Theoretical Hedging: the Scope of Knowledge in Translation Process Research

Theoretical Hedging: The Scope of Knowledge in Translation Process Research

A dissertation submitted to

Kent State University in partial

fulfillment of the requirements for the

degree of Doctor of Philosophy

Álvaro Marín García

August, 2017

Dissertation written by Álvaro Marín García B.A., University of Granada — Granada, Spain, 2007 M.A., University of Granada — Granada, Spain, 2008 Ph.D., Kent State University — Kent, OH, USA, 2017

Approved by

Gregory M. Shreve , Chair, Doctoral Dissertation Committee

Erik B. Angelone , Member, Doctoral Dissertation Committee

Sue Ellen Wright , Member, Doctoral Dissertation Committee

David Pereplyotchik , Member, Doctoral Dissertation Committee

Jocelyn Folk , Member, Doctoral Dissertation Committee

Accepted by

Keiran J. Dunne , Chair, Department of Modern and Classical Language Studies

James L. Blank , Dean, College of Arts and Sciences

TABLE OF CONTENTS

PREFACE ...... v DEDICATION ...... vii ACKNOWLEDGEMENTS ...... viii CHAPTER 1 ...... 1 Introduction ...... 1 Statement of the problem ...... 3 Purpose of this dissertation ...... 6 Significance of this dissertation ...... 7 An ongoing epistemological discussion ...... 9 The challenges of interdisciplinary research ...... 12 Overview of this dissertation ...... 16 CHAPTER 2 ...... 17 The empirical grounding of CTS ...... 18 Empiricism and the hierarchy of the sciences ...... 20 Research traditions in the cognitive sciences ...... 24 The Kuhnian revolution in CTS ...... 30 Paradigms in CTS ...... 32 Incommensurability ...... 38 From paradigms to research traditions ...... 44 CHAPTER 3 ...... 48 Scientific progress ...... 49 Models of scientific development ...... 52 Laudan’s model of scientific progress ...... 66 Scientific problems, empirical and conceptual ...... 69 Empirical problems ...... 69 Conceptual problems ...... 73 Research traditions ...... 77 Criteria for the evaluation of constructs ...... 84 Clarity ...... 86 Adequacy ...... 88 Consistency ...... 89

iii

Simplicity ...... 90 The usefulness of conceptual performance criteria ...... 90 CHAPTER 4 ...... 94 Expertise and its problems ...... 94 Competence as expert knowledge ...... 104 Clarity ...... 107 Adequacy ...... 110 Consistency ...... 113 Simplicity ...... 115 Summary of comparative analysis ...... 118 Expertise and Situated Expertise (STIE) ...... 120 Clarity ...... 123 Adequacy ...... 124 Consistency ...... 128 Simplicity ...... 130 Summary of comparative analysis ...... 131 CHAPTER 5 ...... 132 Conclusions ...... 132 Implications ...... 136 The scope of knowledge ...... 136 Pluralism ...... 141 Theoretical development ...... 145 Pedagogy ...... 148 Limitations and further directions ...... 149 REFERENCES ...... 151 APPENDIX A ...... 183 Glossary ...... 183 APPENDIX B ...... 185 List of acronyms ...... 185

PREFACE

The title of this dissertation appeared causally in a conversation with my advisor, Dr. Gregory M.

Shreve. I used “theoretical hedging” as a way of describing the issue I wanted to tackle in my dissertation: for almost two decades, scholars in Cognitive Translation Studies (CTS) have very rarely approached theoretical issues and disciplinary status in a way that promoted conceptual clarity and open, rational competition among different theoretical traditions in the field. In my view, there is a tendency in CTS for researchers to commit to theoretical positions only implicitly, in a way that avoids putting forward assumptions in the models used and so discussing them. It seemed to me that CTS as a field of scientific practice has been described with models that downplay ongoing competition and promote the establishment of subsequent traditions for the sake of disciplinary unity. Taking for granted an accepted theoretical framework allowed CTS scholars to conduct empirical research without engaging in theoretical discussions. But there is no true scientific development without conceptual clarity and theoretical competition –my argument followed– and, with some exceptions, it seems as if, in the last decade, CTS has focused on methodological development rather than on theoretical debates. It seems as if there was some sort of theoretical hedging in the field.

When I said that the problem looked like hedging, Dr. Shreve, seeing through my point, suggested that the expression could even be my title, and then proceeded to discuss with me how to develop my answer to that problem: I would need to find models of disciplinary evolution and scientific progress in the Philosophy of Science, to develop a model to assess constructs in CTS

and to provide concrete examples from CTS to back up my claims. Eventually, as I think any dissertation should unfold, the proposed answers turned out to be the bulk of the final draft, and the title remained as a trace of the initial statement of the problem. I decided to leave it, though, as it relates to the issues in CTS practice I intend to approach, and also as a reminder of the dialogic process of writing this dissertation, based on frequent informal discussions with all the members of the committee, especially with Dr. Shreve and Dr. Erik B. Angelone, and in-depth engagement with the works of Dr. Shreve and Dr. Ricardo Muñoz, who, in a way, have themselves engaged in an intellectual dialogue that now lasts almost 20 years.

In introducing a common set of criteria for conceptual comparison and assessment, this dissertation also intends to contribute to the establishment of disciplinary dialogues and points of contact among different traditions in CTS. As Gambier (2006:41) remarked, Translation Studies is still not always a coherent discipline, and I would like to argue that there is no need for such coherence provided we critically embrace fragmentation and plurality rather than monolithic, shared views of scientific progress.

DEDICATION

A mis padres, Pedro y Mª Eugenia, que –en su empeño por hacer a sus hijos mejores que

ellos mismos– pusieron el listón inalcanzable.

A Isa, que ha querido medir conmigo el tiempo.

A mi hermano, Pedro, cuya buena sombra me cobija.

vii

ACKNOWLEDGEMENTS

I want to express my gratitude to the members of my dissertation committee. Especially to Dr.

Gregory M. Shreve, who graciously offered his time, patience and support throughout the process of writing this dissertation, and who personifies the paradigm of the true scholar. I am indebted to him for his work and his example. Many thanks also to Dr. Erik Angelone, who co- advised this dissertation and has been a constant source of support, encouragement, sound advice and sanity both in Kentropolis and abroad. I want to thank Dr. Sue Ellen Wright for the long, meandering conversations, and for introducing me to Kuhnian theory. I am grateful to Dr. David

Pereplyotchik for showing me that thinking can –and indeed should– be both rigorous and festive at the same time—for having asked “what do you have to say for yourself?” and having meant it.

I would have never written this dissertation without the far-reaching mentorship of Dr.

Ricardo Muñoz. He has been a fundamental reference in my intellectual development for more than a decade now and remains a source of inspiration to this day. I am also indebted to him for never failing to believe I would rise to the challenge should the opportunity present itself, and then for making me see the opportunity when it appeared.

I also want to express my gratitude to my colleagues in the PhD program, the faculty members, and the administrative staff in the Modern and Classical Language Studies Department at Kent State University.

viii

Thanks to Isa for reminding me where home is, for her comprehension and teamwork.

Thanks to my parents, my brother and the rest of my family for their support and for their contagious faith in me.

I finally want to thank my friends for their cheerful support throughout the four years I devoted to my doctorate. Here’s to you all. In particular, I want to thank Benamí Barros, José

Antonio Espinar and Andréine Clavé for their beneficial presence in my life and for having made it easier than it seemed.

CHAPTER 1

Introduction

By the 1970s a number of Translation Studies (TS) scholars, dissatisfied with the linguistic, prescriptive models of translation then at the fore, began to develop a descriptive framework to empirically analyze translation and translations (Hermans, 1999:7-8). The different approaches that combined these concerns were labeled as Descriptive Translation Studies (DTS) following the taxonomy laid out by Holmes (1988) in the wake of Gideon Toury’s preliminary work. DTS would culminate in the publication of Toury’s landmark Descriptive Translation Studies and

Beyond (Toury, 1995/2012). The descriptive research agenda of DTS sought to establish TS as an autonomous scholarly discipline in good standing, able to extract patterns (laws and norms) and establish predictions grounded in empirical evidence. TS scholars then set out to engage in empirical research that could inform pedagogical approaches and describe professional practice.

The descriptive objective and the pedagogical intent resulted in two discrete disciplinary branches, namely, product- and process-oriented research (see Holmes, 1988; Toury,

1995/2012). By the 1990s, TS product-oriented scholars introduced Corpus Linguistics techniques, leading to a methodological breakthrough with new avenues of inquiry such as the study of translation universals (Baker, 1993; Laviosa, 2002) and to innovative teaching applications (Zanettin, Bernardini & Stewart, 2014).

The descriptive objective that had shifted the focus of interest to processes and their agents had already driven insightful work in Interpreting Studies. Unlike a majority of work in

TS at that point, research on interpreting stemmed originally from professionals’ reflection upon their own performance rather than from existing frameworks drawn from Linguistics or Literary

Studies (Pöchhacker, 2009:129). By the 1970s, Interpreting Studies scholars had drawn from

Psychology to propose information-processing models of simultaneous interpreting (Massaro,

1975; Moser-Mercer, 1978). With regards to translation, researchers started to develop models of the translation process (Vermeer, 1989; Holz-Mänttari, 1984) and hypotheses amenable to empirical tests (Krings, 1986; Lörscher, 1989). They soon found themselves contemplating a

“black box” (Toury, 1982:25; Krings, 1986, 1988), the translator’s mind, and acknowledged the need for research instruments used in Experimental Psychology, Expertise Studies and the

Cognitive Sciences.

A particular case in point is the use of think-aloud protocols (TAPs), which TS scholars started to employ by the mid 1980s to gain indirect insight into translators’ mental processes.

Jääskeläinen (2002:1070 points to 1986 as the beginning of “the official history” of cognitive studies in translation, as it was during this year that TAP-based empirical works in Cognitive TS

(CTS) began to proliferate (Dechert & Sandrock, 1986; Gerloff, 1986; Krings, 1986; Lörscher,

1986). 1 CTS was moving away from Linguistics and towards Psycholinguistics. Throughout the next decade, this evolution would consolidate, as made evident in the foundational work edited by Danks et al. in 1997, Cognitive Processes in Translation and Interpreting, where Psychology,

1 The term Translation Process Research (TPR) has been and still is used to refer to studies on the cognitive processes of translation. However, by its history and evolution, the term has been identified with cognitivist and computational approaches (Muñoz, 2017). I will therefore use Cognitive TS (CTS) (see Halverson, 2017) to refer to the whole subfield of TS that investigates the cognitive processes of translators and interpreters, be it through a cognitvist perspective or not.

Translation, and Interpreting scholars established some cornerstone topics for CTS: the definition of translation competence and its acquisition; the modeling of translation processes as constructed research tools; the psychological theories applied to translation sub-tasks; and the introduction of cognitive demands and the expectations of the parties involved in a translation task as the main constraints of translation performance. All of these issues still engage researchers in the field today (Halverson, 2009). As CTS set out to look into the “black box,” the time was ripe for a new research agenda that would follow the interdisciplinary tradition of TS.

In the following decades Cognitive TS evolved considerably (Shreve and

Angelone, 2010; Alves, 2015) as an interdisciplinary field borrowing from other disciplines to investigate cognitive aspects of the translation process (O’Brien, 2013). A steady growth of research projects, conferences and publications on translation (Muñoz, 2014), methodological and theoretical advances (Mees et al., 2009), and the foundational need to reshape traditional, humanistic forms of inquiry in favor of more empirically-based, scientific approaches (Toury,

1985; 1995/2012; Chesterman, 1993) have made CTS a growing and exciting field of interdisciplinary research.

Statement of the problem

However, this growth may be obstructed in the absence of a set of coherent criteria to analyze and assess which theories and concepts suit our purposes best and how they relate to each other. There are indeed voices pointing out that these issues might already be bringing development in the field to a stall: The CTS community keeps working to advance a discipline at the intersection of multiple research areas, but an imbalance between methodological

achievements (ways of investigating translation processes) and theoretical development (ways of explaining what those methods uncover) have cast doubts upon the validity and limits of the knowledge generated (Jääskeläinen, 2010; Muñoz, 2010; Shreve and Lacruz, 2015). A lively debate on methods and methodology (Neunzig, 2002; Alves, 2003; Mees, et al., 2009) has downplayed discussions at the conceptual level, where CTS has hosted new proposals with the ultimate aim of either complementing (Halverson, 2010) or replacing existing constructs and models (Muñoz, 2014b). Yet issues of general applicability, validity, and the interrelation of research objectives and paradigms have remained a constant and unresolved concern in the field

(Dancette & Ménard, 1996; Malmkjær, 2000; Shreve, 2002; Muñoz, 2010, 2014b; Alves, 2015).

As early as in 1997, Shreve approached the lack of consistency in terminological usages and concept applications, as well as certain misconceptions about theory building in an “attempt to lay out a preliminary conceptual framework for empirical translation enquiry” (1997:41). This conceptual and terminological disarray persists today. The steep growth curve of the discipline seems to have pushed CTS scholars towards the rapid development of new models of the translation process without a rigorous conceptual analysis of existing ones. Thus the relationships between old and new frameworks are, more often than not, unexamined. The problem has been exacerbated and, one might say, circumscribed by the introduction of theoretical frameworks and concepts borrowed from sister disciplines. Thus, the discipline finds itself with both new proposals and borrowed ones appearing at an alarming rate. Although enlightening, these proposals have not generally been followed by the necessary critical discussion and disciplinary debate.

An example of successful interdisciplinary interaction at the theoretical level is the work of Sandra Halverson on the relevance of bilingual conceptual representation models from the

cognitive sciences to understand corpus-based translational phenomena (2010). Halverson makes an explicit call on the need for cognitive approaches to translation to be grounded on theoretical advances from feeder disciplines, particularly cognitive theories of language and bilingualism compatible with the conceptual framework of cognitive translation theory (2010:364). Although it is still recent, Halverson’s work does not seem to have had the positive response its theoretical insights –bringing cognitive models of bilingualism to inform and reinforce translation models– have deserved. Her work addresses one of the major theoretical issues arising from interdisciplinarity, that of the compatibility of borrowed concepts with those of the receiving framework. New proposals and theoretical approaches can often resonate in a community of researchers, spreading quickly. Yet a detailed examination of emerging and borrowed theories and concepts to determine which ones better fit a given research context and explanatory purpose are scarce. It is worth noting that the last five years have yielded several works that reviewed accomplishments in the field and integrated them in a panoramic review of state-of-the-art CTS, but those works still do not engage in a critical, in-depth dialogue about the theoretical

“consistency” of the discipline—the compatibility of a number of simultaneously existing research traditions, cognitive paradigms and theoretical frameworks (Shreve and Angelone,

2010; O’Brien, 2013; Muñoz, 2014a; Angelone et al., 2015).

The considerable body of research currently being conducted by CTS scholars may fail to support further development in the discipline if the conceptual tools used to interpret ever- increasing amounts of empirical data are ill defined, ambiguous or simply invalid. In other words, the models and theories our discipline should depend on need to be robust and internally coherent theoretical entities that can be easily differentiated, operationalized, and related to other entities emerging from different frameworks in the field or from other disciplines. Moreover, the

current dearth of discussion on the underlying assumptions of our concepts may distort the validity of empirical results and of our understanding thereof. As Pym remarks when discussing the relationship between Philosophy and TS, in the absence of conceptual clarity and meta- analysis, “a rather quaint empiricism reigns, as in much of DTS, or in corpus linguistics, or think-aloud protocols, which rarely transcend positivist notions of science” (Pym, 2007:38). A plurality of explanatory frameworks is a much-needed fuel for disciplinary progress, but we need a methodology to navigate the issues of incommensurability, conceptual misrepresentation, inconsistency and empirical inapplicability that plurality can bring along with it.

Purpose of this dissertation

This dissertation intends to put forward a methodology to examine coexisting, possibly competing, conceptual constructs in order to determine which constructs are stronger and more suitable as research tools, which ones better conform to pre-existing constructs in CTS and other disciplines, such as Psychology or Cognitive Sciences, and which ones have more predictive and explanatory power while posing fewer validity issues. In sum, this dissertation proposes developing a disciplinary “conceptual performance model” and applying it to certain constructs in Cognitive TS.

The dissertation applies the work of Larry Laudan (1977) and Mike Mulkay (1975) to analyze the evolution of CTS as a field of practice and propose an alternative model of scientific progress different from Kuhn’s (1962) model, which has been the lens through which the evolution of CTS has typically been viewed. This shift in the description and analysis of the

“history of CTS” and its progress will lay out the ground for a discussion in terms of “research

traditions” instead of paradigms—from a hegemonic view of theoretical frameworks, to one that allows for the coexistence of diverging perspectives, which can then be compared. It is for the purposes of that comparison, and in response to the abovementioned scarcity of debate about conceptual constructs in CTS, that the conceptual performance model is introduced.

The utility of the performance model is then illustrated in two case studies in which competing constructs are analyzed in terms of clarity, adequacy, consistency and simplicity. The case studies center around, first, a comparative analysis of “competence” and “expertise,” as two different constructs within the same research tradition, and, second, around “expertise” and

“Situated Translation and Interpreting Expertise (STIE)” as the same construct developed in the light of two different research traditions.

Significance of this dissertation

Translation Studies is by nature, an empirical discipline (Toury, 1995). As such, translation theory should never detach itself from the objective of framing empirical designs and providing for the testable hypotheses required to investigate translation processes and products

(Toury, 1988). In order to do so, TS scholars need to ensure that the constructs they bring to the experimental design are valid, for “empirical facts do not exist independently of the scholar’s viewpoint; indeed, it is the scholar who creates the empirical facts of the analysis by making observable (raw) data relevant to his/her perspective” (Crisafulli, 2002:33).

Therefore, the validity of the existing concepts and frameworks used to interpret data delimits the scope and nature of the knowledge CTS can produce through a variety of methods

and perspectives. Besides helping to properly acknowledge scientific discovery, a better understanding of the knowledge CTS produces can help us to accommodate different approaches to the process of translation, such as psycholinguistic experimental studies and ethnographic work.

CTS has become an arena where different –and sometimes competing– constructs and research traditions overlap in their attempts to describe and explain empirical findings and so expand our knowledge about translators’ cognitive processes. The simultaneous existence of different sets of models and concepts purporting to explain the same phenomena is certainly a trait of disciplinary advancement and is not exclusive to CTS. But disciplinary advancement is also dependent on having a set of common, explicit criteria to support a dialogue aimed at determining which concepts will be most useful in the pursuit of our research and theoretical objectives. It is important to stress that these common, explicit criteria do not necessarily imply integration, agreement, or coherence —but are a shared assessment tool to ensure a minimal degree of common communication to strengthen the epistemology of CTS. Plurality is, it seems, at the core of disciplinary development, so the objective here is not necessarily one of theoretical agreement, but of achieving consensus on the tools used to understand the quality and scope of competing theories and their ability to explain empirical evidence. Otherwise we may condone the entrenchment of a given established theory or the uncritical acceptance of a new one. As

Martín Ruano (2006:47) remarks when talking about the disciplinary status of TS: “in the long run, the current effort at finding a common theoretical basis may result not in strengthening the discipline but in hampering its progress, to the extent that the marginalization of dissenting voices might prevent it from engaging in self-critical reflection and from being aware of its limitations.”

An ongoing epistemological discussion

Evolving from descriptive and functionalist approaches within TS, CTS has inherited some issues from the “mother discipline.” The need for a discussion about the nature and extension of the knowledge the discipline creates and concerns itself with has been subject of a debate as old as the field itself—what is Holmes’s foundational map (1988) if not an initial attempt to take a position with regards to disciplinary epistemology and chart the scope of translational knowledge and the approaches needed to acquire it? Ongoing discussions about the theoretical scope of TS and how to tackle it (Olohan, 2000; Hermans, 2002; Pym, Shlesinger,

Simeoni, 2008; Gambier & Van Doorslaer, 2009) have often identified a certain conceptual frailty in comparison with Comparative Literature, Linguistics, Psychology or Sociology, to name some of the most common knowledge realms from which TS has borrowed. Although I would like to argue that the issues identified –for instance, ill-defined terminology, lack of consensus about basic concepts and their links to empirical enquiry– are not exclusive to TS at all, it is true that the object of study and the general aim of the discipline remain open to disagreement.

A good example of this is the debate in the pages of Target that followed the publication of a paper on “shared ground” in TS by Andrew Chesterman and Rosemary Arrojo (2000).2 This

2 Chesterman and Arrojo’s proposed 30 statements about translation that scholars both from empirical and postmodernist, deconstructionist approaches to TS could agree upon. Most of the replies that followed focused on the epistemological assumptions underlying both camps, especially the empirical, and on whether or not such a shared, common ground or all-encompassing translation theory was needed. I would like to add to the frailties abovementioned the confusion between basic and applied endeavors in TS. In her coda to the 30 theses, Arrojo, who accepts, with certain caveats, the descriptive nature of TS, concludes that: “Instead of trying to make predictions, a theory of translation should attempt to empower translators-to-be and raise their conscience as writers concerning

lack of solid shared ground can be explained as a matter of disciplinary maturity and as a consequence of interdisciplinary development. Mayoral has argued that the dearth of conceptual consensus among specialists in TS, due to the youth of the discipline, has impeded the adoption of even an initial, shared metalanguage, thus thwarting scientific development (2001:45 and 68).

As indicated earlier, interdisciplinarity plays a marked role in this regard. If, in the initial stages of the development of a field, scholars embark on forays over disciplinary boundaries in search of models and methods from neighboring disciplines, it is to be expected that those expeditions create divergent directions, expanding rather than limiting the scope of thought in the field.

These borrowings don’t necessarily foster conceptual convergence or the elaboration and development of a single approach.

Also, most TS scholars, especially earlier in the history of the discipline, were educated as specialists in other fields, and it is mainly the last two generations of TS specialists that have been formally trained in TS as a scholarly discipline. It is only natural then that TS scholars have imported conceptual instruments from those areas of expertise they were most familiar with or that promised them the greatest success. That may also be the reason why the discussion of “our underlying epistemology,” the “philosophical underpinnings [that] will form the essential conceptual foundation” (Kiraly, 2003:4) have been described in terms of “passing fads” (Kiraly,

2003) or dominant tendencies (Gambier, 2006:35). Be that as it may, a problem graver than transient popularity arises when borrowings are not critically discussed or when integration is attempted based on only skewed or partial understandings of the borrowed theory or concept.

Malmkjær (2000: 166) stresses that: “To be able to make a properly informed selection, it is

the responsibility they will face in the seminal role they will play in the establishment of all sorts of relationships between cultures”. (2000:159)

useful to understand the immediate, and perhaps even the wider theoretical context of candidates for selection: the axioms (assumptions) underlying them, and their implications.” As we will see in the case studies in Chapter 4, it can be problematic, for example, to adopt a construct to analyze translation behavior in terms of expertise whose definition and implications do not cohere with the expertise construct as described in Psychology. In these cases, interdisciplinary borrowings can render research fruitless or, what is more problematic, they can inadvertently confound results. Being clear-eyed and critical about interdisciplinary borrowings is not about excluding new ideas, or favoring some existing theoretical perspective without being able to make a case for the choices made, but is rather a way of acknowledging disciplinary complexity while also preventing the proliferation of imported concepts based on misconceptions.

Yves Gambier (2006) argues for confronting transversal concepts, making a case for the metadiscourse of translation as the opposite of a totalizing knowledge—a self-inquiring “co- wandering” of different disciplines in which the researcher does not take translation practice and its observation apart, justifies her theoretical stance and methodological approach, and explicates her position as investigator (2006:34 and 35). As Gambier summarizes, any approach is legitimate provided a neat description of the research project can be produced: “La traduction abordée par les outils de la linguistique pragmatique par example est tout aussi légitime que celle abordée par la psychologie cognitive, à condition que les chercheurs disent leur visée, leur objet, leur unité et niveaux d’analyse.” (2006:36). [Translation tackled with the tools of pragmatic Linguistics, for example, is as legitimate as translation tackled by cognitive psychology, provided researchers state their aim, their object, their unit and level of analysis. My translation]

The challenges of interdisciplinary research

CTS encompasses cognitive and psycholinguistic approaches to the study of translation processes and involves a considerable number of phenomena that can be studied from a great many research perspectives. The variety of objects of study manifests itself in the

“multidisciplinarity” of CTS (Malmkjær, 2000); this entails further theoretical issues that one must bear in mind when selecting and using borrowed constructs and theories. This situation is not exclusive to CTS, of course. Any field that studies varied phenomena, particularly if it is a young one, faces this kind of challenge (see Millet et al., 2008, for a discussion on multidisciplinarity and epistemic pluralism in the study of social-environmental systems).

In CTS, Malmkjær stresses how important it is to understand the assumptions and implications of theoretical tools from “feeder disciplines,” as well as the differences “between borrowing a descriptive or explanatory or investigative tool from a neighboring discipline, on the one hand, and, on the other, the wholesale adoption of a full blown theory in order to explain translation (away) in terms of the borrowed theory” (2000:166-167). Borrowing a concept or methodological assumption allows researchers to inscribe the theoretical tool in a wider CTS approach, bolstering its power—for example, applying the mind-fixation assumption (Just and

Carpenter, 1980) to study effort in a translation task (see Göpferich et al., 2008). However, adopting a theory from another discipline to wholly describe translation phenomena, however insightful it may be, may pervade CTS theory with axioms that can become truisms in the receiving discipline (Muñoz, 2010:179). This was particularly the case in early stages of CTS development; over the last decades CTS scholars have become increasingly aware of the challenges of interdisciplinary borrowing (O’Brien, 2013). An example of the wholesale

adoption discussed here is Kuβmaul’s use of scenes-and-frames semantics in a Think-Aloud-

Protocol (TAP) design (Kuβmaul, 2000) to theorize translational creativity. As Bayer-

Hohenwarter points out (2009:43), the construct of scenes becomes ambiguous and so problematic as “it is in many cases very difficult to objectively trace the nature of scenes evoked in the heads of translators by inferring from the linguistic elements contained in the think-aloud protocols.”

At the same time, the unit and level of analysis to which Gambier referred also present challenges. The multiple sub-processes involved in the translation process have been modeled in componential proposals with multifarious constituents (Gopeferich, 2009; PACTE, 2003). These models subsume a range of hypothesized processes of varied nature –and the relations among them– under a unique theoretical structure that precedes empirical observation and analysis.

Such a top-bottom integration strategy shares some of the problems of computationalist approaches to cognition: if we assume that cognitive processing can be defined without recourse to its analysis, we can identify whether a model predicts a set of phenomena, but not necessarily whether it explains the modeled phenomena under scrutiny, for “pre-analytic conceptions of behavior and its causes may seriously misrepresent or distort what is actually going on”

(Cummins, 2000:131). Further, as Pym contends (2003:488), a multicomponent model, including all aspects of the translation task with associated terms and sub-competencies, may make the search for what is “specific to translation” even more difficult. Besides, it may be misleading when trying to identify what the main objective of the modeled task is: a “multicomponent model (…) tends to accept complexity without critically distinguishing between means and ends”

(Pym, 2003:494).

In the face of the challenges posed both by the network of sub-tasks and by human interaction in the translation process, some authors have advocated the use of triangulation as a methodological best practice to explore translation phenomena (see Alves, 2003), and to accept intersubjective knowledge as a valid source of data through which “we can get closer, perhaps not to an ‘objective’ result, but to shared, replicable experiences and results” (Hansen, 2003:40).

However, no valid relationship can be established between different experiences and/or results if research constructs used to theorize the process are not previously agreed upon, or if the relationships between the constituents of process models are not neatly delimited. Although we can agree that the nature of the object of study of CTS as a social science makes replicability problematic (Saldanha & O’Brien, 2014: 35), it is quite significant that studies seeking replication are so scarce for a field based on empirical analysis that considers replicability (and reproducibility) an advisable best practice (Tirkkonen-Condit, 2002; Shreve, 2006).

CTS has a tendency to evolve without revising or revisiting its previous constructs and assumptions, and again we find that CTS scholars are not the only ones in doing so: A paper recently published in Science (Nosek et al., 2015) has prompted an interesting and pertinent discussion on reproducibility of results in Psychology. The authors analyzed 100 experimental and correlational studies that had been published in prestigious Psychology journals to find that the significance of results dropped dramatically for reproductions. The authors point to the tendency to favor conceptual novelty over revisiting already published ideas as a factor that promotes misunderstanding about replication—actually, a way to tackle the uncertainty proper to any scientific evidence. Surely, in the case of CTS, this scarcity of replication studies has to do with the dazzling appeal of new directions and with the youth of the field (see Mayoral, 2001), but the situation is also rooted in the absence of a point of comparison for research proposals that

tread common territory but entail different assumptions. Every research project brings along presuppositions about its object of study and about the world, which more often than not remain unstated. It is invariably the case that those presuppositions permeate the statements that that research project offers to describe or explain the world (see, for example, Muñoz, 2016b for a discussion of the assumptions underlying the mind-as-computer and its implications for CTS).

Making these presuppositions explicit and open for discussion is considered one of the most important issues yet to be addressed by CTS researchers (Shreve & Angelone, 2010; Alves,

2015; Muñoz, 2017). Hence the essential question, as posed by Ladmiral:

“est-ce que mes présupposés – ou les vôtres – amenent a développer un discours plus

fécond, plus fidèle a la pratique, plus productif que celui de tel ou tel autre? A moins que

vos présupposés soient productifs dans tel ou tel domaine de la traduction, alors que ce

seraient mes présupposés qui permettraient de dégager un discours théorique plus

éclairant, plus efficace dans la meme discipline et dans d’autres domaines.” (Wilhelm,

2012:553)

[“Do my presuppositions –or yours– lead to develop a more fruitful discourse, a

discourse more faithful to the practice, more productive than those of others? Unless your

presuppositions were productive in this or that field of translation, it would be my

presuppositions that would allow to draw a more illuminating and effective theoretical

discourse, both within the same discipline and in other fields.” My translation]

Overview of this dissertation

This dissertation contains five chapters. Chapter 1 includes an introduction to the problem addressed, the purpose and significance of the dissertation, and this overview. Chapter 2 reviews the main conceptual framework applied to the problem this dissertation addresses, namely the work of Thomas Kuhn (1962), and argues against its limitations both from the viewpoint of Philosophy of Science and TS. Chapter 3 presents the conceptual performance model and the framework it is based on (Laudan, 1977), together with the conceptual aspects considered indicative of good performance and problem-solving. Chapter 4 comprises two case studies illustrating the conceptual performance model described in the previous chapter. Chapter

5 discusses the conclusions and the implications derived, as well as the limitations of this dissertation and possible future avenues of investigation.

CHAPTER 2

CTS is a scientific, empirical discipline and, as such, it conforms to the methodologies of empirical investigation in terms of observational theory building, data collection, hypothesis testing, and inductive analysis. These scientific procedures can be understood as requirements of the scientific paradigm.3 The work conducted in CTS is empirical in nature or, if conceptual, like this dissertation, applied to the empirical endeavor (see Williams and Chesterman, 2002:58 for a discussion of conceptual analysis in TS); therefore, CTS is subsumed under the umbrella of scientific empirical research. An “epistemology of CTS” would include the record and analysis of the ways in which knowledge is generated and disseminated in the field, which in this case comprise the scientific paradigm, and empirical methods and research constructs.4 The interest in the cognitive processes of translators and interpreters has led the CTS community to use frameworks from the cognitive sciences that, I argue, represent different research traditions within the same paradigm. In this chapter I will discuss aspects of

3 Paradigm is used here to refer to the overarching framework that models theory building in a field of research; for example, the scientific paradigm requires that certain requirements be met for a theory to be postulated and tested as opposed to the requirements in the humanistic, non-empirical paradigms. A stronger version of this concept that brings along further epistemic consequences (Kuhn, 1962) is discussed in following sections.

4 Stanford Encyclopedia of Philosophy (http://plato.stanford.edu/entries/epistemology/)

empirical research as applied to CTS and the main research traditions adopted before engaging in a discussion of the concept of paradigm in its stronger, Kuhnian sense and its implications and relation to the evolution of CTS.

Before proceeding, it is important to note that the development of TS as an empirical discipline did not preclude, however, the parallel growth of interesting and useful work pertaining to humanistic, non-empirical forms of inquiry. Works in comparative literature, cultural studies and philosophy, for example, have contributed and contribute to the study of translation and remain fruitful sources of inspiration for TS scholars. The divide between humanistic and empirical research in TS has been subject to debates which, as interesting as they are (see, for example, Pöchhacker’s, Gile’s and Chesterman’s contributions in Schäffner, 2004), fall outside the scope of this dissertation.

The empirical grounding of CTS

As we have seen in Chapter 1, TS scholars interested in the cognitive processes of translation and interpreting turned to cognitive science and cognitive psychology in search for tools to conduct empirical research. This movement towards a behavioral, scientific inquiry into translation processes required that TS scholars adopt and integrate the concepts, theories and models of the cognitive sciences to investigate their own object(s) of study. And what is no less important, it also entailed the adoption, along with the theoretical frameworks, of the methodologies and procedures of the scientific paradigm: looking for emerging patterns of generalizable facts in observable phenomena that can be turned into declarative, falsifiable statements (which are the base for theories) (Shreve, 1997). These observed facts are thus

considered to be examples of concepts (Hawes, 1975), which are represented by terms.

