 149, 151–153, 156, 159, 161, 166, 173–184, 186, 189, 201, 204–206, 208–211, 213–219, 222–224, 229–232, 234–260 researchers 2 damage  43, 60, 99 lesion  42, 123 Brodmann area (BA)  80, 92, 184, 227 C causal evidence  33 clinical reports  6, 37 cognate  47, 49, 56, 141, 148, 163–164, 168, 170–172, 174–175, 207, 212–214, 224, 246, 256, 260 status  47, 56, 163, 171–172, 207, 212, 214, 246, 260 cognition  xi–xiii, xvii, xix, 9, 24, 26, 30, 39, 60, 83, 92, 100, 172, 177, 213, 217–218, 221–223, 229–233, 235–236, 239–240, 242–243, 245–251, 253–261 cognitive approaches  2, 5, 12, 25, 217 cognitive effort  18, 21, 23, 50, 58, 71, 144, 146, 150–151, 169 cognitive mechanisms  4, 47, 66, 146, 163, 169 cognitive neuroscience  xvii, 2, 6, 42, 45, 47, 73, 156, 221, 234, 237–238, 246–247, 250, 252, 254, 260

264 The Neurocognition of Translation and Interpreting

cognitive performance  3, 203 cognitive TIS  2, 4–6, 11–13, 25–31, 33, 36, 40–41, 43–44, 57, 75–76, 100–101, 129, 131–133, 150, 154–155, 172, 174–176, 178, 204, 206, 208–210, 213, 216, 218, 223 competence  12, 33, 44, 57, 59, 122, 124, 134, 141, 146, 148–149, 169–172, 175, 178, 187–188, 194, 197, 208–210, 212, 214, 219, 224, 252, 255–256, 259 comprehension  xvi, xviii, 6, 11, 16, 19, 22, 26, 68, 88–89, 91, 93–94, 100, 102–104, 107–109, 112, 114–115, 118, 120, 124, 132–133, 137, 141, 144, 166, 179, 184, 194, 201, 215, 225, 231, 233, 236, 244, 246–247, 249, 251, 254, 259, 261 compulsive translation  6, 100–105, 118, 122, 225 conceptual‑mediation 135, 140–142, 168 concrete  7, 20, 29, 37, 114, 123, 141, 148, 155, 161, 163, 168–169, 173–176, 207, 217, 235 condition‑related variables  76 connection strength  37, 58 connectivity  2, 4, 6, 32, 35, 37, 43–44, 46, 48, 68–69, 78, 85, 93–95, 117, 134, 140–142, 150, 152, 177, 183, 187, 192, 208, 210, 216, 223–224, 232, 237, 239, 242, 247, 256–257, 260–261 consecutive interpreters  15, 195, 197, 204 controlled experiments  3, 16 controlled variables  56 corpus  xi, 11–12, 18, 20, 31, 37–38, 51, 77, 95, 154, 168, 209, 229, 243, 251 corpus‑based studies  32, 38 cortical  xix, 3, 6–7, 10, 35, 38, 68, 70, 79–80, 86–88, 94, 96, 117, 119, 126, 140, 182, 187, 189, 207–208, 224, 233, 235, 244, 253, 257–258, 261 cortical thickness  xix, 10, 182, 187, 224, 244 cued recall  52

D data sources  5 decontextualized stimuli  20 dependent variables  13, 19, 56 deverbalization  16, 121, 133, 155, 174 dichotic listening  37, 51, 189–190 direct electrostimulation  70, 232 directionality  xii, 6, 33, 38, 43, 47, 56, 69, 119, 131–132, 134, 138, 140–142, 144, 146, 148–152, 157, 159, 168–169, 176, 210, 239, 242, 251, 254, 261 disciplines  3, 14, 27, 76, 211 dissociations  7, 32, 48, 59, 99, 101, 107, 117, 119, 135, 209, 212, 237, 255 domain  xvi, xix, 10, 23–24, 52, 58, 68, 72, 81, 93, 102, 117, 128–129, 155, 180, 188, 197, 200, 202–203, 222, 237 double dissociations  32, 59, 99, 101, 237 dual task  53 dysfunctions  31, 101, 103–104, 114, 129 É École Supérieure d’Interprètes et Traducteurs (ESIT)  15, 227 E ecological validity  16, 20, 23, 46 electroencephalography (EEG) 35, 227 empirical  xi, 6, 10–12, 15, 20, 25–26, 31, 34–36, 38, 40, 51, 118, 152, 168, 172, 174–175, 208–210, 217, 219, 230, 233– 234, 242, 246, 251 enhancement 230 equivalent  19, 37–38, 47–50, 62, 114, 120, 159, 161–162, 164–166, 172, 201, 206, 224, 240 recognition  37, 48–49, 60, 117, 162, 164, 209 event‑related potentials (ERPs) 84–86

