The use of communication strategies by Spanish learners of English : the collaborative creation of meaning, language and linguistic knowledge

FACULTADE DE FILO LOXÍA Departam ento de Filolo gía Inglesa THE USE OF COMMUNICATION STRATEGIES BY SPANISH LEARN ERS OF ENGLISH. A STUDY OF THE COLLA BORATIVE CREATION OF MEANING, LANG UAGE AND LINGUISTIC KNOWLEDGE VOLUME I Ph.D. dissertation subm it ted by Ana M. Fernández Dobao and supervised by Dr. Ignacio M. Palaci o s Martínez 2004 Ana M. Fernández Dob ao V o B o (Supervisor) Ignacio M. Martínez Palacios ACKNO WLED GMENT S - iii - ACKNOWLEDG MENTS I was a ble to com plete this dissert ation with the guidance , support and encou ragement of many people and i nstitutions. Without them this work woul d not have b ee n po ssible. I wan t to open this vol um e expre ssing my most sincere gratitude for their help and contribut ions. First and foremost I w ish to express my dee pest g ra titude to the supervis or of this work Dr. I gnacio Palacios Martínez, who has a lways provided me with intellec tual guidance a nd e ndless encourageme nt. I am thankful for his interest in al l the diffe rent stages of t his research, his c areful re ading of each of t he c hapters of this di ssertation, and his insightful a nd invaluable comments on my original m anuscripts. Most of all, I want to thank him for his confidence in my work, the genuine considerati on he ha s alwa ys shown f or all hi s student s, and his unfailing help and en couragement a ll through these years. Likewise , I w ould like to expres s my gratitude to Professor Elaine Tarone. I dee ply apprec i ate her c oncern with my investigation, t he time she de v oted to discus s it with me and the interest s he has constan tly sh owed for my work. Her illu minating comments and ideas have con stituted an invaluable he lp during a ll these years, and the final completi on of this d isserta tion is ve ry m uc h i ndebted to her always a vail able guidance and support . During thes e years I ha ve also ha d the opportunity to disc uss some aspects of my research with other scholars. I wa nt to tha nk Professor M artin Bygate f or his appreci ation of my ini t ial work in t he field of Sec ond Language Ac quisition, and the comments and e ncouragement he offered i n the first s tages of t his rese arch during m y stay a t Leeds University. I a lso wish to t hank Professo r Ann Lazaraton for her guidance in the tra nscription and analysi s of my da ta and Professor Andrew C ohen for fa cilitating my work at the Univ e rsity of Minnesot a. I owe my deepest gratitude to all t he people who helpe d me in the experime nta l phase of my re search. I wa nt to thank Dr. Mario Cal Varela a nd Dr. E lena Se oane Pose for their support, and show my most since re appreciation for the s tudent s who participated in thi s study. This re sea rch would not have bee n possi ble without the free and unconditional collaboration of t he Eng lish and S panis h st udents of the Universida de ACKNOW LEDGME NTS - iv - de Santiago de C ompostela who volunteered to take part in my research expe riments. They did not hesitate to sha re their time w ith me and to coope rate unre serv e dly. I will alwa ys be indebted to all of them f or their generosity. As regards the funds that ma de possible this investigation, I than k U niv ersida de de Sa ntiago de Composte la (Vicerrectorado de Investigación ) and Xunta de G a licia (Secreta ría Xeral de Inv e stigac ión e D e senvolvemento Científico) f or th e scholarshi ps I received f rom them. Re s ea r ch f or this study was also partly conduc ted within the project PGI DT00 PX120407PR, financed by Xunta de Galicia. Finall y, most s pecial thanks to my f a m ily, friends a nd col leagues for t heir unconditi onal support and encouragemen t. The w ork carried ou t t hrough a ll the s e years has g iven me the p ossibility of meeti ng great people at Universidade de Sant iago de Compost ela, University of Minnesota, B rown University and Unive rsité de Montréal. Thanks to all of you for list ening to me. Thanks to my frie nds for the ir constant love and support. Their e ncouragement in the momen t s of des peration helped me to kee p sight of my goals. A final and spec ial word of t hanks goes to m y pa rents, Man olo and Sisi, and my br other, Ferna ndo. Without thei r love and patience t he c ompletion of t his dissertati on would not hav e bee n possib le. TABLE OF C ONTEN TS - v - TABLE O F CONTENTS VOLUM E I Key to abbre viations xi L ist of tables xiii INTRODUCTION 1 THEOR ETICAL BACKGROUND 15 Chapter 1: Liter ature review 17 1.1. CS resea rch 20 1.1.1. Def inition of CSs 22 1.1.2. Cla ssification of CSs 31 1.1.3. Empiri cal research on CSs to date 45 1.1.3.1. The effect of learner-related factors on C S use 49 1.1.3.2. The effect of task- and contex t-related factors on CS use 54 1.1.3.3. CS use in interac tional conte xts 56 1.1.3.4. CS effect iveness 61 1.1.3.5. The teachabilit y of CSs 62 1.1.3.6. Concl uding rem arks 66 1.2. Comm unication models : t he collab orative model 67 1.2.1. The c ollaborative theory 73 1.2.2. The c ollaborative model at work 80 1.2.2.1. The e lementary excha nge 81 1.2.2.2. Contributi ons by turns 82 1.2.2.2.1. Re pairs 84 1.2.2.2.2. Expa nsions 84 1.2.2.2.3. Re placements 85 1.2.2.3. Contributi ons within turns 86 1.2.2.4. Contributi ons via sentence parts 87 1.2.2.4.1. Insta llment contributi ons 88 1.2.2.4.2. Trial constituents 88 1.2.2.4.3. Com pletions 90 1.2.2.5. Failures i n the grounding process 91 1.2.2.5.1. Ignoring 91 TABLE OF CONTEN TS - vi - 1.2.2.5.2. Fei gni ng unde rstanding 92 1.2.2.5.3. Misunde rstanding 92 1.2.3. Summary and c oncludi ng re marks 92 1.3. Intera ctionist approaches to SLA: coll aborative dialogue 96 1.3.1. Interac tionist approaches to SLA 97 1.3.2. Coll aborative dialogue 103 1.3.2.1. Collaborative d ialogue and L2 gr amm atica l development 107 1.3.2.2. Collaborative d ialogue and L2 lexical knowledge acquisition 112 1.3.3. Empiri cal research on SLA from a sociocultural p erspect ive: research and pedagogical im plications 115 EMPIRICA L STUDY 123 Chapter 2: The objec t of study: CSs and strate gic c omm u nicati on in face-to-f ace FL interac ti on 125 Chapter 3: Me thod 137 3.1. Subj ects 139 3.2. Ma terials 145 3.2.1. L a n guage data co llection task 145 3.2.2. Ret rospective data collection tas k 158 3.3. Procedure s 164 3.3.1. Pilot e xperiment 164 3.3.2. Main expe riment 169 3.4. Da ta transcription 171 Chapter 4: Identific ation and classification of CSs 175 4.1. Identi fication of CSs 177 4.1.1. Analyzing per formance fea tur e s for the purposes of C S identification 180 4.1.1.1. Overt strategic behavior 181 4.1.1.2. Explici t s trategy markers 183 4.1.1.3. Implici t strateg y markers 185 4.1.2. Analyzing N L dat a for the purpose s of CS identifica t ion 188 4.1.3. Analyzing retros pective data for the purposes of CS ident ification 191 TABLE OF C ONTEN TS - vii - 4.1.4. Evalua tion of the CS identification ana lysis 196 4.2. Cla ssification of CSs 198 4.2.1. Avoidance CSs 201 4.2.1.1. Topic a voidance 202 4.2.1.2. Messa ge abandonment 203 4.2.2. Achie vement CSs 206 4.2.2.1. Code s witching 206 4.2.2.2. Approximati on 208 4.2.2.3. Use of al l-purpose words 214 4.2.2.4. Complete omission 215 4.2.2.5. Foreignizing 219 4.2.2.6. Grammatica l word coinage 221 4.2.2.7. L itera l tra nslation 222 4.2.2.8. Restructuring 225 4.2.2.9. Circuml ocution 226 4.2.2.10. Semantic word coinage 230 4.2.2.11. Appeal f or a ssistanc e 231 4.2.2.12. Nonverbal strat egies 233 4.3. The c oncept of CS episode 237 Chapter 5: The c ollaborative creation of meaning: anal yzing CS use in FL interac tion as a collaborative activity 241 5.1. Providing evidenc e of CS understandi ng for gr oun ding purposes in C S epis odes 248 5.1.1. Acknowle dgments as a g rounding proce dure in CS episodes 249 5.1.2. Repetit ion as a grounding pr oce dure in CS epi sodes 260 5.1.3. Initiat ing a relevant next contrib ution as a grounding procedure in CS epi sodes 262 5.1.4. Allowi ng t he conve rsation to conti nue as a grounding procedure i n CS epi sodes 268 5.1.5. Demon stration as a grounding procedure in CS episode s 269 5.1.5.1. Offerin g a n a lt ernative CS as a g rounding proce dure in CS episodes 271 5.1.5.2. Offerin g t he intended TL lexical item as a grounding procedure in CS episodes 273 5.2. Providing evidenc e of tro uble i n understanding for grounding purposes i n CS epi sodes 276 5.2.1. Conf irmation checks as evidenc e of trouble in understa nding for groundin g p urposes in C S episode s 277 TABLE OF CONTEN TS - v i ii - 5.2.1.1. Repetit ion as ev idenc e of trouble in und ers tanding for grounding purposes in CS episode s 278 5.2.1.2. Offerin g a n a lt ernative CS as e vidence of trouble in unders tanding for grounding purposes in CS episode s 281 5.2.1.3. Offerin g t he intended TL lexical item as evide nce of trouble in unde rstanding for grounding purposes in C S episodes 291 5.2.2. Cla rifi c ation reque sts as ev idenc e of trouble in understanding for groundin g p urposes in C S episode s 297 5.3. Comm unicative brea kdowns in CS episode s 301 5.3.1. Avoidance CSs and communicati ve breakdowns in CS episode s 303 5.3.2. Feigning understa nding and co mmunicative brea kdowns in CS episodes 306 5.3.3. Misunde rstandings and commun icative breakdowns in CS e pisodes 309 5.4. Summa ry of resul ts and concluding re marks 315 Chapter 6: The c ollaborative creation of language and ling uistic knowledge : analyzing strate gic co m munication as coll aborative dialogue 321 6.1. Colla borative dialogue in NNS-NS CS epis odes 327 6.2. Colla borative dialogue in NNS-NNS CS epis odes 337 6.3. Colla borative dialogue and the complexity of the L 2 le xica l acquisition proce ss 348 6.4. Limitati ons of collabo rativ e dialogue 358 6.5. Pedagogica l implications: the role of CSs in the L2 le arning process and the F L c lassroom context 366 6.6. Summa ry of resul ts and concluding re marks 374 CONCLUSIONS AND SUGG ESTIONS FOR F URTHER RESEARCH 379 General su mmary of t he present study and c oncluding re marks 381 Suggestions f or further research 400 REFERENCES 407 RESUMEN 435 TABLE OF C ONTEN TS - ix - VOLU ME II APPENDIX A: DATA COLLE CTION 1 APPENDIX A.1: Plac ement test 3 APPENDIX A.2: Data collect ion instruments 13 Data c ollection ta sk in 1999 study 15 Instructi ons 15 Story 15 Data c ollection ta sk in the present stud y 16 Instructi ons 16 Practi ce task 18 Mai n task 20 APPENDIX B: DATA TRANSCRIPTION 23 APPENDIX B.1: Transcription c onventions 25 Transcrip tion conventions: simpl i fied sys tem 27 Transcrip tion conventions: extende d system 28 APPENDIX B.2: Transcripts 31 APPENDIX C: DATA ANALYSIS 187 INTRODUC TION - 3 - Whe n t rying t o c ommunicat e in a seco nd or f or eign language (henceforth L 2 or FL) lea rn e rs confront a variety of frequent a nd inevitable com municative di fficulties directly resulting from thei r def icient com mand of t he t arget l anguage (TL). T hey need , i n general , more time than native s peakers (NS s) i n order to e xpress and understa nd mea ni ng. They do not alwa ys comprehe nd t heir interlo cutors’ message s and of ten ha v e to negotiate for meaning in o rder to pre vent m iscom munications. Furthermore, when try in g to make co ntributions to the discourse, they do not a lways find in their interlanguage (IL) s ystems t he linguistic resources that e nable them to con tribute precisel y what t hey intend to. The prese nt inv e stigati on is prompted by a n interes t in t he la tter of t he s e problems. That i s, i t e xamines how FL lea rners m anage to c omm unicate mea ning when the des ired TL it ems or s tructure s that help convey this m ea ning are not ava ilable . Gaps in lan guag e knowledge , lapses of memory or reca ll difficulties, which m ay also come up in first langua ge (L 1) comm unicat ion a re extremely freque nt w hen the command of th e TL is limited. Despite these difficulties, learners a re generally able to manage FL communication. The question is ho w do they do it? In ord er to m aintain stead y communication, learners sometimes circumvent linguistic di fficulties b y c hanging or reducing the con tent of their messages. They avoid reference t o a t opic or s implif y the ir c ontributions in o rder t o o vercom e the lack of the TL t erm o r expres sion nee ded to conve y their m essages a s ori g inally inten ded. M or e often, howe ver, they are a ble t o keep t heir c ommunicative in tentions and develop a n alternati ve means of expressi on to convey the c ontent of thes e m essages. They ma ke use of s ynonyms, de scriptions, native l anguage ( NL) t ransf ers or even ges tures that compe nsate for t he unavailabl e TL form and a llow for t he comm unication of th e orig inal ly i ntended idea. All th ese different t echnique s langua ge l earne rs use t o communicate in a FL despite th eir IL shortcomings ar e known in the field of Second L a ngua g e Acquisiti on (S LA) rese arch as communic ation strategies . Communica tion strategies (h ereafte r CSs) pl ay a maj or role i n L2 a nd FL communication, widely rec ognized by F L te ache r s a nd SLA resea rchers a like. L e arners, even adva nced level le arners, nee d to c onfront rec urrent linguistic diff iculties i n their FL inte ractions and to make frequent use of these stra tegies. The learne rs’ use of a FL is to a consi derable extent s trategic, regardless of whethe r they use it in the L2 cla ssroom INTRODUC TION - 4 - conte xt or in a m ore natural non-aca demic setting, or w hile i nteracting with NSs or wit h other NNSs of similar or different li nguistic background. CSs ha ve attracted a lo t of attention in th e field o f SLA resea rch. The c oncept o f CS a nd its s tudy has now more than 30 years of history a nd during t his ti me a considerable amount of both theore tical and empirical re search has bee n a ccumulate d in the area. An ov e rview of this r e searc h re veals, h ow e ver, that most of t hese s tudies have bee n conducted from a c learly lea rner-centered approac h. Attentio n has be en focused on the lear ner, that i s to say, on the c ognitive proce sses the y e ngage in whe n becomi ng aware of a linguistic difficulty and the strategic utt era nc e s they produce a s a r es ult o f these proce sses. This kind of resea rch has been abl e to describe and explain how lea rn e rs mana ge to compensate for or c ircumvent linguistic deficiencies in ord er to express their mes sages. It ha s no t bee n able to e xpla in, howe ver, how th ese messa ges are actually c omm unicated in f ace-to-face FL interaction. In other words , it has provi de d m uch evide nce on the c ognitive a nd linguistic as pects of CSs , but not on their interacti onal and c ommunicative f unctions –that is, on how, t hrough CS us e, learners and their i nterlocutors manage t o e stablish a mutual agree ment on mea nin g and ach iev e the successful communi cation of their mess ages despite t heir shared, or n on-shared, IL shortcomings. Communica tion of meaning as we underst and it is al ways a joi nt and collaborati ve a ctivity . It involves a t least t wo diff erent individuals trying t o agree on a mea ni ng. This impl ies that, with face-to-face oral interaction, individuals are working together to rea ch a common goal. Communic ation cann ot be u nderstood by looking only a t one side of the page, that is, a t what t he speaker sa y s or at what t he listener understa nds. W e belie v e t hat, in order to understand how c ommun i cation of meaning is ach iev e d, a ttention needs to be paid to the actions, e fforts a nd beliefs of all the interlocutors taking part in the communicative act. We have found that th e role of CSs in t he communicati on process has always bee n a cknowledged. However, they ha v e been gene r ally studied as pa rt of the s peech production process, in other words , as pa rt of