Following The Oxford Dictionary of Philosophy, we can define a “concept is that which is understood by a term, particularly a predicate. To possess a concept is to be able to deploy a term expressing it in making judgments (…).” However, to use a term in a useful way, and to make judgments based on that use, we need to understand the relationships of that concept with other concepts relevant to the theory we are working on. The meaning of a concept lies in its relationships with other concepts. Therefore we can more appropriately define a concept as an inferential role in a network (Sellars 1963). Also, a concept is applicable to theory construction only if operationally defined, that is, if the concept is defined according to the variables we will be testing empirically—hence the paramount importance of terminological clarity: It is not simply a matter of providing working definitions, but making sure these definitions relate back to the conceptual framework within which the empirical test is developed.

In order to describe complex phenomena, theories can then be used to build models, idealized representations of phenomena that focus on particular aspects of the object of study. A popular example of a model in CTS is the competence acquisition model presented by the

PACTE group (1998, 2003), which posits a number of sub-competences and the ways in which these interact. Different models can emphasize different features of the modeled phenomenon based on divergent assumptions, on inferences derived from empirical data, or from other theoretical statements. Therefore, inferences from a model are assumed to apply only to the model itself unless it captures the most important aspects of the actual phenomena under study; that is, unless clearly and solidly grounded on empirical data. In the case of CTS, Shreve and

Lacruz (2015) elaborate on the relevant aspects a cognitive model of translation should deal with.

The idealized description of facts sometimes involves the creation of complex concepts derived either from empirical observation or logical reasoning, and guided by a given framework. We call these concepts “constructs” and they are models to the extent that they propose explicative representations of specific phenomena. Given that constructs are guided by theoretical frameworks, they bear the assumptions of those frameworks, either implicitly or explicitly. Examples of relevant constructs in CTS include competence (PACTE, 2000), expertise (Shreve, 2002; Muñoz, 2014); literal translation (Schaeffer & Carl, 2014) or default translation (Halverson, 2016).

Empirical theories seek to describe and explain observable facts and their relationships, and their validity is linked to the predictive power they exhibit. By the same token, empirical

CTS models aim to represent and describe phenomena that are anchored in actual cognitive processes. Their validity as research tools, therefore, depends on translation process data to support their predictions. That is why hypotheses, or testable predictions about the nature of the object of study, are put forward and rejected or not in the light of empirical results. The ways in which data are analyzed, that is, the methodology used, impose further division within the empirical paradigm.

Empiricism and the hierarchy of the sciences

In describing empirical TS as a scientific discipline, Shreve (1997:42) states “all scientific inquiry proceeds from observation and description of empirical facts; and all scientific inquiry seeks to explain relationships or patterns observed.” The author bases his position on

Kaplan’s (1964) classification of methods of inquiry and concludes that, while all scientific

enterprises are empirical, they are so to a degree—scientific enterprises are more or less empirical depending on how much their theories depend on data or are derived from other theoretical statements and the relationships between their postulates.

The distinction between the sciences and their “degrees of empiricism” (Conant, 1950) is related to the hierarchy between natural and social (or soft and hard, see Storer, 1967) sciences, in which the sciences are placed at the top or at the bottom of the scale depending on their theoretical development, consensus and growth, among other variables. Debates about scientific hierarchies hark back, at least, to the work of August Comte (1896) and have remained open in the philosophy of science up to date.

In principle, CTS should not be concerned with those debates, as, based on its object of study, it is assumed to be a social or soft science that makes use of the social sciences methodologies, ranging from the experimental settings of Psychology to the ethnographic approaches of Anthropology or Sociology, for example. Yet the notion of hierarchy reaches beyond methodology or the object of study to the very notion of scientific progress. Hierarchies distribute the sciences according to levels of maturity and consensus, of theory predictability and disciplinary growth, and those factors are a current concern in CTS. Although there is no absolute agreement on what the dimensions according to which we can analyze the development of different disciplines are, there are some elemental variables that come into play. Cole

(1983:113) summarizes them in the following table:

Characteristics of Different Types of Science Variable Top of the Hierarchy Bottom of the Hierarchy Development Highly developed theory; No theory or low-level of Theory research guided by a generalization; pre- paradigm; high levels of paradigmatic phase; low codification levels of codification Quantification Ideas expressed in Ideas expressed in words mathematical language Cognitive High levels of consensus on Low levels of consensus on Consensus theory, methods, significance theory, methods, significance of an individual’s of an individual’s contribution contribution Predictability Ability to use theory to make Inability to make verifiable verifiable predictions predictions Rate of High proportion of references Low proportion of references obsolescence to recent work as older work to recent work as older becomes obsolete— remains just as important as indicating significant new work—indicating lack of accumulation of knowledge significant accumulation of knowledge Rate of Growth “Progress,” or the rate at “Progress,” or the rate at which new knowledge grows, which new knowledge grows, is relatively fast is relatively slow Table 1. Variables in the hierarchy of the sciences

These characteristics apply across disciplines, with the natural sciences scoring higher than the social sciences on the middle column. Thus, for example, Physics would be in a higher position than Biology, and Sociology would be found further down the hierarchy. The place a science occupies in the hierarchy depends on the strength of the theoretical and methodological tools used, and on the consensus level these tools gather among scientists in each field. It is also an indicator of disciplinary maturity, the higher a science is, the more mature –and so progressive– it is.

This conception of scientific progress and maturity derives from Kuhn’s theory of scientific evolution (1962). Kuhn suggested that paradigms, or overarching epistemological frameworks that model scientific practice mark maturity and set the theories and methods agreed

upon by scientists. We will go back to the notion of paradigm and its application to CTS in the next section, but let me stress that considering the establishment of paradigms (see the first row,

“theory development”) as a variable in the hierarchy of the sciences hints that different approaches within one discipline can, actually do, rank differently. As I will discuss below, I do not consider that having established paradigms as indicators of maturity represents the complexity of scientific evolution and the actual imbalance between different approaches within a given science. In fact, as Laudan remarked, “it is extremely unclear whether the notion of a

“mature” science finds any exemplification whatsoever in the history of science” (1977:151). At any rate, although I have been using “science” and “discipline” interchangeably in this section, and given examples of different disciplines as unified fields of scientific inquiry for illustrative purposes, I do not intend to suggest that all disciplines are monolithic; on the contrary, they very rarely are. CTS is an example of that variety, evidenced in different research traditions.

Apart from issues of maturity of the sciences, any hierarchy or division between disciplines imposes a determined view of science that sometimes may not correspond to actual scientific practice. In pointing this out, Robert Wilson aims beyond traditional divisions of science and offers his own take, introducing the term “fragile sciences,” instead of “human sciences,” to refer to cognitive, biological and social science, as all of them are fields where

“the conceptualization of the individual has become more partial and less encompassing,

but also more closely tied to models, techniques, and research strategies in particular

sciences. It is these ties, rather than the study of human agency or human nature, that

interest me” (2004:8).

Divisions of the sciences and their placement in any hierarchy are dependent on the research interests of those presenting the ranking; but there is no doubt that, as Wilson mentions,

methodology is the research dimension around which different disciplines can be grouped, and that the methods used are a basic variable to measure scientific development.

Before we continue to discuss the suitability of paradigms and research traditions to describe the evolution of CTS, we need to know which are the different approximations to the cognitive study of translation processes that these terms aim to describe.

Research traditions in the cognitive sciences

Since its inception during the last decades of the 20th century, CTS scholars have applied their own variant of Lakoff’s “cognitive commitment” to the study of translation, setting themselves the objective of describing human translation based on –or in accordance with– knowledge about cognition that comes from research beyond TS.5 Because CTS is a younger field than the cognitive sciences, and a borrower from them, the theoretical apparatus of CTS has been modeled out of cognitive science research traditions. It is therefore appropriate to provide a cursory review of the main research traditions in the cognitive sciences before we discuss their application to CTS research.

In discussing the theoretical evolution of CTS, Alves (2015) and Muñoz (2016c) identify three main cognitive science traditions –they both call them paradigms– that roughly correspond with research programs singled out by cognitive scientists themselves (see Piccini, 2012 and

Shapiro, 2012).6 These traditions include cognitivism, or the “information-processing paradigm,”

5 “(…) to make one’s account of human language accord with what is generally known about the mind and brain from disciplines other than Linguistics.” (Lakoff, 1991:54) 6 See also the entry for “cognitive science” in the Stanford Encyclopedia of Philosophy (http://plato.stanford.edu/entries/cognitive-science/#CriCogSci).

based on the computational theory of the mind as a partly modular processor of information

(Fodor, 1983); connectionism, which considers cognition to take place in the form of neural networks (see Churchland, 1989) computing information in parallel distributed processing (PDP)

(McClelland, Rummelhart, and Hinton 1986); and embodied and extended cognition, which assumes cognition to be a biological adaptation that takes advantage of constructed environmental underpinnings and, thus, understands cognitive processing to be a bodily phenomenon that transcends the brain and relies on external tools (Maturana and Varela, 1988;

Wilson, 2002; Anderson, 2003).

These research traditions are actually labels that refer to complex, bundled sets of theoretical positions and axioms accepted to varying degrees by cognitive scientists and philosophers of mind. Indicative of the plurality within traditions are the different names they have been assigned. Cognitivism is often identified as computationalism after the Computational theory of mind (Fodor, 1983), although not all cognitivist approaches are computational to the same extent (see Piccini, 2012 for a review). It is also known as the “information-processing” paradigm, as it is based on the notion that cognition is based on the sequential, lineal, discrete processing of information, very much in the way a computer works (Newell & Simon, 1961).

This idea reflects the impact of initial Artificial Intelligence in the development of cognitive science (Turing, 1936). The conceptualization of thinking as the parsing of symbolic units of information is also related to the “sandwich model” (Hurley, 1998), according to which higher- level processing –cognitive processing– is different in nature and segregated from lower-level processing carried out by the peripheral (sensorimotor) systems. Cognitivist views are usually thought to flesh out the cognitive science standard view or classical paradigm.

In response to the classical view that proposed a “language of thought” (Fodor &

Pylyshyn, 1988) to describe cognition at a level different from neural processing (see Marr,

1982), a new brand of computationalism emerged that aimed at explanations inspired by neural networks (McClelland, Rummelhart, and Hinton, 1986). One of the differences between cognitivism and connectionism is that most variants of the latter rejects syntactic computation and proposes parallel processing instead. The inspiration in neural networks of connectionism does not ensure that the modeled network corresponds to actual neural networks. That is why the model’s capacity to represent actual cognitive processing has been subject to criticism by neuroscientists (Globus, 1992).

An increasing interest in the role the body and the environment surrounding the thinking brain play in cognition has fostered the proliferation of different hypotheses and theses that can be loosely put together under the term 4EA (embodied, extended, enacted, embedded, affective) cognition (Protevi, 2010). Robbins & Aydede use the term “situated cognition” to name this set of approaches, whose main shared elements they summarize as follows:

“First, cognition depends not just on the brain but also on the body (embodiment thesis).

Second, cognitive activity routinely exploits structure in the natural and social

environment (the embedding thesis). Third, the boundaries of cognition extend beyond

the boundaries of individual organisms (the extension thesis)” (Robbins & Aydede,

2009:3).

Situated cognition departs a long way from cognitivist or connectionist approaches in arguing that information processing is not sequential nor symbolical, and so that there is no

“language of thought,” but a distributed and probabilistic continuum by means of which humans construe the world (Spivey, 2007) using their bodies and external tools as part of the process

(Clark, 1997). It is important to stress that situated cognition is a relatively new approach to cognitive science and that it has not yet developed a basic consensus among scholars other than the abovementioned elemental theses, mainly developed as a critique to computational cognitive science. Situated cognition is described more often according to its points of departure from computational views than according to its own proposals. It has emerged as a reaction that is yet to build an alternative. As Shapiro (2012) points out when discussing embodied cognition, there is not even agreement as to how much (in terms of ideas and methods) can be brought in from classical cognitive science.

The debate between computationalists (either cognitivists or connectionists) and proponents of situated cognition is an engaging discussion that tackles the very nature of mind and cognition. Scholars defending situated cognition contend that setting boundaries to the mind, as cognitivists do, reflects a Cartesian –and so dualist– prejudice to the view of mind (Rowlands,

1999; Rockwell, 2005). On the other side of the spectrum, Adam & Aizawa, who tend to be very critical with all forms of situated cognition, claim that “what advocates of extended cognition need (…) is a plausible theory of the difference between the cognitive and the non-cognitive that does justice to the subject matter of cognitive psychology” (2008:11).

The level of acceptance of the positions from these traditions is anything but homogenous across the board, and any attempt to draw overarching neat distinctions would be a misleading oversimplification. More or less strong versions of the main hypotheses put forward by proponents of one tradition or another permeate other scholars’ work: Despite Rowlands’ (1999) and Rockwell ‘s (2005) criticisms, virtually no scholar who rejects the hypotheses of embodied and extended cognition would deny that human cognition is physically grounded in the body and that, sometimes, humans avail themselves of external tools to “offload” their mental load

(Dennett, 1996:134); but, at the same time, they would not accept that those tools are part of the cognitive process and would affirm that cognition is bound to the cranium. Others may accept a principle from the opposite camp as an open possibility but not as a descriptive statement; for example, Adam & Aizawa would accept the parity principle, which states that parts of the world can be part of cognition if they work the way mental processes do (Clark & Chalmers, 1998), because they “think that the possibility of cognitive equivalence between the intracranial and transcranial processes allows for the possibility of extended cognition” (2008:28); yet that possibility has not yet actualized, according to the authors, given that the differences in nature between what happens in the brain and outside the brain disqualify the latter phenomena to be called “cognition.” On the other hand, strong opponents of computational models of mind, such as Spivey, who describes cognition as distributed and probabilistic, would still accept the existence of internal mental representations, albeit not as “an internal mental entity that symbolizes some external object or event to an attentive central executive,” (2007:4) but as

“internal mental processing that is continuous in time, is contiguous in state space, and whose function is to mediate between sensory stimulation and physical action” (2007:5).

For the purposes of this dissertation, I am interested in the reading of these traditions made by CTS scholars in their attempts at building models of the cognitive processes of translation, and how they have informed the two main research traditions currently endorsed by scholars in the field, cognitivist and cognitive translatology.

While, in the light of the preceding sections, we can state that CTS is a scientific, empirical discipline that can be associated with the soft or social sciences, different methods used have led CTS scholars to clam different positions in a hypothetical hierarchy. Research conducted according to computationalist approaches in CTS has traditionally used experimental

methods borrowed from Psycholinguistics that entitled their proponents to claim their research was “stronger” (hierarchically superior). On the other hand, situated cognition research in translation, which, although also embracing experimental designs, has developed mainly ethnographic studies in the workplace, has yielded claims on ecological validity on the grounds that experimental designs were not representative of the translation task as an object of study.

However, using experimental designs does not define a discipline as scientific, and the analysis of non-experimental, empirical data is as valid a source of scientific knowledge. But there is a difference with regards to the possibilities of generalization and replication, and thus with regards to the “hardness” or “softness” of the research conducted, depending on the method used.

Although the different approaches in cognitive science discussed in this section are oftentimes referred to as paradigms, I have deliberatively called them research traditions. Some of the reasons are precisely the heterogeneous nature of these approaches, their variety, and the interrelations established among them, which do not fit into the concept of paradigm as defined by Kuhn (1962). Another reason is that the concept of research tradition, as defined by Laudan

(1977), seems to be a more effective tool to describe the current situation in CTS and provides for the competition of research constructs. In the following section I will discuss the concept of paradigm and its problems. Research traditions will be presented as an alternative in Chapter 3.

The Kuhnian revolution in CTS

The different axioms in cognitive sciences research traditions entail different assumptions that have an impact in developing theories about the translation process. In order to describe these divergences, CTS scholars found in Kuhn’s work (1962) a model of scientific progress instrumental for structuring a diachronic view of the field at a time, roughly the second decade of

21st century, when the development of the discipline invited them to look back and take stock of the progress made. This approach to the evolution of the field, which implicitly parted with the humanities and aligned CTS with the (social) sciences, implies evidences that areas of interest in the field are broadening and branching out (García, 2012; Rojo, Ramos and Valenzuela, 2014;

Risku, 2014; Ehrensberger-Dow and O’Brien, 2015; Halverson, 2016), and that the discussion is no longer limited to interdisciplinary challenges, but also includes the risks to internal coherence of different traditions within one’s own field.7 While Kuhn’s contributions to the Philosophy of

Science show great insight and enticing alternatives to positivist perspectives on scientific progress, we will see in the following sections that his proposals do not fully describe the current state of affairs in CTS and fail to offer a rational ground for the competition of theories.

In 1962, Thomas S. Kuhn published The Structure of Scientific Revolutions, one of the most influential books in 20th century History and Philosophy of Science. In order to describe scientific evolution, Kuhn introduced the term paradigm to refer to the set of underlying assumptions about the phenomena subject to scientific study, and the methods used to investigate

7 See Feyerabend (1970:198) and his discussion of Reagan (1967) about the distinctions of basic and applied science.

them, that pervade and structure scientific enterprises. A paradigm is the epistemological framework that encompasses and shapes research in a field at a given time. According to Kuhn there is normal science and revolutionary science. Normal science refers to the period during which a well-established paradigm guides scientific research. The research conducted does not question the theoretical framework given by the paradigm, but tries to produce evidence in its support by investigating small areas of the world—what Kuhn calls puzzle-solving.

Revolutionary science refers to periods in which a consequential scientific breakthrough severely challenges the existing paradigm and brings the scientific community to question the previously accepted tenets, ultimately leading to a paradigm shift. Examples of paradigms are Ptolemaic astronomy and Newtonian physics (Kuhn, 1962:10). In this case, the former was succeeded by the latter in a paradigm shift prompted by Newton’s laws and the body of observations and mathematical proofs that accompany them, which would become the ‘exemplar’ out of which science would be modeled henceforth. Here we encounter the first problem with the Kuhnian term paradigm—its ambiguity. In the example just mentioned, Newtonian physics is a paradigm, but so are the Newtonian laws that led to it in the sense of exemplars, of paradigms as

“universally recognized scientific achievements that for a time provide model problems and solutions to a community of practitioners” (Kuhn, 1962:viii). A paradigm is not only the framework, but also the work itself that initiated and shaped the framework and so future practice. Kuhn has been criticized for using the term “paradigm” with at least 21 different meanings, some of them overlapping (Masterman, 1970:61-65), in his 1962 book, which is ironic bearing in mind that Kuhn considered paradigms to be incommensurable and self-contained, and hence mutually exclusive and unintelligible (Kuhn, 1962: xx). For the sake of my argument here,

I will stick to the first meaning unless otherwise stated, that is, that a paradigm is the framework that encompasses, informs and shapes research in a given field at a given time.

Paradigms in CTS

CTS scholars have used the notion of paradigm to describe the emergence of different research traditions in the field stemming from those of cognitive science previously discussed.

Alves (2015) points out the existence of three paradigms in CTS, cognitivism, connectionism and situated cognition. He considers that translation is an embodied, situated activity that could be studied within the embodied, situated paradigm, but he notes that the scarcity of studies developed within that paradigm may be due to the quest of CTS for the control of variables in experimental settings and neat operationalizations: “Although TPR is epistemologically related to embodied, situated cognitive processes, paradigmatically TPR needs to anchor itself on foundations that allow empirical-experimental analyses of objective data. And connectionism, I would like to argue, is a paradigm better suited to achieve that goal” (Alves, 2015:23). The author calls for further disciplinary interaction to foster “complementarity and reciprocity” and thereby strengthen CTS (2015:34).

Muñoz, one of the main proponents of the situated cognition tradition in CTS in the form of Cognitive Translatologies (Muñoz, 2010a, 2010b, 2014b) stresses the differences between situated cognition, and cognitivism and connectionism based on the conception of cognition as an emerging phenomenon that the mind develops in interaction with the environment (including interactions with other minds), the physical world and the rest of the body, and that is influenced by affective factors. According to his view, shared by other CTS scholars (see Risku, 2014;

Hubscher-Davidson, 2011), the computational models of translation processes do not represent realistic theorizations insofar as they do not acknowledge situated and extended aspects of cognition. On the other hand, other CTS scholars, skeptical about the potential of situated cognition research to yield results comparable with those from other disciplines, remain interested in the classical computational research tradition as the best way to obtain knowledge valid across levels of analysis (Shreve & Diamond, 2016).

As applied to CTS, connectionism can be considered to be an evolution of the computational postulates of cognitivism (Lörscher, 1989; Carl, 2012; Muñoz, 2017), and has scarcely been applied to CTS. Examples of application of connectionist models include a model of translator’s competence as a stimuli-based process of incremental learning (Alves &

Gonçalves, 2007).

While Alves (2015) focuses on points where CTS intersects with paradigms from other disciplines and calls for interaction, Muñoz takes up the temporary nature of paradigms –the subsequent paradigm shifts that follow “scientific revolutions,” in Kuhn’s terminology– to analyze the current status of CTS as a disciplinary move forwards. Muñoz (2010b) claims that

CTS is at “a pre-paradigmatic stage, the period before a paradigm has been broadly accepted, which seems to describe well the situation today for cognitive translatologies, where disciplinary changes seem to have been faster at generating research questions and methods, than in developing new disciplinary tenets (or challenging older ones) and ways of interpreting results”

(Muñoz, 2010b:170, original emphasis). In more recent work, Muñoz (2016c; 2017) identifies two main paradigms that, he predicts, will mark the future of CTS, computational translatologies

(which Muñoz identifies with cognitivism and connectionism) and cognitive translatologies

(based on 4EA cognition). The use of the plural form is not incidental. It alludes to the varied

complexity of views among researchers that investigate the mental processes of translators and interpreters through cognitive and psycholinguistic lenses even within the same tradition. It also makes reference to the plurality of approaches within each of these traditions in other fields, as pointed out when discussing cognitive science research traditions.8 Concepts such as effort and mental load are used within one paradigm with consistently different yet often overlapping meanings (Hansen, Chesterman, and Gerzymisch-Arbogast, 2009, Dragsted, 2012), sometimes in relation with interpreting studies (cf. Gile, 1999 and Seeber 2011), and with multifarious intents, from text profiling (Jensen, 2009), to research on the translation of metaphors (Sjørup, 2011) and post-editing (O’Brien, 2005). There is also convergence across traditions: Hansen (2010), Risku

(2013) and Ehrensberger-Dow & Massey (2014) concur from different perspectives and approaches on the need to reach beyond the laboratory settings of experimental research to grasp the extended, environmental and social dimensions of translation processes. Milestone concepts such as competence or expertise provide the backbone for many research projects in the field that emerge from distinct paradigmatic camps (Shreve, 2002; Englund-Dimitrova, 2005; Göpferich,

Jakobsen & Mees, 2009 from a cognitivist view; Muñoz, 2014a from a situated perspective). The terms computational translatologies and cognitive translatologies are instrumental as they allow for overall comparison of the main differences while acknowledging the complexity of the views themselves. Therefore, I will apply them for the sake of the discussion in the rest of this section.

The Kuhnian model of scientific progress establishes that, as a new paradigm settles, a new, paradigm-laden view settles in along with it and constraints the perspective of practitioners to what is “afforded” by the paradigm. A Kuhnian scientist does not design research projects

8 It is worth mentioning that this idea of plurality contradicts “paradigms” as explained by Kuhn. A paradigm cannot be plural or include diverging approaches other than those exemplified in the exemplar. Otherwise, normal science would yield to revolutionary science. It is precisely Muñoz’s use of the plural, together with the variety we mentioned in the previous section, that makes “paradigms” insufficient to describe the actual CTS scientific practice.

across paradigms, she does not use the methods of paradigm X to test the theories of paradigm

Y. Not even in those cases where the theories of a given paradigm better suit the phenomena under study, but the methodology allegedly offered by that same paradigm does not seem valid to the scientist for whatsoever reasons. Hence the seemingly contradicting remark by Alves

(2015:23) about CTS being “epistemologically related” to the embodied paradigm but bound to connectionism by methodological requirements. Yet, as Muñoz has contended (2016b, 2017), cognitive translatologies do not recant experimental designs or quantitative analyses. Further, proponents of cognitive translatologies do not disregard previous works in CTS conducted within the computational translatologies paradigm and vice versa. Although proponents of a paradigm do not necessarily build on the works of proponents of another one, they engage and critique concrete aspects of the computational proposals, particularly with regards to ecological validity, without denying their overall value (see Risku, 2014), which evidences interaction and mutual intelligibility. It is worth repeating that there is an asymmetry in the development of these two main traditions. While computational translatologies have informed CTS during the last two decades and have therefore been considered mainstream (oftentimes called the “classic TPR paradigm”), cognitive translatologies, being more recent, are still grappling with the definition of their theoretical statements and so with their operationalization for further empirical enquiry.

It is true that cognitive translatologies make a stronger case for the embodiment and situatedness of translation-related mental processes. The same goes for the need to elicit more realistic process data from professional, in-situ performance. But, while computational translatologists often forbear taking situatedness into consideration when designing their research, and let lab settings take their toll on ecological validity, these scholars do not

necessarily deny the relevance of these issues or deem them impossible to deal with.9

Computational and cognitive translatologies vary in their approaches and assumptions, indeed; but they seem to be more related than the Kuhnian idea of paradigm would make us believe.

Kuhn accepted that there might exist concurrent puzzle-solving traditions in a given field, but that they were ‘quasi-independent’ and were led by their own paradigms and therefore focused on different problems (Kuhn, 1962:363). As per the previous examples, the CTS community pursues some shared problems, such as mental effort or translation problem-solving patterns, whether it is from a computational or a cognitive translatologies viewpoint and so they are not

‘quasi-independent’ as per Kuhn’s definition in that regard. Also, Philosopher Paul Feyerabend

(1970:206) challenges the independence of paradigms offering a case from the history of Physics

(Bohm’s and Einstein’s competing approaches on quantum theory) that provides evidence that the existence of alternative theories in a given field necessarily entails a change in the supposedly firm appraisal mechanisms of a reigning paradigm, thus leaving us with a question applicable to our field of study: “if the existence of competing theories involves a change of argumentative style, must we not then doubt this alleged quasi-independence?” Kuhn contends that paradigms are independent and incommensurable not only in their descriptions of the object of study, but also in the standards by which theories are deemed acceptable, in the ways to assess

9 As a matter of fact, the call for papers of the CTS panel for the 8th EST Congress considers that “Classic TPR has expanded the boundaries of our knowledge of the cognitive processes involved in the act of translating, i.e., the processes involved in the reading of the source text, the reformulation of its meaning, and the production of the translation. These processes do not operate in vacuo, however. They are situated; and therefore CTS must continue to be concerned with studying the way in which external (environmental) factors impact processing, especially processes associated with translators’ increasing interactions with technology, and the way internal factors such as personality, affect, and experience influence the translation process. From a theoretical perspective, there is also the issue of reactivity, of how CTS methodology itself may affect the processes we wish to study.” Further, down on the same page: “The panel proposers [Arnt Lykke Jakobsen, Isabel Lacruz, Fabio Alves & Kristian T. Hvelplund] believe this is an opportune time to strengthen CTS by also considering how to move beyond classic CTS.” < http://bcom.au.dk/research/conferencesandlectures/est-congress-2016/panels/2-expanding-the-boundaries-of- translation-process-research/>

science. However, if, as Feyerabend hints, a theory from another paradigm imposes a change on the argumentative style, that is, in the ways theories are appraised, then the paradigm ceases to be independent.

Even if there seems to be an increasing interest in cognitive translatology, CTS scholars are moving progressively into it. The migration it is not taking place in the form of a scientific revolution in which an incomparable framework supersedes another without any prior contact, without any online consequence other than “the blurring of a paradigm and the consequent loosening of the rules of normal science” (Kuhn, 1962:84). If a paradigm shift entails a reconstruction of the established assumptions and theories to date, as well as of the standards by which we considered them valid, then the new paradigm imposes its own rationality, making the comparison of competing paradigms irrational or, as Imre Lakatos puts it, “a matter for mob psychology” (1970:178 original emphasis). Moreover, the strictures of paradigms and the isolation they force upon theories can be detrimental for further scientific progress as they reduce the scope of scientific action to the window of one paradigm—an issue that was the main concern of Feyerabend’s (1970, 1978), who maintained that normal science as the practice described by Kuhn hinders scientific discovery and that discussions of the foundations of science bolster progress.

Psychology, for example, one of the most relevant feeder disciplines for CTS, has benefitted from lively debates about fundamental tenets and paradigms (see Feyerabend,

1970:198 footnote 4; Lincoln, Lynham & Guba, 2011). For Feyerabend, scientific progress does not happen within a paradigm, he would rather have scientists badgering their own assumptions into new, revolutionary discoveries than promoting Kuhnian puzzle-solving. Feyerabend admitted that his most famous work, Against Method, where he puts forward his views on

epistemology and the philosophy of science, was meant to shock the philosophy of science community (1995). His anarchic view on epistemology advocates, among other beneficial things such as the critical appraisal of established ideologies, openness before any possible theory that may show scientific promise; but it also denies the value of any scientific method to gain new knowledge and considers science a practice close to myth (1978:295). His is a valid contribution to CTS insofar as it warns against a unique, constraining paradigm and signals the descriptive shortcomings of the Kuhnian model. There is a mismatch between the concept of paradigm, its emergence and relation with other paradigms on the one hand, and the instantiations of supposed actual paradigms in CTS on the other. In fact, Kuhn’s critics adduce a panoply of cases –mainly from the history of Physics– to illustrate one of the most consequential issues of his model—the acknowledged (Kuhn, 1962:19) rigidity of paradigms that can make the descriptive power of the model falter (Lakatos & Musgrave, 1970).

Incommensurability

According to Kuhn paradigms are mutually exclusive, which is to be expected: no one would rationally guide a research agenda based both on a statement that one considers to be an inherent truth and the opposite statement at the same time. But paradigms not only exclude alternatives, they make it impossible to confront them. Paradigms change the perception of the world, as they are a “prerequisite to perception itself” (Kuhn, 1962:113). Scientists would perceive the world their disciplines describe as a different one according to the paradigm framing those disciplines (Kuhn, 1962:111). Therefore, paradigms are also incommensurable, as noted before, which means that a researcher in one paradigm cannot fully understand what researchers

in another paradigm are doing. This bold statement, also maintained by Paul Feyerabend (1978) with certain differences, has been contested by many philosophers of science (Lakatos, 1970;

Toumlin, 1970; Laudan, 1977) and is subject of ongoing practical concern in the social sciences

(Lincoln, Lynham, Guba, 2011). 10

In CTS, papers written by Muñoz (2010b, 2016c, 2017), Shreve & Angelone (2010),

O’Brien (2013), Angelone, Ehrensberger-Dow and Massey (2015) and Alves (2015), among others, refute to a certain extent the existence of strong incommensurability: all of these publications are disciplinary discussions and comparative reviews of research that include works allegedly pertaining to the different paradigms mentioned above, therefore the authors should not have been able to compare them according to Kuhn (see Toulmin, 1970:43). Further, distinctions between paradigms are not crystal-clear, as we have seen, particularly in cases where two of them coexist (see Kuhn, 1962:ix). Although cognitive translatologies are younger than the classic, computational tradition, some of the initial claims for the adoption of cognitive translatologies (Risku, 2002), which would later inform Muñoz’s and Risku’s proposals, were published almost at the same time as works considered to be cornerstones in the computational translatology tradition, which derived from Psycholinguistics and Expertise Studies (see Alves,

2003; Shreve, 2002).

Over the years after the publication of The Structure of Scientific Revolutions, Kuhn engaged in discussions with his critics to defend his positions on incommensurability, sometimes qualifying some of his previous statements. Thus, the original shift in worldview that would make scientists see the world differently—a switch in perspective (the object of study as

10 Feyerabend (1970, 1978) argues against paradigms as conservative drawbacks to creativity that hamper progress, and questions some of their main qualities as expressed by Kuhn; hence he does not consider incommensurability to take place at the paradigm level, but at the theory level.

understood only in terms of the established paradigm explanations) with no return

(Kuhn, 1962:85)—was boiled down into a linguistic issue in later iterations (1970, 1999). Kuhn draws on linguistic determinism, quoting the Sapir-Whorf hypothesis, as Feyerabend did (1978), to equate frameworks (paradigms) to languages, and the possible ways to try to overcome incommensurability with translation. However, translation does not solve the problem of theory comparison for Kuhn: “reference to translation only isolates but does not resolve the problems which have led Feyerabend and me to talk of incommensurability” (1970:268). Languages would shape the way we understand the world, and so any strategy other than immersing oneself in the language –namely, translation– would fail to fully grasp what is expressed in that language.

Kuhn returns to Quine’s example of radical translation (1960) to elaborate on the impossibility of establishing reliable equivalents between two paradigms. This indeterminacy-based example brings along a paradox that translation scholars are familiar with: the possibility of determinist theories of language to lead to indeterminist theories of translation (see Pym, 2014:86-113 for a discussion on uncertainty and indeterminacy in TS).

When introducing the incommensurability of paradigms, Kuhn was reacting to the logical positivist view that claimed that theories can be reduced to a neutral, logical expression that would easily render them comparable (Popper, 1959). This was Feyerabend’s position as well, one that parallels the poststructuralist and deconstructionist ideas that permeated TS in the second half of the 20th century in response to models of translation that assumed the stability of meaning and its transfer (Derrida, 1985; Arrojo, 1993). Kuhn was rightly problematizing the meaning of theories, linking it to their instances of use (see Wittgenstein, 1953) and embedding theories into a closed net out of which their meanings simply ebbed away.

Incommensurability, therefore, creates terminological anomalies that, when clustered, stymie understanding. In a Kuhnian pre-paradigmatic stage, such as the one Muñoz portrayed in

CTS, new theories, models, or discoveries challenge the established paradigm and gain ground to the point of coexisting first and then succeeding after a “scientific revolution.” During the period of coexistence, the qualities of the conceptualizations referred to by terms are altered while the term itself remains the same, thus resulting in incommensurability: “Though most of the same signs are used before and after the revolution (…) the ways in which some of them attach to nature has somehow changed. Successive theories are thus, we say, incommensurable”

(Kuhn, 1970: 267).