executive functions  10, 12, 23–24, 38, 43, 52, 54, 104, 135, 149, 180, 183, 189, 202, 211, 232, 236, 250–251 experiment(al)  6, 43–44, 47–49, 56, 63, 71–73, 99, 138, 161–164, 166, 169, 200 design  42, 217 methods 6 paradigms 46 expert  4, 28, 31, 70, 154–155, 178, 197, 203, 211, 218, 231, 244, 260 expertise  xix, 7, 12, 25, 31, 33, 38, 44, 52, 178, 181, 188–189, 193–194, 197, 200–202, 209–210, 214–215, 230–231, 234, 236–237, 241, 244–245, 251–252, 259, 261 F formal linguistics  6, 41 form‑level processes  140–142 forward translation (FT)  3, 136, 139, 141–143, 145, 158, 170–171, 193, 224, 227 free recall  185 frontal (lobe)  xvii–xviii, 62, 64, 77–78, 80, 82, 88–90, 92–95, 100, 102, 104, 106, 116, 118, 122–123, 126, 135–136, 138–140, 142–143, 150, 159, 163, 167, 184, 189–190, 192, 199, 238, 250, 252, 255, 258 frontostriatal  xvi, 81–82, 89–90, 95, 116, 119, 122, 124, 130, 135, 140, 149, 157–158, 162, 172, 174, 176, 180, 183, 207, 223–224, 231 functional  xvi–xvii, xix, 2, 4, 6–7, 9–10, 12, 14, 29, 32–33, 35, 37–38, 44, 46, 48, 53, 61, 67–69, 75, 86–88, 93, 107–108, 115, 117–119, 122, 125, 127–128, 133–134, 137, 140, 144, 146, 149, 152, 154, 157–158, 162, 177–178, 183, 186, 188–189, 192, 201, 205–210, 216, 221, 223, 227, 230, 232–234, 237, 239–240, 242–247, 249, 254–255, 257–259

Index 265

aspects 183 autonomy  6, 119 correlates  10, 233 functional connectivity  2, 4, 32, 35, 37, 44, 48, 68–69, 93, 117, 140, 152, 177, 192, 208–209, 216, 223, 237, 239, 242, 247, 257 functional magnetic resonance imaging (fMRI)  2, 35, 227, 232, 245 functional near‑infrared spectroscopy (fNIRS)  158–160 functional neuroimaging  6, 44, 61, 134, 140, 157, 249 functional organization  4, 6, 12, 32, 128, 146, 178, 205, 210 G generalizability  18, 23, 43, 151, 188, 203 gray matter volume  10, 45, 190 H healthy subjects  4, 24, 38, 91, 214, 216 hemisphere  51, 77–79, 87, 104, 106, 111, 113, 115–116, 126–127, 138, 142–143, 192, 223, 227, 231, 233, 247, 249, 258, 260 hemodynamic  2, 9, 31, 33, 37, 60–63, 73, 90, 135, 138, 140, 146, 152, 157–158, 173, 187, 201, 209 techniques  2, 140, 157 hippocampus  80–81, 91, 177 history  xv, 1, 6, 10, 13, 25, 27, 34, 43, 57, 178, 212, 219, 236, 239, 241, 248, 254 humanistic  2, 4–5, 217 I imaging research  4 inability to translate  6, 100– 101, 104, 107–110, 117, 119–120, 157, 225 independent variables  12, 56 inhibitory control  10, 16, 23, 54, 58, 100, 183, 186, 209, 249 interdisciplinary  4, 12, 14, 16, 34, 39, 209, 218, 231, 242