the learner’s us e of the FL, not of the interacti on taking place between this learner and a t least a nother interlocutor. Re s e archers adopting what is known as the in ter actional perspec tive on the study of CSs ha v e point ed out the need of research to move i n t his d irectio n. Tarone has INTRODUC TION - 5 - alwa ys conceived CS use in relation t o “a mutual a ttempt of two interlocutors to a g ree on a mean ing in situations whe re requ isite mea ning struct ures do not seem to be shared” (Tarone, 1981 : 288), but she never des cribed in a s ys t em atic way the spec ific rol e o f the listener in th is mutual endea vor. In 1991 Yule and Tarone s ugg es ted the pos sibility of combini ng CS a nd n e gotiation of meaning fra m eworks of a nalysis in order to look at both sides of the page, that is, at t he speaker’s a nd the listener’s a ctions in strategic communication. These sc holars found tha t negotiati on of mea ning strategies u sed during inte raction t o solve c omprehensibility d ifficulties, su c h as clarif ication r eque sts or conf irmation checks, a re often followed by strategic at tempts to communi cate the initially inte nded m essage through an a lt ernat ive means of expression. M u tual comprehe nsion is thus ac hieved through the combinat ion of the listener’s use of a negotiation of meaning strategy and the s peaker’s use of a C S. Attem pts to describe stra tegic co mmunication a s a n interacti ve activity, involving the le arne r’s a nd interl ocutor’s co ordinate d use of c ommunication a nd negotiation of mea ning str ate gi e s a s sugg e sted by Yule and Taro ne (1991), have bee n made by Suni (1996), W illiams, I nscoe and T as ker (1997), W agner and Firth (1997), or Anderson (1998). Ho wever, this re searc h is sti ll l imited in scope; and it ha s of ten pa id more at tention to com pr ehe nsibility proble ms and negotiati on of meani ng s trategies than t o CSs. Other researchers have looked at the interl ocutor in th eir ana lyses of strategic communication, but c onsider this to be a passi ve receiver. They have a cknowledged the interlocutor’s inf luence on the lea rner’s decision about which CS t o use, but de nied their active participation in the strategic c ommunication of meaning process –cf. for instance Poulisse, Bongaerts a nd Kellerm an ( 1990), or L u ján-Ort ega ( 1997). In part, this is due to the fact that a la rg e amount of the rese a rch carried out in this field has be en centered on t he c ognitive a sp ects of CS use. R e searchers hav e a dopted a predomina ntly psycholingui stic pe rspect i ve to the study of CSs. As al ready mentioned, the y have bee n more i nterested in the c og nitive process es unde rlying CS use and in t he factors t hat may have an inf luenced on these processes tha n in the interacti onal or c ommunicative value of CSs . Cons equent ly , they ha ve pai d attention to the individual a nd mental acti ons of the F L le arner, rat her than to the conte xt of the interaction. INTRODUC TION - 6 - As previously state d, with t his study we a im to unders tand how co mmunication of mea ning is achieved in face-to-f ace FL st rategic in teraction. W e believ e t hat meani ng is not sent and rec eiv e d, but jointly crea ted in ong oing and socia lly s ituated interact ion. L ike Ta rone (1981), we consider that st rategic c ommunication involv e s a mu tual attempt of tw o inte rlocutors to a g ree on a mea ning. O n this b asis, we will a nalyz e FL lea rn e rs’ use of C Ss and their interloc utors’ reacti ons in ord er to try to ide ntify and descri be the acti ons that all the co nversational partici pants take together in the pu rsuit of one common goal: the succes sful co mmunication of thei r messages . We expect to be able to expla in communication of mea ning through CS use a s a colla bor ative acti vity involving the j oint action and effort of al l the interlocutors. For t his purpose, we will dra w on L 1 c ommunicati on studies. Research ca rried out o n NL non-strate gic inter action has been able to outline a theoretical framew ork that account s f or communication of meanin g a s a colla borative activity: the c ollaborative model of c ommunication (Cla rk a nd W ilkes-Gibbs, 1986 ; C lark and Schaefe r, 1987, 1989; Wilkes -Gibbs, 1997). This mode l de scribes how spea kers and addre ssees coordinate t heir individual a ctions and beliefs in order to reach the successful communication o f thei r mes sag es. W e intend h ere to bu il d on the collaborative prototype in or der to develop a paradigm of analysi s a ble to ac count for strategic communication in face-to-f ace FL i nteraction as a c ollaborative creati on of meaning process. The prese nt rese a rch project will be conducted in the context of a Spa nish university ins titution, the University o f Santi ago de Com postela, with Galician lear ners of English as a FL. These lea rn ers begin thei r E ng li sh Philology studie s with an averag e intermediate level of the FL and, after fou r or five years of E nglis h language and literature c ourses, they are e xpected to r e ach an adva nced, h ighly prof icient com mand of thi s language. The learner’s profi ciency l evel in the FL ha s bee n fou nd, i n previous research, to p lay a signific ant role in CS use –c f. Bialystok a nd Fröhlich (1980), Hyde (1982), Paribakh t (1985), Poulisse et al . (1990), Liskin-Gasparro ( 1996) or Jourdai n (2000). On this basis, and tak ing int o account the sp e cific c ontext of our research, we intend to a nalyze FL interac tions involving En glish lan guag e lea rn e rs wit h two different profic iency levels: intermediate and ad vanced. A pro fici ency pre- test w ill be administe red to st udents enrolled in beginner and ad vanced cour ses. We e xpect to be INTRODUC TION - 7 - able to select two homogenous groups of pa rticipa nts that may be considered represe ntative of the assumed intermediate and advance d language leve l of o ur average first and la st year students. Since one of the main ai ms of our research is t o acc ount for t he role of the interlocutor in the strategic com munication of meanin g process, we a lso f eel nece ssary to anal yz e a t least two dif ferent inte r ac tional c onditions: lea rner-learner and lea rner-NS. This does no t mean , howe ver, that we intend to compar e t hese two different condi tions. The purpose of this proje ct is t o dem onstrate that inte rlo cutors, i ndependently o f the ir TL co mmand, c ollaborate with the lea rn er in a mutual at tempt to communicate meani ng through CS use and to des cribe how this mutua l at t e mpt is ac complished. Whe ther or not native sp eaking i nterlocutors, because of their higher c ommand of the T L, ca n off er more eff i c ient support to t he learner is an issue that e xt ends the limits of t his investigation. He re we will try to e stablish a framework of analysis able to descr ibe FL strate gic interac tion as a j oint a nd c ollaborative activity, and that wi ll be general enough to acc ount f or both le arn e r-learner and learner-NS int eractions. If this f ramework proves to f it t he purposes of our researc h, it m ay then be used in f uture i nvestigations t o compa r e native a nd non-native speaking interl ocu tors’ be havior in strate gi c interaction. As e xplained above, FL lea rners make use of a CS whe n the y fi nd t hat the desire d TL lexic al i t em o r structure to convey the ir int ende d me ssage is not ava ilable in their I L system. NSs m ay f ind si milar problems in their L1 comm unications. Sometim es they ar e unable to re trie ve a word –the ti p-of-the-tongue phenom enon, or they ma y lack knowledge of th e vocabulary nece ssary to communicate ab out hi ghly specialize d topic s. These pro blems are, however, ce rtainly less f requent than t he linguistic p roblems the FL lea rn e r expe riences. We may then assume that in l ea rner- NS intera ctio n s, le arners will often make use of a CS to c ompensate for an unav a ilable TL word or expre ssion that is part of t he NS’s knowle dge of the TL. In their attempt to collabora t e with the learner to build a mutual agreement on a meaning, or tr ying to provide c orrect ive f eedback on lea rn e rs’ non-t arget-like strate gic utteranc es, NSs m ay provide this TL f orm. In othe r words, we bel ieve t hat CS us e in f ace-to-face FL int eracti on c an e licit n e w and rele vant input. W hile c ollaborating t o com municate mea ning, interl ocutors may provide the TL item o r struct ure the lea rner is struggling f or. Fur therm ore, we shoul d a lso a ssume tha t other NNSs ca n als o pro vide this ne w input. A word unknown by one l ea rner may be INTRODUC TION - 8 - known by anoth er le arner f rom t he same proficiency level a nd used for the purpose s of agreein g on a meani ng within the context of strateg ic i nteraction. In a second sta ge of our research we will analyze native and non-n ati ve speaki ng interlocutors’ res ponses to le arners’ CSs t o i dentify if, in the ir att empts to colla borate in the mea ni ng c reation process, these int erlocutors a re a ble to provide new input f or the FL le arne r. T hat is, the y ma y col lab orate not o nly to c rea te mea nin g but al so to construc t acc urate and appropriate T L f orms to convey t his me aning. I f this oc curs, as we ex pect it wil l, we wil l pa y at tention to t he nature of t he st rateg ic interacti onal exc hang e , trying to fi nd out whether thi s also offers the necessary conditi ons for this new input to become intake . In s um, we wil l anal yz e stra tegic communica tion as a possible occasion fo r L 2 developme nt a nd we w ill try to demonst rate that C S us e in face-to-f ac e FL interact ion can be a source fo r SL A. Previous rese ar c h on CSs ha s overlooked the ir potential role in the S LA p roce ss. In the origins of this f ield of st udy s chola rs had to define the limits of the fuzzy and somet i mes ambiguous conc ept of CS. With this in mind, Tarone (1981) distinguishe d CSs from lea rning strategies on the grounds t hat the la tter a re used by the le arner i n order t o lea rn the language while the form er a re employed to use the language. Although Ta rone made explicit that a CS used to c ommunicate meaning may, consciously or sub consciously, he lp the learner to ac quire la nguage –cf. Tarone (1981: 67-68), the def inition of C Ss on the ba sis of the la nguage us e criteria prompted subseque nt re search t o approac h the st udy of the se s trategies as a study of lan guag e us e, not of language le arning. Some claim s have bee n ma de on the pos sible relationship between CS use a nd SLA. It has be en argued, for inst ance, that transfers ca n be used as a st rategy t o communicate m eaning but a lso to tes t hy pot heses (Corde r, 1978). It h as a l so been postula t ed that appea l for ass istance str ategies can be used as generators of new input (Færch and Kasper, 1980), and tha t stra t e g ic generalizati ons and word c oinages may b e incorporated to the learner’s IL s ystem when t hey e lic it p ositive f eedback (Færch and Kasper, 1980). It has been s ugg e sted tha t C S use ca n make already e xisti ng FL knowledge more accessible (Færch and Kasper, 1986), and th at, by enabling lea rners to kee p t heir conversatio ns continuous, these strate gies pr ovide them with mor e opportunitie s for ne gotiatio n of meaning and compre hensible i nput ( L arsen-Free man INTRODUC TION - 9 - and L ong, 1991). Finally, other sc holars, suc h as K a spe r a nd Kelle rman (1997), have also m aintaine d that cer tai n psyc holinguistic strategies d e velop l earners’ ana lyzed lexica l knowledge. Ho w ever, all thes e di fferent cla ims h ave been ma de f rom a theoretica l approach without clear attempts to test them empirically. In general, most researche rs have acknowle dged t hat C S use may have a n indirect inf luence on the S LA proce ss, a lthough arguments have al so bee n raised in the opposite direction –cf . for instance Kel lerman (1991). Ye t, none of the se researchers has attempted t o i nvestigate this issue in a systematic way . They se em to hav e followed Poulisse et al .’s claim that “ when c onducting a study of c ompensatory stra tegies [a subset of C Ss] it is irrele vant whether or not the us e of particular comp ensatory strate gies happens to have an additional learning e ffect” (Poulis se et al . , 1990: 8). As a result, Cook’s c laim that “at present the resear ch is second language res earch rat her than second la nguage acquisition resear ch” is still completely valid (Cook, 1 993: 137). We belie ve, however, tha t it is t ime to pay a ttention to the role of CSs in the SLA proce ss. By a ppr oaching the study of these strateg i es a s an analysis of FL interacti on, many and i nteres ting new i nsights c an be obt ained on this issue . The s tudy of inte raction as a so ur c e fo r L2 de velopme nt i s one of t he mos t dynamic a r eas o f research i n the field of SL A. In rec ent years s ociocultural appr oaches in this fiel d have reveale d t hat one form of i nteract ion directly pr omoting the SLA proc ess is what Swa in (2000) ha s c alled collaborat ive dial ogue . Collaborati ve dialogue is descr ibed a s a communicative an d c ognitive a c tivity, i n whic h spea kers and i nterlocutors use language as a tool to c ollaborate not onl y in the c reation of mea ning but als o in the c onstruction of accurate an d a ppropriate TL f orms to co nvey this m ea ning . This c ollabora tiv e construc tion of la nguage bui lds new T L k nowledge f or the FL l earner a nd ca n thus be seen as an occasion for the learne r’s IL system t o develop towards the TL norm. During this in vestigation we will analyze CS use as a possible t rig ger for collaborati ve dial ogue. As suggested above, we will try to demons trate that in FL interacti ons, when the T L items o r struc tures de sired to co nvey a m essage a re unavaila ble, learners and int erlocutors c ollaborat e to crea t e meaning. Wi thin th i s process, interl ocutors m ig ht also colla borate to gener ate TL and TL knowl e dge. In other words, the int erlocutor might be able offe r to the l earner previousl y un available TL items or struc tures. T hey m ay thus c ollaborate with th e speaker to create accurate a nd INTRODUC TION - 10 - appropriate TL f orms in orde r to co nvey the ir messages, a nd to build new TL knowledge f or the learner. In sum, wit hin the lim its of this research we i nte nd to a nswer two different but closely rel ated questions: how is com munication achi eved through C S use in face-to- face FL intera ction? Can C S us e in an interactional con t ext be an occasion for L2 developme nt? Specifically, we aim t o demons tr ate that com munication is a chieved through the joint collaboration of t he learner a nd the ir native or non-native spe akin g interlocutor; and that this c ollaboration c an al so be a sou rce of new T L knowl edge f or the F L le arner. For this purpose we ar e going to draw on thre e d ifferent areas of knowledge that have al ready be en mentione d throughout thi s introduct ory c hapte r: CS res earch, L1 communication studies and in teractionist perspectives on S LA . In Chap ter One a survey is presented conce rning the most relevant literature in these domains. We study thos e theoretica l frameworks that will constitute the basi s of our research. The fir st sect ion of this chapte r conce ntr a tes on CS res earch. Attention i s paid to the theoretical di scussion existing in thi s f ield around the issue of the defini tion a nd classificati on o f C Ss. Th e relatively wi de range of definit ions and taxonomies of CSs that have been pr oposed b y pre vious sc holars is re viewed a nd e valuated in a n a ttempt to obtain a better un derstanding of the c oncept of CS a nd of the di fferent types of CSs availa ble. W e describe previous empirical studies on C