While Kuhn was making a very consequential case about the indeterminacy of scientific communication and alerting about un-challenged assumptions on the cumulative nature of scientific growth, his strong version of incommensurability led him to a blind alley similar to

Zeno’s dichotomy paradox. As we have seen in examples from CTS, communication and comparison do exist between paradigms. In TS, for instance, Chesterman and Arrojo (2000) put forward a lucid comparison of deconstructionist and descriptive/empirical paradigms, where they neatly outline the shared ground and the points of divergence. Anthony Pym structured a whole book on translation theory in terms of paradigms, their takes on translation and the possibilities of productive theorizing:

When the paradigms clash, people are often using the word “translation” to refer to quite

different things. Debate then becomes pointless, at least until someone attempts to go

beyond their initial paradigm. Only then, when an attempt is made to understand a new

view of translation, can there be productive public theorizing. (2014:4)

What Kuhn disregarded was that willingness to compromise, to understand each other, in the scientific community. Even without resource to translating, the history of science shows that

“scientists are often adept at ‘scientific bilingualism,’ switching from one framework to another” for the sake of mutual understanding (Godffrey-Smith, 2003:92). In the specific case of TS, it can be further argued that the youth of the discipline, its interdisciplinary growth, and even the nature of the object of study, a multifaceted phenomenon that requires different explanatory tools, have made CTS scholars more disposed to compromise than scholars in other fields.

If a dictionary fails to bridge incommensurability in the absence of equivalences (see

Lakatos, 1970:179, footnote 1), we can always bring indeterminacy and uncertainty to the fore and subject them to scrutiny. Thus we could present ourselves with an opportunity to address

Ladmiral’s question with which I closed the first chapter—which are better your presuppositions or mine (Wilhelm, 2012:553)? It is beyond the scope of this chapter to discuss in which ways debates about incommensurability can benefit from translation theory; suffice it to say, though, that TS has turned uncertainty and indeterminacy (issues that often arise from terminological ambiguity) into core subjects, borrowing, precisely, from Philosophy. The German hermeneutic tradition (Schleiermacher, 1813/2012), Deconstructionism (Derrida, 1985/2012, Arrojo, 1993) or

Social-Constructionism (Kiraly, 2000) are some of the paradigms that have brought their attention to these issues. Other authors, not necessarily translation scholars, have made great contributions to the discussion; for instance, Walter Benjamin, who made a point of distinguishing “the intended object from the mode of its intention” when he analyzed the differences between the German (Brot) and the French (pain) words for bread

(Benjamin, 1923/2000:18). Indeterminacy and uncertainty have not only been topics for abstract discussion and theoretical characterization in TS. CTS scholars have found in them an

explanation for intrinsic phenomena of translation problem-solving processes, and so have studied uncertainty and uncertainty management as relevant aspects of translation expertise development (Angelone, 2010).

Moreover, as already mentioned, the indeterminacy of translation leading to incommensurability can be embraced as an opportunity to explore how our own assumptions and tenets, where many conceptual problems brew, affect theory generation. If there is a gap between the knowledge within a paradigm and the knowledge within another, “awareness of indeterminacy would be well served by any mode of translation able to extend internal knowledge as far as possible into the external sphere” (Pym, 2014:109). Again, we need to candidly put forward our theoretical apparatus to make it available for intelligibility and comparison as, for instance, Muñoz did (2010).

So much for the linguistic incommensurability of paradigms. But although Kuhn characterized it in linguistic terms, the incommensurability of paradigms is far more ingrained in them than a change in language use might make apparent. As we mentioned earlier, paradigms are also incommensurable in the standards by which they assess science. A paradigm may impose certain requirements for theories to be scientific that another paradigm may deem irrelevant. This would be the case if we accepted that scientists working within the parameters of a given paradigm cannot see beyond the limits of their framework or cannot comprehend other paradigmatic standards. This is the case to some extent, as we have seen, incommensurability does exist indeed, but it is not insurmountable and so “insofar as the protagonists of two rival points of view are successful in understanding one another, it must be the case that they share standards of rational evaluation” (MacIntyre, 1981:370).

Further, such a view on incommensurability as marking a winner-takes-all succession of paradigms would underestimate the role of methodology as a parameter for rational theory comparison and choice. “If a scientist has good grounds for accepting some methodology and if some scientific theory violates that methodology, then it is entirely rational for him to have grave reservations about the theory” (Laudan, 1977:61). I should say that, instead of a cause for incommensurability, the fact that different paradigms or traditions favor varying methodologies rather provides an opportunity precisely to collate those traditions, it provides ground for comparison: for example, any theory about the translation process that is not falsifiable is not useful to CTS scholars as it cannot be used to describe or make predictions about empirical facts.

From paradigms to research traditions

CTS scholars have availed themselves of Kuhn’s work to describe processes of scientific development and change, and the relationships between the traditions within the discipline. This work importantly identified patterns of evolution in scientific development, invited scholars to acknowledge the role of social practices in science as a community of practice, problematized scientific communication, and intimated that the acceptance of theories and methods can be relative. These are all insightful contributions that helped Philosophy of Science challenge positivist views on scientific knowledge. But Kuhn’s theory is too categorical to accurately describe CTS. In the light of the examples reviewed, we might say that he went too far in delimiting normal and revolutionary science, and the consequences of incommensurability.

Incommensurability is to be found in any kind of communication in varying degrees; yet it is a fact that incommensurability issues do not necessarily impede mutual understanding. The

change from one paradigm into another does not seem to be sudden; pre-paradigmatic stages and periods of coexistence seem to abound for extended periods of time, leading to the interaction of scholars and ideas, and to rational choices between paradigmatic proposals. While it is easy to devise a Quinean radical translation example applied to incommensurable paradigms, real interactions are usually far more nuanced and the fact remains that difficulties are frequently overcome. Then, faced with Kuhn’s solid argumentation in favor of incommensurability and rigid paradigms, we cannot but marvel at instances of inter-paradigmatic communication. Much like Zeno of Elea’s audience, when he illustrated that movement is illusory, or as the legend has it, like Galileo after recanting, we find ourselves ruminating, convinced that practice goes beyond analysis, that “yet it moves.”

Further, the Kuhnian model only offers a solution to the main problem CTS faces – underlying assumptions as an obstacle to discerning between competing theories– in cases of revolutionary science. But then, it is not an open debate that tackles the issue, but the total acceptance of an alternative.

I would like to argue that the issue lies in Kuhn’s dichotomist view of paradigms. A weaker version of paradigms and of incommensurability may prove useful as descriptive tools if we accept that their differences are not a matter of fundamental nature, but of degree. 11 Systems of belief do not wholly overlap or differ, but concur or diverge in concrete aspects, such as the construct referred to by a term or the application of a given methodology, and that range of convergences and divergences can be studied as a cline against the tertium comparationis of

11 Much influenced by Karl Popper, Lakatos (1970) proposes “research programmes” as a weaker version of paradigms, comprising a “core” of fundamental tenets and a “protective belt” of theories subject to heuristic methods of falsification. Lakatos does not solve, however, the problem of competition between theories or research programmes (see Laudan 1977:76-78 for a discussion).

conceptual performance, that is, how many problems does a theory solve and how many does it generate while doing so. By transcending binary oppositions we turn the degree of difference into a matter of study.

Although they are closer to actual instances in knowledge systems (see Pokorn,

2009:142) distinctions of degree allow for more uncertainty than discrete categorizations. So it would be instrumental to apply a metric of conceptual performance to structure a comparison that by necessity becomes a composite.

The degrees of separation can also be characterized in terms of resemblance (see

Wittgentein, 1958:32). Lincoln and Guba (2011) compared different paradigms in the social sciences, identifying the possible points of confluence among them. They introduce the notion of

“axiomatic resemblance” to refer to the “proximity” of those axioms or tenets that can be compared across paradigms. Thus, in the case of TS, Deconstructivist literary theory may have little if any axiomatic resemblance with, say, localization theories, DTS and CTS. However, different traditions within CTS could be far more easily compared since they show axiomatic elements that “strongly resonate between them” (Lincoln and Guba, 2011:174): an empirical, descriptive agenda, shared foundational agreements borrowed from Psychology, Cognitive

Science and Sociology, a pragmatic, functional conception of translation processes, etc.

The problem remains, though. CTS has already accumulated a considerable amount of evidence about the translation process, yet one of the most relevant conclusions we can extract is that “it has forced us to be more specific in our definitions of the concepts and categories used”

(Jääskeläinen, 2010:214). Appropriate definitions are necessary, and so is a methodology to identify usage inconsistencies. Otherwise, we expose CTS to a proliferation of ad hoc operational definitions of the same or similar concept, making comparison difficult in a field

where the nature of the phenomena studied, from human factors to textual variability, is already challenging.

But then, the question arises, if CTS traditions partake, with differences, of common axioms, and even share a great deal of the “paradigmatic bottom-line” with other descriptive traditions within TS, how useful is it to retain paradigms together with Kuhnian theory to describe issues of epistemology, scientific progress and conceptual performance within CTS?

Not much, I would say, particularly when paradigm as a notion prevents rational comparison.

Rather, I suggest taking advantage of a model of scientific progress that allows us to assess theories in the field in terms of their problem-solving capacity and that does not bind scientific projects to the hieratic requirements of paradigms. Such a model can be found in Larry Laudan’s research traditions and his work on conceptual problems and theory appraisal, which I will discuss in Chapter 3.

CHAPTER 3

In comparing two scientific theories for the same purpose, we seek to know which one represents an advance relative to the other in terms of predicted and explained empirical facts. Which one is more coherent with the other theories and models in the domain? Which one is simpler, more elegant? In sum, we want to discern which one sheds newer and broader light on the studied phenomena. The idea of theories representing advances, or even movements forward, relative to one another is related to the understanding of progress as the ultimate goal of science.

Philosophers of science have linked scientific success to progressive theories, and have become increasingly concerned with scientific progress over the last 60 years (Popper, 1959; Kuhn,

1962; Lakatos, 1970; Stegmüller 1976; Laudan, 1977, Kitcher, 1993).

If scientific progress marks success in competition between theories, then we need both a definition of the concept of progress and a methodological explanation on how to assess it before we embark in any kind of inter-theoretical comparison. Any definition of scientific progress, in turn, depends on a previous model of scientific development: on how we idealize the ways in which scientific disciplines emerge and evolve; how scientists do and/or should behave; what

inter-theoretical relationships and links can be established; which constructed social dynamics best describe the professional networks in concrete fields, etc.

Thus, different models of scientific development can entail different requirements for progress, as we shall see in this chapter. We need to establish a descriptive model of scientific development and define scientific progress prior to elaborating on how theories, concepts and models relate to each other and to the wider frameworks guiding research in a field. In the following sections I will briefly introduce a discussion of scientific progress and a model of scientific development that, I claim, appropriately describes the evolution of CTS. In the remainder of the chapter, I will apply Laudan’s model of scientific progress to put forward a set of criteria for conceptual comparison in CTS. Laudan defines progress as conceptual and empirical problem-solving, thus allowing us to compare the progressiveness of two theories on rational grounds: we can know which theory is more progressive if we compare its problem- solving ratio –problems solved minus problems posed– with its competing theories. This conception of progress will be applied in developing the set of criteria for the assessment of constructs in CTS.

Scientific progress

Up to the second half of the 20th century, a positivist view had prevailed that considered science to be a progressive, rational endeavor that accumulates knowledge (a view inspired in the works of thinkers such as Bacon, Descartes, Comte). After World War II, philosophers of science started to investigate scientific progress as a diachronic phenomenon, drawing evidence from case studies in the history of science to support the claim that science does not progress by

rational accumulation of new (true) knowledge on top of received (true) knowledge (Popper,

1959; Feyerabend, 1962; Kuhn, 1962). This conclusion led to the problem of recasting a definition of scientific progress and its relations to rationality and truth, where truth is minimally seen as justified beliefs about empirical facts.

Thus, if progressive (that is, good) science does not grow simply by accumulating knowledge that builds on previous knowledge, how can we distinguish good science? Popper proposed falsification of theories as a rational method for aiming at truth about empirical facts, that is: a statement about a fact is true –and scientific– if supported by empirical evidence and will remain so until a new statement appears that better explains the fact while meeting the same empirical and conceptual requirements. But the truth of a statement, as we saw in our discussion of Kuhn’s and Feyerabend’s work in Chapter 2, is relative to the theoretical assumptions and methodological standards of the framework within which that statement, or theory, is developed

(see also Quine, 1953). In fact, observation of empirical facts is permeated by the language

(theory, framework, etc.) in which we display our description, it is “theory-laden” (Hanson,

1958). The moment we abandon the general tenets ruling a framework or introduce a change in method, the “alethetic” –or truth related– virtues (Steven Horst, 2016:93) of our models or theories change or are replaced by others. How do we rationally compare two competing statements if they belong to different frameworks, to two different languages? How do we identify the more progressive one?

Given this problem, and in the absence of a neutral observational language (a language that would allow us to describe empirical facts without being bound to a previous theory of the fact, i.e., not theory-laden) that could enable inter-theoretical appraisal, Kuhn resorts to problem- solving as an indicator of mature, progressive science. For Kuhn, scientific progress develops

within established paradigms through normal science, a practice consisting of making predictions by means of empirical “puzzle-solving” (Kuhn, 1962). This move solved the problem of progressive science partially by reducing the notion of truth –as relevant to progress– to the alethetic requirements of the paradigm at hand. But it was still problematic as the model didn’t provide grounds for a rational comparison of theories and only postponed the problem of relativity: in Kuhnian terms, the success of a puzzle-solving process is solely relative to and dependent on the framework –or paradigm– that enables it and does not consider any proposal from without. In addition, since conceptual controversies are thought to be a sign of an immature science for Kuhn, progressive puzzle-solving relies mainly on empirical evidence, which can be problematic as empirical adequacy may not be sufficient to accept a theory—we can never be absolutely certain about our observations (Duheim, 1954, see Laudan 1977 for a discussion of the implications of the Duheim-Quine thesis for Kuhn’s theory). We still need a way to rationally identify scientific progress that goes beyond the limits of a given framework, that allows us to compare two frameworks against each other. We will need a model that does justice to the development of science as a field of practice and that provides for comparison between theoretical frameworks. I would like to argue that that model can be found in the comparison of the problem-solving rate of theories and constructs proposed by Laudan (1977), to which I will turn my attention in the next sections. First let us identify a model of how scientific disciplines develop as fields of practice.

Models of scientific development

By highlighting the relevance of paradigms as sets of agreed-upon theories and methods in a community of scientists at a given time, Kuhn (1962) importantly introduced a social dimension into the evolution of science. For Kuhn, science does not evolve in a vacuum where no factor other than theory falsification operates. Scientific evolution is rather the outcome of interactions in networks of practitioners and the dynamics within a scientific community.

Research communities accept or reject ruling principles and (best) practices, and do not always do so in terms of rational comparison (let us remember Lakatos’s criticism of the paradigm model, which he called “mob psychology” (1970:178)). Instead, acceptance criteria can be devised according to extra-scientific factors, such as social prestige, preeminence of established agendas, employment prospects, funding allocation, ethical concerns, or technological advancements.

The way these factors impact the development of scientific disciplines has been studied by sociologists of science, such as Mulkay (1975), who proposed three descriptive models of scientific development, “the model of openness,” “the model of closure,” and “the model of branching” to “show how social factors operating within the pure research community contribute to the development of scientific development” (1975:509). These models run in parallel to the main descriptions of the evolution of science (as a field of practice) and of scientific progress (as attaining reliable knowledge: justified true beliefs about empirical facts as analyzed through the lens of the scientific method) in the Philosophy of Science during the second half of the 20th century.

The “model of openness” describes the evolution of scientific disciplines as ruled by a set of norms (originality, communality, disinterestedness, universalism, skepticism) assumed in folk notions of scientific practice (Mulkay, 1975: 510) and is very much influenced by Popper’s ideas on confirmation, falsification and scientific progress. This model assumes that the process by which scientific progress is measured is neutral, exempt from any charge or influence alien to the scientific method. The scientific community would therefore be free of the pressures of social dynamics. Their behavior, searching for descriptions and explanations of empirical facts and submitting their findings to the professional scrutiny of their peers, corresponds to the logico- positivist stance on the assessment of scientific progress defended by Popper. Science as modeled here is an ‘open’ endeavor; there aren’t any overarching frameworks, orthodoxies or social biases other than the scientific norms mentioned above, and inter-peer criticism reigns supreme as a way to identify and evaluate progress. However, this widely accepted model of scientific practice has not been backed up with actual data from the study of scientific disciplines

(Mulkay, 1975:511). Although the characterization of science it proposes is easily accepted by scientists themselves, it is rarely the case that actual practice matches this description.

Mulkay takes up Kuhn’s reaction to these limitations as he introduces the “model of closure” to describe a tendency towards uniformity in science in terms of trainee education, methodology, topics of interest and allocation of resources. These factors, when combined, funnel scientific practice into an orthodoxy of sorts, which is shaped according to the reigning paradigm. Mulkay also acknowledges the relevance of social networks in constantly reinforcing the procedures in place and the answers already given to the central questions of the discipline at hand. The model of closure is a considerable advance towards a more accurate description of actual scientific development, where hierarchy and precedent establish pathways for the

practitioner to follow both at social and professional levels. However, this model exhibits the same issues Kuhnian theory did to account for the constant evolution of science. Despite Kuhn’s theory of paradigmatic shifts, all the factors that can possibly affect scientific evolution evolve themselves without entailing a revolutionary paradigm shift. New empirical and conceptual problems are regularly generated or found, technological innovations appear on a daily basis, and individual backgrounds or social connections among practitioners vary in ways that make scientific disciplines change. Science does not leap from one static state to another, it is mobile throughout.

Aware of this, Mulkay presents the “model of branching,” a variation of the model of closure that considers that science is –much like any other human endeavor– at once bound to conformity with the status quo and in constant evolution to move on away from it. In the light of this model, consensus on central matters is not easily achieved and entails a great deal of negotiation and viewpoint comparison and change. New approaches to theory and method replace established ones little by little. The revolutionary shifts predicted in the model of closure are considered an oddity in the model of branching and not a normal mechanism of scientific evolution—a phenomenon that unfolds under very precise circumstances, namely:

in networks where the availability of significant problems and of recognition is declining;

in networks where movement of researchers in and out is difficult, for example, due to

the need for esoteric technical skills; and in networks where cognitions are highly precise

and where, consequently, the possibility of gradual intellectual redefinition is limited

(Mulkay, 1975:523).

Some of these circumstances may apply to certain, highly specialized realms of the natural sciences such Physics, but they hardly apply to the evolution of CTS. The availability of

relevant problems has steadily increased in the recent years, finding new objects of study, for example, at the neuro-physiological level (García, 2012; Tymoczko, 2012; Diamond & Shreve,

2010; Shreve & Diamond, 2016). The scholarly recognition of CTS scholars is also on the rise, with an ongoing growth in the number of specialized panels and contributions in international TS conferences, so that what originally was a small network of practitioners has become a central subdiscipline.12 A considerable number of CTS researchers come from other fields, and if it is true that the technical skills required by empirical research and the scientific method may impose a learning curve for scholars educated in the humanities, the tools of the CTS trade (empirical design, basic statistical analysis, familiarity with the foundations of cognitive science sister disciplines) may be diverse, but are not beyond the means of an average dedicated graduate student.

Mulkay’s (1975:522) model of branching describes the evolution of a new scientific discipline from the perspectives of its practitioners and their interactions. At the inception of a new discipline, researchers from various fields seek to explore new, promising problems as their own fields decline, their personal interest in their work peters out, or the incentives to remain on their previous disciplinary tracks lose appeal for economic, professional or social reasons. When the new interests of these practitioners converge on a topic or set of problems, small communities and networks of scientists start gathering around them, initially in a dispersed way that favors overlapping of topics, methods and maybe of results (e.g., the TAP paradigm in early

CTS), and progressively in a more coordinated, structured way as channels of communication, specialized venues, publications and even university departments begin to be established. This

12 A look at the number of CTS contributions to EST or AIETIS conferences in the last five editions will reveal a growing tendency.

account, particularly the creation of communities in dialogue around shared scientific interests, derives from Crane’s influential concept of invisible colleges (1972) and corresponds to a sociological reading of Kuhn’s work. Both the model of branching and the invisible colleges provide a good explanation for the evolution of CTS in particular and TS in general. In fact,

Hermans (1999) has already applied Crane’s concept to analyze the rise and growth of

Descriptive TS as a discipline that parted with previous, purely humanistic, prescriptive work in

TS to adopt an empirical, “diagnostic” stance. Hermans might as well have been talking about one of the small scale networks Mulkay mentions as characteristic of the branching model when he introduced the notion of invisible colleges to describe the evolution of Descriptive TS .The concept fits “with almost uncanny ease”: “an invisible college constitutes a personal and intellectual network, with regular informal contacts, joint ventures and publications, frequent cross-referencing in articles and books, and, for the central players, a long-term commitment to the field and to the basic ideas” (1999:11). And that is also precisely the way CTS grew as a subfield within TS, a small group of scholars in communication with each other that established several basic ideas and problems as their main concern.

As we saw in Chapter 1, the first patterns of publication indicating an emerging new subdiscipline within TS with an interest in the cognitive processes of translation and interpretation began appearing around the early 1980’s. In the preceding years, translation and interpreting scholars coming from other disciplines such as Literary Studies, Anthropology or

Psychology (it is worth mentioning that we are only now seeing a generation of TS scholars educated in TS programs) had started to come together around their common interest on the mental processes underlying cross-language communication. The initial migration of scholars from other disciplines may have not been necessarily due to a decline in their respective fields of

practice, but to a combination of several factors. A dissatisfaction with traditional modes of inquiry in the Humanities, the new impetus of Descriptive TS and its empirical agenda, the incipient proliferation of translation training programs that would soon mushroom, especially in

Europe, and the richness of the topic, where different disciplines intersect, attracted the attention of soon-to-be early CTS scholars. By the mid-, late-1980s Wolfgang Lörscher, Barbara Moser-

Mercer, Danica Seleskovitch, or Sonja Tirkkonen-Condit, among others, had initiated fruitful careers investigating translation and interpreting cognitive processes. Cognitive TS was being established. The quick pace at which cognitive sciences and psycholinguistic methods developed in the upcoming years as well as an accelerating technological innovation rate laid the foundations of CTS, appropriating conceptual and methodological research tools and using them to analyze the translation task. As per the branching model, a confluence of initially scattered interests was emerging.

As Mulkay points out, there is an overlap in terms of the techniques and variables applied by the pioneers of a new field. We can see a CTS example in the TAP paradigm: Initially applied by Herbert & Simon (1980), it dominated most of the empirical research conducted in the 1980s and early 1990s (Krings, 1986; Kussmaul & Tirkkonen-Condit, 1995). Despite the doubts TAPs elicited as the most convenient data-gathering method for translation tasks and as a not fully reliable, non-ecologically-valid method (Toury, 1991; Bernardini, 1999), subjects’ self-reporting continued to be a common source of data in CTS in the intervening years, also as a point of support for triangulation, or the use of different methods of inquiry to answer a common research question (Jakobsen, 2003). Something similar happened when the next methodological breakthrough took place at the end of the 1990s with the invention of Translog (Jakobsen, 1999), the keylogging software that would dominate the field for many years as the data-gathering tool

of choice for capturing translation processes. Translog perfectly suited the methodological needs of a series of cognitivist postulates (sequential processing that can be analyzed in discrete phases, typing pauses evidence mental effort and translation processing chunking, experimental conditions ensure valid, reliable data…13) and hypotheses that CTS scholars then began to coherently apply in empirical research around aspects of the translation task, originating what would later be called the “Classic TPR paradigm.”

Work conducted at the turn of the century on basic ideas and concepts taken in early in the development of the field would be subsequently revisited and regarded as very important to the field, e.g., mental effort, translation unit, expertise and competence, among others attracted the attention of empirical CTS scholars who shed new light on these concepts. Either eliciting agreement or disagreement, these works would shape the CTS research agenda to come. For instance, the constructs and translation competence and translation expertise would be adopted by a majority of CTS scholars to investigate the development of translation skills in the intervening years. Some scholars would even propose their own redefinition of those constructs

(Pym, 2003; Muñoz, 2014).

Reflection about methods raised doubts about the validity of TAPs and welcomed

Translog as a powerful tool to bolster experimental applications. Empirical problems such as performance differences between subjects with varying degrees of translation experience or the phases of the actual translation process would be established as relevant (e.g., Tirkkonen-Condit

& Jääskeläinen, 2000; Alves, 2003; Jakobsen, 1999; Jakobsen and Jensen 2008; Shreve, 2002).

In sum, these early ideas and approaches helped refine a budding CTS, empirical and conceptual

13 Please note that many of these postulates are not necessarily exclusive of cognitivist translatology as a research tradition.

problems were identified that would inform future research. It would be the next generation of

CTS scholars that would find a whole body of works exclusively related to the cognitive processes of translation and interpreting. And they did so because approximately a decade before that, the original network of CTS scholars had started to gradually take on their own medium- and long-term commitments, such as areas of interest, problems worth pursuing, and suitable methodologies.

Over the 1990s, the small network of CTS scholars, originally based mainly in Germany and Scandinavia (Krings, 1986, Gerloff, 1986; Lörscher, 1991; Tirkkonen-Condit, 1989), had grown into a wider, richer community. Concepts borrowed from other fields would be applied to translation process phenomena (Gutt, 1991; PACTE, 1998) and specialized conferences and other research venues started to generate collections of articles in the form of edited books

(Danks et al. 1997). CTS as a discipline had emerged.

What would be known as the “Classic TPR paradigm,” the cognitivist research tradition, was settling in as a broadly and vaguely accepted framework, but I would like to argue that CTS did not emerge exclusively as a cognitivist field. Cognitivism surely informed most of the CTS output, but that doesn’t mean it was yet considered as a set of guiding principles, as the model of closure or Kuhn’s work might have us believe. As Muñoz (2016c:9) remarks:

Big ideas pop up unconnected but simultaneously in several places when they start to

ripen. Around the time when Shreve organized the 7th Psychology forum in

Millesbury [sic], OH (Danks et al. 1997), criticisms on received views on cognition were

being presented by younger scholars like Halverson (1996), Muñoz (1994), and Risku

(1994).

The fact is that the two main research traditions in CTS, cognitivist translatology

(identified as the Classic TPR paradigm) and cognitive translatology (based on situated and embodied cognition) did not succeed one another: cognitivism was not established as a reigning framework to be overhauled by cognitive translatology. Pretty much as the model of branching describes, they were coexistent from almost the beginning and the reasons why one has come to be considered the mainstream tradition in the field may be related as much to their undeniable scientific merit as to funding opportunities, social links, hiring prospects, and established academic prestige. It is worth noting that the development of CTS from the early 1980s up to today spans the time of a professional academic career, and so authors who established the field, mainly from a cognitivist approach, are still active and represent figures of authority in the discipline. Another factor may have been the reinforcement of cognitivist perspectives by borrowing from Psychology and Psycholinguistic cognitivist models due to a dearth of 4EA cognition models14.

The backgrounds of these “younger scholars” (the adjective may be misleading for although they are younger than some of the most reputed cognitivist authors, their most relevant work is almost contemporary) made them approach translation processes from differing traditions, theorize them under different premises, and draw models that made differing assumptions as well. So, going back to the model of branching, it is not only that CTS did not produce a revolution or a break with the preceding work in TS; the fact is that, within CTS, cognitive translatology, which has evolved, partly, out of a combination of Halverson’s (1996),

Muñoz’s (1994), and Risku’s (1994) work, does not pose a revolutionary switch, but a

14 Let us remember how 4EA cognition scholars have been more successful in critically revising cognitivist positions than in proposing alternatives.

contemporary, diverging development. As the discipline crystalized, different views concurred from as many directions as concurrent scholars started to take an interest in translation processes.

The “Classic TPR paradigm” was not a paradigm yet, and there was no previous paradigm CTS could replace.

The establishment of consensus is not necessarily a smooth, cumulative process. In many

cases it involves various changes of perspectives, redefinitions of central problems and

strong disagreement among participants. But it does not normally entail an intellectual

revolution, because there is no generally accepted orthodoxy to be overthrown.

(Mulkay, 1975: 521)

It is by negotiation that practitioners reach agreement. Scholars compare their position to those of others’ and, as they work on their own theoretical positions and on the conceptual baggage they bring to bear, reflect on the possibilities and incompatibilities of their views vis à vis the rest of the field. This kind of dialogue has been evidenced in CTS with different takes on central issues: See, for instance, Pym’s (2003) and Muñoz’s (2014) positions on competence and expertise respectively, Schäffer &Carl’s view of literal translation, explained by a monitor model

(2014), as opposed to Halverson’s (2015) explanations based on linguistic entrenchment;

Jääskeläinen’s doubts about the accepted categories of expert and professional (2010). The negotiations that follow the introduction of new ideas or problems are the hallmark of a thriving discipline and that, I would like to claim, is the point CTS is at today.

After a steady growth in terms of conferences, books, journal special issues (see Muñoz,

2014a for a survey of the last ten years) and the recent creation of a journal specialized in

translation and cognition (Translation, Cognition and Behavior15), CTS has gained momentum and visibility. The small network of scholars is now a worldwide community of hundreds of researchers interested in an array of topics ranging over bilingualism, mental processes, philosophy of science and philosophy of mind, ergonomics—and the list goes on. There coexist two main interrelated approaches that draw from different traditions, coincide in some of their commitments and clash on others. But as we have seen, none enjoys the privileged position of an established paradigm and none imposes any orthodoxy over the other. One may eventually replace the other as a more widely accepted tradition, or certain commitments taken by one of them may be accepted by the other, but it will be by the slow adaptation of the views of researchers on traditions usefulness, progressiveness and development potential. There is constant competition, not an overthrow. Science is not a matter of generational gaps, but of sibling rivalry.

For Mulkay the models of closure and branching are not mutually exclusive. The two models differ mainly in the role and need of scientific revolutions and in the strength and autonomy of frameworks. The model of branching predicts scientific revolutions as a possibility under very specific circumstances that don’t apply to CTS, and describes disciplinary evolution as a permeable process. Scientific fields depend on other, neighboring fields to evolve, particularly in the case of disciplines that are borrowing from many sister disciplines. This dependency on other fields is particularly important to TS in general and therefore to CTS evolution as well. Among the recurrent concerns of CTS scholars we find issues such as points of contact with other empirical disciplines (Malmkjær, 2000), borrowings from said disciplines

15 https://benjamins.com/#catalog/journals/tcb/main. The journal Translation Spaces, which has devoted one of its two annual issues to cognitive aspects of translation and interpreting processes, was not specifically –nor wholly– dedicated to the topic.

(O’Brien, 2013), the scope and nature of CTS interdisciplinary efforts (Muñoz, 2010b), and its relation to other non-empirical approaches such as liberal arts explorations and the fruitful frictions this may generate (Moser-Mercer, 1994).

At the level of TS as a wider field, interdisciplinarity is one of the central topics, which has attracted much attention in the last two decades and whose discussion exceeds by far the scope of this section. Suffice it to say that TS scholars have concerned themselves, among other issues, with the unity of their field (Bowker et al, 1998), theoretical integration (Martín Ruano,

2006), the need to “look outward” (Bassnett, 2012:23) to promote our own research challenges, and beyond our borders (Hermans, 2002) to understand our own ways to produce knowledge.

Besides the challenges, the opportunities posed by interdisciplinarity, such as the dialogues it can promote with scholars outside TS has also drawn attention (Gambier & Van Doorslaer, 2016).

Gambier (2012:74) pertinently advises us not to hide or misrepresent how fractioned the field is, but to focus on the issues that define the discipline in its complexity. Gambier also makes reference to Chesterman’s “consilience” (2007) as a way to provide shared ground and promote collaboration among TS scholars who may have quite diverging views. I would like to argue, however, that the kind of consilience, or concurrence of views on central topics, Chesterman advocates can hardly take place at the disciplinary level (as his use of the term, borrowed from

Wilson, 1998 seems to indicate), for the very reasons Pym (2001) suggested in his contribution to the famous debate about shared ground: all of them being valid, it is difficult to reconcile the theoretical frameworks and epistemic views of TS subfields that are rooted in very distant traditions, such as post-modernist humanism and scientific empiricism. This doesn’t mean, of course, that an engaging and productive dialogue can’t be established, or that work can’t be done that by ambition and originality straddles traditions, but it will be at the expense of

compromising either the scientific or the humanistic method and even reshaping the metaphysical commitments of each tradition (Halverson, 2001). This is of course not a TS issue.

The dramatic growth of cognitive science and experimental psychology in the last 50 years have made many of its developments reach across disciplinary boundaries. TS is by all means not the only discipline originating in the humanities that has drawn from cognitive sciences to grow into the social sciences, where concepts and theories are not subject to competition by argument only, but by empirical adequacy as well (see Mark Seidenberg 2017 on science and education).