interlingual  1, 8, 143, 164, 168, 173, 219, 221–222, 227 interlingual reformulation (IR) 1, 221–222, 227 interpreter advantage hypothesis 7, 33, 178, 240 interpreter  xii, 1, 7, 16, 33, 45, 53–54, 132, 177–178, 180, 182, 184, 189, 193, 195, 224, 227, 240, 242 interpreting  xi–xiii, xv, xvii–xviii, 1–12, 15–17, 20, 24, 26–31, 33–35, 37–40, 42–46, 51–52, 54–57, 62, 71, 73–76, 97, 101, 117–118, 120–124, 126, 131–133, 137–138, 140, 152, 154, 156, 167, 172–173, 175–176, 178– 190, 192–195, 197, 199–204, 206, 208–209, 211–212, 214– 215, 217–218, 222–224, 227, 230–231, 234–236, 238–240, 242, 245, 247–249, 251–252, 254–257, 259, 261 intracranial recordings  38, 43, 70, 144, 207 introspection  30, 174 invasive brain stimulation  3, 35 invasive techniques  xii, 6, 60–61, 69 K keylogging  11, 19–23, 27–28, 30, 41, 150, 173, 207, 210 known unknown  xi–xii, 1–2, 34, 39 L language  xv, xvii–xix, 4–10, 13–15, 18, 23, 31, 36–38, 47–51, 57, 60, 62–64, 66–67, 69–70, 72, 74–78, 80–81, 86–89, 92–96, 100–104, 106–109, 112–114, 116–120, 123–127, 130–137, 140–144, 146–148, 151, 154, 156–159, 161–162, 164–168, 170–173, 175–176, 178–180, 183, 185–186, 189–191, 193, 195–196, 202, 204, 206–207, 209, 212, 214–216, 221–225, 227, 229–261 switching  xviii, 229–230, 249, 254–255

lateralization  6, 37, 115, 122, 190, 205, 238, 250, 253 lesion  4, 6, 9, 42, 59–60, 70, 72, 89, 99–100, 105–111, 113, 115, 122–123, 125–126, 128–129, 149, 157, 241, 250, 256 lexical  xix, 6, 8, 11, 14, 16, 19–20, 23, 25, 49–51, 58, 91, 93, 97, 104, 109, 121, 123–124, 135, 142–144, 147–149, 154, 156–157, 159, 161– 163, 165, 170, 172, 192, 194, 203, 209, 223, 235, 237, 239–240, 242, 248, 251, 254, 256, 258, 260–261 access  8, 16, 23, 25, 49–50, 172, 251, 260 decision  49–50, 58, 109, 194, 209, 223, 235, 260–261 mechanisms 142 linguistic subsystems  8 M mechanisms  2, 4, 6–8, 12, 15, 19, 23–25, 29, 32–33, 36–38, 42–43, 45–47, 49–52, 54, 56–57, 59–60, 66, 68, 71–73, 75–76, 80, 87–89, 92, 94–97, 99–101, 104, 115–121, 123, 126, 128, 133–135, 138, 140–142, 144, 146, 149, 151–153, 155–157, 159, 162–164, 169, 173–177, 180, 184, 186, 194, 202–207, 210, 215, 229–230, 254–255 memory  xix–xx, 4, 10–11, 19, 23, 60, 78, 80–81, 91, 94, 100–101, 121, 124–126, 134, 146–147, 157–158, 165, 172, 185– 186, 196–197, 202, 213, 225, 227, 230–231, 233–236, 243–245, 247–251, 253–260 mind  3–4, 11–12, 20–22, 25, 29–32, 46, 100, 131, 153, 178– 179, 206, 208, 211, 219, 231, 245, 247, 249 mental  4, 11–13, 16–18, 21, 23–25, 29, 31–32, 40, 54, 60, 68, 96, 129, 135, 151, 185–186, 197, 209–210, 214–215, 242, 244, 248, 252 flexibility  23, 54 operations  11, 13, 25, 32, 186 systems 4