S use conduc ted within the field of SLA. We then e valuate t heir findings, discuss their contributi ons to the development of this are a of study and appraise the current state of empirica l research on CSs. In t he second a nd third sections of Cha pter One w e di scuss L1 c ommunica tion studie s a nd the theoretical a nd empi rical i nvestigation of the relati onship be tween interacti on and S L A. It i s not ou r int ention to re view all the studies or the oretical proposals ex i sting in these two are as of knowledge. We introduce s ection two c larifying the relationship existing betwe en L 1 communic ation a nd L 2 CS re searc h. On t his ba sis, we provide a j ustification for t he use of L 1 c ommunicati on models i n order to ex plain L 2 strate gic communication. We then describe th e collabora tive model of communication as we ll a s giv e an ov er view of the most rele vant empirical studi es supporting the sui tability of this archetype for the descripti on of face-t o-face oral interacti on. In sectio n th re e of thi s c hapter we intr oduce inter a ctionist p e r specti ves on INTRODUC TION - 11 - SLA by pro viding a n o verview of the developm ent of thi s a rea of resea rch. S wain’s (1997, 2000) th eoreti cal a pproach to t he a nal ysis of L 2 interac tion, th e concept of collaborati ve dialogue a nd the re sults o f previous e mpiric al research w ithin t his theoretica l framework are explained and as sessed. In Chapter T wo we int roduc e ou r em pirical research on CSs describing the object of s tudy: C Ss a nd s trate gic c ommunication in f ace-to-face F L inte raction. Building on the the oretical background outli ned in C hapter One, a working definit ion of CSs is propose d. This working de fini tion unde rlies and guides the methodology a dopted for the colle c tio n of the data, the analyses conducted on t his d a ta an d the c oncl usions drawn. It illustra tes not only wha t we cons ider t o be a CS and strategic co mmuni cation, but a lso our the oretic a l approach to the study of the phe nomena. It thus helps t o clarify the objectives of th is r esearch and how we in te nded to a chieve them. In this chapte r the scope of t he research is a lso n arrowed dow n to t he study of CSs used to so lv e lexi cal diffic ulti e s, that is, the unavailabili ty o f a T L le xical item or structure, therefore le aving out of o ur investigation phonological, syntactical or pragmatic problems. Cha pter Three is de voted to methodol ogical issues rel ated to the col lection of the data . Th e f irst section describes ho w we selected a nd grouped the English language lea rn e rs a nd En glish la ng u age NSs that took pa rt i n our proj ect. We giv e de tails of the instrument s used f or the elic it ati on of the data . The se include a communicat iv e ta sk aimed at elicit ing represe ntative samples of FL interac ti on and a pos t-i nterview to obtain learners’ and interlocutors’ retrospe ctiv e c omment s on their inte ractional performance. I n the f inal section o f this c hapter we explain and e valuate our d at a collec tion procedure. The analy sis of the data i s c onducted in C hapter Four, Chapte r Five and Ch apte r Six. F i rst we a nalyze the transcripts of the FL intera ction samples elicited for th e purposes of our research in orde r to ident ify lexical d ifficulties and subse quent CS use s. The proce dure empl oyed for this purpo se is based on a t riangulation of three diffe rent sources of evidence : performance features or proble m indicators, NL data and retrospect iv e comments . This proc edure is outlined a nd c onsidered i n t he first s ection of Chapter Four. The i dentified CSs are also des cribed and clas sified int o a coherent taxonomy of C Ss. In the sec ond section of Cha pter Four, we describe this ca tegorization and t he procedure employed to assi gn e ach C S t o the proposed c ategories. I n the final CHAPTE R 1: Literature re view - 19 - This fir st c hapter prese nts the theoretical backg rou nd o f t he current i nve s tigation. As explaine d in the int rodu ction to our work, thi s study draws on three primary are as of knowledge: CS researc h, L1 comm unication studies and interact ionist perspectives on SLA. We inte nd to build on previous a dvances made in these t hree di fferent but closel y related f ields, show that t he ir theoretical and e mpirica l frameworks can fi t together, a nd relate t hem to the analysis of CS use in f ace- to-face FL int eractio n. In t his cha pter a ttention is therefore p a id to pr evious work c arrie d out in the se three different areas . In th e f irst s ection we rev iew the mos t inf luential literature publishe d to d ate on CSs within the field o f S L A. Theoretical and em pirical a pproaches to the s tudy of s trate gic c omm unication are here surveye d with the aim o f gaini ng a better and m ore c omprehensive unde rstanding of the concept of CS, the different types of CSs available for the F L learne r, t he use that t his learner m akes of these problem- solving device s and the current state of the research on this is sue. The next s ection is centered on L1 c omm unication stud ies. We highlight the relationship existing betwee n this area of study and L2 C S resea rch, justifying the use of ana lyt ica l frameworks originally devised to a ccount for L 1 non-strategic communicati on in the a n alysis of L2 s tra tegic interaction. As als o explained i n the introducti on, for the purposes of the prese nt study we will dra w on the collaborative the ory of communication. We intend to appl y the c ollaborati v e mode l –as devised by Cla rk a nd Wilkes-Gibb s (1986), C lark a nd Schaefer (1987, 1989) a nd Wi lkes-Gibbs (1997), to the ana lysis of CS use in F L inte raction, and t o exp lain thus stra tegic communic ation a s a collaborati ve c reation of meaning proce ss. In t he sec ond s ection of the present c hapter we discus s the pr inciples of t he c ol laborative t heory and de scribe the c ollaborative model. Some attention i s also pai d to previous e mpirical resea rch carried out withi n t his theoretica l paradigm. In the final sect ion we review the work c onducted on interaction withi n the field of SL A. We intend to build on s ocioc ultural a pproaches to the study o f inte raction in order t o try t o f ind eviden ce of SL A processes oc curring within th e limits of stra tegic communication. More specifically, we will draw for t his purpos e on Swain’s t heoretica l proposals an d the con cept of collab orative dialogue (Swain, 1997 , 2000). In this cha pter, we provide f irst a general overv i ew of t he cu rrent state of the research in t his area, locating thus t he collabo rative dialogue theory in c urrent views on the role of CHAPTE R 1: Literature re vi e w - 20 - interacti on in SLA. We then c oncentrate our attention on th is theoret ical c onstruct, reviewing both theore tical and empir ical studies conducte d within this framework. 1.1. CS RESEARCH The c oncept of CS and i ts study dates back to 1972, w hen Sel inker, in his classical article “ Int e rlanguage”, introduce d for the first time the notion of strategies of L2 communic ation . One year la ter Váradi carried out the f irst theoretical and em pi rical investigation spec ifically de voted to this iss ue. 1 This would b e fol lowed by the work of Tarone who, in collaborati on with other scholars, c oined the term CS and prepare d the ground for wha t would soon become a fruitful a nd prol ific area of S L A re search –cf. Tarone, Cohe n and Dum as (1976), Tarone, Frauenfeld er and Selin ke r (197 6), a nd Tarone (1977). However, in orde r to arr ive at a f ull understa ndin g of the origin of CS res earch we need to go one d ecade backwa rds. Ce rtain a dvances made i n the f ield of SL A at the end of the 60s m ad e it p ossible th e appeara nce in the 70s of this new area of study. The strong developm ent o f me ntalistic th e ories played a major role in this iss ue. Behaviorism princ iple s st arted to be a bandoned an d empha sis shifted towards the mental processes at work in the learner’s us e of th e language. A s a result, the notion of strate gy , eithe r learning or com munication s trategy, beca me to be consi dered a relevant object of research. The c h ange in the view of errors produce d i n 1967 w i th the publi cation of Corder’s artic le “ Th e significance of lea rn e rs’ errors” al so c ontributed to rea djust the direction o f SL A resear ch, at this mome nt dominated b y Contrast ive Analys is an d Error Analysis s tud ies . In general, the inc reasing interest in learne r perf ormance, c omprising both erron eous a nd non-erroneous l anguage use , and the quick expansion of IL investigations favored the development of CS research. At a different level, t he ris e of c ommunicative approac hes to FL teac hing a nd the emphas is on the d e velopment o f communicative skills a lso prompted the i nterest for CSs. Schola rs and t eachers alike, realizing t he impossi bility of prov i ding L2 le arners with all t he linguistic resources t hey m ay ne e d for present and futur e co mmuni cation, 1 Váradi’s (19 73) paper is general ly considered t he first systematic i nvestigati on on CSs, but it i s not the first stud y p ublished in this are a. This paper was present ed at a c on fer ence in 1973 and, althou gh it circulate d among researc hers durin g the 70s, it did not beco me pu bli s hed until 1 980. CHAPTE R 1: Literature re view - 21 - starte d to recognize the value of CSs, to enc our age their use among L 2 le arners, and t o call f or mor e an d systematic work in thi s area of SLA. At the beginning of the research scholars were mainly conc erned with theoreti cal issues . It was necessa ry to esta blish a c lear def inition of the not ion of C S, a nd t o identify and cl assify the diffe r ent t ypes of CSs available in orde r t o delimit t he object of study and thus se t it apart f rom other relate d I L phe nome na, such as product ion or lea rning st rategies. Initial studies we re therefore basicall y dire cted a t defining the concept of C S and developing ta xonomies that could be us ed to cla ssify these strateg ies –cf. Váradi (197 3), Tarone, Cohen a nd D um a s (1976), Tarone, Fr a uenfelder and Selinker (1976), T arone (1977), C order (1978), or Færch a nd Kasper (1980, 1983a, 1983b). Whe n empirical observation was carri ed out , it was subordinated to theory. I t was usually intende d to c ollect da ta that co uld sust ain the defini tions or taxonomies proposed f rom a theoretical poi nt of view. In s ome of these initial studies re s e archers pointed o ut the possi bility that certain factors, such as prof iciency level, L1 bac kground or persona lity cou ld affect CS us e, but without a ttempting to study thes e iss ues in a systematic way – cf. Ta rone (1977) or Corder (1978 ). Once the s cope of CS research had b een es tablished, t here started to appear more e mpirical investigatio ns, a im ing at studyin g these and ot her related questi ons that could help t o unde rstand CS use in FL co mmunicati on –cf . Ervin (1979), Palmberg (1979, 1982), B ialystok a nd Fröh lich (1980), Hyde (1982), Bi aly stok (1983), or Haa strup and Phillipson (198 3). In 1983 Fæ rch a nd Ka sper pub lished a volume, Strate gi es in Interlanguage Communication , that incorporated the m ost influential pape rs written on CSs to that da te and that serv e d t o c onsolida t e this ar ea of study (Færch a nd Kaspe r, 1983a). This publica tion was foll owed by a c onstant a nd increase d interest in strate gic communication, and nowadays a considerable amount of l i terature, reporting on a variet y of a spects c oncerning both t he nat ure a nd t he use of CSs, has accumu lated in the field of SLA rese arch. In the pre sent cha pt e r we review the most influential work c onducted in this area, wit h the aim of providing a f air a nd comprehe nsive view of t he development of CS research from its very be ginnin g up to dat e. To clari fy the notio n of CS and t he s cope of this area of study, we first concentrate o n the theoretical iss ues of the defini tion and classificati on of CSs. We the n turn to C S empirical resea rch in orde r t o ide ntify a nd CHAPTE R 1: Literature re vi e w - 22 - asse ss what is known so fa r a bout the use that the FL learner makes of CSs, and the role that t h ese stra tegies play in the processes of L 2 use and acquisitio n. 1.1.1. DEF INITION OF CSs The questio n of the definiti on of the CS conc ept is obviously c entral f or the developme nt of the f ield. A clear and c onsistent cha racterization of what cons titutes a CS is not only nece ssary to de scribe the limits and nat ure of t he object of study, but it is also impor tant to esta blish the ba sis of a ny at tempt to i dent ify a nd c lassify C Ss, and t o dete rmine the me thodological approach t hat needs to be t aken to the analysis of the FL lea rn e r’s strategic la nguag e use. This issue ha s th erefore received a great de al of attent ion, particularly, as alrea dy pointed out, in the early years of the research. However, despit e the amount of work conduc ted and the rel evance of th e que stion, scholars have not yet b een able t o reach a consens us on the c onceptua lization and description of the CS c onstruct. In fac t, the definiti on of the CS con cept has become one of the most c ontroversial is sues in this are a of SLA studies. Since the notion of CS fi rst appeared in the literature, the re has been an ope n deba te among schola rs centered on the c ri teria t hat should be a dopted to dis tinguish strate gic from non-strategic comm unication. As a result of t his c ontrov ersy, there exists nowada ys a c onsiderabl e variety of defi nitions for t he con cept of CS that s om etimes cover different ranges of de vices and account for different aspe cts of the strategic communication process. The m ain reason f or thi s lac k of agree ment among sc holars is tha t the study of CSs ha s been a pproached from two different, though g ene rall y com patible, t heoretical perspecti ves: t he inte ractional and t he psycholi nguistic. 