Postmodern relativism as an approach informing post-colonial research on the translation of literature is a perfectly valid –and interesting– angle, but of very limited use for a CTS scholar. There need be a certain degree of “axiomatic resemblance” or points of contact among methodological and ontological requirements in each framework (Lincoln and Guba, 2011): we need a series of very basic agreements about the object, level, methodology and limits of research: what is the translation task, what is the level (behavior, psychological phenomena, neurological) at which we are analyzing it, how far can we go in our claims about what we have discovered about translation. These agreements are conceptual commitments, precede actual research and set what is acceptable as research by the community of practice according to the axioms of the domain (see Mulkay, 1975). Even if we agree that any empirical observation is theory-laden and that absolute objectivity is unattainable, any CTS scholar commits to the idea that reality can be explored, and that thoughts and mental processes are real; and therefore, mental processes can be explored, even if indirectly. Any research tradition that were to assume otherwise would be too far away from cognitive or cognitivist translatology to allow for a fruitful dialogue that would not entail an earthshattering rearrangement of the axioms of any of the relevant research traditions. Hence my concerns with consilience as a possible or even desirable

aim for a whole field. I would like to propose a different definition of consilience as the number of objects of different kinds explained by a theory or construct (Whewell, 1840; Thagard, 1978), a desirable attribute to which I will return in the next sections.

It is the competition of theories and concepts within and across CTS research traditions that interests me in relation to the evolution of the field. Particularly, how is that plural and unavoidable competition envisioned, how to clarify the negotiations Mulkay referred to? More importantly, how to engage in a discussion not based on a shared framework, on shared ideas or assumptions, but on a shared method of evaluation for different ideas, for, as Gile (2004:126) pointed out, “Scholars in the same TS field appear to be speaking the same language but they may be using different rules.”

The intellectual mobility that the model of branching provides allows for scientific variance in terms of methods and theories. It also provides for competition without incurring a

(necessary) revolution; at the same time, it allows for the “conservative” forces inherent to scientific conformity. Thus, it provides for a more permeable/flexible framework—one that enables the mutual influence of interdisciplinary scientific growth.

As we have seen, the model of branching suitably describes the evolution of TS and of

CTS without falling into the pitfalls of logico-positivist views –as represented by the model of openness– or into the strictures of Kuhnian orthodoxy—as visible in the model of closure.

Having a model of disciplinary evolution, we now need a model of scientific progress that accounts for problem-solving and social interaction, but that provides a comparable, and less rigid, description of frameworks so that we can pinpoint inconsistencies among research constructs and provide rules for competition. We can find that kind of framework in Laudan’s model of scientific progress.

Laudan’s model of scientific progress

In the model of branching, we find a depiction of the evolution of CTS as a discipline from a sociological perspective, describing how a community of practitioners progressively establish a field by building bridges to other disciplines while constantly negotiating their own discrepancies and conceptual differences. However, the models of scientific progress and rationality that we saw in the first section of this chapter, mainly those of the logico-positivist thinkers and Kuhn, didn’t offer a comparable explanation, being related to the models of openness and closure respectively.

Philosophers of science had proposed models of rationality and theory confirmation that they themselves acknowledged did not represent actual scientific evolution in the light of historical evidence (see Carnap, 1962: 243). The proposed methods to confirm truth and rationality in science also failed, as Kuhn (1962) and Feyerabend (1978) demonstrated from a historical perspective. It followed then that science is not necessarily a rational endeavor, that choice between theories is often irrational and that even progress is compromised as the gain of knowledge is not cumulative and every paradigm shift entails losses. Either we accepted the rigid framework of rationality requirements that didn’t match actual science, or we adopted “the general arguments of cultural relativism to the effect that science is just one set of beliefs among many possible ones (…)” (Laudan, 1977:5), a view that leads to postmodernist views and that, in terms of methodology and theory comparison, would be well described as Feyerabend’s

“anything goes” (1978).

In seeking to preserve the rationality of science as a virtue, Laudan (1977) acknowledged it is not bound to a given method of confirmation or to the unveiling of truth as contrasted

against the physical world. Firstly, because we do not corroborate hypotheses against the world, but against evidence, which is theory-laden (Hanson, 1958), and secondly, because it is ultimately impossible to corroborate or reject a hypothesis or theory without doing so with the whole system of beliefs and assumptions that support it (see Quine, 1953; Duhem, 1954). This doesn’t mean, however, that an empirical agenda doesn’t make certain metaphysical commitments, such as the idea that there is a world out there that we can explore despite our theory-laden and context-bound observation methods: “even granting that they arise only in certain contexts of theoretical inquiry, even granting that their formulation will be influenced by our theoretical commitments, it is nonetheless the case that we treat empirical problems as if they were about the world” (Laudan, 1977:15, original emphasis). This view, adopted by CTS scholars, has been linked to the philosophical stance identified as “embodied realism”

(Halverson, 2013; Muñoz, 2016).

According to Laudan’s view, in terms of truth, there is no difference between a true theory and one that is empirically adequate. We cannot then measure whether a new theory is progressive or not in comparison to a previous one by looking at how true it is. Science is not about finding truth, Laudan stresses, but about solving concrete problems. Scientific progress depends on the solution to concrete, important problems set by scientists. One of the problems with rationality and truth is that they both soon lead to situations where we cannot measure them, and we need a definition of progress that depends on solutions whose success can be measured.

If a theory solves the problems that it was devised to solve and does so in the face of another theory that failed to do that, then our first theory “shows progress” with respect to the second one. Rationality will lie on the side of choosing the most progressive theory. In this way, Laudan makes progress quantifiable at the same time that he subordinates rationality to progress. In this

model there are two kinds of problems, empirical and conceptual. The progress of a theory is measured by the analysis of the problems (both empirical and conceptual) solved and their importance, and the conceptual problems the theory introduces and their importance. The progressive effectiveness of a theory is a function of the problems solved and introduced at both levels. So, for example, if a theory, T1, about translators’ mental processes during the translation task, explains all the facts so far documented in the literature, and preexisting T2 only explains some phenomena under certain circumstances, then T1 is more progressive even if it poses some minor inconsistency issues or introduces more anomalies. However, if T1, besides explaining all the facts in the literature, is totally inconsistent with everything we know about translation, the brain and mental processing, then T2 remains the most progressive theory. It is important to stress that progress as such is inter-theoretical and comparative. In avoiding the pitfalls of confirmation and truth, we forfeit the possibility of testing the progress of a theory against the world, for it may be empirically adequate, yet not progressive. The same applies to the absolute standing of a given theory—it is impossible to determine, as it is by necessity intertwined with the rest of theories in the conceptual framework of the domain. In the following sections I will discuss the two types of problems and their implications for progress assessment as well as the overall conceptual frameworks—which Laudan calls “research traditions” and considers the unit of analysis of scientific progress. The cost-benefit analysis of how progressive a theory is, together with the analysis and clarification of research constructs, will allow us to set forth the necessary criteria to compare different perspectives or epistemic traditions.

Scientific problems, empirical and conceptual

Empirical problems Laudan considers that the success of a theory depends on whether it is the right answer to the relevant problem in the domain it was set to solve. He also argues that a theory’s adequacy as a solution to a problem is more important than determining whether the theory is either corroborated or true. This follows naturally from his view that while truth and corroboration as measurements of progress render science irrational, problem-solving allows us to identify which theories are more progressive. While Newtonian theory is largely corroborated by evidence, it is not adequate for contemporary physicists. It solves the problem of explaining a series of facts, but it exhibits irreconcilable inconsistencies with today’s Physics, and so is less progressive.

The first problems any theory faces, or in the light of which any theory is devised as a tentative solution, are empirical problems. Considered “first order problems; they are substantive questions about the objects which constitute the domain of any given science” (Laudan, 1977:15, original emphasis). Anything from the natural world which merits an explanation is an empirical problem, but not all facts are empirical problems. Only those facts known to us which are deemed worthy of being solved are empirical problems. To offer an example of relevance to TS, it is widely known that fatigue may impact the quality of a translation. But that “fact” had not been considered interesting or relevant enough by TS scholars until very recently, by turning their attention to it, they turned it into an empirical problem. It is important to note that there is a terminological discrepancy between the use of “fact” here and the way I used it in Chapter 2.

“Fact,” as used in Chapter 2, can be considered an empirical problem in Laudan’s terms as

opposed to his use of “fact” as any phenomenon or occurrence in the natural world known (or not) by and relevant (or not) to the interests of the scientific domain.

There are three types of empirical problems 1) unsolved ones—which have not been solved by a theory; 2) solved ones—which have; and 3) anomalous ones—which a given theory has solved but its competitors have not. Solved problems are obviously an indicator of empirical progress. Solutions are relative, a problem solved by a theory doesn’t mean the theory is true.

Consider, for example, the possibility of finding empirical evidence for the theory of sense

(Seleskovitch, 1978), which can then be thought to solve the empirical problem of meaning transaction in interpreting. However, that doesn’t mean that the theory is true as it has been indicated (Muñoz, 2016a:157).16 The solving of a problem, also, is always transitory and context-bound. Transitory in that other theories may emerge that offer a better explanation than the first, successful theory or construct (Einstein’s Physics, Darwinian theory, the Mercator projection or translation expertise, for that matter, are all solutions to problems that had already been (tentatively) solved in one way or another. And those previous solutions became less interesting and applicable, less progressive, and even maybe useless, with the advent of the newcomers). Context-bound in that the solution, unless otherwise evidenced, only applies to a particular construction of the problem (which certainly depends on the nature of the object and the theoretical framework wherefrom the solution derives: the effect of typing on mental effort in translators might cease to be an empirical problem worth pursuing if voice recognition spreads to the point of being overwhelmingly common practice). That is why according to Laudan no solution is definitive, but approximate: “a theory solved a problem as long as it entails even an

16 “(…) when Seleskovitch and her colleagues were faced with a notion of logical, rational, truth-conditional meaning that was indelibly stamped onto linguistic units, they shied away from challenging it and resorted to sense and deverbalization (Seleskovitch 1978; Wirl 1958, 22-32; Lederer, 2003, 9-18), with negative consequences for the field and for the development of the discipline” (Muñoz, 2016a:157).

approximate statement of the problem” (1977:22 original emphasis). As Spanish author José

Luis Cuerda cunningly put it “relativity does absolutely not exist” (2016:149).

While solved problems indicate the disciplinary focus of a domain, what the discipline is about or what scholars care about, unsolved problems are, so to speak, a-disciplinary. That is why they are detrimental to the progress rate of a theory but not so much as anomalous problems.

For instance, here is a valid empirical problem to which tentative solutions have been offered in

CTS: Does the layout of information in a CAT tool’s interface affect the translator’s decision making processes? CTS scholars (Teixeira & O’Brien, forthcoming) have already tackled this problem, posing exciting avenues of research pointing towards its resolution. But the problem underlying their question is not yet solved. We assume the answer is a yes because of what we know about cognition, translation processes and ergonomics, but we cannot yet say whether or how the interface affects decision making with any certainty. The point here is that many empirical problems like this one could well be solved by scholars working in cognitive ergonomics, or by information scientists, or by experimental psychologists, because the general problem is one of how interfaces affect decisions and it is not intrinsically about translation.

Unsolved problems can be addressed by multiple disciplines because they embody a more general problem that can be tackled from different perspectives, angles or areas of interest.

So the issue of CAT tool interfaces and translator decision-making is a more specific instance of the problem “software interfaces and user decision-making,” and solving the one may lead to solving the other.

Unsolved problems potentially belong to different camps as much as the solutions to solved problems settle topical identity and move across disciplinary boundaries to prompt further solutions. The border crossing depends, however, on whether there are enough points of contact

–resemblance among the axioms of the domains and their research traditions (Lincoln & Guba,

2011)– between disciplines so that the receiving discipline would consider the issue and its solution a relevant scientific problem.

Anomalous problems exert way more (negative) weight on the progress of theories than unsolved ones, and they are one of Laudan’s major points of departure from traditional views in the philosophy of science. Traditionally, anomalies have been considered to be a sufficient reason to discard theories. For empirical data to represent an anomaly, it needed to be logically inconsistent with the theory at hand. Laudan takes issue with this position as he contends, elaborating again on the Quine-Duheim thesis, that anomalies, while generating serious doubts about theories do not necessarily invalidate them. He points out that most theories in the history of science have exhibited anomalies. Also that the ambiguity of testing precludes absolute certainty about knowledge derived from empirical data: every empirical test concerns a set of theories needed to put forward the prediction tested, and hence it is not possible to accurately relate the data to the prediction, to locate the inconsistency, point of failure or success. This ambiguity and the appearance of any anomalies found both impel us to be cautious and have doubts about a theory, but not necessarily to abandon it (Laudan, 1977:28).17 Also, there may be the case that a theory is consistent with data for which it offers no explanation. If a second theory offers that explanation, the data automatically represent an anomaly for the first theory.

17 Laudan bypasses Quine-Duhem’s conundrum avoiding concrete falsity and therefore its allocation, and so making the consequences of anomalies reach the entire network of theories, the research tradition: “Within the problem- solving model, however, we make no assignments of truth or falsity; there is nothing in the structure of deductive logic which precludes the localization of properties such as problem-solving effectiveness. When we say that a is an anomaly for a theory T1, we are not saying that a falsifies T1 (to claim that would open oneself to Duhemian objections); rather, we are saying that a is the sort of problem which a theory such as T1 ought to be able to solve. That, of course, does not prove that T1 is false; but it does clearly raise doubts about the problem-solving effectiveness of T1 (and, for that matter, about every other Ti in the complex that failed to solve the empirical problem a)” (Laudan, 1977:43).

The weighting of empirical problems becomes thus a complex process in which we need to take into account aspects such as how central the problem is for the discipline, how general the problem is (does it encompass other problems?), the conceptual consequences of the theory or theories proposed to solve it… the whole field as well as the network of theories that sustain the field are relevant to the assessment of progress in terms of problems. A theory may exhibit a long series of anomalous problems yet be the only one in the field to solve a given problem, which should deter scientists from doing without it, but encourage them to replace it with another one.

Conceptual problems Conceptual problems are inextricably linked to the theory that engenders them and cannot exist apart from them. They are “higher order questions about the well-foundedness of the conceptual structures (e.g., theories) which have been devised to answer first order questions,” that is, empirical problems (Laudan, 1977:48).

Conceptual problems can be either internal (there are inconsistencies or vagueness within a theory) or external (when a theory conflicts with another one). Clear examples of internal problems for a theory are poorly defined concepts or lack of terminological precision. External conceptual problems arise from the cognitive relationships established between theories, which

Laudan (1977: 54) summarizes as follows:

• Entailment: T entails T1

• Reinforcement: T provides a rationale for (a part of) T1

• Compatibility: T entails nothing about T1

• Implausibility: T entails that (a part of) T1 is unlikely

• Inconsistency: T entails the negation of (a part of) T1

Of these relations, entailment represents the optimal relationship two theories may have and inconsistency the most serious of problems. External conceptual problems arise when a theory is in conflict with another, well-developed one.

Entailment: Of the cognitive relationships identified by Laudan, entailment is the only one that does not represent a conceptual problem, it is rather a strong a positive relationship between to theories. “In principle, any relation short of full entailment (…) could be regarded as posing a conceptual problem for the theories exhibiting it” (1977:54). A good example from CTS are the monitor model (Tirkkonen-Condit, 2005) and Schäffer and Carl’s (2014) model of literal translation, which entail each other; the latter is built on the monitor model’s postulate that automatic (literal) translation processing is a first mode of processing that becomes more effortful, conscious problem-solving when the monitor spots a problem.

Reinforcement: Relations of mutual reinforcement often come in the form of analogies

(Laudan, 1977:230; see also Brooke, 1970: 378 and ff.), correspondences established between different objects or their constituents, sometimes based on perceived resemblances.18 For example, the use in CTS of models from research on monolingual writing processes. The similarities between translating and monolingual writing allow to apply writing models based on pauses and text production chunks as indicators of effort to translation even when the two tasks are not the same and the results or the theoretical implications do not entail each other across disciplines (Immonen, 2006; Immonen & Mäkisalo, 2010; Dragsted & Hansen, 2008). Relations of mutual reinforcement are not a conceptual problem by necessity; but they may create mild inconsistencies, for instance, when borrowings from other disciplines are imported wholesale.

18 A full-fledged discussion of the role of analogy in scientific thinking and theory development exceeds the scope of this chapter. For a panoramic, fascinating view of analogy see Hofstadter & Sander, 2013.

Compatibility: This cognitive relationship pertains to the interdisciplinary nature of science and to the “hierarchical systems of interconnection between various sciences” (Laudan,

1977:53). When a theory or construct that should reinforce another more basic theory or construct, is merely compatible with it, then compatibility becomes a problem. An example in

CTS would be developing a theory of translation cognitive processes that is compatible with cognitive science theories but does not use any of its concepts or constructs or postulates.

Implausibility: two theories are jointly implausible “when the acceptance of either one makes it less plausible that the other is acceptable” (Laudan, 1977:52). A research agenda that proposes theories to investigate translation processes beyond information processing in experimental settings (Risku, 2010) does not entail that laboratory studies should be jettisoned, but its theories of translation and its modeled translation tasks are jointly implausible with those proposed by scholars who remained committed to experimental investigation of information processing.

Inconsistency: This is the most consequential conceptual problem. Although not necessarily the most common, logical inconsistency is the result of a contradiction between theories or their implications. In CTS, Muñoz’s model of situated expertise (2014) is inconsistent with the Monitor model (Tirkkonen-Condit, 2005): Muñoz’s model is deeply rooted in the cognitive translatology research tradition, which contends that processing is distributed and parallel. This model precludes discrete processing and so is inconsistent with alternate automatic and problem-solving processing modes.

External conceptual problems originate in cases where theories from different domains or within the same domain are in conflict, when there are discrepancies between a theory and the

methodological norms accepted by the scientific community, or between a theory and the established world view, that is, between a scientific theory and general beliefs.

The case of conceptual problems originated in methodological dissent is particularly interesting for CTS scholars, given that a major point of friction between proponents of cognitivist and cognitive translatologies is how to come to terms with the implementation of the methods used in one tradition into the other without compromising theoretical assumptions.

World view conceptual problems are also at the center of current theoretical affairs in

CTS: the models of the mind inspired in extended and embodied cognition, which assume that cognition exceeds the limits of the brain and the human body, seem to be counter intuitive, or at least way more so than more traditional, cognitivist models of mind, which generally conform to the general belief of how mind works and where it is located. Sometimes doubts about a theory that is inconsistent with the accepted world view come in as caveats by appeal to self-evident

“common sense:” “If the mind happens in space at all, it happens somewhere north of the neck”

(Fodor, 1999:69).

As with empirical problems, the weight of conceptual problems is relative, and depends on the comparison of two or more theories, on how confident we are about the empirical problem solving capacity of those theories, as well as on the potential conceptual problems those theories may pose.

Precisely because of the lack of absolute reliability of empirical data and the impossibility of totally falsifying a theory with it, as discussed earlier, “a conceptual problem will, in general, be a more serious one than an empirical anomaly” (Laudan, 1977:64, original emphasis). Laudan gives the example of how Newtonian mechanics was not attacked because it couldn’t predict how the moon moves; but because several thinkers thought it went against the

grain of the metaphysical commitments of the day. In an example closer home, the construct of competence and its acquisition has been subject of an extensive debate in CTS and been the object of a number of empirical tests (see, for instance, PACTE 2000, and the subsequent results of their longitudinal study on the subject). The scope of the empirical results is limited yet, and still the most telling criticism has come from scholars who have either revised the internal and external conceptual problems of competence (Pym, 2003) or have, in the light of those problems, directly abandoned the construct for another one (Shreve, 2002). Competence can be considered empirically adequate as it fits the data obtained in the longitudinal study conducted by the

PACTE group, but questions about competence have arisen less from empirical shortcomings than from conceptual ones. It has not been challenged for its relation to empirical data, but on a conceptual basis because it does not resolve or generates conceptual problems.

In sum, we can say that scientific progress is appraised in terms of problems solved, either conceptual or empirical, and that

The overall problem-solving effectiveness of a theory is determined by assessing

the number and importance of the empirical problems which the theory solves and

deducting therefrom the number and importance of the anomalies and conceptual

problems which the theory generates (Laudan, 1977:68).

Research traditions

There are different types of theories: sets of concrete, empirically testable statements

(what we referred to as theories in Chapter 2), and wider, more complex theories which do not lend themselves to empirical testing and comprise assumptions and implicit tenets. We will explore the latter as described in Laudan’s model in this section.

These wider, more complex theories are what philosophers of science have looked at to describe scientific progress in the second half of the 20th century, and they correspond to Kuhn’s paradigms (1962) and Lakatos’s research programmes (1970). I have already elaborated on the limitations of paradigms as the model of choice of CTS scholars in Chapter 2, and the discussion of Lakatos’s work exceeds the scope of this dissertation. Suffice it to say that research programmes are an alternative to paradigms that provide for comparison between programmes and establish an explicit set of possible relationships between the research programme and its concrete theories. Research programmes are, however, strict and rigid in the ways these relationships can be established and rely on empirical content and logic to measure progress.

Laudan shares Kuhn’s and Lakatos’s conviction about the relevance of these theories, yet models them differently under the name of research traditions, in the remaining of this section I will introduce them for their application to the study of CTS evolution.

Research traditions exhibit explicit assumptions and beliefs about a given domain, for instance, cognitivist translatology assumes that cognitive processing is bound to the brain and is sequential in nature. Therefore, the theories and models developed within that research tradition

–often implicitly– bring to bear these assumptions, for example, Angelone’s modeling of problem solving as bundles entails sequential processing (2010:19). As such, every research tradition is committed to certain metaphysical, ontological and methodological positions; we could say that a research tradition is made up of a set of entities that can represent problems and the methodology to solve those problems, incorporating them to the system of knowledge of the tradition. As opposed to paradigms, research traditions are changing, extend over time and allow competing theories to coexist within them. In fact, the model assumes that competition is the rule, rather than the exception, in science. As Laudan defines them (1977:81, original emphasis):

A research tradition is a set of general assumptions about the entities and processes in a

domain of study, and about the appropriate methods to be used for investigating the

problems and constructing the theories in that domain.

The research tradition offers its concrete theories a blueprint, a roadmap of what problems to solve and how; but it does not entail any specific concrete theories or dictate their formulation. As a research tradition, cognitivist translatology states that translators’ mental processes are a set of relevant problems, it can assume that the workings of the mind are computational and discrete, and favor a given experimental methodology, but it does not specify what a model of the translation process should look like19. That is why there may be a number of theories within the same research tradition that can contradict each other. What research traditions do is to set which phenomena are problems and which are pseudo-problems, thus delimiting the scope of the domain and the techniques available to approach solutions to those problems.

There is a good example of this in the current debate between cognitivist and cognitive translatologies about what is the fundamental unit of analysis in our cognitive subdiscipline, and about how far the translation task extends into other processes not traditionally considered as meriting a translation scholars’s concern.20 The increasing interest in ergonomics and interpersonal communication in the workplace as sources of empirical problems, for example, is generating new conceptual problems, such as opening a discussion about the convenience of

19 This, however, does not preclude the development of models of translation process based on cognitivist perspectives. The point here is that research traditions are not prescriptive. 20 Interestingly, this was one of the main topics of argument in the roundtable that closed the Translation Process Research Workshop 5 in Graz in December 2016.

revisiting already well-established constructs and models to include new problematic aspects

(Angelone & Marín, forthcoming). These situations are normal in periods of change in which scholars often straddle two research traditions, but are not exclusive to these periods. It is not unusual for research traditions to create conceptual problems for their theories and vice versa, for example on account of methodology. This does not entail that the theory ceases to belong to the research tradition or that the inconsistencies render it useless. It will depend, again, on the problem-solving ratio the theory exhibits and the ratio of competing theories.

How well a research tradition fares is intimately linked to the problem-solving effectiveness of its individual theories as much as every single theory depends on the rest of the theoretical framework—Duhem’s (1954) positions about how intertwined individual theories are to the theoretical network within which they were developed are born in mind throughout every aspect of this model. Research traditions also impose constraints on the kinds of theories that can be developed, for example by delimiting which assumptions are allowed, or which ones need further justification if they reach beyond the scope of the research tradition. The same goes for methodological choices.

Research traditions are not immutable. Anomalies can trigger minor changes and the tweaking of perspectives (Mulkay’s negotiation), but they can also change central ontological or methodological items. There is a core of basic beliefs and assumptions to them, but this core can change over time: TAPs were at the core of cognitivist translatology during its initial years and were later replaced by keystroke logging and eyetracking as methods of choice, not only because of the advantages these methods present, but because of the increasingly problematic standing of

TAPs themselves. The validity of TAPs was not only questioned because of an empirical issue

(they were not really a way to gain knowledge about translators’ mental processes, but about

their own introspection about those processes); but also because a central tenet of the information-processing tradition that sustained TAPs as a methodology increasingly fell out of favor among cognitivist CTS scholars: namely, that the thoughts of a subject performing the task are available through verbalization (Ericsson & Simon, 1980) and therefore always conscious, assuming there is some sort of overarching monitoring device external to the task that surveilles and can report on its own task processing. Preferences of one methodology over another are related to empirical problems, for example, the validity and reliability of the instruments; but can also pose conceptual problems as the one just described, which may vary from being negligible, to introduce inconsistencies with the core postulates of the tradition. Some of these core elements, though, even when changing in importance over time, cannot be modified without discarding the research tradition. While cognitivist translatology was related to TAPs in the

1980s and evolved conceptually and methodologically to consider that mental processes are not verbally available, it never ceased to postulate that there is internal representation of information in the brain. In doing so, cognitivist translatology might evolve into another research tradition.

By the same token, cognitive translatology would radically change if the idea of extended cognition were to be abandoned.

We cannot test a research tradition against empirical data, but we certainly can compare its overall problem-solving capacity to that of the next research tradition and so appraise them.

Research traditions can be assessed according to two main criteria, their adequacy and their progress. Adequacy relates to the problem-solving effectiveness of the latest theories in the research tradition, and so it is bound to a certain time, it is synchronic. Progress can be measured as the general progress of a research tradition, or the diachronic comparison of the adequacy of the initial theories and the latest. We can also obtain a progress rate by analyzing how adequacy

has changed during a specific period of time (Laudan, 1977:106-107). Scientists can choose one research tradition over another, that is, consider it more acceptable for a number of reasons not necessarily rational (let’s remember Mulkay’s description of scientific network evolution), but there is ground to affirm that it is the rational choice to accept as valid, or provisionally “better,” the research tradition that solves more problems than another. Scientists can also choose a given research tradition not so much for the problems it solves or has solved overall, but for those which it will solve or bears promise of solving, that is, for its pursuitability (Laudan, 1977:109-

111). Good indicators of pursuitability, or of “how much future” a research tradition has, are the progress or rate of progress of the research tradition. In fact, Laudan acknowledges, scientists are not bound to any research tradition and might, under certain specific circumstances work in two different research traditions: when a research tradition has a good level of acceptance but its progress has stagnated, at the same time that an emerging research tradition, perhaps with more internal problems and definitely with lower acceptance has a promising progress rate. Some scholars in CTS might see their field described in this situation. Cognitivist translatology still offers a better problem-solving record and so is more acceptable, but cognitive translatology, despite showing a shorter publishing record and showing methodological inconsistencies (such as the lack of a clear methodological apparatus with which to empirically test its hypotheses), shows a considerable growth and a good rate of progress in the last decade. This is so particularly on a conceptual level, cognitive translatology has offered new takes on the task

(human computer interaction, dynamics in the workplace), proposed new constructs and theories to better accommodate blind spots or phenomena not accounted for in previous cognitivist theories (new models of the translation process, new definitions of expertise that blur the lines between the mental and the social dimensions of translation, investigations on the role of

emotion in translation processes) and even alternative methodological recommendations (new thresholds for pause measurement in keylogging reports analysis, ethnographic methods as a methodology of choice). Cognitive translatologists, very much like embodied cognition proponents have so far been more effective at unveiling inconsistencies than at elaborating their own alternative, full-fledged research methodology (see Shapiro, 2012; Muñoz, 2016b). It seems to be the case that up until around 2010 there were very scant and widely dispersed papers focusing on extended and embodied cognition as applied to the translation process. Since then, and heavily inspired by the work of Ricardo Muñoz (2010b), an increasing number of CTS scholars, particularly Hanna Risku and Sandra Halverson, but also Ana Rojo and Marina Ramos,

Celia Martín, Adolfo García, Maureen Eherensberger-Dow and more recently Erik Angelone, seem to be at least partially interested in some of the problem-solving opportunities this research tradition has to offer. This does not mean, however, that all of them endorse at the same level

(some or all of) the postulates of cognitive translatology, or that they deny the advances cognitivist translatology poses in many aspects of research, but as we have seen, that is not a requirement for them to work on theories or contemplate problems whose solutions are inspired by that research tradition. Research traditions, unlike paradigms, do not impose an exemplar that needs be accepted as a gold standard or ruling principle, but propose a set of coherent assumptions about the domain of practice and the methodologies that best suit its exploration.

Theories and methodologies with varying degrees of inconsistency can coexist within a research tradition, in fact, research traditions allow for a relation of degree rather than of kind.

In sum, research traditions offer a more flexible account of how networks of theories develop and interact with their constituent theories. They also allow for the rational comparison of one network of theories or framework with the other. Further, research traditions provide a

useful model to describe the development and current state of affairs in CTS. Laudan’s model of scientific progress as based on problem-solving and articulated in the concepts of research tradition and empirical and conceptual problems, represents a useful improvement over Kuhnian theory, which until now has been the framework applied in TS to analyze its evolution with very limited results.

The comparative framework that I propose in this dissertation, however, is not complete.

Now that I have introduced a model of scientific evolution as a network of practitioners sharing interests (Mulkay, 1975), and a model of scientific progress that allows us to identify it rationally and thus make inter-theoretical comparisons (Laudan, 1977), I will need to specify the criteria I will be using to identify internal inconsistencies in CTS research constructs and clarify them. As mentioned in my discussion of conceptual problems, the lack of clarity of the concepts and the terminology used are fundamental issues in the appraisal of scientific progress and its comparisons. We will hardly be able to identify which construct is better or which research tradition is more progressive until we are clear about what is it that these theories propose by means of the concepts they use, their relation to empirical data and to other concepts in the same and in other research traditions.

Criteria for the evaluation of constructs

Let me begin with a terminological clarification myself. In the course of the preceding chapters I have been using “concept” and “construct” without adequate distinction. This is only because, according to the definitions used in this dissertation, constructs are kinds of concepts. A concept is an abstract idea or notion that is associated to other ideas or objects. A concept is,

according Sellars, an inferential role in a network (1963). For instance, the concept ELEPHANT is associated both to the animal and to other concepts, such as PACHYDERM. In special languages, concepts are labeled by specialized terms, related to other terms and concepts in the field of knowledge. The Oxford Dictionary of Philosophy, on the other hand, defines construct as a

“concept based either on empirical observation or theoretical argumentation that is guided by a particular framework and subject to its application in a particular research design or model.”

Thus, in this section I will talk about constructs as they are referred to by the terms in the field and guided by their research traditions, that is, how the construct EXPERTISE, is used in CTS with the term expertise.

A conceptual performance model to appraise the problem-solving effectiveness of constructs in CTS needs to draw from and be consistent with Laudan’s and Mulkay’s models of disciplinary evolution and scientific progress, and also offer a set of coherent criteria for the evaluation of constructs.

Although Laudan did not elaborate on the appraisal and evaluation of constructs –or concepts– it is not difficult to introduce them into his model of scientific progress. First, constructs are used to design research proposals and elaborate models and theories and therefore participate in the ontological and methodological requirements of research traditions. Second, constructs are models themselves –abstract representations devised to fit empirical data– and as models, they are subject to relative comparison in terms of conceptual and empirical problem- solving, not of absolute truth. Constructs are essential instruments in the elaboration of theories and models and so key elements in their capacity to solve problems, but they can also produce potential points of problematic friction. As such, constructs bring to bear both their beneficial and problematic implications wherever they are applied, and so they can potentially pose a

conceptual problem of inconsistency within a research tradition. At the same time, testing a construct empirically implies testing the rest of the constructs and theories in the complex network of beliefs and assumptions that we call a research tradition. I suggest clarity, adequacy, consistency and simplicity as the most suitable criteria we can use to evaluate research constructs and their problem-solving effectiveness.

Clarity

The clarity of concepts and constructs has been related to scientific progress since the

19th century. William Whewell (1840) already talked about the “explication of conceptions” and how concepts could be increasingly refined by scientific experience. For a construct to be clear it needs to be well defined in such an explicit way that there is no doubt as to what evidence or empirical or conceptual problem it refers to and with which assumptions. It is relevant to note that just as Whewell’s conceptions, constructs evolve, are improved and change. We only need to look at the basic concepts in CTS to see that they have been given quite many different or slightly diverging definitions. There is no use pretending that scientists stop honing constructs; in fact, it is detrimental to progress. That is why instead of imposing rigidity on them, we need to require clarity. A construct is useless if poorly defined or if defined in an idiosyncratic way that precludes further applicability, but those are not the only sources of ambiguity. Research traditions are explicit about their assumptions and so should be the use of constructs for clarity purposes, otherwise they would become problematic. Consider again Malmkjær’s hypothetical example (2000:166):

if one wants to borrow the idea that translation is indeterminate (Quine 1957-8; 1959;

1960) –perhaps to explain why several translators working with the same text typically

produce different text versions (which is not actually Quine's point at all)– then it is

useful to be aware that within its original setting, the notion of translational

indeterminacy is grounded in a combination of behaviourism, holism and empiricism,

and that it implies that there cannot be a theory of translation, in Holmes’ (1972/1988:73)

desired sense of a system which can “explain and predict what translating and

translations are and will be.”