266 The Neurocognition of Translation and Interpreting

mental‑set shifting  74, 205, 217 metacognitive  4, 12, 17 methods  2–4, 6, 13, 16, 18–19, 22, 27–29, 32–33, 42, 63, 69– 70, 73, 127, 144, 150–151, 207, 210–211, 217, 237, 245, 249 milestones  19, 25, 27, 34–38, 213, 223 mind  3–4, 11–12, 20–22, 25, 29–32, 46, 100, 131, 153, 178– 179, 206, 208, 211, 219, 231, 245, 247, 249 modality  8–9, 18, 23, 53, 109, 117, 132, 149, 154, 163, 179–181, 183, 185–188, 190, 197, 200– 201, 203–204, 212, 215, 219, 233 model  xix, 12–15, 19–20, 38–39, 42, 63, 97, 100, 121, 123–127, 132–134, 146–149, 156, 168, 173, 199, 222–224, 227, 232, 234–237, 242, 248, 256, 259 modulations  6, 31, 48, 55, 60, 63, 65–66, 72, 93–94, 130, 135, 138–140, 142–144, 150, 159, 163–165, 172, 174–175, 191, 204, 207–209, 213–215, 224 morphosyntactic  6, 78, 87, 95, 100, 123–124, 126, 138, 172, 214 mechanisms 100 processing  78, 87, 95, 138 multimodal semantic processing 8 N N400  65–66, 93, 143–144, 150–151, 163–164, 166, 190–191, 224, 248 native language (L1)  127, 131, 133, 136–137, 141, 147, 161, 193, 195, 261 naturalistic  46–47, 61, 138, 156, 213, 236, 241, 259 paradigms 46–47 tasks 46 neocortex  76–78, 80, 84–86, 259 networks  xx, 7, 12, 32–33, 52, 60, 62, 66–69, 75–76, 81, 86–91, 93–95, 97, 116, 118–119, 122–124, 126, 128, 133, 141, 157,

162, 172–174, 176, 183, 192, 201, 205–206, 215–216, 230, 234, 251, 258, 261 neural  xi, xvi–xvii, 2, 4–6, 9–10, 12–14, 25–28, 30–34, 38, 40, 42–43, 46, 48, 50, 55–56, 59, 61–62, 66, 68, 70–74, 76, 78, 87, 91–92, 95–96, 100, 117–119, 123, 126, 128, 130, 134, 136–137, 140, 144, 149–151, 157, 159, 163, 167, 174–176, 181, 183, 187–188, 192, 200–201, 204, 206–207, 209–211, 213, 215–217, 219, 223, 229–230, 233–234, 236, 244, 246–247, 255–256, 259–261 correlate 9 processes 30–31 regions  10, 157, 163 signatures  4, 136–137, 163, 223 structures  4, 78, 150, 200 Neuroarchitectural Translation Model (NTM)  123, 125, 222–223, 227 neurocognition  xi–xiii, xvii– xviii, 3, 5, 7, 31, 42, 55, 57, 59, 65, 70–71, 73, 97, 99, 216, 229 neurofunctional changes  7 neurological data  2, 29 neurology  5–6, 74, 76, 97, 211, 221, 231, 237–239, 242–245, 253–254, 256 neuron  65, 82–87, 223, 235, 247, 254–255, 258 neuropsychology  xvii, xix, 5, 43, 211, 230, 233, 241, 243–244, 246, 258, 260–261 neuropsychological assessments 2 neuroscience  xiii, xvi–xvii, 2–3, 5–6, 10, 31–32, 34, 39, 42, 44–45, 47, 73, 76, 87, 100, 152, 156, 187, 208, 211, 213, 217–218, 221–222, 230–232, 234, 236– 252, 254–255, 257–261 neuroscientific evidence  6, 9, 174 neuroscientific research  4, 31, 38, 73, 214, 217