2 The se two perspec tives do not only un derlie the variet y of defi nitions ava ilable f or the concept; as we wil l s oon see , they also accou nt f or differences in the taxon omies proposed to classify CSs, the methodol ogical design of the empi ri ca l investigations ca rried out in thi s field and the position a dopted by different sc holars on the issue of the possible teachability of CSs. 2 In 1997 Y ule a nd Tarone re ferre d to these tw o ap proaches as the Pr os an d the Con s. The y w ere t hus accountin g for t he fact that whereas i nteractional re searc hers ha ve been very prolific i n th e number of CS types the y ha v e ide ntified, p s ycholin guistic scholars, s uch as P oulisse et al . (1990), have tended to be more con servati v e , tr y i ng a lwa y s to red uce the number o f CS ca tegories as much as possi ble. CHAPTE R 1: Literature re view - 23 - From a n interactional perspect ive CSs are viewed as “discourse strategies tha t are e vident i n i nteraction i nvolv i ng l ea rners” (Ellis, 199 4: 396). CSs resul t from the “ne gotiation o f an a greement on meaning” be tween t wo interlocutors (Taron e, 1981: 288) and ther efore the discuss ion concerning CS use is considere d to belong to the study of learner i nteraction. On this b asis, int eractionist res earchers ha ve f ocused the ir s tudies on the analysis of variability in linguistic performance. Thi s approa ch be gan with Váradi (1 973) a nd Tarone (1977, 198 0, 1981). R esearchers adopting an int eracti ona l perspecti ve, such as Hyde (1982), Paribka ht (1985), a nd mo re recently Wil liams et al . (1997), W agner a nd Firth (1997) , or Anderso n (1998), have gene ral l y drawn on Tarone’s (1980, 1981) definition of the c oncept of CS f or their em pirical analyses o f this I L phen omenon. Withi n the f ramework of the psycholinguistic app roach, CSs a re treated a s “co gnitiv e p rocesses i nvolv e d in t he use of the L2 i n rece ption a nd produc tion” (Elli s, 1994: 396). More spe cifica lly , CSs a re c once ived as “ mental plans implemented by the L 2 learner in res ponse to a n internal signal of an immine nt p roblem, a form of se lf-help that did not have t o eng a ge t he i nterlocutor’s support for resolution” ( Kasper a nd Kellerman, 1997: 2) and located within c ognitive mode ls of speech production. This is the ap proach taken by authors such as Færch a nd K asper (1980, 19 83b, 1984), Bi alystok (1990), Poulisse e t al . (1990), Poulisse (1993, 19 97), or Kelle rman and Bialyst ok (1997). Here we int end to provide a general overview of the de v elopme nt of t his theoretica l is sue, reviewing some of the m ost infl uential def initions pro pose d t o date from both pe rspectives. Special c onsideration i s given to Tar one’s (1980 , 198 1), Færch and Kasper’s (1983b) and Poulis se et al .’s (1990) def initions of the CS c once pt. Thes e three differe nt c onceptualizati ons of the notion of CS illustrat e t he principles that have bee n g e nerally used to distinguish strat egic f rom no n-strategic communication a nd, in some w ay or anothe r, th e y have i nfluenced the whole of the research carried out in t hi s area. We have already explained tha t the concept of CS has its ori gin i n t he not ion of strate gies of L 2 co mmunicati on, introduced by Sel inker in 1972 a nd acc ounted f or as a lea rn e r’s attem pt to expre ss meaning with a limited IL system. Selinke r, howev e r, never tried to dev e lop this notion or to go into detai l about the nature of C Ss. CHAPTE R 1: Literature re vi e w - 24 - The first at tempts to provide a system atic definition for the CS conc ept were made by Ta rone, Frauenf elder and Selinker (1976), an d Tarone, Coh en a nd Dum as (1976). A syste matic attempt by the learner to ex press m e aning i n the t arget lang uage, in situ ations whe re the appropriate sy stematic target language r ules have not been formed. (Tarone, Frauenfelder a nd Sel inker, 1 976: 100) A syste m atic atte mpt by the lear ner to expres s or deco de me aning in the target la ngua g e, in situ ations where the ap propr iat e systematic target lan guage r ules have not been f or m e d. (Tarone, C ohen and Dumas, 1976: 78) These two defin it ions rely on the sa me princi ples. The proposal made by Tarone, Cohen and Dumas (1976) repre sents a n e xtension of the previous de finition, i .e. Ta rone, Frauenfelde r and Selinker (1976), aim ing at relating the not ion of IL strate gies t o language c omprehension as well a s language production. These authors’ interpretation of the CS conc ept is sti ll fi rmly rooted i n the traditi on of Error Analysis. At this momen t a ny la nguage f orm that w a s erroneous compa r ed to the T L norm, irrespective of whe ther thi s was in acc ord ance with the lea rn e r’s I L s ystem or not, was being take n as evide nce of CS use. This e xplains why the not ion of CS appears restric t ed t o le a rners’ use of the langua ge. One year l ater Tarone, realizing that comm unicativ e problems “ may occur when one sp ea ks i n on e ’s L 1, a s well as w hen o ne a ttempt s to communicate in a n interlanguage” (Tarone, 1977: 195), proposed a m ore de veloped de fi nit ion f or the term CS. Conscious communic atio n strategies are u sed by a n indivi dual t o overc ome the crisi s wh ich oc curs w hen language s tructures are inadequat e to co nvey the in dividual ’s thought. (Tarone, 1 977: 195) This was not, however, Ta rone’s last proposal on thi s i ssue. Wit h the intentio n of rec ognizing t he previo usly overlooked interactional functi on o f CSs, thi s schola r built a ne w def inition of C Ss –pr esente d in T arone (1980, 1981), that als o a ccounts f or the role pla yed by the interlocutor in the stra tegic communication process . A mutu al atte mpt of tw o int erlocutors to agree o n a meanin g i n situa tions w here re quisite mea n i ng structures do not seem to be shared. ( Tarone, 1 981: 288) This de finition make s reference to two features that Ta rone considers are the two basic de fi ning c haracteristics of C Ss: la nguage use and i nteractional functi on. CSs a re CHAPTE R 1: Literature re view - 25 - seen as dir ect ly rela ted to la nguage u se and not a s p art of the spea ker’s linguistic knowledge. Quoting Tar one, “t hey are de scriptive of the learne rs’ pat t ern of use of what they know as they try to c ommunicate w ith speakers of the t a rget l anguage” (Tarone, 1981: 287). This int erpretation also implies t hat CSs a re used only whe n a communication problem arises a nd tha t, th e refore, Tarone c onsiders t hem to be problem-oriente d. CSs are also said to b e cha racterized by ha ving a n interactional function. They are used to ne got iate a n agreem ent on a mea ning between two interlocutors a nd, in t hi s way, ensure t hat both of them are talking about the s ame thing (Tarone, 1981: 288). The not ion of c onsciousness ha s been e xclu de d from the defin ition on the grounds that it i s not cl ear whethe r all instances of CS use real ly occur with the aware ness of the speaker. Tarone argues on this basis t hat it is preferable not to spe cify the de gree of consciousness in a de scription of CS s (Ta rone, 1981: 287). Finall y, t aking al l t he previous cons iderations into account , the a uthor propose s the following cri teria to characte riz e CSs: (1) a speaker desires to co mmunicate meanin g x to a listener; (2) the speake r believes the lin guistic or s ocioling uis t ic structur e desired t o commu nicate m ea ning x is unavail able, or is not sh ar ed with t he listener; thus (3) the speake r cho os es to (a) avoid – not atte mpt to c ommunicate meani ng x –or (b) attem pt alte rnate m e ans t o com municate m e aning x. T he s peaker s tops tryi ng alternatives when it see ms cle ar t o the spe aker that t here is sh ared meani ng. (Tarone, 1 981: 288) This set of cri teria serves not only to c larify what is meant by t he t erm CS, but also to differentiate these devices from other related IL phenomena , such as lea rning, production an d reception strate gies which, at this stage of the research, we re not ye t clearly de limited. CSs are disting uished f rom production str ate gi es –defined as “an attempt to use one ’s linguistic system efficie ntl y and clearly, with a minimum of effort”, e.g. the us e of prefabricat ed patterns, di scourse planning or rehe arsal (Tarone, 1981: 289); on the grounds t hat al though the lat ter a lso be long to language use the y ha ve no t an interactional f unction, i. e. criteria 3b is absent. L ea rning strate gies –“ an at tempt to develop li nguistic and sociolinguistic c ompetence in the target langua ge” (Tarone, 1981: 290); are dif ferent iated from C Ss on the basis of the f irst criterio n, i . e . motivation. The basic force behind a learnin g strat egy is the desire to learn the TL, whe reas the ba sic motivation of a CS is t o communicate meaning, alth ough Tarone recog nizes tha t C Ss CHAPTE R 1: Literature re vi e w - 26 - can als o i ndirectly c ontribute to IL de velopment . Fi nally, the first criterion c an also be used to distinguish CSs from p ercepti on s trat egies –de scribed a s “ the a tt empt to interpre t incoming utte rances efficiently, wit h the l east effort” (Tarone, 1981: 291). These defining cri t eria also ha ve the a dvantag e of be ing valid f or the description of L2 performance as well as for L 1 comm unicat ion. In g e nera l th e three co nditions est ab lished by Tarone f or the pre sence of a CS have bee n widely e xploit ed in late r research and proved t o be extremely he lpful for the identification of C S us es in IL performa nce. H owever, thi s does not m ean tha t Tarone’s definiti on has bee n general ly acce pted without be ing questi oned. It h as be en widel y criticized by thos e aut hors who prefer to a dopt a psycholinguis tic view o n C S use. Færch and Ka sper (1983b) a rgue, f or exam ple, a gainst the interact ional conditi on because the y co nsider it exc lud es the po ssibility of using CSs in a monologue, a nd Dörnyei and Sco tt (1997) bel ieve tha t, by introducing an int eractional perspe ctive, Tarone has b lurred the distinction betwe en CSs and negotiatio n o f meani ng de vi ce s. Whe r ea s Tarone (1980, 1981) de scribes CSs in interactional terms , Færc h a nd Kasper (1983b) loc ate these strategies i n the individual us er. They a ssume t hat the speaker i s the pe rson who expe rien c es t he proble m and, subsequentl y, the onl y one involved in de veloping a stra t egic pla n for its s olution. For t hese a ut hors, CSs are mental plans tha t belong to the speech production proce s s ; they need therefore to be explaine d in terms of a speech production mode l. On t his ba sis, Færch and Kaspe r (1983b) lo cate C S s i nto Cla rk an d Clark’s (1977) c ognitive mode l of s peech productio n. In this model the proce ss of la nguag e production is d ivided i nt o two pha ses: the planning and the exe cution phas e. In the planning stage t he speaker sets up a communicative goal and develops a plan to reach it. This m eans that the s peaker decides the ve rbal a ct the y wis h to perfo rm , the con tent they want t o communicate and how t o define t he relationshi p wi th t he interlocutor. That is, they es tablish the actional goal, th e propo sitional goal a nd the modal goal. Once th e goal ha s bee n i nstituted, the spea ker assesse s the situational c onditions and sele cts, from the communicative resources available, the rules and ite ms c onsidere d most a ppropriate to reach this goal, i.e. the speaker develops the plan. This plan i s then c arried out in the executi on phase. At this s tage t he ne urologic a l a nd physiologic al processes whic h will CHAPTE R 1: Literature re view - 27 - lea d t o the v e rbal behav i or expect ed to s atisfy the original g oal take place. Thes e include th e articulation of speech organs as well as the use of gestures. During all these sta ges spea kers monit or t he communicat ive proce ss and, if they find that t he original plan c annot be carried out successfully, the y either change their communicative goal or de velop an alternative plan; tha t is , t hey a ctivate a C S to solve the problem encountered – a reducti on or a compensatory s t rategy . CSs are usually placed in t he pl anning phase , when t he ind ividual f inds that the requi red linguistic mea ns to construct t heir pla n a re not a t t heir dispos al. Neverthe l ess, problems can also appear when the speake r attempts t o execute the pla n. These problems usually ha ve t o do with retrie ving t he it ems or rules c ontained in the pla n. On the basi s of this descr iption of the speec h p roduct ion process, Færc h a nd Kasper (1983b) propose the followin g de f inition f or the CS c oncept: Communic ation s trategie s a re potentiall y co nscious pla ns f or s olving what to an ind ividual presen ts it self as a proble m in reac hing a p articular c ommunic ative go al. (Færch and Ka sp er, 1 983b: 36) We can see in this definition tha t Færch and Kasper (1983b) c onsider the criteria of pro blematicity and consciousness a s the t wo basic f eatures to be taken into acc ount in a descr iption of CSs. C Ss are pro blem-oriented beca use they are e mployed w hen the speaker does not ha ve the res ources required to express the intended m eaning. The criteri on o f consciousness is seconda ry because it is de rived from the fi rst one . In order to expe rience a problem, one nee ds to be conscious that some kind of difficult y e xists . Færch and Kaspe r (1983b) re cognize, a s Tarone (1981) ha d previous ly poi nted out, t hat consci ousness is a graded ph enome na and t hat stra teg i es c an be used i n a mor e or le ss c ons cious way. It can also be the case that an i nitially c onscious pla n bec omes automatic a fter a time. For th is re ason t hey talk a bout “ potentially conscious plans” (Færch and Kasper, 19 83b: 35-6). This de scription also a ccounts for the two mai n categories of CSs, i.e. reducti on and com pensatory st rategies, expla ined as the r esul t of two di f ferent processe s. It shows as w ell that strategic la nguag e us e is more com plex than non-strategic languag e production, becau se it always involves the e xtra proces s of de veloping and exe cuting a second plan, i.e . a sub-plan. Pouli sse (1993), however, considers that the explanation is too gene ral and vague. It c annot a ccount fo r the different types of CSs within e ach category, e .g . for the differences bet ween paraphrase and transfer strateg i es, a nd it CHAPTE R 1: Literature re vi e w - 34 - This experime ntal categorizati on c onstitutes, however, the starting poin t of Tarone’s 1977 tax onomy of le xic al C Ss. 4 Bui lding on this p revious wor k and on the ana lysis of empirical data obtained from a small sa mple of L 2 stude nts this s cholar was able t o deve lop a c lassi fica tion s ys tem tha t has proven robust and thor ough acros s numerous e mpirical studies of CS use. TABL E 1.1. T ARONE’ S TAXONOMY OF C Ss STRATEG Y DESCRIP TION AND EX AMPLES 1. Avoidance a) Topic avoida nce The l e arner simpl y d o e s not t alk a bout c oncepts for which the vocabular y i s unknown b) Message aband onment The learner begins to talk about a c oncept but is unable to continue and begins a new sentence (e. g . ‘ the w a t er (mumble)’ for ‘the water spill s’) 2. Paraphr ase a) Appr oximation The lear ner uses a si ngle T L vocabula r y i tem o r s truc ture , which t hey know is not c o r rect, but w hich sh are s en ough semantic features wi th t he de sired i te m t o s atisf y t he le arner (e.g. ‘w o r m’ f or ‘silkw orm’) b) Word coinage The learner makes up a ne w word i n order t o c ommunicate a desired c oncept (e. g. ‘per son w orm’ to ref er t o ‘an anima ted caterpillar’) c) Circ umlocution Th e learner de scribes t he c harac teristics or ele ments of t he object or action instea d of using t he a ppr opriate TL structu re (e.g. ‘I d on’t know what’ s its na me. that’s u h, Pe rsian, and we use in Tur key, a l o t o f ’ t o describe ‘a waterpi pe’) 3. Consciou s transfer 5 a) Litera l translation The le arner tr anslates word for wo rd from the NL (e. g . ‘he invites hi m to drink’ in place of ‘t he y toast one anot her’) b) L a nguage switch The lear ner s i mpl y uses the NL t erm w i thout bothe ring to translate (e. g. ‘bal on’ for ‘balloon’) 4. Appeal f or assistanc e The le arner as ks f or t he c orrec t term: asks the expe rimentor, any NS, o r even r efers to a dic tionar y (e. g. ‘what is thi s? What called?’) 