The implications, and even the definition, of the construct undermine our purpose as they deny a core axiom of Descriptive TS (and therefore of CTS research traditions), namely, that behavioral patterns can be extracted and predictions made through the analysis of translation products and processes. There are other, less blatant examples that do not entail denying the whole research tradition, maybe not even a single one of the core elements or axioms of the research tradition, but that nonetheless introduce internal inconsistencies among the elements of a theory because of lack of clarity. If, aiming to study the acquisition of expertise in translation from the viewpoint of cognitivist translatology we use competence defined as expert behavior

(PACTE 2003), we are introducing ambiguity as competence brings along its own assumptions, which do not necessarily match with those of expertise as defined in Expertise Studies literature

(a certain number of hours of deliberate practice, the existence of informative feedback, etc.). It is not only a matter of terminological choice; it is a conceptual problem. And it is not necessary that competence be inconsistent with the axioms of cognitivist translatology. Actually, I would like to say it is not, and neither is expertise if defined according to the requirements of the

research tradition. The conceptual problem arises because there is ambiguity, because there is a breach between the definition of our object of study and that of the construct we are using to model and study the object.

Adequacy

Adequacy is easily summarized in the following question: is the construct definition adequate to the purpose for which it is being applied? A construct can be clear yet fail to solve a problem or maybe generate a new one if not correctly applied. The previous example of the use of competence is a case in point. Adequacy does not refer to the construct in isolation, but to its relation with other of the constructs in the theory or model and to the research tradition commitments. In cases in which there is only one construct available to scholars, the adequacy of the construct may be the first criterion to suffer. The result most probably will be the modification of the construct in the long run or the abandonment of it for a new one. In a highly common case scenario, the construct will have changed but not the term used to label it, see for instance the evolution of “translation competence” from a linguistic construct to a cognitive one

(Wilss, 1976, Campbell, 1991, PACTE, 2003, Pym, 2003). An obvious example in CTS are legacy concepts such as competence and now also expertise, whose adequacy has been evolving alongside their definitions, but not necessarily at the same pace. As CTS broadens and the profile of cognitive translatology becomes more defined on the one hand, and translation processes begin to be investigated from different levels of implementation on the other, the “research focus” becomes blurry (Shreve & Diamond, 2016:152) and constructs as elemental as translation task can easily become inadequate. Does the translation task include only the process of reading

the source text and typing the target text? Does it include liaising with other stakeholders? Can an experimental lexical decision task be considered a translation task without conceptual consequences?

Also, for constructs whose aim is to solve empirical problems, they need to be empirically adequate to the relevant problem. The definition of the construct or of any claims derived from it must be substantiated with evidence that the empirical problems have been solved. So, for instance, expertise, defined as consistently superior performance at a set of tasks described in a translation performance model can be apply to solve the empirical problem of how is translation expertise developed.

Consistency

Although closely linked to the two previous criteria, consistency refers to the relationship of the construct with the other items in the theoretical network where it is embedded. These relationships generate conceptual problems external to the construct, and as such come in the same varieties as theoretical external relationships: entailment, reinforcement, compatibility, implausibility and inconsistency (Laudan, 1977:54). A construct can be consistent but not adequate, as we saw in our example about using competence to analyze expertise characteristics.

Not introducing major inconsistencies with regards to cognitivist translatology competence is mainly compatible with that research tradition by omission, as we do not find explicit endorsements of cognitivism. Even when this could be an issue due to lack of strong entailment, competence remains consistent. It is however, not as adequate as expertise to empirically describe translation skills development due to its definition.

Simplicity

The more problems, either conceptual or empirical a single construct can solve, the better it is. I use simplicity instead of parsimony or economy of means given that the solution to a problem does not depend on the construct per se, but on the complex set of assumptions established in the research tradition as well. Also, if we adopt problem-solving effectiveness as a measure for rational choice, nothing prevents us from accepting a less economic construct or theory if it solves more problems or poses fewer ones. This view is in line with Thagard’s criteria for theory choice (1978) and Whewell’s (1840) notion of consilience, which I mentioned previously in quoting Chesterman (2007). Consilience, which refers to the scope and variety of elements explained by a theory, is here understood as the scope and variety of problems solved by a construct. For example, a complex construct such as metacognition can provide for a range of empirical problems (Shreve, 2009).

The usefulness of conceptual performance criteria

These criteria allow us to determine how useful and adequate constructs are with regards to our purposes, but also to identify any possible conceptual problem that might arise between the construct and the research tradition within which it is being used. When weighting the problem-solving effectiveness of a construct it is important to bear in mind that constructs are never set into a definition that escapes evolution, and that it is precisely their changing nature that accounts for scientific progress. There may be the case where different construct definitions

compete for adoption within a research tradition, or a change in a construct may trigger or be the first consequence of the emergence of an alternative research tradition, particularly if it solves previously unsolved or anomalous problems. The redefinition of matter in Einsteinian Physics is a classic example, and so can be translation performance in CTS if we accept a model that assumes social interaction is part of cognitive processing.

Constructs, as opposed to what some scholars in other interdisciplinary fields maintain

(see Morse et al. 1996 for a discussion of the evaluation of concepts in nursing) do not mature, they do not reach an established status comparable to the elements of a reigning paradigm. Their progression is a matter of degree and not of kind, and their assumptions can be made explicit, which is why incommensurability needs not be unsolvable. Strong incommensurability precludes understanding across different frameworks as the assumptions upon which theories are based color our views in a way that make them incomparable. But, if we can put forward the assumptions and compare them against each other –see example Malmkjær’s (2000) example above about the assumptions underlying Quine’s indeterminacy– we can certainly “co-measure” constructs.

In any interdisciplinary field, bringing in or exporting constructs is of the essence, a major force of progress. The migration of scientists and ideas across disciplinary boundaries is, besides, something that naturally happens in science, as we saw in the discussion of Mulkay’s model, but caution needs to be exercised to avoid inconsistencies. To repeat a fundamental claim in this dissertation, it is inevitable that multiple diverging perspectives occur at the same time, and it is also a source of progress in science, thus the need to avoid models of science that impose unity in one way or another. Any unity is to be reached by negotiation of the plurality;

but plurality without a methodology to identify inconsistencies and promote rational decisions can be detrimental.

Such a methodology provides the ground for rational comparison between constructs.

Whenever there is more than one construct for the same phenomenon or problem, either within a given research tradition, or across different ones, we can apply the methodology described here to identify which of them presents more inconsistencies and solves more empirical problems while posing fewer conceptual ones, thus identifying which one is more progressive. This problem-solving ratio, in any case, is to be obtained by analysis of the domain and the problems every construct was set to solve. In the next chapter I will apply this methodology to the comparison of expertise and competence, and I will also identify points of contact and divergence between expertise as a construct brought to cognitivist CTS from Expertise Studies and expertise as a construct consistent with the axioms of cognitive translatology.

A good indicator of how well a new construct will fare if brought into a new research tradition is the degree of “axiomatic resemblance” between research traditions (Lincoln and

Guba, 2011), or how close the assumptions of the receiving research tradition are to those the new construct presupposes. While cognitivist and cognitive translatologies share considerable ground in terms of assumptions (empirically driven research traditions that have taken in the major tenets of cognitive science), and migrating constructs from one of them may easily evolve into the other; it would take more effort to apply Derrida’s différance (1968) to a project based on either one of the translatologies. This does not mean, of course, that new constructs or definitions that might upset a given research tradition should not be introduced; on the contrary, it means that those changes should be adopted if they will not produce inconsistencies that might outbalance the gains in terms of problems solved.

In this chapter, we’ve discussed an effective model of disciplinary evolution as applied to

CTS, a model of scientific progress and a methodology on how to rationally measure it, all of which makes for an alternative to the Kuhnian model generally accepted in TS. Empirical and conceptual problems and how to appraise their solutions have been introduced, along with research traditions as more flexible theoretical frameworks. These discussions have prefaced and inspired the introduction of conceptual performance model, a set of criteria to evaluate research constructs in terms of problem solving and their relationship to their research traditions. In order to take rational decisions, we need to clarify constructs, check their adequacy and assess their standing with regards to the explicit constraints of the research tradition within which they have developed.

In the next chapter, I will put these criteria into practice in a case study analysis of competence and expertise as two of the core constructs in CTS, both for cognitivist and cognitive translatologies.

CHAPTER 4

This chapter applies the conceptual performance model introduced in chapter 3 to the construct of expertise in CTS by comparing its most relevant variants, namely translation competence, translation expertise and situated translation expertise. The general cluster of ideas around expertise has been important (as a general notion of competence) for a long time, but they have diverged and there now seem to be three competing constructs. Thus, these constructs have been chosen as a case study to illustrate the comparison of competing constructs coming to CTS from different disciplines within and across research traditions.

Expertise and its problems

Since the turn of the 21st century, there has been increasing acceptance in CTS of the construct of expertise as developed by Ericsson & Charness (1997:6), that is, as “consistently superior performance on a specified set of representative tasks of the domain that can be administered to any subject.” The borrowing of the construct from Psychology and Expertise

Studies (generally considered a research subdiscipline within psychology) was a consequence of

a research agenda that sought to empirically investigate translation performance and the cognitive processes leading to its success. Even though CTS itself has a relatively brief history,

CTS exhibited quite early on a keen interest in developmental differences in translation, for instance by comparing students, lay subjects and professionals (Krings, 1988; Jääskeläinen and

Tirkkonen-Condit, 1991; Kiraly, 1995, PACTE 2000). Against that background, expertise represented a useful tool to investigate those differences from an interdisciplinary cognitive perspective. Introduced into CTS by Shreve, who discussed the implications of the construct for empirical research on the cognitive processes of translation (2002, 2006, 2009), this theoretical construct has been widely put to the empirical test and been subject of theoretical debate (Sirén

& Hakkarainen, 2002; Göpferich, Jakobsen & Mees 2008, 2009; Mees, Alves & Göpferich,

2009; Shreve & Angelone 2010; Jääskeläinen, 2010; O’Brien 2011; Muñoz 2014). However, the assumptions behind translation expertise have often been implicit in empirical designs (Muñoz,

2014) with no clear (operational) definitions or consensus on the nature and extent of the construct so far in CTS.

Expertise, as a borrowed construct applied to translation, is not devoid of problems, related both to its definition and its application. By definition, expertise depends on the measurement of a subject’s performance on a task that represents the field of interest. These requirements pose no excessive challenge to the definition or application of expertise in domains such as chess (De Groot 1965), for instance, where the rules are explicit and the success of movements or strategies in winning a match are clear. Chess is a well-defined task where performance can easily be measured. But the same cannot be applied to translation, which is not a neatly defined task (Sirén & Hakkarainen, 2002), but a highly complex cognitive task that subsumes many other complex sub-tasks, such as reading and text production in two different

languages, strategy coordination, problem-solving procedures, and metacognitive, regulatory processes (cf. Muñoz 2014, and Shreve & Lacruz 2015). Besides, as Pym pointed out (2003), translation has evolved into a remarkably technologized problem-solving task. As such, the practice of translation is related to specific, changing and historically-bound technological applications that shape the nature of the task and impact how representative of the domain sub- tasks are. So, for instance, the nature of the task is different depending on whether we use voice recognition-enabled systems to produce the TT or not. Also, sub-task processes such as reading may differ in translating and post-editing, or new sub-tasks may become integrated into the process: file conversion and translation memory alignment were not part of the trade at the inception of CTS. Therefore, the first step to take to approach translation expertise is to identify which are the sets of tasks that are truly representative of the translation domain; which are the tasks that define translation and hence translation expertise.

This core difficulty of expertise is the result of a problem of adequacy: the definition provided for the construct is not adequate to the purpose it is applied for. When designing empirical projects, the construct tested may have been translation expertise, but the actual experimental task implied a task or set of tasks that were representative of translation sub-tasks such as lexical decision or reading, but not of translation proper. Hansen (2010) warns us against the risks of “oversimplifying translation,” of uncritically decomposing translation for the sake of subjecting it to empirical scrutiny. Shreve and Diamond (2016) also signal the importance of clarity about the translation task, about the unit and level of investigation. Not only to avoid inadequate use of research constructs, but also to allow for convergence of research outcomes at different levels, for example, behavioral and neurological. As quoted in chapter one, Gambier

(2006:36) reminded us that any approach to the study of translation is valid if the object, unit and level of analysis are clearly set forth.

The representative tasks of the domain thus need to be measured in an empirical way.

Again, this is not particularly problematic for domains such as chess or when trying to assess the performance of athletes, but assessing translators’ behaviors requires a performance model of translation, and any such model will be necessarily subject to debate as there may be as many performance models of translation as concepts of translation exist. For the purposes of our discussion here, that means that any translation performance model ultimately depends on the research tradition that hosts or inspires it.

The assessment of performance is essential to recognize and measure translation expertise, to operationalize it. As Shreve (2002) reminds us in quoting House (1997:1), any evaluation brings along a theory of the evaluated object, hence the need to start with a model of translation that would define the knowledge and skills required by and the expected outcomes of a translation task. Setting out from such a task model, a performance model could be developed

(Shreve 2006) so that actual subject performances could be assessed accordingly. Research into expertise has been mainly developed by CTS scholars interested both in expertise and in its development. Thus, it is useful to identify, qualify and evaluate behavioral indicators of expertise development. A model of translation performance should include aspects of this development process as evidenced by shifts or progressive –diagnostic– changes in indicators such as increased proceduralization, strong-problem-solution patterns, metacognitive skills application

(Shreve 2002, 2009, Muñoz 2014), differences in segmentation (Dragsted 2005) or uncertainty management procedures (Shreve & Angelone 2011, Angelone 2010). Such assessment criteria, which would need to be agreed upon by the CTS scholars’ community, bring to the fore one of

the challenges that also applies to expertise research in terms of validity and reliability of results: the need for shared methods of process research data management, coding and storage (Shreve &

Angelone 2010) and the need to address potential consistency problems between the theoretical assumptions of different research and methodological traditions.

Shreve and Lacruz (2015) propose a model of translation that not only answers to the needs stated in the previous paragraph, but that tackles two major issues in translation expertise discussions in CTS: the need to build a model in which transfer is the central item (see Pym

2003) and to acknowledge and bolster interdisciplinary integration. Following Lakoff’s cognitive commitment, it is now a long-standing requirement of cognitive TS (Shreve 1997, Muñoz 2010,

Halverson 2013) that any investigation into the cognitive processes underlying translation and interpreting must be in accordance with the models and tenets existing in other cognitive sciences, such as bilingualism, expertise studies, neuroscience, psycholinguistics or cognitive ergonomics. Only by borrowing and integrating existing models (cf. O’Brien 2013) of the subtasks of translation as a complex cognitive phenomenon can we account for the development and nature of translation expertise in a way that is empirically valid.

The model by Shreve and Lacruz stems from Shreve (2006). Shreve presents an expertise-oriented model of translation that includes linguistic knowledge in both languages, knowledge of the two cultures involved, knowledge of both source and target textual conventions, and translation knowledge, which stands for the usage and application of translation strategies and procedures. This model conveniently accounts for the asymmetry of expertise development –also pointed out by Muñoz (2014)– as the development of expertise is bound to the nature of the task (Dragsted et al. 2009) and the fields of specialty or the personal background and practice history of the subjects may vary (Jääskeläinen 2010). Also, given that

the difficulty of a task is the intersection of the task requirements with the subject’s cognitive resources (cf. Shreve 1997, Shreve 2002), it might be the case that different subjects develop different strategies and procedures to deal with lack of resources of skills, for instance, a translator may be a poor typist but nonetheless retrieve whole translated segments from memory, thus coping with the possible impact of his typing on overall task time.

Following the development of cognitive translatology, Muñoz (2014) proposes a construct of situated translation expertise where he defines expertise as a research construct to investigate observable behaviors that identify inter-subject differences. Muñoz’s approach to expertise and to the five dimensions through which expertise is observable seems to be based precisely on the difficulty of recognizing and measuring expertise. Muñoz moves away from the circular notion of expertise as a performance characteristic of those considered experts in a community of practice to use the term to refer to a construct applicable to empirical research design: a set of observable behaviors in specific dimensions of translation performance that are to be matched against a task model (2014:18).21 Muñoz constructs a three-layer model including, the task model, the component sub-tasks and the actual cognitive processes underlying the observed performance. This way he offers a possible path towards operationalization and measurement of expertise provided the observable processes at the base of the model are operationalized and empirically observed in a valid way.

From an empirical perspective, translation expertise has often been approached not as a single construct, but through the investigation of sub-constructs that intersect meaningfully with expertise. Especially, through constructs whose empirical investigation might yield evidence on

21 A stance rooted in folk conceptions of expertise Pym already took issue with (1996), and Shreve parted with in his paper on expertise in 2002.

diagnostic changes in indicators of translation expertise development. Cases in point are metacognition, self-regulation, self-awareness and uncertainty management phenomena

(Tirkknonen-Condit, 2000), which, if not unique to expert translators (cf. Angelone 2010), are key for the development of expertise and a measurable indicator thereof in their allocation and extent (Shreve & Angelone 2011). The empirical study of metacognitive skills in translation tasks can provide valid insights into expertise evolution since metacognition is present in novice problem-solving as well. The fact is that studying phenomena such as metacognition or uncertainty management is instrumental to learn about translation expertise development. But that does not mean that any of those phenomena can be part of the definition of expertise.

Defining translation expertise by pointing to the indicators or to the changes in the indicators of its development would be proposing a circular definition.

Also, approaching translation expertise indirectly, either by examining related phenomena, such as metacognition, or considering the requirements for expertise specified in the literature (Ericsson et al. 1993), can make us miss the forest for the trees. If we operationalize the definition of translation expert as someone who has been engaged in deliberate practice for at least 10,000 hours instead of doing so per her performance as assessed against a performance model of the task, we are introducing ambiguity and inconsistencies. Besides, it is not only the case that it is virtually impossible to find a valid sample in which every subject meets deliberate practice requirements, it is also that these requirements are thought to be conducive to expertise, not a guarantee of it. Still, there is general agreement on the requirements needed for expertise to develop and the conditions leading to it, which has induced some scholars to operationalize expertise by its constituent constructs, requirements or by the outcome of its development (see, for instance, Englund-Dmitrova, 2005:15).

100

According to the Ericssonian rule, becoming an expert entails 10 years of deliberate practice (practicing activities directed to improve performance) in a well-defined task. The task must pose the appropriate difficulty for the subject at her developmental stage and be accompanied by informative, relevant feedback and the opportunity to repeat the task so that self-awareness and self-regulation, that is, metacognitive skills, are enhanced (Ericsson et al.

1993; Shreve 2006).

The focus on those required characteristics as a threshold beyond which an expert is to be found has led CTS scholars not so much to look for the representative tasks in the domain, but for subjects whose acquisition history (cf. Shreve 1997) matches the above-mentioned requirements. It can, besides, lead to the assumption that professionals who show that record of practice are experts and, as Jääskeläinen points out (2010), the use of professionals as a research sample to be compared to students can be somewhat problematic, since not all professionals are necessarily experts or the field of expertise, their specialization, may be too specific vis a vis the translation task used in the research design. Let alone the difficulty in identifying a professional.

Who qualifies as a professional translator? Someone who earns a living by translating only?

Someone whose main source of income comes from activities in the language industry?

Someone who used to translate for direct clients but ceased to do so and has not translated for the last five years? These subjects might or might not be professionals depending on how strict we are when drawing the line, and yet that line would tell us very little about subjects’ performance.

Even if we come up with an operational definition of professional translator, their performance would not be comparable to expertise by default because it aims at meeting concrete requirements (the translation acceptance criteria or translation brief, client satisfaction, ethical considerations), which may coincide with those of the representative tasks of the domain

101

or not.22 If we included the case-specific task requirements in the performance and task models, we would be defeating the purpose of observing and measuring translation expertise, we would be measuring expertise at a given job description or at meeting an employer’s needs. This is clear if we look at the notion of quality, which is central to measuring the output and therefore the performance of translators in the industry: “In the marketplace, quality is not viewed as an absolute but rather is framed in terms of customer satisfaction. Thus, ‘quality’ is properly understood not as degree of excellence, but rather as an intrinsic characteristic, property or attribute that influences the ability of a product to meet a buyer’s requirements (identified needs) and expectations (unidentified needs)” (Dunne, 2012:145, original emphasis).

CTS scholars have addressed the scarcity of “textbook” expert translators by looking for indicators of performance variance at different developmental stages of translator trainees

(Göpferich 2009, Göpferich et al. 2011) or by comparing translators’ performance and approach to the task with those of non-translating bilinguals (PACTE 2008, 2011). These works have been fundamental in advancing our understanding of the translation process as well as in the application of methodologies new to CTS such as longitudinal research designs (TransComp,

Göpferich 2009). However, they haven’t advanced our ability to identify a set of representative translation tasks, to measure the progression of expertise or recognize the cognitive indicators of expertise. In the case of the competence acquisition models put forward by PACTE group and

Göpferich, this stagnation in terms of results may be due to their use of competence as a multi- componential attribute to be acquired rather than developed. Interestingly, these models spell out what translators need to know to translate, what translation takes, but not how to identify

22 Given the multifarious tasks and responsibilities that are nowadays associated with professional translation (translation, editing, post-editing, project management, desktop publishing tasks…) calling practitioners professional translators may fall short to describe actual professional performance. A possible, yet admittedly vague alternative might be “language industry professionals.”

102

superior performance at translating. Although competence is defined as expert knowledge, it is not expertise (Shreve, 2002), but a legacy concept (cf. Shreve & Angelone) that brings along connotations of its original usage as a linguistic and pedagogical term (Muñoz 2010a, and

Mayoral in Kelly 2005).

Other approaches have looked at the differences between experienced, or professional, and non-experienced, or student, translators in specific task conditions such as the use of speech recognition tools in a translation task (Dragsted et al. 2009), whose results proved the importance of being familiar with the task for performance quality, yet were not matched against an explicit performance model of the task.

Despite the efforts of scholars who have allowed us to peek into the translation process and the development of the required skills, recognizing and measuring expertise has eluded CTS researchers. Partly, because of the pervasive use of legacy concepts (e.g., “translation competence”) and the lack of consensus on translation models, operational definitions, representative tasks and methodological approaches. These differences in conceptualization and approach appear even if the basic assumptions and axioms about the object of study are shared, as evident in the apparent co-existence of competence and expertise to model the cognitive development of translation skills.23 The development of these two constructs in CTS is ingrained in cognitivism and even when they may exhibit mutual inconsistencies neither of them presents a challenge to cognitivist translatology.

But it is not always the case that constructs compete among themselves without presenting inconsistencies to the research tradition within which they were developed. The

23 These assumption include translation understood as a multifaceted task comprising several subtasks which, on their turn, prompt cognitive processes observable through behavioral indicators.

103

current situation of CTS, as we have seen in previous chapters, indicates there are two research traditions in the field that inevitably fuel the development of new constructs and shape and influence the ways existing constructs evolve and are understood. That is the case of expertise, which has been readdressed as a construct by Muñoz (2014) from the viewpoint of cognitive translatology. In the remainder of this chapter, I will apply the criteria of the conceptual performance model outlined in Chapter 3 to compare competence (PACTE 2003) and expertise

(Shreve, 2002) as two different constructs competing to model translation cognitive skills development within the same research tradition. I will then compare expertise (Shreve 2002) and

Situated Translation and Interpreting Expertise (STIE) (Muñoz, 2014) as the case of the same construct brought from another field into CTS and developed in to two different research traditions.

Competence as expert knowledge

As we have seen in previous chapters, the definition, use and application of a construct cannot be studied in absolute terms, but only in relation to the research tradition that supports its assumptions and always in comparison with other construct(s) aiming to solve the same problem(s). A model, theory or construct can be empirically adequate but still exhibit a poorer problem-solving rate as compared to another construct. In fact, a construct may be empirically adequate and yet not solve the problem it is set to solve as correctly, completely or efficiently as another construct, or maybe not solve it at all. Bearing that in mind, I will address the case of translation competence in comparison to translation expertise. Both constructs have been applied

104

in CTS to empirically investigate translators’ development and have often been considered comparable or even almost synonymous (Muñoz, 2014:6) in the literature.

The genealogy of translation competence in TS can be traced back almost to the infancy of the discipline (Wilss, 1976). Its theoretical and pedagogical implications have had a powerful impact on the field (Pym 2003). It was the construct of choice for pioneering CTS scholars to identify successful performance and set the aim of pedagogical proposals (Shreve, 1997). This long-standing construct has been common in TS since the 1990s and over the first decade of the new century, usually employed in conjunction with empirical approaches to translation process and product (Campbell, 1991; PACTE 1998; Englund-Dimitrova, 2005; Göpferich, 2008).

In his interesting review of translation competence, Pym (2003) differentiates mainly three concepts referred to by the term competence as applied to translation:

 a mode of bilingualism, open to linguistic analysis—a notion that

originates in Chomsky’s generative Linguistics (Chomsky, 1965; see discussion

also in Shreve and Lacruz 2015);

 a multi-component competence made up of sets of skills (bilingual,

extra-linguistic, knowledge of translation, instrumental and strategic

subcompetencies) that are assessable and therefore entail pedagogical and

descriptive implications, oftentimes matching competence with market demands

(PACTE, 2003; Göpferich 2009);

 and a “supercompetence” that covers the rest of

“subcompetences,” the one Pym proposes, a minimalist construct defined as the

ability to generate and choose among viable alternative TT candidates, a way to

pinpoint the essence of translation as an activity, i.e., the transfer of meaning. I

105

would like to argue that despite Pym’s doubts about expertise (1996), his

approach to competence is the one closer to expertise in the conceptual issues

identified and the alternatives proposed. Identifying the translation-specific

requirements for a successful performance and trying to describe it in the most

simple, elegant construct possible, without engaging in detailed enumeration of

sub-tasks requirements, has been a concern to be found in Shreve’s (2002) and

Muñoz’s (2014) works to varying degrees.

Among the translation competence and translation competence acquisition models in the literature, the ones that have had the greatest impact are those developed by Göpferich (2009) and the PACTE group (1998 and their research output onwards). Translation competence is misleadingly considered to overlap with translation expertise as “expert knowledge.” The construct, however, presents several differences once compared with expertise. We can argue that an expert is by necessity competent in the representative tasks of the domain –translation, in this case–; but that does not mean the constructs are interchangeable, for a novice can exhibit full competence if she is able to translate, but not full expertise: competence is “what one has to know (and by implication, what one has to learn or be taught) to become a translator” (Shreve,

2002:154), while expertise is a conceptual tool to further empirical research, to generate hypotheses and research questions about the development of skills required to attain outstanding performance in translation (Shreve, 2002:168).

Let’s see the two constructs, translation expertise and competence, in the light of the criteria of the conceptual performance model developed in Chapter 3: clarity, adequacy, consistency and simplicity.

106

Clarity

For this analysis, I will take the definition of competence that is most widely accepted in

CTS, that of the PACTE group. According to PACTE, translation competence is “the underlying system of knowledge” that enables the translator to translate (2003:47). Translation competence clearly derives from generative grammar, where competence was an innate ability, the underlying system that allowed us to perform language, hence the dichotomy between competence and performance. It is worth noting that in a Chomskian view, performance is not so interesting as a matter of enquiry, it is a consequence of the intriguing aspect—the underlying system (Chomsky, 1965). However, per its definition, translation competence is not an innate ability, but expert knowledge that can be acquired during training and practice and that “should be defined in terms of declarative and procedural knowledge” (PACTE, 2003:58). Translation competence is then the knowledge acquired through practice and/or instruction that improves performance. The definition of competence as expert knowledge is problematic, especially from the perspective of expertise as a cognitive psychology construct. As per the introduction to this chapter, experts are those who consistently exhibit outstanding performance at a given task.

Expertise is not only a combination of procedural and declarative knowledge necessary to perform a task, but a combination of hypothesized cognitive processes, knowledge and its rearrangement (Shreve, 2002; Muñoz 2014) leading to outstanding performance as observable in translators’ behavior and measured against a performance model. Translation competence is the underlying system of knowledge that makes translators able to translate, either with consistently superior performance or not. Competence is being assumed to be expert knowledge without

107

assuming the entailments of expertise as described in the literature: equating competence to expert knowledge without considering superior performance or the need for a performance model is problematic. A competent translator’s performance exhibits traces of translation- specific training or experience vis à vis lay bilinguals according to a pre-existing list of criteria of what is needed to translate (knowledge of two languages, strategic competence, heuristic competence…). She does not exhibit shifts in measurable indicators of expertise development such as size of translation units processed or uncertainty management. The translation-specific experience and training undergone by the competent translator may or may not have caused changes in expertise development indicators, but translation competence does not help us to know or measure it.

Competence is not expert knowledge, at least not in the sense of the definition of the construct as a research tool meant to generate hypotheses. In fact, the term expert as used by

PACTE seems to be closer to the functionalist experts of Holz-Mänttaäri (1984), the translator as a person socially recognized to have more knowledge about the domain than lay people do, a

“folk” notion of experts that outsiders usually have, oftentimes about liberal professionals, against which Pym (1996) has already argued in TS.24 While this kind of expertise is not at all detrimental for translation competence as a pedagogical construct, it can be problematic when using competence as a construct to hypothesize and explore expert behavior through a cognitive lens. Hence the impact on the clarity of the construct.

24 In discussing the inaccuracies of so-called economic experts’ predictions about stock market fluctuations, Taleb (2010) points at the very nature of expertise as a socially accepted status given to certain individuals without further consideration, which often results in a mismatch between (supposed) expertise and skill.

108

In sum, when applied to hypothesize and empirically investigate translators’ cognitive development and expert behavior, there are several issues that prevent translation competence from exhibiting as much clarity as translation expertise:

1. Translation competence, as a construct devised to explain and

describe translation performance derives from Chomskian competence, which is

not related to performance, but to the underlying system enabling it. Defining

translation competence as an underlying system shifts the focus away from

performance as the central point of investigation.

2. Translation competence is not “Expertise-Studies” expert

knowledge.

3. Translation competence does not allow to fit empirical data elicited

through measurement of changes in translation expertise development behavioral

indicators.

4. Not being able to directly relate expertise indicators to expertise

makes translation competence-derived empirical hypotheses problematic.

I would like to argue that, in view of the application of the term competence and the empirical problems it intends to solve, there is a lack of clarity in the definition of competence as opposed to that of expertise as constructs applied to the investigation of translators’ performance.

109

Adequacy

Taken at face value, translation competence and its sub-competencies (bilingual, extra- linguistic, knowledge of translation, instrumental and strategic) are useful tools to make explicit what translators need to know, and thus –in the spirit of laying out the learning outcomes in a syllabus– reverse engineer translation processes. The problem, again, is that while specification of desired outcomes or objectives is an efficient approach for pedagogical purposes, it may not be as convenient for empirical research. This does not mean that the competence is not valid, or that it does not merit consideration, but it means that as a research construct to investigate translators’ behavior it is not as adequate as expertise. Translation competence is deductive, it sets out from the premise that there is an underlying “expert” system of knowledge that enables concrete instances of performance, rather than inductive, as expertise is, analyzing behavior against a performance model while looking at observable patterns in the meantime to extract potential generalizable statements. It might be argued that it should not be an issue for competence to be deductive if it can be hypothesized and we can provide an operational definition of it that is empirically testable. The issue arises when the operationalization of translation competence is expert knowledge needed to translate (PACTE 2003:48). If we accept the definition of expertise, to identify expert knowledge we need to identify first what the requirements of the representative task of the domain are, what the behavioral indicators of cognitive changes are, and how to measure them. If, as competence does, we assume that expert knowledge can be defined by stating a priori the procedural and declarative knowledge needed to translate, we can do without a model of task requirements, and behavioral indicators will be

110

reduced to retrospective assessment instruments to know whether subjects meet expected outcomes, not whether there is progression in terms of cognitive development. Therefore, there is a clear gap between expert knowledge as understood in Expertise Studies and by CTS scholars borrowing from them (Shreve, 2002), and expert knowledge as operationalized competence. We could opt for the operational definition of competence as expert knowledge, taking expert knowledge for what Expertise Studies tells us it is, thus detaching it from original competence assumptions, and test it empirically. But then, the results of the test could hardly be related to translation competence. I would like to argue that the operational definition of competence does not do justice to the construct and vice versa.

Further, the operational definition assumes that anyone able to translate possesses translation competence (expert knowledge). And, at the same time, that same premise is applied to look for differences between translators with different levels of expertise, this time as a variable operationally defined as an amount of professional experience, not as consistently superior performance, or the system of knowledge that enables translating: “As there are no external criteria or standardised tests to establish expertise in translation, we started from the premise that translators with a certain amount of professional experience are more expert than those without it” (PACTE 2014:88).

It is important to note that this example is not presented here to signal an apparent faux pas in operationalization, but to bring to the fore one of the most conspicuous and pervading problems of adequacy CTS concepts have faced and continue facing. The final claims of quite interesting empirical contributions have seen their strength limited by conceptual problems behind their operational definitions, partly because “researchers have had to formulate their

111

questions in relation to feasible methods” (Alvstad, Hild & Tiselius, 2011:1), and partly because of conceptual inadequacy.

CTS literature evidences that these issues apply both to the use of competence and expertise as research constructs. For example, Englund-Dimitrova weighs in the possible impact that time of experience and training may have on competence to conclude that “when studying aspects of translator competence, a possible solution is to operationalize the concept in terms of a certain amount of experience and/or training” (2005:16). But then, she considers that the solution to the problems of identifying experts is “operationalizing expertise in translation in terms of a certain length of experience, in this study10 years or more” (2005:19).

We face two conceptual problems here:

1) In the light of the definitions of competence and expertise those

operationalizations may not be representative of the constructed phenomena, and so

can compromise the adequacy of the construct; for example, in the case of

competence just explained: operationalizing competence as expert knowledge either

compromises the definition of translation competence or of expert knowledge as

defined in Expertise Studies. Accepting an operational definition of expertise as

(years of) experience amounts to accepting that a series of modifications in cognitive

skills can be reduced to one of the conditions potentially leading to them. And not

even that, for it is not years of experience, but of deliberate practice that are relevant

to expertise development (Ericsson, 1996).