non‑neural approaches  12–29, 170–171 non‑participant observation  35 nouns  20, 37, 50, 109, 114, 120, 134, 141, 148, 159, 162–163, 174, 176, 191, 207, 236, 258 O observational  xi, 6, 12, 15–17, 151 approaches  12, 17 occipital  62, 78, 80, 87, 92, 95, 109, 111, 141, 159, 162 oscillatory activity  66–67, 223 oscillatory dynamics  2, 93, 167, 224 P P200  92, 248 paradoxical translation behavior 6, 100, 112–113, 117, 119, 225, 254 parietal  77–78, 80–82, 87, 90–96, 102, 106–107, 111–113, 115–116, 122, 124, 126, 130, 135, 141–143, 157, 162–164, 167, 172, 176, 182–183, 207, 223, 255 lobe  78, 80–82, 91–92, 106, 111 pathways  6, 81–83, 89–90, 118–119, 124, 146–147, 149, 153, 172, 176, 206, 223, 229, 247–248 performance  xviii, 3–4, 9, 15, 18–20, 22, 29, 36–37, 42–44, 48, 52, 54, 56, 58–59, 71, 97, 108–109, 112, 118, 120–121, 137, 146–150, 157–158, 161–162, 164, 168, 170, 174, 178, 183–185, 187–188, 192–195, 198–203, 212, 216, 225, 237, 240–242, 245, 247, 249, 259, 261 perisylvian  78, 88–89, 94, 109, 111, 118, 124, 157, 172, 180, 187, 201, 206 phonology/phonological 260 processing  8, 238 production  78, 123–124, 126, 185 recognition  78, 123

Index 267

physiology  xii, 244, 248, 254, 257 plasticity  xix, 204, 206, 237, 242, 244, 260 positron emission tomography (PET)  2, 35, 227, 243 practice  xv, 1, 4, 7, 13, 29, 38, 45, 55, 62, 70, 117, 178, 180–183, 185–190, 192, 195, 200–204, 208–209, 212, 215, 221–222, 234, 236, 252, 254 practitioners  5, 7, 52, 132, 185, 199, 202 pragmatic mechanisms  6, 25 pre/post designs  45, 73, 188 process  xv, xviii, 2, 4, 8–9, 12–13, 16–18, 21–23, 25, 27, 32, 43, 51, 58–59, 61, 71–72, 82–85, 87, 94, 126–129, 134, 140, 144, 151, 153–154, 168, 174, 179, 204, 207, 223, 229, 233, 241, 245– 247, 249, 252, 257, 259, 261 production  xiii, xvii–xviii, 4, 14, 16–18, 21–22, 26, 29, 36, 47, 78, 88–89, 91, 95–96, 100–104, 107, 109, 112–113, 118, 123–126, 128, 132, 142, 180, 185, 206, 210, 215, 223, 229, 235, 240, 245, 249–250, 258, 260–261 product  18, 21, 101, 154, 229, 233 professionals  3, 8–9, 121, 132, 148, 155, 178, 180, 185, 194, 200, 203–204, 209 proficiency  xix, 57, 72, 94–96, 100, 102, 121, 125–126, 135, 144, 146–150, 159, 163, 178, 185, 187, 203, 234, 236, 259–260 psychobiological  2, 8–9, 178, 186, 215 psycholinguistic(s)  28, 156, 252–253, 256, 261 evidence  6, 134, 146, 168 paradigms  11, 19–20, 23–24, 28, 30 Q qualitative  4, 18, 21, 137, 231 quantitative  xi, 12, 19, 21, 25–26, 58, 104, 232, 249