5. Mime The learner uses a n onverbal device to refer to an object or event (e. g. ‘and e ver y b od y sa ys [ clap hands] ’) ( Tar one, 1977: 197- 199) 4 Notice t hat, although i n t heir defi nitions of CSs researcher s d o not r estrict the sc ope o f t hese d evices to lexicalit y , i . e. to solve pr oble ms a rising from lexical deficits in t he lear ner’s IL s ystem, in their a ct ual anal y ses of strate gi c l an guage behavi o r atte ntion is g enerall y f ocused on t his p artic ular kind o f CSs. As a result, most of the taxonomies we are goin g t o revie w her e and most of t he empirical stu dies we will anal y ze in the next secti on are restricted to lexical CS s. 5 In 1981 Tar one refers t o this group of strategies a s bor rowing . CHAPTE R 1: Literature re view - 35 - Although Tarone (1980, 1981) makes in her definition of CSs a binary distinction betwee n a voi dan ce and ac hievement or com pensatory s trate gies –c f . present cha pter, p. 25 , her taxonom y, as illustrat ed i n table 1.1, is c onstituted around fi ve maj or categories. These reflect five dif ferent sorts of dec isions on how to solv e a lexical communicative problem and five diff erent sources of knowl edge. Corder (1978), adopt ing an approach to the d e scription and class i fication of CSs that ca n be c onsidered quite similar to Ta rone’s (1 977), sets out, li ke Váradi (1 973), a productive means of a ssig nin g CSs to two broad c ategories. In Cor der’s (1978) m odel when a spe aker finds a problem to a chieve a com municative goal they have two possibil ities: eit her to c hange the original goal or to c hange the route t o this goal. If they choose the fir st option, tha t i s, if t he speake r dec ides to “tailor his message to the resourc es he ha s available”, they will be usin g a mes sage adjustment strategy (C order , 1983: 17). If the y dec ide to change t he route bu t not the goal, that is, if t hey c hoose t o “atte mpt to increase his resourc es by one m eans or another in or der t o reali ze his communicative intentions”, they will be making use of a resource expansion strategy (Corder, 1983: 17). These two broad cate gories are furt her subdiv i ded into different t ypes of strate gies that more or less c orrespond to the CS types identi fied by Tarone (1977). Corder, h owever, di stinguishes four types of me ssage adj ustm e nt strategies whe reas Tarone recognizes only two. To th e t opic avoidance and m e ssage a bandonment categories Corder adds semantic avoidance –“saying some thing slightly di ffere nt from what you int ended but still broadly re levant to the t opic of discourse”, a nd me ssage reduction –“ saying less, or less precisely, what you intended to say ” (Cord er, 1983: 17). Data c ollected by diffe rent resea rch e rs for different purpose s ha s confi rmed the cohe r enc y a nd uti lity of Tarone ’s (1 977) classif ication. Eithe r fo llowing Tar one’s combination of CS t ypes in to five major c ategories or Corder’s bina ry di vision, ma ny scholars in the f ield ha ve adopt ed Tarone’s fra mework of ana lysis, obt aining res ults that support the c onsist e ncy and rel ia bil ity of t his mode l. Furthe rmore, alternative taxonomies proposed by oth er resea rche rs –cf. for in stance P ariba kht (1985), ca n usuall y be traced to Tarone’s original ca tegori es. As Bialystok (1990: 61) ex plains it, there is a core group of strat egies that appears i n variably in a lmost all taxonomies a nd Tarone’s is probably the one that be st captures these strategies . CHAPTE R 1: Literature re vi e w - 36 - From a different, th oug h not incompat ible pe rspec tive, Færch a nd Ka sper (1983b) propose , in line with t heir conceptualization of the notion of CS, a ta xonomy that represe nts a n attempt to distinguish CS categories reflect ing int erna l m ental processe s rather than external linguistic f eatures. The c l assifi cation that Færch and Kasper (1983b) su ggest is based on the m odel of speech produc tion that also informs the ir concept ualization and def ini tion of CSs , i.e. Cla rk and Clark’s (1977). We have a lready seen t hat this frame work ma kes it poss ibl e to draw a disti nction between t wo main types of CSs: re duction and com pensatory strate gies. T hese t wo categor ies ope rate by al tering di ff erent aspects of the produc tion process. In the first c ase the lea rner changes or m odifies the ir c ommunicative goal; in the sec ond, they change the initially devised pl an to reach this goal. TABLE 1.2. FÆRCH AND KASPER’S TAX ONOMY OF C Ss STRATEG Y DESC RIPTION 1. Reducti on strategies 1.1. Forma l reducti on strate gies a) phonol ogical b) morp hological c) s yntactic d) lexical The lea rner com municates b y means of a ‘re du ced’ sy s te m, i n orde r t o a vo i d produc ing non-flue nt or i ncorrect utter ances by realizi ng insuffic ientl y automat ized or hypot h etical rules/ite ms. 1.2. Func tional reduct ion strate gies a) ac tion al re ducti on b) modal reducti on c) reducti on of the propositi onal c on te nt: topic a voidance message a bandonme nt meaning r eplaceme nt The learner r educes t heir c o mmunicative goal in order to avoi d a problem. 2. Achieve ment strate gies a) com pensator y strate gies code s witch interli ngual tra ns fer inter-/in tralingual tr ansfer IL based strate g i es: generaliza tion paraphrase word coina ge restructuri ng cooperat ive strate gies non-li nguistic strate gies b) retri eval strategies The learner at tempts t o s olve a c ommunicative proble m b y expandin g t heir communicative re sources. (Færch a nd Kasper, 1 983b: 52-3) CHAPTE R 1: Literature re view - 37 - As c an be observed in table 1 .2, within these two groups of CSs Fæ rch a nd Kasper (1983b) ma ke further subdivisions. The res ult is q uite a complex t axonomy containing mor e distinc tions and m ore s ubtype s tha n a ny of the pr evious ones. However, a c aref ul a nalysis re veals th a t most of the C S t ypes proposed by Færch and Kasper (1983b) can be traced to Tarone ’s (1977) initial ones. A subset of the functional reductio n strateg ies co incides with T arone ’s avoidance strategie s a nd wi thin the group of achieveme nt strategies most of th os e included in the CpS category ha d al ready bee n identified by Tarone back in 1977. Whe r ea s the ov erall CS c ategories p roposed by Færc h a nd Kasper (1983b) are based on a cognitive analysis of st rategic beha vior, the s ubtypes they es tablish seem to reflect linguistic rat her tha n cogniti ve f eatures. In fac t, a s a lso pointed out whe n discuss ing these authors’ de finition of the CS conce pt, Poulisse (1993) has criticized Færch and K a spe r’s (19 83b) approac h to the description of strategic languag e use for being too general to account , in a psychologically plausible way, f or different C S types within ov e rall c ategories. To a certa in ex tent we ca n s ee Færch and Kasper’s (19 83b) taxonomy as a c ompromised positi on be tween t he interac tional and the ps ycholinguistic approach, g e neral ca tegories re f lect ing the ps yc holinguistic perspe ctive a nd specific subtypes the interactional one. Adopting a st rict psycholi ng uistic approach to the desc ription ad classification of CSs, Poulisse et al . developed in 1990 a new taxonom y f rom the ba sis that any CS classificati on sy ste m needs to be parsim onious –i.e. to use a minimal number of CS categories because when there are too many str ategy t ypes genera lizations c an be easily missed; gener alizable to differe nt individuals and different co mmunicative situa ti ons; and psychologically plausible –that is, it has to f it into a psychological model o f speech production (Kellerm an, 1991: 145). These scholars reject all previous taxonomi es without e xception, cri ticizing them for be ing produc t-oriented, i.e. base d on the surface stru ctures produc ed rather than on underlying psychological proce sses. In order to show the fallacy of these classifi ca tions they re view some of T arone’s (1977) c a tegories. For inst ance, in their s t udy a ‘hairdress er’ is referre d by one learne r a s ‘a p ers on who cuts pe ople’s hair’ a nd by othe r as ‘a haircutter’. The f irst example would be classif ied by Ta rone as a c i rcumloc ution and the second as a word coin age strategy . Acc ording to Poulisse e t al . (1990: 29), the CHAPTE R 1: Literature re vi e w - 38 - process unde rlying t hese two utterances is t h e same. They are similar in terms of semantic content and they both respond to an attempt m ade b y the learne r t o communicate the intended ref erent by mention ing some of it s most distinctive attribute s. Therefore , they nee d to be c onsidered as different lin guistic realizations of one only type of strategy (Bongaert s and Poulisse , 1989: 254; Poulisse et al . , 1990: 29-31). Acc ording to Poulisse et al . (1990), th is f ocus on li nguistic form i s c h a ra c t eristic not only of Tarone’s (1977) tax onomy , but a lso of the f ollowing ones. B uil ding on their concept ualization of the C S notion, on L evelt ’s (1989) model of speech production, a nd on the empirical a nalysis of a c onsiderable a mo unt o f data obtained f rom 4 5 Dutc h lea rn e rs of English, these sc holars develope d a cla ssification system that, they argue, strictly reflects differences in c ognitive proces sing . This sc heme is t he result of quite a comprehe nsive research proj ect ca rried out during the sec ond h alf of th e 9 0s at Nijmegen University, us ually known a s The N ij mege n project . For thi s rea son, this classificati on is usually refe rred to as the Nijmegen tax onomy . We n ee d t o bear i n m ind tha t, s ince, as we ha ve al ready se en, the se s cholars limited thei r def in ition of C Ss t o the subs et of CpSs, t heir taxonomica l p roposal is restricted to only this particu l ar type of str a tegies. This taxonom y consist s of two archistrategies that are based on L evelt’s (1 989) psychological mode l of speech producti on, more specifically on two o f the components of this m odel, i.e . the c onceptualizer and the f ormulator, whic h ar e supposed to draw on two differe nt sources of knowledge. The c onceptualizer draw s o n conceptual knowledge, whic h c onstitutes the e ncy c lope dia and contains informati on a bout th e propertie s of the concept a nd its relati on to other conce pts. The f ormulator draws on linguistic kno wledge, i .e. on the knowle dge of the m orpho-phonologica l feature s of the items o f the fi rst, the s ec ond or a ny oth er possi ble language; a nd the syntactic, morphological and phonological rules of all these languages. These two sources of in forma tion consti tute t he basis for the d istinction between the two archistrategies. Dependi ng on whethe r the s peake r predominantly expl oits the concept ual know l edge sourc e, i.e. manipulates the c oncept, o r the linguistic kn owledg e source, i.e. manipulates their l inguistic k nowledge, they wil l be m aking use of a concept ual or a linguistic str ategy . T able 1.3 illustrates the s pecific n ature of each o f CHAPTE R 1: Literature re view - 39 - these t wo ma in c ategories and how the y n eed to be subdivided into on ly t wo more strate gies ty pes to e xplain all the poss ible realizations of strategic behavior. TAB L E 1.3. POULIS SE, BONGAERTS AND KELLERMAN’S TAXONOMY O F C Ss STRATEG Y DESC RIPTION AND E XAMP L E S 1. Conce ptual strate gies 6 The spea ker r efers t o the inten ded concept b y listing (some of) its pr operties or b y s ubstituting t he word f o r a relate d concept which s hares som e of the crite rial p r operties a) Anal y tic strate gies The spea ker r efers t o the inten ded concept b y listing (some of) its prop erties a ssuming t hat th e listener will be a ble t o infer the intended c oncept f rom the proper ties which have been mentione d (e. g. ‘this you use fo r a bab y so uh, tha t it c an’t uh, make uh, his cl othes erm uh dirt y ’ for ‘bib’) b) Holistic strate g ie s The spea ker r efers t o a c oncept b y using the wor d f or a re la ted concept (e. g. ‘ v e geta bles’ f o r ‘ peas’ or ‘tabl e’ f or ‘desk’) 2. Linguistic strate g ie s The spe aker manip ulates their linguistic knowle dge. T his w ill be knowledge o f the sy n tac tic, morph ological a nd phonologi cal rul es that ap ply in the L1, s ome knowle dge of these r ules i n t he L2 ( and possi ble in a L3 7 ), and kn owled g e o f similari ties an d dissimila rities betwee n the L1 and the L2 ( and L3) a) Transfer strate gies The s pea ker e xploit s t he similaritie s between la nguag es. Words o r phra ses ma y be tran sferr ed from one la nguage t o t he other. S ometi mes, the utterances r esulting are ph onologica ll y and/or morphol ogicall y a dapted t o the L2 (e .g. ‘pa prika’, Dutch word for ‘ green pe pper’) b) Morph ological creati vi t y 8 The spea k er uses L2 rules o f mor phological derivation to cr eate (what the y as sume t o be) compre hensible L2 lexis (e .g. ‘re presentator’ for ‘representa ti ve’ or ‘ir onize’ for ‘t o ir on’) (Poulisse et al ., 19 90: 60-62) Despi te its limited scope, i.e. re stri cte d to C pSs, the p roposal made by Po ulisse et al . (19 90) has be en widely acc epted as a valuable theoretical cont ribution to the field. It ha s as well been freque ntly used for the ana lysis of em pirical da ta by th ose researchers who have adopte d a psy c holinguistic approach to t he study of s trategic 6 In a pre vious a pproac h to the tax onomic description of CS use Kellerman et al . (19 87) had distin guished three different t y pes of c onceptual stra tegies: holistic st rategies , when the c oncept is s een as a wh o le rather than in t erms o f it s pa rts; partitive st rategies , when an obje c t o r figure is seen a s consis ting of t wo o r more parts ; and l inear s trategies , whe n the fi gure is divided i nto its ultimate one-di mensional components, su ch as lines, a ngles or spatial relations, and described in terms of these as if giv ing instructions on how to dra w it. The partitive an d l inear strat e g ie s correspond now t o t he anal y ti c strate gy cate g ory. 7 The third langua g e ( L3) involve s an y la nguage known b y the learner ot her t han the L1 or the L2. 8 Kellerma n an d Bia l y s tok (1997) su ggest tha t a more a ppropriate ter m f or morph ological crea tivit y should be gramma tical derivat ion . CHAPTE R 1: Literature re vi e w - 40 - communication. However, we a lso need to point out t hat this taxonom y does not even cover all the ra nge o f CpSs identif i ed by previous re searchers. Nonverbal dev i ces are being here ignored, al though, in 1991, K ell erman re name d linguistic st rat egie s as code strategies in order t o includ e under their scope non verbal m echanisms. Appe als for assis t ance could not, howev e r, be incorporate d into the system. They are not considered compe nsatory be cause they are not learners’ a uto nomous solut ions to their own l exical problems. The strongest critic ism to t his sy s tem has c om e f rom Poulisse herself who, in 1993, argues tha t i t fa ils to meet the psychological plau sibility c riterion when Lelvelt’s (1989) cognitive mode l is a dapted to account for L1 and L2 use. Po ulisse (1993) reviews L evelt’s pa radigm in o rder to prop ose a bilingual speech producti on model involving a bil ingual le xicon. Resea rch on the notion of the bi lingual lexicon suggests that c onceptual a nd transfer st rategies ar e largely sim ilar. In both ca ses the learne r substitu t es the i ntended a nd unavailable le xical item with a n alter native word retrieve d from their lex icon, which contains words fro m the L 1, the L 2 and the L3. 