2) We end up having two different constructs with virtually the same

operational definition. However well-defined a construct is in theory, it is its

operationalization that counts when using it as a variable in an empirical test. If, faced

112

with the need to apply competence or expertise to an empirical design, we

operationally define competence as “a certain amount of experience” (2005:16) and

expertise as “a certain length of experience” (2005:19) we are testing empirically two

different constructs as if they were the same, which not only can create ambiguity

issues, but also interfere with empirical results analysis.

Beyond the potential conceptual problems that this lack of adequacy might pose, translation competence can be empirically adequate, with results substantiating the premise, for example, that lay bilinguals (availing themselves of bilingual competence) process translation tasks differently than translation students (who, in their turn, have acquired translation competence). But such empirical testing would tell us about the influence and potential success of the training programs or of years of familiarity with the task specified in the definition of competence and its sub-competencies, rather than about expertise, expert knowledge, the translation process per se or the task. Again, the adequacy of a construct is relative to that of other constructs in solving a concrete problem. Translation expertise, in comparison, can exhibit greater adequacy in terms of the exploration of expert behavior, but provided it is adequately operationalized and anchored to a task model and a performance model.

Consistency

The origin of translation competence in generative Linguistics (Chomsky, 1965) entails some possible inconsistencies. The first one appears with regards to the original Chomskian notion of competence. Originally, competence was devised to refer to an underlying, innate system that enabled linguistic performance. Translation competence, however, is used to explore

113

the performance leading to the acquisition of a system of translation knowledge that enables it. In other words, translation competence does not pertain to competence, but to performance. But this would be a conceptual problem between translation competence and the Chomskian research tradition, not necessarily with CTS and its research traditions, and therefore is of no interest here.

Translation competence models are compatible with cognitivist translatology axioms (for instance, the implied assumption that translators engage in sequential processing when translating), and it can be argued that the projects based on translation competence have thrived within that research tradition. The construct of competence presents some inconsistencies by implication with cognitive translatology. For example, aspects of the translation competence model such as the strategic sub-competence (PACTE, 2003:57) assume –without explicit reference to any indicators– that translation processing is based on a monitor model, and therefore that processing is sequential. This inconsistency, however, is not a conceptual problem: both expertise and competence as discussed in this section align with cognitivist axioms and derive their assumptions from them to solve problems.

The problems of clarity and adequacy derived from considering translation competence as expert knowledge generate inconsistency issues with the literature in Expertise Studies and

Psychology. Further, problems in operationalization both of competence and expert knowledge make comparison of empirical results across fields or even within the same discipline or research tradition hardly possible. As mentioned before, the operationalization problems and the subsequent replicability issues they may entail are not related exclusively to translation competence, but to expertise as well.

114

Simplicity

Simplicity is related to the number and variety of relevant problems solved by a single construct. In this comparison, it is worth remembering the difference in the nature of the potential empirical problems solved by translation competence and expertise in the light of the previous discussion of adequacy: while competence may solve problems about what it takes just to translate, expertise aims at generating hypotheses about how superior translation performance is developed. Adequacy issues aside, and bearing in mind that the ultimate problems translation competence and expertise are meant to solve converge on the development of successful abilities and skills applicable to translation performance, we need to look at the structure of the constructs compared as individual items.

Expertise is a simple construct, with a clear and neat definition and without a very detailed extension of elements. A composite of other constructs feed into expertise. These “sub- constructs” represent the cognitive changes and patterns that make possible the observable behavior associated with superior performance (e.g., memory, bilingual lexicon access, or metacognition). It is important to note that these are not components of expertise, but sub- constructs that model phenomena that change in their nature as expertise develops. As changes are observed in the phenomena modeled by these sub-constructs, we argue that the result of the interaction of the relevant changes of state is a performance level we can label “expert performance.”

A reason why these constructs are not constituent parts of expertise as a construct is that expertise development is assumed to be asymmetric (Shreve, 2002:157): any of those individual

115

“sub-constructs” or any combination thereof may exhibit different levels of development in concrete instances depending on the “acquisition history” of the subject at hand. A subject does not need to evidence changes in metacognitive skills, problem-solving skills, linguistic proficiency and IT savviness all at once and at the same level for her performance to be an expert one. Rather the opposite, subjects may compensate their lack of one skill with another and, at any rate, the evolution of skills will depend on the requirements faced in subjects’ previous experience.

Also, the absence of diagnostic shifts in any of those sub-constructs does not change the definition or applicability of expertise as a construct at all. While competence is intrinsically linked to its sub-competencies (PACTE, 2003), expertise remains consistently superior performance at the relevant task as observed through behavioral indicators, independently of the prominence of changes in, say, the general problem-solving skills of the subjects under study.

Competence is by definition the knowledge necessary to translate as spelled out in the sub- competencies; therefore, if you lack one of the sub-competencies you do not fully have the knowledge needed to translate and therefore you are not competent or you are in the process of acquiring the relevant expert knowledge.25 On the other hand, the definition and components of expertise remain the same disregarding the dimensions on which the cognitive changes are observable or the improvement in knowledge (either procedural or declarative) that may be entailed. In fact, that is why when the pre-requisites for expertise such as deliberate practice conditions, or the consequences of expertise such as improved processing speed are

25 I am aware that this line of reasoning could be considered a stretch by proponents of competence, arguing that they do not assume all the sub-competencies are strictly necessary to identify competence. I would like to say that, if the multi-componential is not to be assumed an enumeration of the not only sufficient but also necessary kinds of knowledge needed to translate, then the issues of ambiguity discussed in the clarity heading of this section become more problematic than a potential issue of applicability.

116

operationalized and tested we are not actually testing translation expertise or the representative set of tasks of the domain. In those cases, we are indirectly looking at expertise development by empirically investigating the pertinent sub-constructs.

However, when applied to CTS, expertise cannot be ignored. Researchers interested in expertise need a defined model of the task as well as a task performance model to operationalize expertise (Shreve 2002:152). Therefore, any problem solved by expertise solved in conjunction with, at least, two other constructs.

In principle, translation competence does not require any other construct to solve any problems, which would make it a simpler construct than expertise. Following the logic in its definition, competence is knowing translation, and it specifies in its sub-competencies the types of knowledge one must have. It is tempting to accept that you may be either in the process of acquiring that knowledge or in possession of that knowledge. But then, how do you know in which category you fall if not by measuring performance? Would we not need a translation knowledge assessment instrument? We would, indeed. Although the translation competence model does not explicitly mention either an instrument or a construct to that effect (PACTE,

2003) it is a fact that the translation products and processes need to be evaluated, even when the conditions and metrics of such evaluation are not stated and the construct sustaining them is, therefore, implicit (PACTE, 2014:89-90).

Both constructs are dependent on other constructs to be useful for problem-solving purposes, which makes them comparable in terms of simplicity. As mentioned in Chapter 3, the complexity or the number of elements in a construct are not relevant for the model of conceptual performance applied here. Of course, Occam’s razor applies to CTS constructs all things being equal, but it is problem-solving efficiency that drives any comparison in this model. I add this

117

caveat because it may be the case that the internal complexity of a construct is detrimental to its problem-solving efficiency, for example, translation competence.

The multi-componentiality of translation competence has already been addressed by Pym

(2003) as a potential issue that can blur the boundaries of relevant aspects of the task. By adding sub elements to account for every single aspect of translation, even those that are not exclusive of translation tasks, we may lose aim at the object of interest –the transfer of meaning from one language to another– or fall into commonplace if we end up being too comprehensive, e.g., problematizing basic elements of the task that are not specific to translation: basic IT literacy or linguistic processing.

In that regard, there is a risk in including sub-competencies more consequential than mere elaboration beyond necessity: “A multicomponent model, on the other hand, tends to accept complexity without critically distinguishing between means and ends. And that, with all due respect and comprehension, is a recipe for perdition.” (Pym 2003: 494).

Summary of comparative analysis

The analysis of competence and expertise according to the criteria of the conceptual performance model leaves us with the following main comparative points:

 Clarity: The definition of translation competence as expert knowledge is

unclear in the light of Expertise Studies, and so makes it problematic to derive

hypotheses and to relate the construct to empirical data on expertise indicators.

Competence seems to be closer to a functionalist definition of translator expert, which is

118

instrumental to translator training. Expertise definition, on the other hand, lends itself to empirical testing of the relevant indicators.

 Adequacy: The ambiguities in competence definition with regards to expertise as a psychological construct challenge its adequacy as a theoretical tool useful to empirically investigate translators’ cognitive development. Even the operationalization of competence as expert knowledge can generate problems given those very ambiguities.

Although inadequate operationalizations of expertise may also be problematic, expertise is overall more adequate than competence to empirically research translator expert behavior.

 Consistency: Both translation competence and translation expertise as discussed in this section are constructs embedded in the same research tradition, cognitivist translatology, and as such they are consistent with it. Inconsistencies, in both cases, appear when operationalizations fail to correctly portray the research construct or, in the case of competence, where there are preexisting problems of clarity and/or adequacy.

 Simplicity: Expertise and competence are comparable in terms of simplicity, but not in terms of economy. Even when it also needs “auxiliary” constructs such as a performance model, competence is a more complex construct than expertise, which can be problematic if this complexity makes us lose aim at the main object of study.

119

Expertise and Situated Expertise (STIE)

In the previous section, we have compared expertise as introduced to CTS (Shreve, 2002) with competence as a CTS research construct (PACTE, 2003), that is, as most widely accepted as comparable or almost synonymous with expertise. The previous comparison has been a good example of how the criteria of the conceptual performance model can be applied to compare two competing constructs which share the basic assumptions of a research tradition, in this case, those of cognitivist translatology.

In this section I will turn my attention to the comparison of two approached to the same concept, translation expertise, as consistent with two different research traditions in CTS. There are therefore two constructs that model phenomena in close terms yet exhibit inconsistencies between them.

On the one hand, expertise as introduced by Shreve to CTS is an example of a borrowed construct that has traveled across fields but not necessarily across research traditions. The cognitivist research tradition in which the Ericssonian construct was originally produced is basically comparable with cognitivist translatology and its assumptions in terms of information processing, mental representation, level of processing, and methodology of choice (see

Chapter 2). Shreve elaborated on and analyzed the potential of expertise for empirical research into translation-related cognitive processes. He showed that expertise could be a very useful theoretical instrument for CTS scholars to pursue their interests, which importantly brought CTS research closer to what had been done in other relevant fields, such as psychology. Yet Shreve was aware that the proximity attained might be misleading: he warned us against uncritically

120

applying expertise “wholesale to the translation activity” and reminded us that “It is important to assess some of the claims made by the expertise ‘experts’ against the unique nature of translation” (Shreve, 2002:169). In other words, a useful tool may not be used in the same way across disciplines, hence the recurrent appearance of task and performance models as the warhorses of expertise implementation in his work.

Muñoz (2014) elaborated on Shreve’s (2002) and Risku’s (2010) work to develop a situated construct of translation and interpreting expertise (STIE) that heeds Shreve’s call to adapt the construct to the specifics of translation as a cognitive task, while making the construct consistent with cognitive translatology. STIE derives from translation expertise as envisioned by

Shreve and it is explicitly indebted to it. Both constructs share much of the implications and structure of Shreve’s approach, but they differ mainly in the situatedness of STIE –expertise as grounded on the circumstances surrounding the task and on how it is perceived by the practicing community– and in its explicit statement of dimensions through which the scope of expertise development can be observed by measuring subjects’ performance. Also, Muñoz distinguishes his construct from previous takes on translation expertise by logically extrapolating the theoretical consequences of underscoring the importance of the task and its environment for expertise development and analysis, and by discussing the dimensions where expertise is observable. He proposes a typology of research objects (translation behaviors, translation task models, component subtasks, and cognitive processes) (2014:13) that caters to the need for linking expertise to task performance in a way germane to empirical study. He stresses the economy of the construct by considering that the improvement in processes and skill that translation expertise entails is not necessarily made up of subcomponents, which parallels Pym’s minimal approach on competence (2003). Muñoz envisions dimensions as “scopes” or

121

perspectives into complex behavior (knowledge, problem-solving skills, adaptive psychophysiological traits, regulatory skills, self-concept) that “do not presume any ‘internal’,

(fully) separate mental activity” (Muñoz, 2014:18).

When summarizing the differences of STIE with comparable constructs in the field,

Muñoz (2014:35) is also addressing several recurrent issues in the study of translation expertise that have already been mentioned in this chapter. Among other things, STIE does not contemplate experience, but rather its effects as a relevant dimension; translation-specific task aspects are put to the fore relative to linguistic aspects; the construct contemplates translators’ perceptions of the task difficulty and the relevance of differences induced by task factors. In the light of these differences, Muñoz enumerates some advantages of his construct, such as the fact that it can be applied to all kinds of translation processes and phenomena, it avoids inconsistencies originated in borrowed models of memory, and it is more economical. Also, and importantly from the point of view of research tradition comparisons, Muñoz’s STIE “attempts to redress the views on the controlled/uncontrolled, conscious/unconscious dichotomies” and

“seems internally consistent with some current understandings of cognition” (2014:35). In other words, it makes translation expertise consistent with cognitive translatology axioms about the continuity of mind and some other of cognitive translatology understandings of cognition.

Let us see now the specific differences between expertise and STIE according to the criteria of the conceptual performance model.

122

Clarity

Not surprisingly, the definitions of expertise and STIE exhibit small, subtle differences.

Shreve (2002:150) borrows Ericsson’s & Charness’s (1997:6) definition of an expert as the subject exhibiting “consistently superior performance on a specified set of representative tasks of the domain that can be administered to any subject,” which hints at expertise as the cognitive resources and skills enabling that performance. Muñoz does not offer a definition stricto sensu, but rather a description of the construct: “Translation expertise can be described as the bulk of cognitive resources and abilities leading to behaviors that yield superior performance in translation tasks” (2014:35). The definitions are, apparently, the same, it is a matter of wording. I would like to argue, however, that STIE differences noted above are present in this

“description.” While the first definition focused on the behavioral aspects –consistently superior performance– STIE focuses on an indeterminate composite of resources that enables various behaviors (mind the plural), which lead to superior performance.26 The set of three elements – cognitive resources, task-related behaviors, performance– derives from STIE typology of research objects. The description evidences Muñoz’s attempt to establish a hierarchy of levels of study for clarification purposes while introducing translation as the representative task of the domain. It is important to note, however, that the requirements of that desired hierarchy and the need to adapt Ericsson’s construct to the specifics of translation has been a constant in Shreve’s work as the continued references to his publications should have made clear this far

26 The different definitions for bulk in the Merriam Webster Dictionary revolve around the idea of a shapeless, ponderous, indivisible mass or (organized) structure: https://www.merriam-webster.com/dictionary/bulk

123

(Shreve, 2002, 2006, 2009, Shreve & Lacruz, 2015, Shreve & Diamond, 2016). But it is

Muñoz’s construct that specifically addresses those issues in a way that qualifies the existing definition, thus adding clarity to the construct.

Adequacy

Adequacy depends on how much the definition of the construct fits the purposes to which the construct is applied. In this case, both constructs, expertise and STIE, are potentially applied to the same problems, and their definitions are very close, with differences that pertain more to how the construct is modeled than to its purpose. In this sense, it is difficult to find a clear difference in adequacy if one does not investigate empirical adequacy by using specific examples.

An important point of divergence between expertise and STIE is how the environment surrounding the task is envisioned, the way expertise is situated. The different levels of specificity between the constructs also play a role here. Expertise as developed by Shreve does not specify the importance of interaction with the task environment as a central aspect of the process, but it does not deny it either. The construct omits, as it were, the environment because those aspects of the task are not of immediate relevance for the study of translation expertise cognitive processes:

124

This cognitive focus does not deny that there are socio-cultural, economic, and political

aspects to expertise and its attribution to individuals or groups, but focuses on the

empirical discovery of knowledge structures and cognitive processes that might

characterize expertise (Shreve, 2002:150).

The differences in attribution seem to tie back to the folk notions of expertise Pym concerned himself with (1996). STIE parts with expertise in the understanding of the “empirical discovery of knowledge structures and cognitive processes” as distinct and independent from the situated or interpersonal dimensions of cognitive processing. STIE assumes that translation performance and therefore expertise are indissolubly linked to their circumstances. This is not new, but something already tacitly assumed by Shreve (2002) when proposing expertise as non- transferrable and task-bound, an idea that appeared also in discussions of competence:

the nature of translation competence as it is acquired in real practice may also have much

to do with the exact types of translational tasks one is called upon to perform. Thus,

translators undergoing socialization also develop strategies for coping with specific types

of problems that are likely to recur during actual translation (Toury, 2012:251).

What is new about STIE is the core assumption that the physical and social surroundings of the task are not external influences on translators’ cognitive processes, but part of them as far

125

as expertise is situated. This entails a different approach when using either expertise or STIE to explain translational behaviors, for the interaction with the tools, the importance of ergonomic factors or of interpersonal communication all acquire a completely different status as research objects depending on which construct is used. Thus, from a situated viewpoint translators’ interaction with peers within the organizational context of a translation team would be on a par with problem-solving as a relevant dimension to study translation expertise (Muñoz, 2014:30).

STIE further entails the blurring of the line dividing ‘translation acts’ and ‘translation events’

(Toury, 2012), that is, the division between cognitive and sociological processes (see Muñoz,

2016c for a discussion). It also poses new methodological issues: if one commits to explaining all the factors of translational cognition including social dynamics and physical interaction, then you need to find a way to observe and measure these factors and their interactions as variables.

This presents an additional layer of complexity as the number of variables and their relations get increasingly intricate when the whole situation is considered. We ran the risk of stating the obvious if our empirical questions are not well defined.27 This methodological challenge may account for the scarcity of studies taking this approach so far as also noted in previous discussions of the situated and embodied research traditions (cf. Shapiro, 2012).

Again, expertise (Shreve, 2002) is not necessarily inconsistent with the view that processes are situated, but it is not explicitly committed to it either, and the task models that may be used can be inconsistent themselves: much of the empirical research on the topic has isolated performance in experimental settings (Dragsted, Hansen & Sørensen, 2009; Jakobsen, 2003;

Alves, Pagano & Silva, 2009; Jakobsen & Jensen, 2008). It is a case of mere compatibility (see

27 Pym (2003) already warned against uncritically complex models of translation competence that “could lead anywhere.”

126

Chapter 3) in which the construct does not reinforce or deny a theoretical position, which can eventually lead to inconsistency. What it certainly leads to is to a different understanding of the translation task and of the empirical problems it poses, and therefore of the analysis of empirical results. As discussed in Chapter 3, mere compatibility is not a consequential problem in and of itself, but it is not negligible either. It may impose taking positions or aligning one’s work with one research position or another depending on the claims we want to derive from our research, for example, that the immediate physical surrounding of the translator can be part of the cognitive processes in her mind.

At this stage of the conceptual analysis, adequacy can benefit from the input of empirical testing of the assumptions or implications of research constructs.

The impact on the empirical adequacy of the differences in the conceptualization of expertise and STIE can be discussed in the light of recent findings about the perception of translation expertise and the conditions for deliberate practice in translation in the workplace.

Angelone & Marín (forthcoming) present results of a small scale exploratory study indicating that the impact of peers’ feedback and of interpersonal communication is considered key to improve translators’ performance in the absence of other conditions for expertise associated with deliberate practice, such as the repetition of the task, the increasing level of difficult or the informative feedback. Their paper is a pilot study aiming at tentatively exploring expertise from an emic perspective, and validating a measurement instrument for that purpose. And still, it poses an intriguing set of questions for expertise and STIE adequacy: social interaction and peer- to-peer communication seem to be the coping mechanisms to enable performance improvement in the absence of the hypothesized conditions for translation expertise development, so it is only natural to ask: Are social interaction and peer-to-peer communication part of the phenomena

127

potentially conducive to translation expertise? Is interaction with the surrounding environment part of the cognitive resources called for to develop expertise or just external aspects that shape expertise acquisition? Further research needs to be conducted to properly answer these empirical questions, but the existing literature seems to indicate that STIE, by explicitly committing to an empirical problem expertise is just compatible with, may be more applicable in exploring the role of the task environment in translators’ performance.

Consistency

Expertise could be discussed according to as many different perspectives and definitions as instances of the construct we found (cf. Muñoz, 2014:7). This challenges consistency, of course, although not necessarily consistency with a more basic theoretical framework or research tradition. Many of the different takes on expertise or expert knowledge (Shreve, 2002; PACTE

2003; Englund-Dimitrova, 2005) may present inconsistencies among themselves and still fit well into cognitivist translatology. In the case of expertise and STIE there is a clear alignment with cognitivist and cognitive translatology respectively. As apparent throughout this discussion,

STIE develops some aspects that expertise only hints at, such as the relevance of interpersonal communication, and offers alternative solutions to matters already addressed by expertise; for example, how to structure the different levels of observation and analysis.

Expertise is an “empty” construct in the sense that it depends on a performance model, on a model of the task, and on models of the cognitive processes and resources that enable superior performance. This “emptiness” makes expertise an extremely ductile and flexible construct that can be applied to virtually any task as modeled in virtually any way. Therefore, expertise as

128

envisioned by Ericsson (1996) and implemented by Shreve is a tremendously strong construct in terms of consistency. In principle, it is like water poured into a vase: the construct will assume the constraints of the research tradition that adopts it just like liquids adopt the form of their receiving vessel. The more evolved and specific expertise construct becomes, the more consistent with the research tradition it is, and therefore the less flexible it is. The liquid solidifies.

The tensions between general and task-related features of expertise arise (Muñoz,

2014:10). Expertise is consistently associated with a series of phenomena and changes in cognitive processing across domains, such as problem-solving capacity, task- and self-awareness and monitoring, increased processing speed, ability to successfully implement informative (self) feedback… These might be considered generic features of expertise, as Muñoz points out referring to Sirén & Hakkarainen (2002:71) who, in a seminal paper, envisioned “expertise in translation as a specific case of expertise in general.” Then, the task at hand imposes its own features in the form of requirements. In the case of translation, those may include bilingual proficiency or knowledge of textual and cultural conventions in the target language and culture.

But they may also include desktop publishing skills, terminology management, knowledge of subtitling procedures, or note-taking for interpreting purposes. The generic features will be mediated by the task-specific ones, and the more prominence these have in our use of the construct, the more concrete it will be as well. There is therefore a tension resulting from the need to account for task-specific items while still trying to pinpoint identifiable characteristics of expertise in general. That can account for the difficulty to find “textbook” experts, and that is why even more specific instances of the construct, such as STIE, are thought of as “a bunch of

129

tightly knit constructs linked by family resemblances, rather than a unitary construct” (Muñoz,

2014:10).

That is the way expertise and STIE have developed and stand in terms of consistency.

The original pliability of expertise made it possible to bring it from psychology and successfully apply it to cognitivist CTS. Thanks to that same flexibility, expertise could be taken and used by cognitive translatologists, in some instances adapting it to specific purposes –the need to account for the task environment in a new way, for example– to the point of being reshaped into a markedly cognitive translatology construct, STIE. Any gain in consistency can entail potential consistency losses. So, from a cognitive translatology viewpoint, what expertise lost in consistency towards the research tradition in the process was gained by STIE, and vice versa from the perspective of cognitivist translatologists (or any potentially alternative research tradition).

Simplicity

The flexibility of expertise might be related to an increased consilience. Being

(potentially) consistent with more frameworks or theories should open more avenues to solve more numerous and varying problems. However, the less specific a construct is, the more complex or research tradition-heavy the auxiliary constructs need to be, and those (the task model, the performance model…) may introduce their own inconsistencies and, more importantly, restrictions with regards to the explained object. In that sense, STIE is more specific, more consistent with its research tradition –and therefore more inconsistent with others– and more consilient, since it includes a typology of objects of study and the dimensions through

130

which expertise can be analyzed. That it is more consistent and simpler, does not mean that STIE is more economical than expertise, which I would argue it is not. Again, economy, while being an important overarching best practice, is less important for problem-solving effectiveness than simplicity. Muñoz argues that STIE has been pared down to bare bones and represents a minimal approach (2014: 17), which is true if we consider that it adds new explanations without necessarily multiplying the constituent elements of expertise.

Summary of comparative analysis

Using the conceptual performance model to compare competing constructs can shed light into their validity for our purposes. For example, in the light of the review in this section, expertise seems to be a more suitable construct than competence for empirical research on the cognitive processes of translation; and STIE seems to better accommodate empirical questions about the role of the translation task environment in cognition. At the same time, potential issues such as inconsistencies between constructs and research traditions or sources of ambiguity can be more easily identified.

The implementation of the criteria discussed here is an illustration and by no means an exhaustive analysis of the potential and problems of the research constructs discussed. Any analysis of conceptual performance is necessarily bound to the concrete research tradition, hypotheses and empirical questions formulated, as they will determine which factors are considered in making the case for one construct or another. It is also because of that relative applicability that the results of a comparative analysis as the one conducted in this section cannot be considered as absolute, but as props to make rational, progressive theoretical choices.

131

CHAPTER 5

Conclusions

The discussion of the case studies in Chapter 4 provides evidence for how a conceptual performance model can be instrumental in identifying hidden assumptions, ambiguity, potential inconsistencies and lack of adequacy in the use of research constructs. It also shows how the applicability and validity of a construct depends on its problem-solving capacity as compared with another construct.

The model presents a methodology to clarify constructs even when those constructs have had considerable longevity in the discipline. It does not matter how clear, how neatly fixed the definition of a construct is in principle—usage is flexible, and unstated assumptions can reshape the construct every time it is used in a study. As we saw in Chapter 4, competence and expertise, clearly different constructs, have been assumed to be comparable to the extent of being operationalized as the same conceptual entity (see Englund-Dimitrova, 2005). One and the same construct may have different implications, even mutually exclusive ones, if developed and used based on the axioms of one research tradition or another, as we saw in the discussion of expertise and STIE. This bears clear implications for conceptual clarity and empirical data interpretation, which underscores the importance of being explicit about what we mean by our particular uses of

132

research constructs and what cognitive models relate to that usage—“our most important endeavor in the next decade will and should be to clarify the model of the mind we adhere to in every single research project” (Muñoz 2017:556). In order to present a comparative methodology as the conceptual performance model, I needed first to depart from a model of scientific progress that did not allow for the comparison and competition of theories and research traditions without entailing a paradigm shift, that is, an absolute overturn of one theoretical system by another leading to the abandonment of the first one (Kuhn, 1962). Although critiques to the Kuhnian model have been common currency in Philosophy of Science quarters since the 1970s (Toumlin,

1970; Masterman, 1970; Lakatos, 1970; Laudan, 1977), the concept of paradigm has been, and still is, at the core of the epistemic discussion in CTS (Alves, 2015; Muñoz, 2010; Muñoz, 2017;

Halverson, 2017 to cite just some of the most recent examples). In that regard, introducing a new model, such as Laudan’s research traditions, is a contribution to the debate about the evolution of

CTS and of TS in and of itself. Most congruently, that contribution does not represent a

“revolutionary shift,” but a change in perspective that does not fully break away with Kuhn’s work, but that redefines it to accommodate the realities of how science progresses in actual practice (see the discussion of Mulkay, 1975). Instead of trying to make CTS progress fit the

Kuhnian model, Laudan’s idea of progress as based on problem-solving is certainly Kuhnian in origin, and even the plurality of perspectives and competing research traditions that the model affords is not that distant from Kuhn’s views on revolutionary periods. As Hasok Chang (2012) pointed out, scientific plurality is not dramatically different from what one might expect of a

Kuhnian period of revolutionary science, with the difference that that plurality is not here assumed to be a sign of immature science (see Longino, 2002:188).

133

The main points of departure from Kuhnian views in this dissertation are:

 Incommensurability is not unsurmountable

 Scientific unification is not considered an aim per se

 Normal science, research uncritically accepting the theoretical

monopoly of a paradigm, is no longer deemed an ideal scientific practice

 Problem solving is not restricted to normal science confines

 “Kuhnian losses” are not necessary, for there is no need abandon

old theories to embrace new ones across the board, but depending on a process of

case-by-case comparative assessment.

These differences entail some important implications beyond the immediate benefit of allowing comparison of theories and research traditions. I will discuss these implications in the next section.

The criteria for the performance model, clarity, consistency, adequacy and simplicity, are inspired in and derived from the work of Larry Laudan (1977). They conform to the main traditional epistemic virtues that are broadly accepted in Western History and Philosophy of

Science and that, as such, apply as general rules when creating and analyzing scientific constructs.28 CTS is no exception, but as in every field of practice, its practitioners have favored some epistemic virtues over others. As pointed out when discussing Mulkay’s models of disciplinary evolution, the values, aims and best practices of a research community vary. Relying on Laudan, who had also pointed out this variation (1984), Longino concludes that “The so- called epistemic virtues, then, are really, at best, standards around which a cognitive community

28 Kuhn, for instance, mentioned simplicity, consistency, fruitfulness, and scope (1977:322)

134

can coalesce, standards that its members adopt as theirs, but not standards that hold universally

(2002:185).”

In the case of CTS, empirical adequacy and methodological consistency seem to have been the central concern of researchers, sometimes to the detriment of discussions about theoretical clarity and potential inconsistencies brought along by unchallenged borrowings.29

I would like to argue that, by highlighting the relevance of these epistemic virtues as criteria, the conceptual performance model contributes to the disciplinary discussion in CTS while providing a methodology for theoretical comparison. Even when debates about the objectives and tools, both theoretical and empirical, of CTS are as old as the field itself, no specific methodology for the assessment of constructs and their comparison with regards to these epistemic virtues has yet been proposed. Besides, the criteria jointly contribute to another epistemic virtue not included in the model: that of the principle of charity. It “constrains the interpreter to maximize the truth or rationality in the subject’s sayings” (Blackburn, 2016), that is, avoiding misrepresentation of the speaker’s statement or opting for irrational interpretations when a rational one is available. The conceptual performance model allows us to unpack construct use, to look at the assumptions and at the empirical applications, and to examine inconsistencies and the logical relations with other constructs. Thus, the model clarifies usage and avoids misrepresentation and the uncritical acceptance or rejection of the construct at hand.

The model, therefore, can be applied beyond the critique of existing or vying constructs. It can

29 Studies describing the evolution of CTS in stages concur that the field has been dominated by a methodology- driven momentum, fueled by technological development, but that there seems to be a change of tendency with an ever-increasing interest in theoretical and metatheoretical, disciplinary discussions (see, for instance, Muñoz, 2017; Alves & Hurtado, 2017).

135

be used by proponents of new constructs as a methodology to build their case while avoiding ambiguity.

CTS has evolved into two main research traditions and there is an increasing recognition that “the analytic lenses and methods of process research can and should vary” (Angelone,

Ehrensberger-Dow, Massey, 2015:51). However, given that the development of CTS has been mainly methodological up to quite recently and that “the selection of relevant constructs or theories is not always straightforward” (Halverson, 2017:209), a methodology such as the conceptual performance model seems pertinent. In the face of the existence of a plurality of perspectives and research traditions, having a method to analyze which constructs are better for our specific research purposes, aiding our theoretical choice, is the best possible –maybe the only possible– shared ground.

Implications

The scope of knowledge

The possibility of comparing constructs, theories and research traditions by looking at their problem-solving efficiency allows for a shift in the focus of attention from paradigms to the constructs themselves, that is, to the extent to which we can use constructs.30 It is a fact that the same constructs have been used across research traditions, and that even some of them straddle

30 This position is related to that of Feyerabend’s (1978), who maintained that progress is more easily identifiable at the theory level than at the paradigm level.

136

across theoretical frameworks promoting methodological plurality (Halverson 2017). This is an example of why we should be looking at constructs instead of looking at paradigms.

The knowledge generated by a scientific discipline arguably coincides with that of the total constructs successfully applied to describe and explain empirical facts in the discipline. But, what can we say about the limits of our knowledge if construct use is quite often implicit, if the applicability and validity of our constructs is by necessity local and bound to the concrete scientific tasks, models, research traditions we are analyzing? The solution may lie not on focusing on the limits of our knowledge, but on the scope of what we know.

What delimits the scope of the construct is related to its external usage, the term that labels it, the discipline of origin, the research tradition… Useful as all that information is, accepting it uncritically can be misleading. To take again the examples previously discussed in the case studies, if we focus on the limits of competence, a term linguistic in origin used in cognitivist empirical studies of translation skill development, it really seems comparable with expertise, a term originating in psychology used in cognitivist empirical studies of translation skill development. However, if we focus on the scope of the constructs, i.e., on their clarity, adequacy, consistency and simplicity, on the overall potential for fact explanation and description as integrated into a set of research tradition axioms, competence is quite a different kettle of fish. As we saw in Chapter 4, by its definition and structure competence does not lend itself to the empirical investigation of expertise development in the same way the construct expertise does.

The same applies to expertise. Shreve (2002) already warned us about the risks of taking expertise at face value, that is, focusing on what delimits and distinguishes the construct from other constructs, without challenging the scope of expertise as applied to the investigation of

137

translation. Muñoz’s expertise construct (2014) can be understood as an exercise in scope redefinition.

Focusing on the limits amounts to corralling constructs into a research tradition, distinguishing them in kind from other constructs corralled into other research traditions.

Focusing on the scope entails analyzing the applicability of the construct to each phenomena vis

à vis its nature, assumptions and implications, as well as those of any potentially competing constructs and its research tradition. It is because of these different dimensions of scope that it conforms better to differences of degree and not of kind, as a limit-centered approach might do.