R rationalist approaches  12, 14 reading  xiii, xvi, xix, 4, 21, 33, 43, 47–48, 56, 61–62, 65, 67, 71, 88, 91, 93, 97, 102–103, 107, 109–110, 117–118, 121, 135, 138, 141, 153, 158, 166, 170–172, 185, 194–196, 202, 207, 209, 213– 216, 224, 233–236, 241–242, 246–247, 249, 257, 261 repetition  63–64, 88, 90–91, 101, 103–104, 109, 118, 134, 136, 166, 199 research design  28, 71 researchers  xii, 2, 4–5, 11, 17, 21, 27–28, 39, 49, 55–58, 60–61, 66, 68, 70–71, 74, 151, 179, 209 response time  3, 58, 170–171, 193, 195, 207, 261 resting‑state paradigms  30 retour  132, 242 Revised Hierarchical Model  19, 121, 146–149, 224, 232, 248 S segments  51, 154–157, 173–176, 179, 214 semantic(s)  80, 231, 236, 241, 260 access  10, 47, 78, 143, 147, 164, 237 decision  50, 190–192, 209, 224 processing  8, 51, 64, 81, 90, 135, 183, 190, 194 sentence  xvi, 20, 26, 37, 47–48, 56, 60, 65–66, 68–69, 88–89, 93–94, 107–110, 112–113, 115, 117, 122, 124, 128, 130, 138, 154–159, 162, 174, 184, 187, 190, 214, 224, 231–233, 235, 245, 248, 258, 260–261 translation  37, 47, 56, 107–110, 112–113, 115, 117, 122, 124, 157–158, 174, 224, 232, 248 short‑term memory (STM)  72, 200, 205 sight translation  8, 46, 109–110, 119, 163, 211, 243

simultaneous interpreting  7, 9, 15, 35, 37–38, 45, 52, 55, 117, 121, 126, 137, 140, 156, 175, 178–186, 188–190, 192, 194–195, 197, 199–204, 211, 223–224, 231, 234–235, 245, 248, 251, 255, 257, 259, 261 source language (SL)  8, 13, 227, 236, 241 spatial resolution  61, 63, 70 spatiotemporal correlates  157 ST processing  14, 22, 25, 124, 150 ST segment  6, 172 stimulus  6, 18–20, 22, 43–44, 48–49, 51, 55–58, 65–67, 71–72, 91–93, 107, 124, 136, 150, 158, 165, 167–168, 198, 212, 218 design 6 stimulus‑related variables  76–77 STM span  52 Stroop test  24 structure  xv, xx, 13, 25, 30, 38, 49, 81, 83–84, 86, 133, 135, 178, 211, 230–231, 248, 255, 260 structural correlates  xvii, 10, 252 students  3, 5, 9, 24, 33, 44, 51, 56, 62, 148, 157, 167, 170–171, 178, 181–187, 189–190, 196–197, 203, 209, 217, 221, 224–225, 230, 257 subcortical  6–7, 70, 77, 79–81, 96, 108, 117, 126, 135, 140, 187, 201, 206–208, 223, 229, 238, 258 subjective  3, 16–17 subject‑related variables  77, 170, 189 subsystems  8, 100–101, 118, 128 Sylvian fissure  77–78, 81, 88 synapse  83–85, 223 syntax  80, 125, 173, 234, 242 syntactic operations  8, 89, 162 syntactic parsing  8

268 The Neurocognition of Translation and Interpreting

T task  xvi, 3, 9–10, 21, 23–24, 26, 32, 43, 47–48, 50, 53–56, 58–62, 68, 70, 72–74, 90–91, 107–108, 121, 123, 128, 135, 138, 140, 149–150, 152, 154, 158–159, 161–164, 166–167, 169, 171–172, 178, 180, 183, 185, 187–188, 192, 195, 198–199, 201–202, 206–209, 214, 217–219, 223, 234–235, 237, 253, 257–258 temporal dynamics  6, 32, 68, 94, 134, 140, 142, 150, 157, 205 temporal lobe  77–78, 80–81, 89, 91–92, 100, 106, 111, 113, 115, 138, 159, 206, 234, 236, 247 temporal resolution  4, 21, 61–63, 65, 71 text  xiii, xvi, 4–5, 8–9, 11–12, 17–18, 21–23, 154, 179, 194, 214, 219, 227, 234, 245, 252, 259 theory  xv, 13–14, 16, 68, 100, 133, 221–222, 234, 240–241, 243–244, 246, 249, 252–253 theoretical formulations  3, 28, 33, 214 Théorie du sens  15–16, 25, 31, 121, 132, 155, 174 theta band  69 think‑aloud protocols (TAPs) 37