9 This scholar also c onsiders t hat the two s ubtypes o f c onceptua l strategies, i.e. holistic a nd analytic , differ s ubstantially from a c ognitive point of view. W hereas holistic devices rely on a substitution proce ss, a nalyti c ones req uire th e reconce ptualization of the e ntire preverbal m essage. These two s trategies s hould therefore be treate d a s two differ ent ca tegories. On t his basi s, Pouliss e (1993: 163) argues that “al though different knowledge source s do affect the choice between a concept ual and a linguistic st ra tegy, the y can not be considered criterial in the d istincti on of st r ate gy type s” and reviews t he initial Nijmegen taxonomy to prop ose a new classificati on system for CpSs. In Poulisse ’s (1993) model when a speaker cannot ac cess an intended lexical item from their mental lexic on the y ha ve three options: (1) to stop the speech produc tion process and giv e up thei r mes sage, i.e . to make use of a mess age abandonment s trategy; (2) to implici tly or e xplic itly appe al for help f rom the ir i nterlocutor; or (3) to use a CpS to com pensate for their le xical problem. In this res pect Poulisse (1993) agrees with Poulisse et al. (1990) – c f. present chapter, p. 30. Howeve r, a s the following t able 9 The bili ngual l exic on st ores lexical it ems and t heir features o r specificati ons. For instance , f o r t he word ‘boy’, the specif ication s are [+ human], [+ male] , [ + child], [+noun] and [ + English]. T he English language s pecificati on is there fore treate d as just one more o f the word features ( P o u lisse, 199 3: 176). CHAPTE R 1: Literature re view - 41 - illustrates , within CpSs Poulisse ( 1993) distingui shes three instead o f two diffe rent types of st rategies. TAB L E 1.4. POULIS SE’S TAXONOMY OF CSs STRATEG Y DESCRIP TION AND EX AMPLES 1. Substitut ion strate gies The speaker r eplaces the i ntended lexica l i tem with a nother one. This ca n be either a r elate d ite m (e .g. ‘ animal’ f or ‘rabbit’) or the corres ponding L1 item (e.g. Dutch ‘voorw oord’ for ‘pref ace’) 2. Substitut ion plus strat egies The s peaker applies a L1 or L2 morpholo gical an d/or phonological e ncodin g procedure to a L1 or L2 lexic al ite m (e.g. t he creati on o f the ver b ‘ironize’ on the basis o f th e noun ‘ir on’) 3. Reconce ptualization strate gies The spea k er ma y ch oose to enc ode the conceptual fea tures of the inte nded lexical i tem one b y one, for insta nce b y listing them (e. g . ‘it ’s gree n, you ea t it wit h potatoes a nd Pope y e eats it’ for ‘ spinach’) ; or ma y d eci de to select two lexical ite ms fr om the le xico n which ca n b e c ombi ned int o one new word (e.g. ‘c ookin g apparatus’ f or ‘cooker’ ). A speaker ma y also decide to a dd further bac kground informati on to t heir m essa ge, so that t he li stener w il l be a ble to e xploit the c ontext i n int erpreting thi s messa g e (e .g. sa y i ng that three quarters of a swi mmi ng p ool i s f or people who wa nt to dive and pl ay, and one quarter is for peopl e who want t o swim in this wa y , in order to c ompensa te for ‘lengths’ as in ‘t o swim l engt hs ’). Finall y , t he speake r ma y opt for gesture s rather t han speech or a c ombin at ion of speech an d gestures (Poulisse, 1993: 180-1 83) Kelle rman and Bi alsytok (1997) c onsider this new a pproach a nd the e xpansion of the original two basic types of CpSs int o three diffe r ent cate gor ies unnece ssary. The y therefore prefer to kee p the original binary divi sion. However, Dörnyei and Kormos (1998) take advanta ge of t his new mode l to propose a new CS taxonomy that, as we have already poi nted out, is not only the most updated classification syste m av a ilable but al so th e m ost comprehensive one. On the ba sis of an e mpirical anal ysis o f data collect ed from 44 Hungarian lea rn e rs of English, Dörnyei and Korm os deve lope d in 1998 a fra mework of anal ysis that a im ed t o prov i de a comprehe nsive and systematic descripti on of the different ways in which learn ers ma nage problem s and dif ficulties in F L com munica ti on. These scholars we re t hus t rying to bri ng to gether phenom ena that had been t raditionall y CHAPTE R 1: Literature re vi e w - 42 - treated in a n in d ependent way in diffe rent areas of SLA res ea rch. The s ystem developed distinguishes fo ur ma in sources of proble ms –resourc e deficits, proc essi ng time pressure, perceived def iciencies in one’s own language out put a nd perce ived deficienci es in the interlocutor’s pe r f ormanc e– and four types of devices associate d with them – CSs, s talli ng mec hanisms, se lf-repairs and self -rephrasing, and negotiat ion of mea ni ng strat egies. Here we focus on their a nalysis of CSs a nd, wit hin this g roup, on those str ategies us ed to solve le xical problems , althou gh their mo del a lso account s for CSs used to manage grammatical and phonological deficits. In 1997 D örny ei a nd Scott ha d carrie d out a re view of previous theoretical a nd empirical a pproaches to the c las sification of C S s tha t r evealed that most of the taxonomies prop os e d in t his fie ld of research, e ither fr om an interac ti onal or a psycholinguistic p erspective, differ mai nly in term s of terminology a nd organizin g principles, but display the same s e t o f categories . Dra wing on this previous work, Dörnyei and Kormos e laborated in 1998 a t axon o my of CSs –or le xical problem-solving mec hanisms a s they c alled them, t hat int ends to com bine previous a pproa ches. This taxonomy involves all t he C Ss t ypes generally iden tified a nd acc ounted for in previous classificati on systems and ma kes it possible to est ablish a r elationship be tween interacti onal and psycholinguistic categories. 10 TABLE 1.5. DÖRNYEI AND KORMOS’ TAXO NOMY OF CSs 11 STRATEG Y DESCRIP TION AND EX AMPLES 1. Content r eduction strategies a) Mes sage abandonm ent Leaving a me ssage unfinis hed becau se of some languag e difficult y ( e.g. using ‘ that is a f lat … in a hous e…’ to avoi d ‘housing state’) b) Message reducti on Reducin g the message b y avoidi ng certai n languag e structures or topics c onsidere d pr oblematic langua gewise or by lea ving out some i ntended e leme nts f or a lac k of linguistic res ources (e. g. ‘ he is responsi ble… f or t he…for the cleanne ss of the h ouse an d er … he locks t he d oor … at night an d ope ns it … in the m o r ning’ for ‘he was resp onsibl e for the ru nning of t he house’) 10 In 1997 Yule a nd Tar one had alread y il lustra ted the r elations hips exist ing betwe en the psy c holi nguistic classification o f CS s pr oposed by P ouli sse et a l. (19 90) and pre vious CS taxonomie s de veloped by interactioni st resarc hers. 11 As ab ove explai n e d, Dö r n yei an d Kormos’ (199 8) typology a ls o accou nts for CSs used to solve g r a mmatical a s well a s p hon ological and art iculator y proble ms. H owever, since most of the previous taxonomies have been de vised f o r le xical CSs a nd the purp ose of our r esearch is t he st udy thi s s pecific kind of CS s, we prese nt here onl y t hose cate gories r elated t o lexical problem-s olving me chanisms. CHAPTE R 1: Literature re view - 43 - c) Mes sage replace ment Su bs t itut ing t he original messa ge wit h a ne w one b ecause of not fee ling capable of exe cutin g it (e. g. ‘ y ou can … sta y her e until … midnight or … how do you want’ for ‘until da wn’ or ‘until morni ng’) 2. Substitut ion a) C ode switchi ng Includin g L1 or L3 w o r ds wit h L1 or L3 pronunc iation in L2 speech; t his may inv olv e stretches of discour se rangi ng f rom single words t o whole ch unks an d e ven c omplete turn s (e .g. ‘fer rum’ for ‘ir o n’) b) Appr oximation Using a si ngle alter native le xical item, such a s a superordina te or a related ter m, t hat share s se mantic feature s with t he targe t word or struct ure (e.g. ‘and er … takes th e apple in its mouth’ for ‘bea k ’ ) c) Use of al l-purp ose words Extending a general , “empt y ” l exical item to c onte xts where specific words ar e lac king (e.g. the o veruse of ‘thing’ ‘ stuff’, ‘make’ or ‘do’) d) C omplete omissi on Leaving a gap w hen not knowi ng a word an d carr y in g on a s if it had bee n said (e. g . ‘the n … e r … the sun is is … h m sun is … and t he Mic key M ouse’ for ‘s hine’) 3. Substitut ion plus a) Foreigni zin g Using a L1 or L3 word b y a djustin g it to L2 p honolog y , i.e . with a L2 pr onunciati on or morp holog y ( e. g. ‘m y guest f rom the … ministeriu m’ [ w ith a n Eng li sh pronu nciation]) b) Gra mmatical w ord coinage Creating a non-e xisting L2 word b y a pply i n g a s upposed L2 rule to an e xisting L2 w ord (e. g . ‘de junktion’ a nd ‘unjunkti on’, created fr om ‘junk’ a n d the prefi xes ‘ de-’ a nd ‘un-’, f or ‘clear the junk’ and ‘street cle aring’) c) Litera l translation Translati ng literall y a lexical item, a n i diom, a compou nd word, o r a struc ture from L1 or L3 t o L2 (e . g . ‘s nowma n’ as a literal t ranslation of a Hunga rian wo rd) 4. Macr o -re conceptualiza tion a) Restruct urin g Abandoning the e xecution of a v er bal plan be cause of language difficulties, l e avin g the utterance unfinished a nd communicati ng the inte nded message acc ording to an alternative plan (e. g. ‘she has t o care about t he house, t o car e about the g ar bage, an d t o car e about t he … or t o c lean th e house’ f or ‘cleanness’ ) 5. Micro-r econceptual ization a) Circ umlocution Exemplif y i ng, ill ustrating, or d escri bing the p r operties o f th e tar g et object or ac tion ( e.g. ‘it become s water’ for ‘it melts’) b) Semantic word c oinage C reatin g a non-e xisting L2 word b y c ompoun din g word s (e.g. ‘sn owsculpture’ for ‘s nowman’) 6. Appeals f or help a) Direct appeal f or help Turnin g t o the inter locutor for a ssistance by aski ng a n explicit q uesti on c oncer n in g a g ap i n one’s L2 knowl edg e (e.g. ‘i t’s a kind o f o ld clock s o w hen it strucks e r … I don’t know, one, t wo , or t hree o’cloc k t hen a bi rd is c oming out. What’s the na me? ’) b) Indirec t appeal for help Tr y ing to elicit hel p fr om the interl ocutor i ndirectl y b y expressing lac k of a nee ded L2 item eit her verba ll y or nonverball y (e.g. ‘I don’t know the name …’ [ rising intonati on, pause, e ye contact] ) (Dörn y e i a nd Korm os, 1998: 359-361) CHAPTE R 1: Literature re vi e w - 50 - French bel onging to thr ee different proficienc y le vels reveals that, a s expec ted, low level learne rs resort to the borrowing of lexi cal items from th eir L 1 more often than high level learners; whe r ea s these l atter use a significa nt ly hi gher proportion of L 2-ba sed strate gies, i.e. strategies based on the speaker’s man ipulation of their resource s in the FL. Although sma ll sca le and fa irly explorat ory in n ature these studies establishe d the main l ines of research for following more comprehe nsive anal yses. Pa ribakht’s (1982, 1985), Cor rales a nd Cal l’s (1989), Ma nchón’s (1989), Si-Qi ng’s (1 990) or L iskin-Ga sparro’s (1996) studie s, conducted in a varie ty of diff e ren t settings and a lso using diffe rent methodological designs, obtained similar r esults and prov i ded further support to the hypothesi s tha t, both in terms of f requenc y a nd choice, CS us e correlates with profi ciency le vel. However, c ontradict ory evidence to this assum ption ha s a lso been f ound. The researchers inv olved in the already me ntioned Nij megen project cond ucted rigorous quantitati ve and sta tistic a l a nalysis on m ore than 4,000 insta nces of CSs. This da t a was obtained from a total of 45 Dutc h learners of English, organized in three differen t groups accordin g to their profi ciency level, who were asked t o per f orm four diffe rent types of oral tas ks. The analyses carrie d out reveal a significa nt inverse rela tionship between the absol ut e number of CSs used a nd the proficienc y le vel of the speaker. However, i n te rms of choic e of p articular CS types, th e r es ults obtain ed sugg es t tha t the profic iency level f actor has qui te a limited influe nce, ov e rruled by tha t of ot her features, such as the nature of the c ommunicati ve task used in the elicitati on of t he data –cf. Poulisse a nd Schils (1989) and Poulis se et al . (1990). Whe n considering these fi ndings we need to bear in min d that, first, the se researchers paid attention only to a subset of C Ss, i .e. CpSs; sec ondly, they classified these s trategies a ccording t o a taxonomy emphasizing pa rsimony a nd psychologicall y plausi bi lity whic h onl y considers f our different types of strate gies –cf. table 1.3 in p. 39; and, thirdly, although the amount of quantitative data is impressive, no qualitative ana lyses were c omplete d by these researc hers on the que stion of the profi cie ncy l evel infl uence. In recent years research has widened its scope to focus on t he infl uence that profic iency has on CSs not onl y in t e rms of number a nd types of strateg i es use d, but CHAPTE R 1: Literature re view - 51 - also in thei r linguistic reali zation. Evidenc e ha s be en found th a t one’s ability to make use of certain CSs, in pa rticular paraphra se stra te gies, inc reases wi th profic iency level. That i s, higher p roficient stude nts seem t o be more na ti ve-like and, conse quently, more effective and successful i n their strategic b ehavior (Jourdain, 2000). 13 R esearch ha s als o bee n c onducted trying to relat e specif i c a spect s of la nguage prof iciency , suc h a s, for insta nce, vocabulary size, with the choice betwee n m ore and less de manding C Ss (Ba ba, 2001, 2004); obtaini ng a lso positive resul ts that show the import ance o f this learne r- related fact or on strategic communication . Nowada ys, t he influence of the le arner’s proficiency level on th eir use of CSs is a wide ly a ccepted f act. H owever, s chola rs a lso general ly re cognize that thi s is a more complex is sue than initiall y thought a nd that the e ffect of p roficiency le vel on C S use alwa ys needs to be considered in relation to other persona l and context ual variable s. Together with the possibility of a prof i cie ncy level inf luence on t he le a rner’s use of CSs, Tarone (1977) a nd Corde r (1978) hypothe sized, a lso f rom a theoretical perspecti ve, th at differe nt learners may resort to different CS ty pes because o f diffe r ences i n their learning st yles and pers on a lity profiles. This issue was f irst invest igate d in an empirical way by Ha as trup and Phillipson (1983). This preli minary stu dy i s based on t he analysis of CS us e b y le arners of English as a F L chose n f rom t hree di fferent Da nish sc hools. It is a ssumed that learners attending diffe r ent types of sc hools , with di fferent mo tiv ations a nd pursuing dif ferent ac ademic goals al so pos sess different lea rning profiles . Learne rs’ styles a re esta bli she d on the basis of conversa tional feat ures, such as fl uency or pa tterns of c ommunicati on disruptions, and the results obtained reveal tha t eac h lea rner has t heir own conversati onal a nd strategy st yle. This is take n as evide nce of differences in learni ng style, but no te st to confirm this a ssumption is administered. More re cently, the rela tionship betwe en personalit y and strat egy use has be en investigated in a m ore sys t ematic way. Thus, Luján-Ortega and C lark-Carter (2000) use a psychometric IQ test (Heim, 1970) and the Myers-Briggs Type Indic ator test o f persona lity /learni ng style (Myers and McC aulley, 1985) to ass e ss the e ffect s of individual differences on CS use. Li ttl emore (2001) a lso uses a standardized tes t, Riding’s (1991), f or the same pu rpose. The re sults of both st udi e s re veal a significant 13 This had a lso bee n rec ognized in Pa ribakht’s (1 982, 1 985) r esearch o n the r elation ship b etween CS u se and proficienc y le vel. CHAPTE R 1: Literature re vi e w - 52 - stat istical rel atio nship bet ween cognitive or learning st yle and prefe rence f or different CS t ypes, ide ntified on the basi s of Pouli sse et al .’s (1990) ta xonomy. Furtherm ore, L uj án-Ortega a nd Clark-Ca rter’s (2000) a nalyses suggest t hat indi vidual characte ristics infl uence CS use, a nd also tha t these t wo factors take n together predict achieveme nt in forma l languag e e xaminations. From a psy choli nguistic pe rspective the study of cognitiv e or learning style diffe r ences i n relation to CS use has become an important re search question. If it ca n be demons trated t hat learners with d ifferent cognitive character istics prefer diff er ent types of CSs, t his should giv e light i nto the psychologica l processes underlying the use of those CSs –cf. L i ttlem ore (2001). Howev er, mor e work is still needed, assessi ng how these individual variable s inte ract with othe r persona l or c onte xtual facto rs , in order t o reach a definit e conclusion on this i ssue. A third major top ic in CS re search has been the exa mination of t he possibl e infl uence of L1 a nd c ul tural bac kground on the learner’s use of CSs. Initial studie s, such as Tarone (1977) or Palmbe rg (1 979), found n o rel ationship at al l between t he subjects’ NL and th eir patterns of FL C S use . Tarone’s (1977) res earch o n CSs involves learners of English as a FL wit h three different NLs –Spanis h, T urkish a nd Manda rin. It therefore makes poss i ble t o draw som e conclusions on the issue, but these need to b e take n wit h c are beca use the study was not s peci f ica l ly des ig ned for the e xamination of diffe r ences in strategic behavior. Palmberg’s (1979) work examines CS use by Finns, Finland-Swedes and Swedes. This st udy is s pecifically devoted to t he analy s is of the L1 infl uence on L 2 strat egic c ommunica t ion, but the fin dings gaine d ar e similar t o t hose in Tarone (1977). On this ba sis Palm berg conclude s tha t “strategy prefe rence is primaril y a questi on of proficienc y le vel and personality, ir respective of lea rn e r L1” (Palmberg, 1979: 73). However, con tradictory e vidence ha s a lso been f ound in subseque nt research. Tarone a nd Yul e (1987) a nalyzed CS use be twee n learne rs with diffe rent NL s, i.e. lea rn e rs w ho only sha re the ir knowledge of the TL , f inding tha t th is affects C S choic e. The results of this study re veal t hat languag e switch a nd appeal for as sistance strategies are not used in NNS-NNS interactions as opposed to NNS-NS one s. Si-Qing (1990 ) f ound that the l ea rn er’s l inguistic background ha s a cer tain infl uence on t heir use of some types of CSs, more specifically on t he use of L1-based CHAPTE R 1: Literature re view - 53 - CSs a nd on t he use of CSs t o make ref erence to culture-specific noti ons . Chinese lea rn e rs show a strong tende ncy t owards a voi dance strategies tha t this s cholar e xplains as a result of their soc iocultural background. They al so te nd to a voi d L 1 -based strategies and this i s accounted for a s an e ffe ct of the learners’ pe rceived differenc es betwee n th e two languages, i.e. Chinese and English. Si mil ar results w e re also obtained by Wongsawa ng (2001). T his s cholar ana lyzed the use of C Ss by Thai learne rs of Eng lish as a FL referring to culture-specif ic notions. A lthoug h t his part icular ki nd of r e ferents may se em, ini tially, m ore s usceptible of e liciting L1-based str ategies, the subjects involved in the study tended to pre f er the us e of approxi mation or circumloc ution strate gies. Wongsa wang e xplained the lac k of forei gnizing and lit eral t ranslati on strate gies in the data as a result of the learners’ pe rceived dist anc e between Thai a nd Eng lish. In g e nera l the res ults o f t hese different studies sugg es t that, more th a n the NL itse l f, it is the relati onship pe rceived by the learne r bet ween their N L and t he T L what has a direct influence on the ir use a nd choice of CSs. Tha t is, the closer the learner perceives th e L1 a nd t he L2, t he more li kely they will resort t o th e use of tran sf er strate gies –cf. a lso Paribakht (1985), Tarone and Yule (1987), or Chen ( 1990). In the ana lysis of face-to-fa ce interacti on we also need to take into accoun t, as pointe d out by Tarone and Yule ’s (1987 ) stud y, tha t the NL of the inte rlocutor may a s wel l have a strong influenc e on the learner’s use of certain CSs. Other resea r che rs have trie d to lo ok at other factors that coul d directly or indirectly a ccount for some of the i ndiv i dual difference s f ound in CS patterns of use. Thus, L ab arca a nd Kha nji (1986) designed a c omparative s tudy to examine th e use of CSs by s tudents who had been instructe d with two differe nt teaching m ethods –th e Strategic Interactive Method (Di Pietro, 1982) a nd t he Tota l Physic al R esponse Met hod (Asher, 1969; 1977). Th e resul ts obtained suggest tha t t he method of instruction has an importa nt bearing o n the a mount and kind of CSs use d and, as a result, al so on profic iency attai nment. Hersc hensohn (2003 ) c ompared the CSs used by two English speaking le arners of French: one studying French a s a F L in an academic sett ing and the other one lea rning the language in a na tural context, i .e. i n Fran ce. However, the results of th e analyses c onducted reveal no signif i c ant diff erences between the two l earners’ use of CSs, suggesting th at the learning setting has not a relevant i nfluence on this iss ue. CHAPTE R 1: Literature re vi e w - 54 - Hersche nsohn a ccounted for these r e sults following Bia lystok’s (1990) and K ellerm an’s (1991) c laim that CSs a re u sed in the L1 a nd do not nee d to be learnt again i n the L 2 . The scholar conc luded, on this basis, that there is no nee d to teach CSs in aca demic FL lea rning settings . Nowada ys, it i s widely recognized t hat indi vidual d ifferences have a s igni ficant infl uence on the way l earners m anage proble ms in FL c ommunicati on. Howe ver, the diffe r ent and sometime s e ven contra dictory re sults of the research he re reviewe d seem to sugge st that these different fac tors inte rrelate in c omplex wa ys wi th each other a nd probably a lso w ith tas k- and con text-related f eatures. More w ork nee ds the refore to be carried out i n thi s a r ea befo r e a cl ear and definite understa nding of the use t hat the FL lea rn e r makes of CSs can be a ttained. 1.1.3.2. The effect of task- and contex t-related factor s on CS u se The possi bil ity t hat the nature of the task being accomp lished coul d affect C S use has bee n c losely ex amined from the very beginnings of the research. B ialystok and Fröhlich (1980), and Bialystok (1983) paid c ertain attention t o this issue, com paring writte n versus ora l modes of discourse, but they di d not f ind a s ig nif icant relationship be tween this t ask-related featu re a nd CS us e. However, in 1979, a dif ferent s tudy conduc ted by Galván and Ca mpbell yi elde d some evidence s uggesting tha t certain CS types are depe ndent on the nature of the tas k in which they are used. The most c ompr ehe nsi ve analysis of this is s ue is , without doubt, the one conduc ted by t he re sear c her s involved in the Nijmegen project. The se scholars compa r ed CS us e ac ross three differe nt types of t ask: a c oncept i dentification task, where subjec ts had to refer to t w e nty photograp hed object s; a story-retell ta sk, wher e they had to re t ell in Englis h a story previously he ard in their L1; and a co nversation, where t hey ha d to interact for f ifte en m inutes with an English NS. 14 W e h ave al ready pointed ou t th a t the results obtained i n this s tudy show t hat task-rela ted fa ct ors play a major role in the lea rner’s use of CSs that e ven o ve rrules the infl uen ce of other factors, such as prof iciency leve l. This study suggests that the choice of s pecific types of CSs –identif ied and classified on the basis of the already discus sed Ni jme gen taxonomy, se e 14 A f ourth tas k c onsisting of a descr iption of ab stract s hapes is a lso used in thi s pr oj e ct to compa re L1 and L2 CS s. Lear ners’ perf o r mance in this t as k is not, h owe ver, ta ken into acc ount in the analysis o f ta sk- related fa ctors and their influenc e on FL C S use. CHAPTE R 1: Literature re view - 55 - table 1.3 in p. 39; depends on the demands imposed by the ta sk a nd the situ ation. This is explaine d in relation to com municative principles, i.e. Gri ce’s (1975) coop er ative principle a nd the prin ciple of eco nomy. In the f irst ta sk, s ince it i s im portant t o get the message a cross, spea kers are eager to e xpend a cons iderable deal of e ffort in their strate gies and therefore t he y ma ke fre quent us e of time-consuming an alytic strategies. In the story-rete ll task and the conve rsation the general idea is mor e important than the details. Furthe rmore, thes e two ta sks are c ognitiv e ly more de ma nding a nd impose certain time c onstrai nts on the spea ker. Subse q ue ntly, in these two t asks l ess effort dema nding strat egies, such a s holis tic or tra ns f e r ones, are m ore f requently use d. Finally, t he conv ersation task g ives the speake r an opportunity to check for the comprehe nsibility of the ir strategies , which is genera lly exploi ted by the lea rner . Speake rs tend to resort first t o less demanding strate gies, such as the use of tr ansfer, and, only if th es e fail , they ta ke t he effort to elabora te more c omplex s t rategies. Tha t i s, they try to achi eve maximum comprehe nsion wit h the minimum possible effort . 15 Luján (1997) rep licated the previous s tudy with E nglish spea king lea rners of Spanish a s a FL , o btai ning s imilar results and also conc luding tha t the various fac tor s that d i fferentiat e communicative sit uations predict diffe rences in CS use. Other projects c onducted on this iss ue, b ut f r om di fferent appr oaches, a r e Yarmohammadi a nd Seif (1 992), o r Khanji (1993). T he f ormer compa red, a s ha d be en previously done b y B ialystok and Fröhli ch (1980) , and Bi a lystok (1983), CS use in oral and written tasks. These scholars did not find sig n ific ant differences between the two modes of d iscourse i n the le a rners’ choice of reducti on versus a chievement strategies, but the y ide ntified a relat ionship be tween tas k ty pe a nd the use of spe cific CS t ypes within e a ch of thes e tw o main cate gories. Khanj i ( 1993), building o n L a barca and Khanji (1986), compared CS c hoice in two different types of conversation tasks –D i Pietro’s (198 7) Strate gic Interaction tasks, based on the notions of scenario and rol e play , and lea rner’s interview tasks. The f irst f orm of conversatio n le d t o t he use of more ach iev e me nt strategies, sugges ting that interview t a sks , be cause o f their answer- response pattern, do not prov ide go od, natural- lik e i nteraction practice. The use of C Ss in i nterview tasks but in a Spanish se tting ha s been a pproached by Pons Sanz (2001). This scho lar analyzed the s peech of Spa nish learne rs of English as 15 Fo r a detailed account a nd discussion of these r esults, and t he st udy of t he influence of comm unicative principles on CS u se, see also Poulis se and Schil s (1989) or P oulisse (1997). CHAPTE R 1: Literature re vi e w - 56 - a F L in terviewing English NSs. The resul ts of h er a nal ys es reveal a s ignificant e f fect of conversati onal role i n the performance and strategic be havior of t he NN S, suggesti ng that the role ad opt ed by the learner may have a strong er i nflue nc e on their use of ce rtain CSs than t heir proficiency leve l. This l ast s tudy a lso il lustrates a relatively rece nt tende ncy a mong rese archers towards the us e of sociolin guistic perspectives and fra m ew orks of anal y sis in the exa mi nation of FL strate gic communic ation –cf. Kasper a nd K elle rman (1 997) or Safont (2001). CSs have s tarte d to be explored in a uthentic n atura lly occ urring communica tiv e situa t ions and more attenti on is being paid to t he so ciocultural factors inv ol ved i n communication, such as conve rsational role or l in guistic and social ide ntity –cf . W agner and Firth (1997 ), Will iams et al . (1997) or Rampton (1997). This new line of research shows tha t an in teractional perspective, ta king into a ccount the lear ner ’s and th e interlocutor’s a ctions, thei r relations hip, an d the social sit uation c reated by the interacti on it s elf, can offer new a nd interest ing poss ibilitie s f or the de velopment of CS research. 1.1.3.3. CS use in interac tional conte xts In al l the studies rev i ewed so far, c onducted eit her from a psyc holing uistic or a n interacti onal perspecti ve, t he a nalysis of CS use has been e ssentially ce ntered on the FL lea rn e r. D epending on t he pe rspec tive adop ted, rese archers have conce ntrated eit h er on the linguistic or on the co gnitive feat ures of CSs, but they have all de scribed and explaine d CSs as products of t he learner’s individual use of t he l anguage. Howeve r, i n the las t few yea rs, new studie s have a ppeared adopti ng a more stric t intera ctiona l approach and trying to account f or CS use as a truly interacti onal phenomenon; i. e. involving the ac tive participation of the learner and their interlocutor in a joint, socially situa t ed mea ning crea tion proc ess. They have thus gone ba ck t o Ta rone’s (1 980, 1 981) concept ualization of strategic com munication and rea ssessed the benef its of an interacti onal perspective to the st udy of C Ss The i nteractional defini ti on of CS use as “a mut ual attempt of two i nterlocutors to agree on a meaning in situa t ions where requisite meaning structure s do not seem to be shared” was first pr e sented by Ta rone i n 1980 ( T arone, 1980 : 419, 1981: 288 ). In her 1977 study of CSs, how e ver, Tarone ha d already poi nted out t he inte racti ona l functi on CHAPTE R 1: Literature re view - 57 - of these st rategies. This pa per includes the analysis of a n excerpt of fac e- to-f ace ora l conversati on that clearly il lustrates the use of CSs as interactional to ols use d by the speaker t o e ngage their interlocutor in a joint negotiati on of meaning proces s. Unable t o recall the T L ite m ‘waterpipe’, a l earner uses a paraph rase strategy tha t is not dire ctly understood by the interloc utor. This lea d s both of them to a relatively long excha nge in which se veral clarif ication reque sts, confirmat ion chec ks a nd reformulations of th e orig inal CS ne ed to be used before the two interlocutors ma nage to reach the desired mutual agreement on the m eaning ‘wat erpipe’. This ext ract c le a rly shows t hat strategic communication of meaning in face-to-face inter a ction i nv olves the j oint act ion and mutual eff ort of both the speaker and the listener. As j ust menti oned, subsequent i nvestigations tended, nev e rthe less, to focus exc lusiv e ly on t he F L learner’s act ions in strategic communicat ion, i gnoring the possible rol e of t he i nterlocutor in thi s process. The inte r actional a pproach to t he study of CS s defi ned its a gen da a s a n a nalysis of t he linguistic f eatures of CSs, versus the ana lysis of cognitive proce ssing pursued b y psy cholinguistic scholars. Frui tful as this line of re s ea r ch m ay have been in order t o cla rify the l inguistic na ture a nd de fining features of CSs, it doe s not however succee d at mee ting the re quir ements of a true interacti onal perspective, which can e xplain the function and role of CSs in interactio n. In 1991 Yule a nd Tarone ma de a n ex plicit cal l for resea rch ‘l ooking at the two sides of the page’; t hat is, inte g ratin g within one single framew ork th e a nalysis of speakers ’ and interloc ut ors’ ac tions in fa ce-to-face strategic interacti on. The study of communicative proble ms in speech production had so far been approa ched by C S investigations, whe reas comprehe nsibility p roblems in t he reception of messages had bee n the foc us of a dif fer ent a rea of SL A re search, i .e. negotiation of meaning studie s. These two areas had traditionally worke d inde pendent ly, rec ognizing e ach oth er’s work, but not at tempting to build on it. However, as t hese two schola rs clea rly illustrate, in interaction, CSs a nd negotiation of meaning s trategies tend to be us ed in c ombination. Meaning ne g otiati on moves, such as clarification requ ests o r conf irmation checks, used by an interl ocutor when una ble to c omprehend received input , tend to elic it C Ss from the speaker. The s e moves initiate the negotiation of meaning proces s, but for a mut ual ag r eeme nt on mea ni ng t o be reac hed the initial input needs t o be mo dified thro ugh CS use. That is, the CHAPTE R 1: Literature re vi e w - 58 - speaker n ee ds to ref ormulate th e ir initial utt e rance in o rder t o make it comprehe nsible for the ir interlocuto r, using f or this purpose an a lternative m eans of e xpression, i. e. a CS. Yule a nd Tarone (1991) a rgue on t his basis that to und e rstand how communicative problem s are resolved and meaning i