Let’s take, for example, automatization as a research construct. Associated with the monitor model in CTS (Tirkknonen-Condit, 2005) and cognitivist approaches to CTS, it is often used to refer to fixed, quick and unconscious cognitive processing as opposed to variable, slow and conscious problem-solving processing.31 The construct is certainly linked to computational models of processing (see Marín 2016 for a discussion) and so with cognitivist translatology. But that doesn’t mean that the construct belongs to that tradition. Automatization is present in models by authors endorsing the main axioms of cognitive translatology (Halverson, 2015, 2017). Rather than assigning the construct to a research tradition, of labeling it, it would be interesting to apply the conceptual performance model to study the potential differences in scope across usages.

As we saw when discussing Laudan’s model of scientific progress and Mulkay’s model of disciplinary development, research traditions do not usually overthrow each other in clear-cut revolutions, scientists do not defect from their research agendas overnight, new constructs do not

31 Unconscious or, at least, not readily available to verbalization.

138

appear to instantly replace earlier ones on their own, suddenly, or even perfectly.32 More than abrupt shifts, we find gradual change. The progression of scientific knowledge is an evolution on a cline and it is the case that sometimes constructs and research traditions thought long jettisoned cast long shadows over newcomers; for instance, the revival of literal translation in CTS

(Schaeffer & Carl, 2014). Science changes “Not with a bang but a whimper.”33

Comparing two constructs entails looking at their problem-solving rates (number of problems solved minus problems posed) through the assessing prism of a set of several criteria.

This means that the advances posed by a given construct will always be potentially subject to asymmetry: a construct may suppose an overall advance while not performing better than another construct in every criterion. Competence may be a more useful construct than expertise for pedagogical purposes, but it is more problematic as a research construct to derive hypotheses from. STIE presents an advance with regards to expertise on the grounds of the analysis of the environment of the task as part of the cognitive processing. This means that rational theoretical choice is ultimately not based on a binomial comparison or on a question of discrete differences of kind, but of degree. And even when two theories or constructs are pitted against each other in terms of problem-solving rates, the decision is not taken because one is absolutely better, but because it is better than the other construct for this particular purpose.

A methodology that focuses on the limits of constructs does not do justice to the gradual and comparative nature of construct assessment. Changes in scope, on the other hand, can be as

32 In a very interesting presentation, the historian and philosopher of science Hasok Chang convincingly argues against the abandonment of phlogiston in favor of oxygen in the 18th century: https://www.youtube.com/watch?v=zGUsIf9qYw8 Researchers at Google Brain took inspiration from several models, some of which had been abandoned for years, to develop the AI models with which they improved the quality of their Google Translate service considerably. 33 “The Hollow Men,” T.S. Eliot, 1925.

139

piecemeal as the problem-solving capacity of a construct: first one anomaly arises, then several problems are found that the construct fails to solve, then a hitherto accepted axiom is dramatically challenged and debunked… until the construct ceases to be applicable without compromising applicability, adequacy and consistency.

Moreover, by focusing on the limits of our constructs, we can easily identify what our constructs are not, but not necessarily what they are. To do that we would need to look at their scopes. Focusing on the limits requires that these are already set, and assumes that the limits, instead of conceptual performance criteria, mark the confines of what is acceptable.

Analyzing the construct first we are also defining it, and so finding its limits because of its definition. Otherwise we are using the limits to define the concept a priori. The scope of a construct is a function of the logical relationships (that can be) established between that construct and the network of constructs or beliefs operating in its contextual use—in the case of expertise, for instance, it is cognitivist translatology, the axioms of cognitivist psychology (see Chapter 2), a commitment to empirical enquiry… Therefore, they cannot be determined just by “placing” the construct in a discipline, research tradition, or empirical agenda. When the limits of a construct are redefined without having critically analyzed the scope first, the construct becomes a legacy concept (Shreve & Angelone, 2010), a received, unexamined, often ambiguous construct.

By analyzing the scope of our constructs, we are analyzing the scope of our knowledge.

This is particularly important when treating interdisciplinary fields, for by analyzing the scope we are also acknowledging the long-standing caveat about interdisciplinary borrowing: do not import constructs wholesale, ensure they fit into the receiving research tradition and conceptual framework, make sure that there are no hidden inconsistencies. The result may be the discouraging discovery that our tools are less powerful or valid than expected. On the other hand

140

we may also identify blind spots in the disciplinary theoretical system and new research avenues to pursue, or find innovative, previously unthought-of empirical questions.

Pluralism

Unlike paradigms, there is no reason why one research tradition should have the monopoly of a discipline. In cases where a research tradition shows a considerable advance in progress over another, it is rational for scientists to “migrate.” But committing to a new research tradition does not prevent those scientists from still working in the old one if it is empirically adequate and fits the purposes it is intended for. Different research traditions may complement each other illuminating different aspects of the same phenomena, the cognitive effort of translators, their social dynamics with other project stakeholders, for instance. Competition between research traditions can be productive too, promoting the refinement of constructs, see for instance the case of expertise and STIE.

Also, there are often phenomena so varied and complex that they cannot be accounted for from a single scientific perspective, a situation calling for pluralism (Kellert et al. 2006).

Translation and interpreting processes are good examples of this complexity, and a look at the methods applied to investigate them bears witness to this plurality: from ethnographic research on translators’ workplaces, to experimental projects to observe translators’ behavioral indicators of cognitive processing. There is a whole range of different methodological approaches, sometimes mutually exclusive, that on their own provide interesting and solid data about the

141

translation process.34 CTS scholars have embraced this methodological plurality (Angelone et al.,

2015; Halverson, 2017). On theoretical terms, however, the same scholars who celebrated having a varied panoply of methods, called for a common set of assumptions, a basic disciplinary ground, where different perspectives would converge, enable comparison and replicability of results and strengthen the field (see also Shreve & Angelone, 2010).

I would like to contend that while a theoretical unification of CTS might bring some benefits, it is not a necessary pursuit. On the contrary, I believe that plurality is not a transitory, immature phase of CTS and that there is rational ground to actively embrace it.

Understood as “the general tolerance of different kinds of thing, or more particularly of different and perhaps incommensurable descriptions of the world” (Blackburn, 2008), pluralism is often compared to irresponsibly accepting one view and its opposite at the same time, and so is related to relativism and “the more sinister doctrine that no view is true, or that all views are equally true” (Blackburn, 2008). However, the kind of pluralism I endorse is not relativism:

The most fundamental difference is that relativism involves a renunciation of

judgment and commitment at least to a degree, which pluralism most definitely does not.

The mature pluralist attitude is to engage productively with what one disagrees with,

which is very far from the feared caricature of relativism in which one says “Whatever”

(Chang, 2012:261).

Chang defends the idea that pluralism in science is not only to be tolerated, but actively sought as a normative approach beneficial to scientific progress, he calls it “active normative epistemic pluralism” (2012:268). Chang does not accept just “any” possible method or theory,

34 See Horst (2016) for a discussion of how mutually exclusive models can provide valid empirical data, contributing to epistemic pluralism.

142

but only those that abide by the requirements set up by science, those deemed acceptable; hence the normativity. Therefore, one needs first to make explicit what it is that the scientific enterprise aims for. Explicitly inspired by Feyerabend’s anarchism, Chang intends to benefit from having many –but not just any– theoretical and methodological approaches: “(…) unlike Feyerabend, I want to do this in a systematic fashion, by surveying all the various things that one might think science should desire to achieve” (2012: 269).

Like many other thinkers (Mill, 1859; Peirce, 1877), Chang points to humility, to the acknowledgement of our limitations and those of our capacity to understand and explain reality, as the main rationale for pluralism (2012:255). And he extends on the benefits both of toleration and of interaction between different systems of knowledge (constructs, theories, search traditions…) (2012:269-284), which I will summarize in the following.

Toleration entails protecting ourselves from the unpredictability of scientific results and the possible failure of individual projects. It also divides labor in the domain: as mentioned above, different theories or models (ethnographic, experimental…) can explore different aspects of translation phenomena, and help us meet the same aim via different pathways provided there is no logical contradiction (see Schaeffer & Carl 2014; Halverson, 2015). Also, the preeminence of epistemic virtues and aims vary in scientific practice, and there is no reason why all scientists in a domain should be committed to the same ones, thus “tolerant pluralism” helps meeting multiple aims.

Interactive pluralism evidences how the interrelation of theories and approaches strengthens them. A benefit derived from interaction is integration—combining the results and practical applications of different, maybe inconsistent systems of knowledge in one single project. Chang exemplifies integration with the global position system (GPS):

143

by means of satellites kept in space by Newtonian physics, and atomic clocks ruled by

quantum mechanics and corrected by special and general relativity, this system maps the

spherical surface of the round earth on a geometric grid (or rather, a geostatic grid), and

gives advice to people on the ground from a flat-earth point of view. (2012: 266)

The fact that two or more systems of knowledge provide antagonistic explanations on the same phenomena, or set out from widely diverging views of the world in their respective endeavors, is by no means an impediment to take into consideration the knowledge they separately produce and jointly apply it for the benefit of scientific success.

Integration is beneficial as well in the case that one “system of knowledge,” a discipline, a research tradition, takes some constructs, empirical results or data-gathering tools from another. Borrowing from other disciplines is one of the defining practices in CTS in terms of models, methods, techniques, materials… (O’Brien 2013). This borrowing also happens within

CTS, taking constructs from cognitivist translalogy into cognitive translatology and vice versa.

The boundaries between cognitivist and cognitive translatologies are permeable, constructs –and methods, too– migrate (e.g., expertise) or happen to be at the core of both traditions (mental load, pauses and time as behavioral indicators).

And, finally, there is competition—not necessarily aimed at overthrowing another system, but competition as a source of benefits, refinement and improvement, all of which disappear once systems cease to compete. Competition is so useful that Chang even recommends finding new competitors for a knowledge system that supersedes another, vying system. Here is

144

the main difference with regards to non-pluralist approaches, the argument for competition as an end in and of itself. Chang argues that even if a system is failing on its own, without being in competition with another, it should not be abandoned—It is a reminder of the unsolved problems. This view concurs with the one expressed in setting up the comparative framework of the conceptual performance model: there is no rational motive to abandon a theory or construct unless there is another one with a better problem-solving efficiency rate.

In order to compete, the rules of the field need to be specified, we need to put forward the objectives “science should desire to achieve,” as Chang put it. Introducing a methodology that sets criteria for the comparison of constructs, and by implication the comparison of research traditions, in the light of their specific purposes, arguably is a contribution to engaging in active normative epistemic pluralism in CTS and so in reaping its benefits

Theoretical development

Until now, the general aim has been unification—to combine borrowings into a strong, shared CTS tradition. But, if we have a methodology to compare traditions and constructs rationally, if we see the benefits of competition, and we discard the need to hastily abandon unsuccessful theories, why should be not apply integration of different models of translation, coming from cognitivist and cognitive translatology? Why not take heed of the GPS example given by Chang? Patterns of cognitive effort described by cognitivist models of the translation process can be used to accommodate neurological findings (Shreve & Diamond, 2016; Muñoz,

2017), they can also be useful as a template to develop CATs, which could as well benefit from research on human-computer interaction (O’Brien & Teixeira forthcoming) and ergonomic and

145

ethnographic models (Risku, 2014; Ehrensberger-Dow, 2014). We could be using cognitivist and cognitive translatology models, computer science and neuroscience models all at once.

In a similar vein, Shreve (forthcoming) argues that research on translation expertise would reap considerable benefits from “bridging” computational and connectionist models across levels of research, from functional descriptions of task-bound processes to neural implementation. In fact, combining models would not only be a way to expand or diversify the object of study, but a way to actually delve further into questions already posed in the field: “We need to move forward—but how? One way forward is to introduce approaches that (at least for now) ‘bridge’ the gap between higher-level paradigms of translation and translation expertise to more implementational (presumably connectionist) analytic perspective” (forthcoming: 13).

The wealth of translator UAD (“User Activity Data;” Carl, Schaeffer, Bangalore, 2016) can be used to extract general behavioral patterns, cues about the translation cognitive processes that can be further explored from radically opposite perspectives. And the other way around, models of the translation process that include affect and the impact of translators’ beliefs (Rojo,

Ramos and Valenzuela, 2014; Rojo and Ramos, 2014), the interaction with the task environment at all levels (Risku, 2014; Ehrensberger-Dow, 2014) or evidence supporting an alternative, non- linear progression of translation cognitive effort (Martin & Muñoz, forthcoming) can help develop new tools, challenge accepted pedagogical practices or establish new, promissory links with the language industry.

The kind of interaction I am calling for is not a form of anarchism in which “anything goes” (Feyerabend, 1978), but one in which, as Chang (2012) put it, “many things go.” The case studies included in Chapter 4 evidence that inconsistencies do arise between research traditions, weak forms of incommensurability exist, for we can be certain that cognitivist and cognitive

146

scholars may not be referring to “the same thing” when the talk about expertise. But that just means we have at least two explanations for the same set of phenomena, which as a matter of fact is quite useful provided we are clear and explicit about our position.

Chang (2012) stresses that pluralism requires us to put forth the rules of the field of competition and the desired scientific outcomes. In CTS, Shreve (2002) warns us against the uncritical adoption of constructs, Gambier (2006) highlights the importance of clarity about the object and level of analysis across TS, Muñoz (2017) stresses the need of identifying the models we endorse or implicitly assume as the central task of the field in the next decade… The concerns of CTS scholars can find an answer –and potentially fruitful approach– in going beyond descriptive pluralism. Assuming that translation phenomena are simply too complex for only one scientific account (see Kellert, et al., 2006) spurred CTS scholars to take in plurality from other disciplines. The time seems to be ripe to adopt normative pluralism within our own discipline: setting a set of common, accepted criteria for the comparison, competition and integration of different, maybe incommensurable, perspectives. Using the conceptual performance model, it would be possible to engage in discussion with others on the same terms, and to challenge our own positions for the benefit of clarity and consistency, which are aspects of theory development and debate of paramount importance.

Also, by being able to collate constructs in the light of concrete criteria and uses, we would avoid the risks of skipping differences, misapplications or opportunities for further application.35 Accepting and acknowledging the existence of diverging research traditions and the criteria to compare them enables us to stop hedging when choosing a construct or adopting a

35 E.g., comparing the “performance” of competence and expertise in terms of clarity, adequacy, consistency and simplicity, against the objective of hypothesizing and observing translators’ skills development.

147

research tradition, to do without a loosely defined common ground that makes “it possible to sweep any conceptual differences they may have under the rug” (Muñoz, 2016c:8).

Pedagogy

Although we can say that there are no direct applications of the discussion here presented to translators’ training, the indirect application is clear: CTS is an applied endeavor with clear pedagogical implications. This has been true since the birth of the discipline—changing the way we relate our concepts and the benefits of pluralism can have therefore an impact on research- informed reaching.

We are not only translator trainers, but also educators with critical thinking skills as one of the main objectives to meet in the classroom. As a pedagogical tool, the conceptual performance model is instrumental in approaching the CTS theoretical corpus and in providing students a useful method for clarification, comparison and analysis.

This application does not need to restrict itself, of course, to CTS, but can easily be applied to other cases in the growing literature in TS. It’s important to note that at a time when more and more theoretical discussions are being held in the field (see Salah, 2012) and interdisciplinary status is discussed (Gabier & Van Doorslaer, 2016) the basics of conceptual analysis can be a convenient aid for TS students entering graduate school.

148

Limitations and further directions

This dissertation introduces a model of scientific progress as an alternative to the most current and widely accepted model in CTS. The new model provides for the comparison of research constructs and a set of criteria for comparison are discussed and applied to concrete case studies. The analysis of the implications and consequences of adopting Laudan’s model is restricted to its benefits with regards to the Kuhnian model; and the discussion of the criteria of the conceptual model are limited to the cases of expertise and its developments. These limitations can be overcome by additional, more extensive work comparing Laudan’s model of scientific progress and its benefits to those of other models. Such an additional comparison, for example with Lakatos’s research programmes, has been hinted in this dissertation, and would merit further analysis. The number of case studies can be expanded to illustrate the variety of conditions and the specific nature of conceptual comparison in terms of problem solving. It would be interesting, for instance, to analyse the use of constructs that bridge over different methodological positions, such as mental effort (Halverson 2017), or those which have definite implications for our definition of the object of study, such as translation unit or translation task.

The social dimensions of knowledge-generating enterprises are taken into consideration in this dissertation when adopting a model of disciplinary evolution; however, the ultimate role of social dynamics in choosing one construct over another has remained unexplored in the discussion. The Philosophy of Science and Science Studies offer exciting ways to explore CTS from the perspective of the impact of social, extra-scientific factors in scientific evolution, the

149

social dimensions of theory choice and the distributed nature of scientific knowledge. (Kellert et al., 2006; Longino, 2006, Thagard 1993).

The conceptual analysis here proposed can be complemented and enhanced with empirical research straightforwardly aimed at validating research constructs. Angelone & Marín

(forthcoming), for example, empirically investigate expertise and its implications from the emic perspective of practitioners in a way germane to the analysis proposed in this dissertation.

Empirical validation of research constructs remains scarce in CTS, arguably because of the complexity of the constructs themselves (Alves & Hurtado, 2017: 540). That complexity, I would like to argue, is not an empirical problem, but a conceptual one. Probably the result of a combination of conceptual problems that can be identified with the conceptual performance model, thus combining conceptual analysis and empirical research.

The pluralist implications discussed in this chapter merit further work, maybe an analysis on their own that might contribute to the general epistemic debates in TS and not only to CTS.

Shifting from a descriptive pluralist stance to a normative pluralism is a liberating move that empowers scientific practice. A more in-depth analysis of its implications seems crucial as the evolution of CTS seems to indicate that we will start seeing wide-ranging, integrated research programs including diverging sets of data and analytical tools in the near future.

150

REFERENCES

Adams, Fred, and Kenneth Aizawa. The Bounds of Cognition (Malden, MA. (2008). Print.

Albir, Amparo. “La Competencia Traductora y Su Aprendizaje: Objetivos, Hipótesis y

Metodología De Un Proyecto De Investigación”. Póster, IV Congrés Internacional Sobre

Traducció. 1999. Universitat Autònoma de Barcelona, Bellaterra.

Alves, Fabio. “Translation Process Research at the Interface: Paradigmatic, Theoretical, and

Methodological Issues in Dialogue with Cognitive Science, Expertise Studies, and

Psycholinguistics.” Psycholinguistic and Cognitive Inquiries into Translation and

Interpreting. Eds. Aline Ferreira and John and Schwieter. Amsterdam: John Benjamins, 17-

40, 2015. Print.

---. Triangulating Translation: Perspectives in Process Oriented Research. 45 Vol. John

Benjamins Publishing, 2003. Print.

Alves, Fabio, and Amparo and Hurtado-Albir. “Evolution, Challenges, and Perspectives for

Research on Cognitive Aspects of Translation.” The Handbook of Translation and

Cognition. Eds. John Schwieter and Aline and Ferreira.Wiley Blackwell, 2017. 537-554.

Print.

151

Alves, Fabio, and José Luiz Gonçalves. “Modelling translator’s Competence.” Doubts and

directions in TS (2007): 41-55. Print.

Alves, Fabio, Adriana Pagano, and Igor da Silva. “A New Window on Translators’ Cognitive

Activity: Methodological Issues in the Combined use of Eye Tracking, Key Logging and

Retrospective Protocols.” Methodology, Technology and Innovation in Translation Process

Research: A Tribute to Arnt Lykke Jakobsen. Eds. Inger M. Mees, F. Alves, and S. and

Göpferich. Copenhaguen: Samfundslitteratur, 2009. 267-291. Print.

Alvstad, Cecilia, Adelina Hild, and Elisabet Tiselius. Methods and Strategies of Process

Research: Integrative Approaches in TS. 94 Vol. John Benjamins Publishing, 2011. Print.

Anderson, Michael L. “Embodied Cognition: A Field Guide.” Artificial Intelligence 149.1

(2003): 91-130. Print. Angelone, Erik, Maureen Ehrensberger-Dow, and Gary Massey.

“Cognitive Processes.” Researching Translation and Interpreting. Eds. Claudia Angelelli

and Brian and Baer. 2010 Vol. London: Routledge, 2015. 43. Print.

Angelone, Erik. “Uncertainty, Uncertainty Management and Metacognitive Problem Solving in

the Translation Task.” Translation and Cognition. Eds. Gregory Shreve and Erik and

Angelone. 15 Vol. Amsterdam: John Benjamins, 2010. 17-40. Print. American Translators

Association Scholarly Monograph Series .

Angelone, Erik, and Álvaro Marín García. “Expertise Acquisition through Deliberate Practice:

Gauging Perceptions and Behaviors of Translators and Project Managers.” Translation

Spaces 6.1 (forthcoming). Print.

152

Angelone, Erik, and Gregory M. Shreve. “Uncertainty Management, Metacognitive Bundling in

Problem Solving, and Translation Quality.” Cognitive Explorations of Translation. Ed.

Sharon O’Brien. New York: Continuum, 2011. 108-130. Print.

Arrojo, Rosemary. Tradução, Desconstrução e Psicanálise. Imago Ed., 1993. Print.

Baker, Mona. “Corpus Linguistics and TS: Implications and Applications.” Text and technology:

In honour of John Sinclair 233 (1993): 250. Print.

Basalamah, Salah. “En Deçà Des Méthodes Et Des Théories, l’horizon d’une Philosophie.” TTR:

Traduction, terminologie, rédaction 25.1 (2012): 13-49. Print.

Bassnett, Susan. “TS at a Cross-Roads.” Target.International Journal of TS 24.1 (2012): 15-25.

Print.

Bayer-Hohenwarter, Gerrit. “Translational Creativity: How to Measure the Unmeasurable.”

Behind the Mind: Methods, Models and Results in Translation Process Research. Eds. S.

Göpferich, A. L. Jakobsen, and I. M. and Mees. 37 Vol. Copenhaguen: Samfundslitteratur,

2009. 39. Print.

Benjamin, Walter. “The Translator's Task.” The TS Reader. Tran. Walter Kaufmann. Ed.

Lawrence Venuti. 3rd ed. New York: Routledge, 2012. 75-83. Print.

Bernardini, Silvia. “Using Think-Aloud Protocols to Investigate the Translation Process:

Methodological Aspects.” RCEAL Working Papers in English and Applied Linguistics 6.

Ed. Jenny Williams. 6 Vol. Cambridge: Cambridge University Press, 1999. 179-199. Print.

153

Bowker, Lynne, et al. “Unity in Diversity.” Current Trends in TS (1998): 25-37. Print.

Brooke, John Hedley. “Organic Synthesis and the Unification of chemistry—a Reappraisal.” The

British Journal for the History of Science 5.04 (1971): 363-92. Print.

Campbell, Stuart. “Towards a Model of Translation Competence.” Meta: Journal des

traducteurs 36.2-3 (1991): 329-43. Print.

Carl, Michael. “A Computational Cognitive Model of Human Translation Processes.” Emerging

Applications of Natural Language Processing: Concepts and New Research. Eds. Sivaji

Bandyopadhyay, Sudip K. Naskar, and Asif and Ekbal. Hershey, PA: IGI Global, 2012.

110-128. Print.

Carl, Michael, Moritz Schaeffer, and Srinivas Bangalore. “The CRITT Translation Process

Research Database.” New Directions in Empirical Translation Process Research. Eds.

Michael Carl, Srinivas Bangalore, and Moritz Schaeffer. London: Springer, 2016. 13-54.

Print.

Carnap, Rudolf. Logical Foundations of Probability. 2nd ed. Chicago: University of Chicago

Press, 1962. Print.

Chang, Hasok. Is Water H2O?: Evidence, Realism and Pluralism. 293 Vol. Springer Science &

Business Media, 2012. Print.

Chesterman, Andrew. “Bridge Concepts in Translation Sociology.” Constructing a Sociology of

Translation. Eds. Michaela Wolf and Alexandra Fukari. 74 Vol. Amsterdam: John

Benjamins, 2007. 171-185. Print.

154

---. “From'is' to'Ought': Laws, Norms and Strategies in TS.” Target.International Journal of TS

5.1 (1993): 1-20. Print.

Chesterman, Andrew, and Rosemary Arrojo. “Forum: Shared Ground in TS.”

Target.International Journal of TS 12.1 (2000): 151-60. Print.

Chomsky, Noam. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press, 1965. Print.

Churchland, Patricia Smith. Neurophilosophy: Toward a Unified Science of the Mind-Brain.

MIT press, 1989. Print.

Clark, Andy, and David Chalmers. “The Extended Mind.” analysis (1998): 7-19. Print.

Clark, Andy. Being there. MIT Press Cambridge, MA, 1997. Print.

Cole, Stephen. “The Hierarchy of the Sciences?” American Journal of Sociology 89.1 (1983):

111-39. Print.

Comte, Auguste. Positive Philosophy. 1 Vol. Bell London, 1896. Print.

Conant, James Bryant. Harvard Case Histories in Experimental Science Volume I. 1 Vol.

Cambridge: Harvard University Press, 1950. Print.

Crane, Diana. Invisible Colleges; Diffusion of Knowledge in Scientific Communities. Chicago &

London: University of Chicago Press, 1972. Print.

Crisafulli, Edoardo. “The Quest for an Eclectic Methodology of Translation Description.”

Crosscultural Transgressions. Research Models in TS II: Historical and Ideological Issues.

Ed. Theo MLA Hermans. Manchester: St Jerome, 2014. 26-43. Print.

155

Cuerda Martínez, José Luis. Me Noto Muy Cambiá. Logroño: Pepitas de Calabaza, 2016. Print.

Cummins, Robert. “How does it Work?” Versus” what are the Laws?”: Two Conceptions of

Psychological Explanation.” Explanation and cognition (2000): 117-44. Print.

Dancette, Jeanne, and Nathan Ménard. “Modèles Empiriques Et Expérimentaux En

Traductologie: Question d'Épistémologie.” Meta: Journal des traducteurs 41.1 (1996): 139-

56. Print.

Danks, Joseph H., et al, eds. Cognitive Processes in Translation and Interpreting. 3 Vol.

Thousand Oaks: Sage Publications, Inc, 1997. Print.

De Groot, Adriaan D. Thoughtand Choice in Chess. The Hague: Mouton, 1965. Print.

Dechert, Hans W., and Ursula Sandrock. “Thinking-Aloud Protocols: The Decomposition of

Language Processing.” Experimental approaches to second language learning.Oxford:

Pergamon (1986): 111-26. Print.

Dennett, Daniel C. “Kinds ofMinds: Toward an Understanding ofConsciousness.” (1996)Print.

Derrida, Jacques. “La Différance” conference Pronounced at the Société Française De

Philosophie.” Théorie d’ensemble. Ed. Tel Quel. Paris: Seuil, 1968. Print.

---. “Des Tours De Babel.” Difference in translation 167 (1985)Print.

Diamond, Bruce, and Gregory and Shreve. “Neural and Physiological Correlates of Translation

and Interpreting in the Bilingual Brain.” Translation and Cognition. Eds. Gregory Shreve

and Erik and Angelone. Amsterdam: John Benjamins, 2010. 189-322. Print.

156

Dimitrova, Birgitta Englund. Expertise and Explicitation in the Translation Process. 64 Vol.

Amsterdam: John Benjamins Publishing, 2005. Print.

Dragsted, Barbara. “Indicators of Difficulty in translation—Correlating Product and Process

Data.” Across Languages and Cultures 13.1 (2012): 81-98. Print.

---. “Segmentation in Translation: Differences Across Levels of Expertise and Difficulty.”

Target.International Journal of TS 17.1 (2005): 49-70. Print.

Dragsted, Barbara, Inge Gorm Hansen, and Henrik Selsøe Sørensen. “Experts Exposed.”

Methodology, Technology and Innovation in Translation Process Research. Eds. I. M.

Mees, F. F.Alves, and S. S.Göpferich. Copenhaguen: Samfundslitteratur, 2009. 293-317.

Print.

Dragsted, Barbara, and Inge Gorm Hansen. “Comprehension and Production in Translation: A

Pilot Study on Segmentation and the Coordination of Reading and Writing Processes.”

Looking at the Eyes. Eds. Susanne Göpferich, Arnt Lykke Jakobsen, and Inger Mees.

Copenhaguen: Samfundslitteratur, 2008. 9-21. Print.

Duhem, Pierre. The Aim and Structure of Physical Theory. Tran. Philip Paul Wiener. Princeton:

Princeton University Press, 1954. Print.

Dunne, Keiran J. “The Industrialization of Translation: Causes, Consequences and Challenges.”

Translation Spaces 1.1 (2012): 143-68. Print.

Ehrensberger-Dow, Maureen. “Challenges of Translation Process Research at the Workplace.”

MonTi: Monografías de Traducción e Interpretación (2014): 355-83. Print.

157

Ehrensberger-Dow, Maureen, and Gary Massey. “Cognitive Ergonomic Issues in Professional

Translation.” The Development of Translation Competence: Theories and Methodologies

from Psycholinguistics and Cognitive Science. Eds. John Schwieter and Aline and Ferreira.

Newcastle upon Tyne: Cambridge Scholars Publishing, 2014. 58-86. Print.

Ehrensberger-Dow, Maureen, and Sharon O'Brien. “Ergonomics of the Translation Workplace.”

Translation Spaces 4.1 (2015)Print.

Ericsson, Karl Anders. “The Acquisition of Expert Performance: An Introduction to some of the

Issues.” The Road to Excellence: The Acquisition of Expert Performance in the Arts and

Sciences, Sports and Games. Ed. K. Anders Ericsson. Mahwah, NJ: Lawrence Erlbaum

Associates, Inc, 1996. 1-50. Print.

Ericsson, K. Anders, and Neil Charness. “Cognitive and Developmental Factors in Expert

Performance.” Expertise in Context: Human and Machine. Eds. Paul Feltovich, M. Ford ,

and Robert and Hoffman. Cambridge, MA: The MIT Press, 1997. 3-41. Print.

Ericsson, K. Anders, Ralf T. Krampe, and Clemens Tesch-Römer. “The Role of Deliberate

Practice in the Acquisition of Expert Performance.” Psychological review 100.3 (1993):

363. Print.

Ericsson, K. Anders, and Herbert A. Simon. “Verbal Reports as Data.” Psychological review

87.3 (1980): 215. Print.

158

Feyerabend, Paul. “Consolations for the Specialist.” Criticism and the Growth of Knowledge.

Eds. I. Lakatos and A. Musgrave. Cambridge: Cambridge University Press, 1970. 197-230.

---. Against Method : Outline of an Anarchistic Theory of Knowledge. London: Verso, 1978.

Print.

Feyerabend, Paul, and Peter Lipton. “Killing Time: The Autobiography of Paul Feyerabend.”

Nature 374.6525 (1995): 837-. Print.

Fodor, Jerry A. The Modularity of Mind: An Essay on Faculty Psychology. MIT press, 1983.

Print.

Fodor, Jerry. “Let Your Brain Alone.” London review of books 21.19 (1999): 68-9.

. Web.

Fodor, Jerry A., and Zenon W. Pylyshyn. “Connectionism and Cognitive Architecture: A Critical

Analysis.” Cognition 28.1 (1988): 3-71. Print.

Gambier, Yves. “Pour Une Socio-Traduction.” TS at the Interface of Disciplines. Eds. João

Ferreira Duarte, Alexandra Assis Rosa, and Teresa Seruya. 68 Vol. Amsterdam: John

Benjamins, 2006. 29-42. Print.

---. “Une Traductologie Pour Quelles Pratiques Traductionnelles?” Target.International Journal

of TS 24.1 (2012): 61-82. Print.

159

Gambier, Yves, and Luc Van Doorslaer, eds. The Metalanguage of Translation. 20 Vol.

Amsterdam: John Benjamins Publishing, 2009. Print.

Gambier, Yves, and L. and van Doorslaer, eds. Border Crossings. TS and Other Disciplines.

Amsterdam: John Benjamins, 2016. Print.

García, Adolfo. Traductología y Neurocognición: Cómo Se Organiza El Sistema Lingüístico Del

Traductor. Córdoba: Facultad de Lenguas. Universidad Nacional de Córdoba, 2012. Print.

Gerloff, Pamela. “Second Language Learners’ Reports on the Interpretive Process: Talk-Aloud

Protocols of Translation.” Interlingual and Intercultural Communication.Discourse and

Cognition in Translation and Second Language Acquisition Studies.Tübingen: Gunter Narr

(1986): 243-62. Print.

Gile, Daniel. “Response to Invited Papers.” Translation Research and Interpreting Research.

Traditions, Gaps and Synergies. Ed. Christina Schäffner. Clevedon: Multilingual Matters,

2004. 124-127. Print.

---. “Testing the Effort Models' Tightrope Hypothesis in Simultaneous Interpreting-A

Contribution.” Hermes-Journal of Language and Communication in Business 12.23 (2017):

153-72. Print.

Globus, Gordon G. “Toward a Noncomputational Cognitive Neuroscience.” Journal of cognitive

neuroscience 4.4 1992: 299-300. Print.

Godfrey-Smith, Peter. Theory and Reality: An Introduction to the Philosophy of Science.

University of Chicago Press, 2009. Print.

160

Göpferich, Susanne. Translationsprozessforschung: Stand-Methoden-Perspektiven. 4 Vol. BoD–

Books on Demand, 2008. Print.

---. “Towards a Model of Translation Competence and its Acquisition: The Longitudinal Study

TransComp.” Behind the Mind: Methods, Models and Results in Translation Process

Research. Eds. Susanne Göpferich, Arnt Lykke Jakobsen, and Inger and Mees. 37 Vol.

Copenhaguen: Samfundslitteratur, 2009. 11-37. Print.

Göpferich, Susanne et al. “Exploring Translation Competence Acquisition: Criteria of Analysis

Put to the Test.” Cognitive Explorations of Translation. Ed. Sharon O’Brien. New York:

Continuum, 2011. 57-85. Print.