time course  12, 32, 58, 92, 134, 142, 149–150, 156, 163, 165, 168, 191, 207, 210, 224 TL processing  133, 175 trace  21, 44 trainees  52, 182–189, 192, 195, 197, 199–201, 203–204, 215, 224 training  xix, 1–2, 7–8, 16, 29, 33, 45, 117, 124, 148, 160, 170, 178–179, 181–190, 192, 200–201, 203, 206, 208–209, 217, 221, 224, 230, 234, 242, 244, 259 transcranial direct current stimulation (tDCS)  3, 227 transfer  13, 68, 133 translation  xi–xiii, xv, xviii– xix, 1–14, 17–23, 25–31, 33–40, 42–43, 45–51, 54–60, 62–63, 66, 68–72, 74–76, 97, 100–105, 107–110, 112–126, 130–131, 133– 136, 138–149, 152–176, 178, 183, 190, 192–193, 201, 203–205, 207–215, 217–218, 221–225, 227, 229, 231–257, 259–261 disorders  37, 100, 116 neuropathologies  6, 60, 116 phases 21 units  3, 6, 18–20, 25, 33, 38, 43, 51, 62, 66, 108, 126, 133, 140, 153–157, 163, 175–176, 214, 229, 233

translation and interpreting studies (TIS)  xi–xii, 1, 26, 222, 227, 238, 240, 256 translation without comprehension  6, 100, 114–115, 118, 120, 124, 225 translators  xv, 7, 17, 20, 33, 41, 44, 72, 117, 132, 141, 148–150, 154, 157, 169–171, 197, 216, 224, 232–233, 236, 238, 240, 243, 245–246, 249, 255–257 TT processing  23, 25, 215 U universality 15 unknown knowns  xiii, 2, 34 V variable  6, 20, 44–45, 47, 56–57, 66, 72, 150–152, 157, 174, 210, 213, 250 verbal  xii, 7, 15–17, 23–25, 36, 43, 46–47, 51–52, 54, 58–59, 64, 77, 80–82, 87, 89–91, 94–97, 103, 107, 109, 116–117, 123, 126, 141, 143–144, 152, 178–180, 183, 185–187, 189, 192–194, 196–197, 201–203, 208–209, 223–224, 230–231, 235, 243, 251–252, 254–255, 258, 261

This groundbreaking work ofers a comprehensive account of brain-based research on translation and interpreting. First, the volume introduces the methodological and conceptual pillars of psychobiological approaches vis-à-vis those of other cognitive frameworks. Next, it systematizes neuropsychological, neuroscientiic, and behavioral evidence on key topics, including the lateralization of networks subserving cross-linguistic processes; their relation with other linguistic mechanisms; the functional organization and temporal dynamics of the circuits engaged by diferent translation directions, processing levels, and source-language units; the system’s susceptibility to training-induced plasticity; and the outward correlates of its main operations. Lastly, the book discusses the ield’s accomplishments, strengths, weaknesses, and requirements. Its authoritative yet picturesque, didactic style renders it accessible to researchers in cognitive translatology, bilingualism, and neurolinguistics, as well as teachers and practitioners in related areas. Succinctly, this piece establishes a much-needed platform for translation and interpreting studies to fruitfully interact with cognitive neuroscience.

“Written by a leading neuroscientist and T&I researcher, García’s book raises neurocognitive work in the ield to new, impressive heights. To anyone interested in the topic, this volume will remain the standard reference work.” Arnt Lykke Jakobsen, Copenhagen Business School “A one-in-a-decade contribution, this book extends cognitive translation and interpreting studies with a much-needed, evidence-based neuroscientiic scope. García spells out a research program that may keep several generations of researchers busy.” Ricardo Muñoz Martín, Universidad de Las Palmas de Gran Canaria “A prodigious achievement. This book bridges the gap between neurocognitive research methods and solid model building in translation studies. García does a brilliant job of illuminating the deepest questions about the translating and interpreting mind.” Silvia Hansen-Schirra, Johannes Gutenberg University Mainz “Bold and ambitious, García’s book leaves no stone

isbn 978 90 272 0339 7

unturned and seamlessly synthesizes the key concepts and discoveries about this exciting ield of research. With its efective structure and conceptual depth, this volume succeeds in making the ‘unknown known’.” Minhua Liu, Hong Kong Baptist University

John Benjamins Publishing Company