Göpferich, Susanne, Arnt Lykke Jakobsen, and Inger M. Mees, eds. Behind the Mind: Methods,

Models and Results in Translation Process Research. 37 Vol. Copenhaguen:

Samfundslitteratur, 2009. Print.

Göpferich, Susanne, Arnt Lykke Jakobsen, and Inger M. Mees, eds. Looking at Eyes: Eye-

Tracking Studies of Reading and Translation Processing. Copenhaguen: Samfundslitteratur,

2008. Print.

Gutt, Ernst-August. Translation and Relevance: Cognition and Communication. Oxford:

Blackwell, 1991. Print.

Halverson, Sandra L. “Conceptual Categories in TS: Moving from Classical to Prototype.” Paper

Presented at the Second International Conference on Current Trends in Studies of

Translation and Interpreting.”.5-7 September, Budapest. 1996. Print.

161

---. “The Fault Line in our Common Ground.” Target.International Journal of TS 2.12 (2001):

356-61. Print.

---. “Psycholinguistic and Cognitive Approaches.” the Routledge Encyclopedia of TS. Tran.

Routledge. Eds. Baker Mona and Saldanha and Gabriela.Routledge, 2009. 211-216. Print.

---. “Cognitive TS: Developments in Theory and Method.” Eds. Gregory Shreve and Erik and

Angelone. Amsterdam: John Benjamins, 2010. 349-369. Print.

---. “Implications of Cognitive Linguistics for TS.” Cognitive Linguistics and Translation:

Advances in some Theoretical Models and Applications. Eds. Anda Rojo and Iraide and

Ibarretxe-Antuñano. Berlin: De Gruyter, 2013. 33-74. Print.

---. “Implications of Cognitive Linguistics for TS.” Cognitive Linguistics and Translation:

Advances in some Theoretical Models and Applications. Eds. Rojo Ana and Iraide and

Ibarretxe-Antuñano. 23 Vol. Walter de Gruyter, 2013. 33-74. Print.

---. “Cognitive TS and the Merging of Empirical Paradigms: The Case of ‘literal Translation’.”

Translation Spaces 4.2 2015. 310-40. Print.

---. “Multimethod Approaches.” The Handbook of Translation and Cognition. Eds. John

Schwieter and Aline and Ferreira.Wiley Blackwell, 2017. 195-212. Print.

Hansen, Gyde. “Controlling the Process Theoretical and Methodological reﬂections.”

Triangulating Translation. Ed. Fabio Alves. 45 Vol. Amsterdam: John Benjamins

Publishing, 2003. 25-42. Print.

162

---. “Integrative Description of Translation Processes.” Translation and Cognition. Eds. Gregory

Shreve and Erik and Angelone. Amsterdam: Amsterdam: John Benjamins, 2010. 189-212.

Print.

Hansen, Gyde, Andrew Chesterman, and Heidrun Gerzymisch-Arbogast, eds. Efforts and Models

in Interpreting and Translation Research: A Tribute to Daniel Gile. 80 Vol. Amsterdam:

John Benjamins Publishing, 2009. Print.

Hanson, Norwood Russell. Patterns of Discovery. Cambridge: Cambridge University Press,

1958. Print.

Hawes, Leonard C. Pragmatics of Analoguing: Theory and Model Construction in

Communication. Addison-Wesley Menlo Park, California, 1975. Print.

Hermans, Theo. Translation in Systems. Ed. St. Jerome. Manchester:, 1999. Print.

---. “Paradoxes and Aporias in Translation and TS.” TS: Perspectives on an Emerging

Discipline. Ed. Alessandra Riccardi. New York: Cambridge University Press, 2002. 10-23.

Print.

Hofstadter, Douglas, and Emmanuel Sander. Surfaces and Essences: Analogy as the Fuel and

Fire of Thinking. Basic Books, 2013. Print.

Holmes, James S. “The Name and Nature of TS.” Translated! Papers on literary translation and

TS (1988): 67-80. Print.

163

Holz-Mänttäri, Justa. Translatorisches Handeln: Theorie Und Methode. Helsinki: Suomalainen

Tiedeakatemia, 1984. Print.

Horst, Steven. Cognitive Pluralism. MIT Press, 2016. Print.

House, Juliane. Translation Quality Assessment: A Model Revisited. 410 Vol. Gunter Narr

Verlag, 1997. Print.

Hubscher-Davidson, Séverine. “A Discussion of Ethnographic Research Methods and their

Relevance for Translation Process Research.” Across languages and cultures 12.1 (2011):

1-18. Print.

Hurley, Susan L. “Vehicles, Contents, Conceptual Structure, and Externalism.” Analysis 58.1

(1998): 1-6. Print.

Immonen, Sini. “Translation as a Writing Process: Pauses in Translation Versus Monolingual

Text Production.” Target.International Journal of TS 18.2 (2006): 313-36. Print.

Immonen, Sini, and Jukka Mäkisalo. “Pauses Reflecting the Processing of Syntactic Units in

Monolingual Text Production and Translation.” Hermes 44 (2010): 45-61. Print.

Jääskeläinen, Riitta. “Think-Aloud Protocol Studies into Translation: An Annotated

Bibliography.” Target 14.1 (2002): 107-36. Print.

---. “Are all Professionals Experts?” Translation and Cognition. Eds. Gregory Shreve and Erik

and Angelone. American Translators Association Scholarly Monograph Series ed. 15 Vol.

Amsterdam: John Benjamins, 2010. 213-227. Print.

164

Jakobsen, Arnt Lykke. “Logging Target Text Production with Translog.” Copenhagen studies in

language.24 (1999): 9-20. Print.

---. “Effects of Think Aloud on Translation Speed, Revision and Segmentation.” Triangulating

Translation: Perspectives in Process Oriented Research. Ed. Fabio Alves. 45 Vol.

Amsterdam: John Benjamins, 2003. 69-95. Print.

Jakobsen, Arnt Lykke, and Kristian T. Jensen Hvelplund. “Eye Movement Behaviour Across

Four Different Types of Reading Task.” Looking at Eyes. Eye Tracking Studies of

Readingand Translation Processing. Eds. S. Göpferich, A. L. Jakobsen, and I. and Mees.

Copenhaguen: Samfundslitteratur, 2008. 103-124. Print.

Jensen, Kristian TH. “Indicators of Text Complexity.” Methodology, Technology and Innovation

in Translation Process Research: A Tribute to Arnt Lykke Jakobsen. Eds. IM Mees, F.

Alves, and S. and Göpferich. Copenhaguen: Samfundslitteratur, 2009. 61-80. Print.

Just, Marcel A., and Patricia A. Carpenter. “A Theory of Reading: From Eye Fixations to

Comprehension.” Psychological review 87.4 (1980): 329. Print.

Kaplan, Abraham. The Conduct of Inquiry. San Francisco: Transaction Publishers, 1964. Print.

Kellert, Stephen H., Helen E. Longino, and C. Kenneth Waters. Scientific Pluralism. 19 Vol. U

of Minnesota Press, 2006. Print.

Kelly, Dorothy. A Handbook for Translator Trainers. Manchester: St. Jerome, 2005. Print.

Translation Practices Explained Series.

165

Kiraly, Donald C. Pathways to Translation: Pedagogy and Process. Kent State University Press,

1995. Print.

---. A Social Constructivist Approach to Translator Education. Manchester: St. Jerome

Publishing, 2000. Print.

---. “A Passing Fad Or the Promise of a Paradigm Shift in Translator Education?” Beyond the

Ivory Tower: Rethinking Translation Pedagogy. Eds. Brian Baer and Geoffrey and Koby.

Amsterdam: John Bejamins Publishing Company, 2003. 3-32. Print.

Kitcher, Philip. the Advancement of Science: Science without Legend, Objectivity without

Illusions. Oxford: Oxford University Press, 1993. Print.

Krings, Hans P. Was in Den Köpfen Von Übersetzern Vorgeht: Eine Empirische Untersuchung

Zur Struktur Des Übersetzungsprozesses an Fortgeschrittenen Französischlernern. 291 Vol.

GNV, 1986. Print.

---. “Blick in Die ‘Black Box’–Eine Fallstudie Zum Übersetzungsprozeß Bei Berufsübersetzern.”

Textlinguistik und Fachsprache: Akten des Internationalen übersetzungswissenschaftlichen

AILA-Symposions (1988): 393-412. Print.

Kuhn, Thomas S. The Structure of Scientific Revolutions. Chicago: University of Chicago Press,

1962. Print.

---. “Reflections on My Critics.” Criticism and the Growth of Knowledge. Eds. I. Lakatos and A.

Musgrave. Cambridge: Cambridge University Press, 1970. 231-278. Print.

166

---. “Objectivity, Value Judgment, and Theory Choice.” The Essential Tension: Selected Sudies

in Scientific Tradition and Theory Change. Ed. Thomas S. Kuhn. Chicago: University of

Chicago Press, 1977. 320-339. Print.

---. “Pandora's Hope. Essays on the Reality of Science Studies.” 1999. Print.

Kussmaul, Paul. Training the Translator. 10 Vol. John Benjamins Publishing, 1995. Print.

---. Kreatives Übersetzen. Tübingen: Stauffenburg, 2000. Print. Studien Zur Translation 10 .

Kussmaul, Paul, and Sonja Tirkkonen-Condit. “Think-Aloud Protocol Analysis in TS.” TTR:

traduction, terminologie, rédaction 8.1 1995: 177-99. Print.

Lakatos, Imre. “Falsification and the Methodology of Scientific Research Programmes.”

Criticism and the Growth of Knowledge. Eds. I. Lakatos and A. Musgrave. Cambridge:

Cambridge University Press, 1970. 91-196. Print.

Lakatos, Imre, and Alan Musgrave. Criticism and the Growth of Knowledge: Volume 4:

Proceedings of the International Colloquium in the Philosophy of Science, London, 1965. 4

Vol. Cambridge: Cambridge University Press, 1970. Print.

Lakoff, George. “Cognitive Versus Generative Linguistics: How Commitments Influence

Results.” Language & Communication 11.1-2 (1991): 53-62. Print.

Laudan, Larry. Progress and its Problems: Towards a Theory of Scientific Growth. 282 Vol.

Univ of California Press, 1978. Print.

167

---. Science and Values: The Aims of Science and their Role in Scientific Debate. Univ of

California Press, 1984. Print.

Laviosa, Sara. Corpus-Based TS: Theory, Findings, Applications. 17 Vol. Rodopi, 2002. Print.

Lincoln, Yvonna S., Susan A. Lynham, and Egon G. Guba. “Paradigmatic Controversies,

Contradictions, and Emerging Confluences, Revisited.” The Sage Handbook of Qualitative

Research. Eds. Norman K. Denzin and Yvonna S. and Lincoln. 4 Vol. Sage, 2011. 97-128.

Print.

Longino, Helen E. The Fate of Knowledge. Princeton University Press, 2002. Print.

Lörscher, Wolfgang. “Linguistic Aspects of Translation Processes: Towards an Analysis of

Translation Performance.” Interlingual and intercultural communication.Tübingen: Gunter

Narr (1986): 277-92. Print.

---. “Models of the Translation Process: Claim and Reality.” Target.International Journal of TS

1.1 (1989): 43-68. Print.

---. Translation Performance, Translation Process, and Translation Strategies: A

Psycholinguistic Investigation. 4 Vol. G. Narr, 1991. Print.

MacIntyre, Aasdair. After Virtue. Notre Dame: Notre Dame Press, 1981. Print.

Malmkjær, Kirsten. “Multidisciplinarity in Process Research.” Triangulating Translation. Ed.

Fabio Alves. 37 Vol. Amsterdam: John Benjamins, 2000. 163-169. Print.

168

Marr, David. “Vision: A Computational Investigation into the Human Representation and

Processing of Visual Information, Henry Holt and Co.” Inc., New York, NY 2. 1982. 4.2.

Print.

Massaro, Dominic. “Language and Information Processing.” Understanding Language: An

Information-Processing Analysis of Speech Perception, Reading, and Psycholinguistics. Ed.

Dominic Massaro. New York: Academic Press, 1975. 3-28. Print.

Masterman, Margaret. “The Nature of a Paradigm.” Criticism and the Growth of Knowledge.

Eds. I. Lakatos and A. Musgrave. Cambridge: Cambridge University Press, 1970. 59-90.

Print.

Maturana, Humberto, and Francisco Varela. The Tree of Knowledge. Boston: Shamballa, 1988.

Print.

Mayoral, Roberto. “Por Una Renovación En La Formación De Traductores e Intérpretes:

Revisión De Algunos De Los Conceptos Sobre Los Que Basa El Actual Sistema Su

Estructura y Contenidos.” Sendebar 12. 2001. 311-36. Print.

McClelland, James L., David E. Rumelhart, and Geoffrey E. Hinton. “The Appeal of Parallel

Distributed Processing.” MIT Press, Cambridge MA. 1986. 3-44. Print.

Mees, Inger M., Susanne Göpferich, and Fabio Alves, eds. Methodology, Technology and

Innovation in Translation Process Research: A Tribute to Arnt Lykke Jakobsen.

Copenhagen: Samfundslitteratur, 2009. Print.

169

Miller, Thaddeus, et al. “Epistemological Pluralism: Reorganizing Interdisciplinary Research.”

Ecology and Society 13.2. 2008. Print.

Morse, Janice M., Judith E. Hupcey, and Maritza Cerdas. “Criteria for Concept Evaluation.”

Journal of advanced nursing 24.2. 1996. 385-90. Print.

Moser, Barbara. “Simultaneous Interpretation: A Hypothetical Model and its Practical

Application.” Language Interpretation and Communication. Ed. Gerver, D. et Sinaiko,

W.H. New York: Plenum Press, 1978. 353-368. Print.

Moser-Mercer, Barbara. “Paradigms Gained Or the Art of Productive Disagreement.” Bridging

the Gap: Empirical Research in Simultaneous Interpretation. Eds. Sylvie Lambert and

Barbara Moser-Mercer. Amsterdam: John Benjamins, 1994. 17-23. Print.

Mulkay, Michael J. “Three Models of Scientific Development.” The Sociological review 23.3.

1975. 509-26. Print.

Muñoz Martín, Ricardo. “La Lingüística Cognitiva, La Teoría De La Traducción y La Navaja De

Occam [Cognitive Linguistics, Translation Theory, and Occam’s Razor]. Paper Presented at

I Taller Internacional De Traducción e Interpretación.”.6-9 December, La Habana: 1994.

Print.

---. “Typos & Co.” Eds. S. Göpferich, A. L. Jakobsen, and Inger Mees. Copenhaguen:

Samsfundliteratur, 2009. 167-190. Print.

170

---. “Leave no Stone Unturned: On the Development of Cognitive Translatology.” Translation

and Interpreting Studies.The Journal of the American Translation and Interpreting Studies

Association 5.2 2010a. 145-62. Print.

---. “On Paradigms and Cognitive Translatology.” Translation and Cognition. Eds. Gregory

Shreve and Erik and Angelone. Amsterdam: Amsterdam: John Benjamins, 2010b. 169-187.

Print.

---. Minding Translation/Con La Traducción En Mente. 1 Vol. Alicante: Universitat Jaume I,

2014. Print.

---. “A Blurred Snapshot of Advances in Translation Process Research.” Minding Translation.

Ed. Ricardo Muñoz Martín. 1 Vol. Alicante: Universitat Jaume I, 2014a. 49-84. Print.

---. “Situating Translation Expertise: A Review with a Sketch of a Construct.” the Development

of Translation Competence: Theories and Methodologies from Psycholinguistics and

Cognitive Science. Eds. John Schwieter and Aline and Ferreira.Cambridge Scholars

Publishing Cambridge, 2014b. 2-56. Print.

---. “Processes of what Models?” Translation Spaces 5.1 2016a: 145-61. Print.

---. “Of Minds and Men - Computers and Translators.” Poznan Studies in Contemporary

Linguistics 52.2 2016b: 351-81. Print.

---. “Reembedding Translation Process Research. an Introduction.” Reembedding Translation

Process Research. Ed. Ricardo Muñoz Martín. 128 Vol. Amsterdam: John Benjamins

Publishing Company, 2016c. 1-20. Print.

171

---. Reembedding Translation Process Research. 128 Vol. Amsterdam: John Benjamins

Publishing Company, 2016d. Print.

---. “Looking at the Future of Cognitive TS.” The Handbook of Translation and Cognition. Eds.

John Schwieter and Aline and Ferreira. Wiley Blackwell, 2017. 555-572. Print.

Muñoz Martín, Ricardo, and Celia Martín de León. “Fascinatin' Rhythm – and Pauses in

Translators' Cognitive Processes” forthcoming. Print.

Neunzig, Wilhelm. “Estudios Empíricos En Traducción: Apuntes Metodológicos.” Cadernos de

traduçao 2.10. 2002. 75-96. Print.

Newell, Allen, and Herbert Alexander Simon. GPS, a program that simulates human thought.

1961. Print.

Nosek, B. A., et al. “SCIENTIFIC STANDARDS. Promoting an Open Research Culture.”

Science (New York, N.Y.) 348.6242 (2015): 1422-5. Print.

O’Brien, Sharon. “Methodologies for Measuring the Correlations between Post-Editing Effort

and Machine Translatability.” Machine translation 19.1. 2005. 37-58. Print.

---. Cognitive Explorations of Translation. New York: Continuum, 2011. Print.

---. “The Borrowers: Researching the Cognitive Aspects of Translation.” Target.International

Journal of TS 25.1. 2013. 5-17. Print.

Olohan, Maeve. Intercultural Faultlines: Research Models in TS I: Textual and Cognitive

Aspects. Manchester: St Jerome, 2000. Print.

172

PACTE Group. “Acquiring Translation Competence: Hypotheses and Methodological Problems

of a Research Project.” Investigating Translation: Selected Papers from the 4th

International Congress on Translation. Eds. Allison Beeby, Doris Ensinger, and Marisa and

Presas. Amsterdam: John Benjamins, 2000. 99-106. Print.

---. “Building a Translation Competence Model.” Triangulating Translation. Ed. Fabio Alves. 45

Vol. Amsterdam: John Benjamins, 2003. 43-68. Print.

---. “First Results of PACTE group’s Experimental Research on Translation Competence

Acquisition: The Acquisition of Declarative Knowledge of Translation.”

MonTI.Monografías de Traducción e Interpretación. 2014. 85-115. Print.

Peirce, Charles S. “The Fixation of Believe.” Popular Science Monthly.12. 1877. 1-15. Print.

Piccinini, Gualtiero. “Computationalism.” The Oxford Handbook of Philosophy of Cognitive

Science. Eds. Eric Margolis and Stephen and Stich. Oxford: Oxford University Press, 2012.

222-249. Print.

Pöchhacker, Franz. “Issues in Interpreting Studies.” He Routledge Companion to TS. Ed. Jeremy

Munday.Routledge, 2009. 128-140. Print.

Pokorn, Nike K. “In Defence of Fuzziness.” The Metalanguage of Translation. Eds. Y. Gambier

and L. and van Doorslaer. Amsterdam: John Benjamins, 2009. 135-144. Print.

Popper, Karl R. The Logic of Scientific Discovery. London: Hutchinson, 1959. Print.

173

Protevi, John. “Deleuze and Wexler: Thinking Brain, Body and Affect in Social Context.” 2010.

Print.

Pym, Anthony. “Ideologies of the Expert in Discourses on Translator Training.” Problems and

Trends in the Teaching of Interpreting and Translation. Eds. M. Snell-Hornby and Y.

Gambier. Misano: Istituto San Pellegrino, 1996. 139-149. Print.

---. “Why Common Ground is Not Automatically Space for Cooperation

.” Target.International Journal of TS 2.12 (2001): 334-7. Print.

---. “Redefining Translation Competence in an Electronic Age. in Defence of a Minimalist

Approach.” Meta: Journal des traducteursMeta:/Translators' Journal 48.4. 2003. 481-97.

Print.

---. “TS and Western Philosophy.” A Companion to TS. Eds. Piotr Kuhlwczak and Karin and

Litau. Clevdeon: Multilingual Matters, 2007. 24-44. Print.

---. Exploring Translation Theories. 2nd ed. New York: Routledge, 2014. Print.

Pym, Anthony, Miriam Shlesinger, and Daniel Simeoni, eds. Beyond Descriptive TS:

Investigations in Homage to Gideon Toury. 75 Vol. Amsterdam: John Benjamins

Publishing, 2008. Print.

Quine, Willard Van Orman. From a Logical Point of View. Cambridge, MA: Harvard University

Press, 1953. Print.

Quine, Willard Van Orman. “Translation and Meaning.” Word and object. 1960. 26-79. Print.

174

Reagan, M. D. “Basic and Applied Research: A Meaningful Distinction?” Science (New York,

N.Y.) 155.3768. 1967. 1383-6. Print.

Risku, Hanna. “Aktive Expertenroutine Oder Reaktive Verhaltensautomatik? Überlegungen Zum

Begriff Der Übersetzungsfertigkeit Bei Wilss.” TEXTconTEXT93 4. 1994. 237-53. Print.

---. “Situatedness in TS.” Cognitive systems research 3.3. 2002. 523-33. Print.

---. “A Cognitive Scientific View on Technical Communication and Translation: Do

Embodiment and Situatedness really make a Difference?” Target.International Journal of

TS 22.1. 2010. 94-111. Print.

---. “Translation Process Research as Interaction Research: From Mental to Socio-Cognitive

Processes.” MonTi: Monografías de Traducción e Interpretación.1. 2014. 331-53. Print.

Robbins, Philip, and Murat Aydede. The Cambridge Handbook of Situated Cognition.

Cambridge University Press Cambridge, 2009. Print.

Rockwell, W. Teed. Neither Brain nor Ghost: A Nondualist Alternative to the Mind-Brain

Identity Theory. Mit Press Cambridge, MA, 2005. Print.

Rojo, Ana, and Marina Ramos. “the Impact of Translator’s Ideology on the Translation process:

A Reaction Time Experiment.” MonTi: Monografías de Traducción e Interpretación.1.

2014. 247-71. Print.

Rojo, Ana, Marina Ramos, and Javier Valenzuela. “The Emotional Impact of Translation: A

Heart Rate Study.” Journal of Pragmatics 71. 2014. 31-44. Print.

175

Rowlands, Mark. The Body in Mind: Understanding Cognitive Processes. Cambridge University

Press, 1999. Print.

Martín Ruano, M. Rosario. “Conciliation of Disciplines and Paradigms.” TS at the Interface of

Disciplines. Eds. João Ferreira Duarte, Alexandra Assis Rosa, and Teresa and Seruya. 68

Vol. Amsterdam: John Benjamins Publishing, 2006. 43-54. Print.

Saldanha, Gabriela, and Sharon O'Brien. Research Methodologies in TS. Routledge, 2014. Print.

Schäffer, Moritz, and Michael Carl. “Measuring the Cognitive Effort of Literal Translation

Processes”. Proceedings of the workshop on humans and computer-assisted translation

(HaCaT). Stroudsburg, PA. Association for Computational Linguistics , 2014. 29-37. Print.

Schäffner, Christina, ed. Translation Research and Interpreting Research: Traditions, Gaps and

Synergies. Clevedon: Multilingual matters, 2004. Print.

Schleiermacher, Friedrich. “On the Different Methods of Translating.” Western Translation

Theory from Herodotus to Nietzsche. Ed. Robinson Douglas. New York: Routledge,

1813/2002. 225-238. Print.

Seeber, Kilian G. “Cognitive Load in Simultaneous Interpreting: Existing theories—new

Models.” Interpreting 13.2. 2011. 176-204. Print.

Seidenberg, Mark. Language at the Speed of Sight: How we Read, Why so Many CanÕt, and

what can be done about it. Basic Books, 2017. Print.

176

Seleskovitch, Danica. “Language and Cognition”. Language interpretation and communication.

Ed. David Gerver.New York: Springer , 1978. 333-341. Print.

Sellars, Wilfrid. “Abstract Entities.” The Review of Metaphysics. 1963. 627-71. Print.

Shapiro, Lawrence A. “Embodied Cognition.” The Oxford Handbook of Philosophy of Cognitive

Science. Eds. Eric Margolis and Stephen and Stich. Oxford: Oxford University Press, 2012.

118-146. Print.

Shreve, Gregory M. “Cognition and the Evolution of Translation Competence.” Cognitive

Processes in Translation and Interpreting. Eds. J. M. Danks, et al. 3 Vol. Thousand Oaks:

Sage Publications, 1997. 120-136. Print.

---. “Knowing Translation: Cognitive and Experiential Aspects of Translation Expertise from the

Perspective of Expertise Studies.” TS: Perspectives on an Emerging Discipline. Ed.

Alessandra Riccardi. Cambridge: Cambridge University Press Cambridge, 2002. 150-171.

Print.

---. “The Deliberate Practice: Translation and Expertise.” Journal of TS 9.1. 2006. 27-42. Print.

---. “Recipient-Orientation and Metacognition in the Translation Process.” Translators and their

Readers: In Homage to Eugene A. Nida. Eds. Dimitriu Rodica and Miriam and Shlesinger.

Brussells: Les Editions du Hazard, 2009. 255-270. Print.

---. “Levels of Explanation: Translation Expertise as an Emergent Phenomenon.” (forthcoming).

Print.

177

Shreve, Gregory M., and Erik Angelone, eds. Translation and Cognition. Amsterdam: John

Benjamins Publishing, 2010. Print.

Shreve, Gregory, and Isabel and Lacruz. “Aspects of a Cognitive Model of Translation.”

Continuum Companion to Translation. Eds. J. Byrne and J. and Kearns. London:

Continuum Press, 2015. Print.

Shreve, Gregory M., and Bruce J. Diamond. “Cognitive Neurosciences and Cognitive TS.”

Border Crossings: TS and Other Disciplines. Eds. Y. Gambier and L. and van Doorslaer.

126 Vol. Amsterdam: John Benjamins Publishing Company, 2016. 141-168. Print.

Sirén, Seija, and Kai Hakkarainen. “Expertise in Translation.” Across Languages and Cultures

3.1. 2002. 71-82. Print.

Sjørup, Annette C. “Cognitive Effort in Metaphor Translation: An Eye-Tracking and Key-

Logging Study.” Cognitive Explorations of Translation. Ed. Sharon O'Brien. 18 Vol. New

York: Continuum, 2013. 197-214. Print.

Spivey, M. “Oxford Psychology Series. the Continuity of Mind.” 2007. Print.

Stegmüller, Wolfgang. The Structure and Dynamics of Theories. New York-Heidelberg-Berlin:

Springer-Verlag. 1976. Print.

Storer, N. W. “The Hard Sciences and the Soft: Some Sociological Observations.” Bulletin of the

Medical Library Association 55.1. 1967. 75-84. Print.

178

Stuart Mill, John. On Liberty. London: John W. Parker and Son, 1859. Print. Reprint, Ed.

Gertrude Himmelfarb, Harmondsworth: Penguin, 1974, 1982 .

Taleb, Nassim Nicholas. The Black Swan: The Impact of the Highly Improbable. 2nd ed. 2 Vol.

New York: Random house, 2010. Print.

Teixeira Carlos, and Sharon and O'Brien. “Investigating the Cognitive Ergonomic Aspects of

Translation Tools in a Workplace Setting.” forthcoming. Print.

Thagard, Paul. “The Best Explanation: Criteria for Theory Choice.” The journal of philosophy

75.2. 1978. 76-92. Print.

Thagard, Paul. “Societies of Minds: Science as Distributed Computing.” Studies in History and

Philosophy of Science Part A 24.1. 1993. 49-67. Print.

Tirkkonen-Condit, Sonja. “Theory and Methodology in Translation Research.” Empirical Studies

in Translation and Linguistics. Eds. Sonja Tirkkonen-Condit and Stephen Condit. Joensuu:

University of Joensuu, Faculty of Arts, 1989. 3-18. Print.

---. “Process Research: State of the Art and Where to Go Next?” Across languages and cultures

3.1. 2002. 5-19. Print.

---. “The Monitor Model Revisited: Evidence from Process Research.” Meta: Journal des

traducteursMeta:/Translators' Journal 50.2. 2005. 405-14. Print.

Tirkkonen-Condit, Sonja, and Riita and Jääskeläinen. “Automatised Processes in Professional

Vs. Non-Professional Translation: A Think-Aloud Protocol Study.” Empirical Research in

179

Translation and Intercultural Studies. Ed. Sonja Tirkkonen-Condit. Tübingen: Günter Narr

Verlag, 1991. 109. Print.

Tirkkonen-Condit, Sonja, and Riitta Jääskeläinen. Tapping and Mapping the Processes of

Translation and Interpreting: Outlooks on Empirical Research. 37 Vol. John Benjamins

Publishing, 2000. Print.

Toulmin, Stephen. “Does the Distinction between Normal and Revolutionary Science Hold

Water?” Criticism and the Growth of Knowledge. Eds. I. Lakatos and A. Musgrave.

Cambridge: Cambridge University Press, 1970. 49-50. Print.

Toury, Gideon. “A Rationale for Descriptive TS.” Dispositio 7.19/21. 1982. 23-39. Print.

---. “Aspects of Translating into Minority Languages from the Point of View of TS.”

Multilingua-Journal of Cross-Cultural and Interlanguage Communication 4.1. 1985. 3-10.

Print.

---. “Experimentation in TS: Achievements, Prospects and some Pitfalls.” Empirical Research in

Translation and Intercultural Studies. Ed. Sonja Tirkkonen-Condit. Tübingen: Günter Narr

Verlag, 1991. 45-66. Print.

---. “A Handful of Paragraphs on'Translation'and'Norms'.” Current Issues in Language & Society

5.1-2. 1998. 10-32. Print.

---. Descriptive TS and Beyond: Revised Edition. 100 Vol. John Benjamins Publishing, 2012.

Print.

180

Turing, Alan Mathison. “On Computable Numbers, with an Application to the

Entscheidungsproblem.” Proceedings of the London mathematical society 2.1. 1937. 230-

65. Print.

Tymoczko, Maria. “The Neuroscience of Translation.” Target.International Journal of TS 24.1.

2012. 83-102. Print.

Vermeer, Hans J. “Skopos and Commission in Translational Action.” Readings in translation

theory (1989): 173-87. Print.

Whewell, William. The Philosophy of the Inductive Sciences. London ed. 2 Vol. John W. Parker,

1840. Print.

Wilhelm, Jane Elisabeth. “Jean-René Ladmiral–une Anthropologie Interdisciplinaire De La

Traduction.” Meta: Journal des traducteursMeta:/Translators’ Journal 57.3 (2012): 546-63.

Print.

Williams, Jenny, and Andrew Chesterman. The Map: A Beginner's Guide to Doing Research in

TS. Routledge, 2014. Print.

Wilson, Edward O. Consilience: The Unity of Knowledge. London: Little, Brown and Company,

1999. Print.

Wilson, Margaret. “Six Views of Embodied Cognition.” Psychonomic bulletin & review 9.4

(2002): 625-36. Print.

181

Wilson, Robert A. Boundaries of the Mind: The Individual in the Fragile Sciences-Cognition.

Cambridge University Press, 2004. Print.

Wilss, Wolfram. “Perspectives and Limitations of a Didactic Framework for the Teaching of

Translation.” Translation Applications and Research. Ed. Richard Brislin. New York:

Gardner Press, 1976. 117-137. Print.

Wittgenstein, Ludwig. Philosophical Investigations. John Wiley & Sons, 1953, 2010. Print.

Zanettin, Federico, Silvia Bernardini, and Dominic Stewart. Corpora in Translator Education.

Routledge, 2014. Print.

182

APPENDIX A

Glossary

Term Definition Source assumption something accepted as true Merriam Webster Dictionary concept an inferential role in a network Based on Sellars (1963) construct Concept based either on empirical observation Based on Oxford A or theoretical argumentation that is guided by Dictionary of Psychology a particular framework and subject to its application in a particular research design or model. design The general plan of an experiment, including Oxford A Dictionary of the method of assigning research participants Psychology or subjects to treatment conditions, controlling extraneous variables, manipulating the independent variable, and measuring the dependent variable. framework (…) set of ideas and approaches that can be (Saldanha & O'Brien, used to view and gather knowledge about a 2014:12) particular domain. hypothesis A tentative explanation for a phenomenon, Oxford A Dictionary of subject to criticism by rational argument and Psychology refutation by empirical evidence. idea Any conception, thought, or mental content. Oxford A Dictionary of Psychology

183

model A deliberately simplified and usually idealized Oxford A Dictionary of and imaginary representation of Psychology a phenomenon, with fundamental properties that are explicitly defined (or physically built) and from which other properties can be deduced by logical reasoning (or by empirical observation if the model is a physical object). Inferences from the model apply only to the model and not necessarily to the reality that it purports to represent, but if the model captures the important features of the phenomenon, then such inferences may apply also to the phenomenon itself. notion A particular idea or opinion not necessarily Based on Merriam Webster shared or agreed upon by a community. Dictionary paradigm A framework of concepts, results, and The Oxford Dictionary of procedures within which subsequent work is Philosophy structured. Normal science proceeds within such a framework or paradigm. A paradigm does not impose a rigid or mechanical approach, but can be taken more or less creatively and flexibly. phenomenon Anything that can be perceived or observed. Oxford A Dictionary of Psychology tenet A principle, belief, or doctrine generally held Merriam Webster Dictionary to be true; especially: one held in common by members of an organization, movement, or profession. theory A proposition or set of propositions offered as Oxford A Dictionary of a conjectured explanation for an observed Psychology phenomenon, state of affairs, or event.

184

APPENDIX B

List of acronyms

 TS: Translation Studies

 CTS: Cognitive Translation Studies

 TPR: Translation Process Research

 DTS: Descriptive Translation Studies

185