s ac hieved, an a lysts cannot f ocus exc lusiv e ly o n either negotiatio n of meaning moveme nts or CSs. They n eed to consider them t ogether. Furthermore , they defe nd th at a CS fr amework of anal ysis allows f or a n effective description of the key mo ves made by both th e learner a nd the inte rlo c utor , that i s, for a n integrate d study of C Ss and negotiation of meaning s trategies taking into account the two sides of the interactional exchange. Following th is a pproac h, Suni (1996) conside red how underst anding pro blems are solved in discou rse be tween native and non-native speaking interlocuto rs. Th e results of this s tudy, in whi ch CSs and n e gotiation of meaning moves are analyzed as related int eractional phenomena , c onfirm the pos tulates o f Yule an d Tarone (1991) . A similar attempt to link CS a nd negotia tion o f m ea ni ng r esearc h was also m ade by William s et al . (1997 ). T hese two studies are de scriptive in na ture and pay mor e attention to comprehensibil ity tha n to production problem s. Thus, t he strategi c categories guiding Wi lliams et al .’s (1997) a nal ys e s corres pond to w hat hav e been traditi onally considere d negotiati on of meaning strate gies, i .e. c onfirmation c hecks, cla ri f ication req uests, comprehension checks , r epetiti ons and reformulations. In 1997, W agner an d Firth made a s trong claim in f avor of the, at that point, still infre quently pursue d intera cti onal approach to the s tudy of CSs. Their a nalysis of natural, spontaneous interactions shows that t hi s perspective can provide light on aspects o f strategic communi cation overlooke d or simply ig no re d in previous research. Wa gner and Firth work on the as sumption t hat communi cation is an inte r ac t ive endeavor, “accomplished a s a sit uated, conting e nt, locally m ana ged a chieveme nt ” (Wagner a nd Firth, 1997: 325); and that e ncoding dif ficultie s enco untered in this process are overcome through t he j oint ac tion of all the i nterlocutors. CSs a re thus seen as e lements of an “ ongoing a nd c ontin gent mea ning c reation proc ess” (W agner a nd Firth, 1997: 325), used to engage the two part i es of t he conversa tion i n a mutua l attempt to resolve a communicative problem . They c an, therefore, be full y e xplained only i f attention is paid to the two s ides of the c onversational exchange. This also leads them to CHAPTE R 1: Literature re view - 59 - argue that f or a no n-target-like element to be co ns idere d a CS it need s to be flagged as such i n speech. Speakers use verbal and nonverba l signals to indica te to their interlocutors how the ir ac tions or uttera nce s need to be interp reted and reacte d upon. Whe n the y make use of a C S, t hey si gnal the strategic nature of their ut terance an d, in this way, they engage their in terlocutors in a m ut ua l attempt to resolve t he encountere d problem. They study from th is perspecti ve authentic, naturally occur ring telephone conversati ons au dio-recorded i n dif ferent business set tings. Thes e conversations involv e employee s of several Danish com mercial companies and the ir foreign partners, speaki ng in a varie ty of la ngua ges –English, German, Fren ch, Sw edis h a nd Da nish; th at somet i mes c onstitute a lingua franca , i.e. th e s ystem of com munication is a FL for bot h interlocutors. Paying a ttenti on to t he soc i al and conte xtual aspects of l anguage production, Wa gner and Firth’s (1997) a nalyses suggest th at non-target-like e lements usually interpre ted a s C Ss in previous rese arch ma y fulfi ll in interaction a variety of dif fe ren t functions. Thus, code switch ing can be em ploy e d a s a CS to negotiate meaning, but als o as a social strategy t o negotiate identity. It can a lso be the result of a chan ge in topic or in si tuational feat ures. Apparen t CSs ma y be ac tually part of private working argots. In interacti ons involving well -a c quainted interlocutors tal king about rout ine topic s, non-target-like element s be come, under certain circumsta nc e s, s table and more or le ss fixed alte rnative expressions to compe nsat e for lexical gaps shared by the two interlocutors. In order t o understa nd these uses, attention nee ds to be paid to the relationship between t he two i nterlocutors a nd t o their idiole cts or p ersona l b iographies. Finally, spea kers fill som etimes their lexica l gaps with un flagge d n on-target-like eleme nts that s erve as du mm y item s. When the appropriate TL e lement is l a cking but is not rele vant for the purposes of disc ourse, speakers may resort to alte rnativ e expres sions that are not necessarily intended for solving the l inguistic proble m. I n the se cas es the non-target-like element function s as a blind spot, w hich will be left unattended by both interlocutors. They keep the proble m unr e solved a nd they only t ry to solve it later on in the co nversation if this item becomes re levant for the meani ng creation process. In s um, Wagner a nd Firth’s (1 997) examinatio n of na turally occurri ng conversati ons shows that C S s a re c ontingent products of interacti on. T he funct ion of a CHAPTE R 1: Literature re vi e w - 66 - up to 4 0 dif ferent c ourse and t eaching res ource books on English as a F L, f inding tha t they offer ve ry fe w e ffective practice ac tivities or t eaching guide line s that ca n be used for the d evelopment of the l e arner ’s s tr ate gic c ompetence. It seems, t herefore, that more work a nd eff ort are sti ll required f rom bot h SL A re searchers a nd F L te achers fo r a real integration of CS instruction in the FL te achi ng practice. 1.1.3.6. Concluding remarks The aim of this literature review is not t o pre se nt a n e xhaustive account of all the research carrie d ou t on CSs within and outside the f ield of S L A –the re are c ertainly significant s tudies that have bee n left of our di scussion because of th eir little releva nc e for the purposes of our investi gation. This survey provide s us, h owever, with a c lear picture of what is k nown so f ar about the us e that different FL learners m ake of diffe rent CSs in dif ferent communicati ve s ituations, and of p resent day c onsiderations a bout the role that the se strategies may play in the FL t eac hing practi ce. A genera l assessm ent of the research cond ucte d i n t he area shows t hat S LA scholars have a pproa c hed the study o f strategic communicat ion mainly with d escrip tiv e aims . T hey have tried to des cribe and e xplain the us e that the FL learner makes of CSs, paying attent ion for this purpose to both individual and c ontext-related facto rs. Wh en other is sues, such a s the effe ctiveness of diffe rent CS types, have be en a pproached, the underlying intent ion has a lso been descriptive. The only line of research t hat has tried to establish some prac tical implications i s th e one f o llowed by those sc holars interested in the poss ibility of int roducing CS instruction in the FL clas sroom. This rev e al s tha t, in g e nera l, th e study of CSs h as b een re garded a s a study of language use, not of l anguage lea rning. S om e s cholars have, howe ver, made some i nitial claims about the poss ib le role that the use o f certai n CSs ma y play on the developme nt of the learner’s IL s ys t em . As a lr e ady me ntioned in the introduction to our work, it has bee n arg ue d that CSs can be used to test hypothes es about the TL ( Corder, 1978), to elicit fe edback on these hypothes es, or to obtain new and comprehensible input (Færch and Kasper, 1980). It h a s also be en suggested t hat the use of certain CSs ca n ma ke already existing knowledge more accessible (Færch a nd K asper , 19 86), help the learner gain cont rol o ver T L lexi cal know l edge (Kasper a nd Kellerman, 1997) a nd, in general, provi de le arners wi th more oppo rtunities f or interaction, negotiatio n of me aning and CHAPTE R 1: Literature re view - 67 - comprehe nsible input ( Larsen-Free man and L o ng, 1991 ). All these different claims have, however, bee n ra ised from a the or etical perspective. T hat is, no em pir ical research specif ically de voted to exami ne t hese issues in a systematic way ha s yet b ee n conduc ted. Thi s explains, in part, why in the last years this field has been unable to attract a s m uch intere st as othe r related a reas of resea rch, suc h as negotiation of mea ni ng or feedback studies, als o foc used on the lea rn er’s u se o f th e FL , but in relati on with the SL A proces s. The pres ent work re present s an attempt to a dvance in this direction. T aking int o account what w e know so far about the use that the F L l earner m akes of C Ss, we w ill try to e stabli sh the pos sible implications t hat this use ma y ha ve f or the dev elopm ent of the learner’s IL s ystem. That is, we w ill ana lyze lea rners’ stra tegic i nteraction in orde r to d escribe a nd ex plain what we believe to be a collaborat ive creat ion of meaning process, but also in order to ide ntify whether this process may have an effect no t only on the learner’s use of the language but also on their IL syste m. 1.2. COMM UNICATIO N M ODEL S: THE COLLABORATIVE M ODEL The s tudy of CSs originate d in the field of SLA r esearch. It has been, the refor e , mostly focused on t he use that lan g uage lea rners make of th ese st rategies t o c ompensate f or their IL d efic iencies w hen tryin g to c ommunicate in a L 2. H owever, fr om the ve ry beginnings it ha s bee n pointed out that CSs are not a L 2 spec ific phenomenon (Tarone, 1977, 1981; Færch a nd K asper, 1983b; Poulisse e t al ., 1990; Yule and Tarone , 1990; Kellerman, 1991). NSs face a range of di fferent comm unicative sit uations which p ose similar pro blems to those encounte red by the FL le arner. Childre n acquiring their L1 have also a limited command of v oca bulary whi c h results i n the same communica tiv e problems L 2 lear ners face, a phasic patients need o ften to compensate for their lexica l gaps in a si milar way a s L 2 lea rners, and unim paired speakers a r e s om etimes tempora rily una bl e to ret rieve a word i n thei r own L1 –the tip-of-the-tongue phenome non– and occasio nally f ace the lack in the ir le xicon of c ertain words o r expressi ons which belong to discou rse dom ains too specialize d or j ust unfami li ar t o them. I n other contexts the NS may know the desired word to communicate their intended messa ge, bu t they ca nnot use it beca use it is unknown f or t heir interlocutor s. This ha ppens t o experts trying to e xpla in technical terms to laymen, teac hers CHAPTE R 1: Literature re vi e w - 68 - introduci ng ne w vocabulary t o thei r s tudents, writers of monolingual dictionaries or simplifi ed rea ding text s, an d a dults trying to c ommunicate to learne rs or chi ldren who have not yet acquired a ful l comma nd of the language –f oreigner t al k and mothere se speech. 16 The source of the problems encounte red by NSs in t hese communica tiv e situa t ions may be different from that of the lexical difficulties t he L2 l earne r usually faces, but the nature of t he problem is c ertainl y the same. In all t hese contexts, as in L2 strate gic communicat ion , the word or e xpression desi red to c ommunicate an intended message i s una vailable and speake rs manage to c ompensate f or it by d eveloping a n alternati ve means of expressi on, i.e. by making use of a CS. NSs may not resort to CSs as ofte n a s L 2 le arners do , b ut they clearly make use of these s tr ategies in the ir ev ery day oral interactions. CSs are thus not exclusive of L2 speech but an as pect of overall com munication. It se ems theref ore reasonable to as sume that, f irst, resea rch ca rried ou t in certai n f i elds of L1 com municat ion could and s houl d inform the study of L 2 strate gic beha vior; and, seco ndly, a fair a nd compre hensive account o f CS u se should be able to provide a common explanation for both L1 a nd L2 strate gies. The need to c ross di sciplinary bounda ries i n order to rem ove the study of CSs from its relatively isola ted posit ion within the field of SLA has bee n frequently pointed out. We have already seen that SLA s cholars ha ve m ade some attempts to adopt psycholinguistic a nd sociolinguistic frameworks of analysis for the purpose of L 2 C S expla nation –cf. present c hapter, p. 47. Att ention has p articularly bee n c alled to the similariti es e xisting betwe en L 2 stra tegic beha vior a nd c hild a nd adult L1 ref erential communication, and to the possibility f or CS rese arch t o benefit form the work being done on refe rential communica tion (Bial ys tok, 1990; Poulisse et a l ., 19 90) . Cla ims hav e also been made t o the potential value of researc h on topic s s uch as child bilinguali sm (Deutsc h et a l ., 1997; Rodino a nd Snow, 1997), mothe r tongue education (Lloyd, 1997) or l a ng uage patholo gy (Stemme r and Joa nette, 1997). As alread y me ntioned, thes e studie s were in clude d i n the la test state of the art c ollection o f l iterature on CSs, i.e . 16 For a more detaile d anal ysis o f these an d other simi lar L1 communica tive pr oble ms a nd their relati onship with L2 strate gic comm u ni cation see Poulisse et al . ( 1990: 9), and Kasper and Keller man (1997: 8- 9). CHAPTE R 1: Literature re view - 69 - Kasper a nd Kell erman’s (1997) w ork; on the grounds that, althou gh not di r ec t ly foc used on CSs, they should be interesting and rele vant to SLA research on this issue. The recogniti on of the work carr ied out i n these related domains ha s influence d the study of C Ss in two different but c losely related w a ys. On the one ha nd, it has m ad e possible the establishment of comparative analyses between L1 a nd L 2 strat egic communication and t o find evidence of the a ssumed s imilarities exi sting be tween L1 and L2 CSs. O n the o ther hand, L 1 c ommunic ation r esearch has p rovided SLA studi es with theoretical models of communi cation ab le to a ccount f or CS use both i n the L 1 a nd in the L 2, and thus to expla in CSs as part of bot h NSs’ and NN Ss’ use of the lan guage. In the previous secti on of the pre sent c hapter, we m entioned t he work o f a group of researchers who, dra w i ng on the existing literature on L1 re fere ntial communication, compa r ed the use tha t s p eakers make o f referential CSs whe n communicating i n their L 1 and in a L2 (Bongaerts et al ., 1 987; Bonga e rts a nd Poulisse, 1989; K e llerman et al ., 1990; Poulis se et al ., 1990; Russell, 1997 ) –cf. pp. 46 -47. T he s tarting point for the s e studie s is the a ssum pt ion tha t, since r esearch on L 2 strategic com munication has been mainly foc used on learners’ use of lexical CSs to compensate f or vocabulary gaps when try in g to e stabl ish ref erence, the analysis of L2 CSs could benefit fr om the work done on L1 ref erential communica tion. These sc holars point out that, i n f act, the tas ks develope d f or t he study of refe r ential c ommunicati on crea te a c ommunicative situati on that shares s ignificant features with the kind o f communica tive context in which L2 C S uses are likel y to occur. These tasks –the abstract figure descript ion task designed by Krauss and Weinheimer (1964) a nd its multi ple ada pted ve rsions– re quire from th e speaker to refer to a pre viously non-e xiste nt abstra ct f igure, s pecifically designed for this purpose and that there fore has not be e n lexic alized in their la nguage. The referential problem th e NS fac es i n t his situation is assumed to be the same ki nd of pr obl em the L 2 speaker fac es when lac king t he a ppropriate TL w ord to n ame an i ntended refere nt. On this basis these aut hors concl ude t hat the results o f a s tudy using this kind of task can shed light on both L 1 a nd L2 stra te gic and non-strategic re fe rential comm unication (Poulisse et al ., 1990: 5 2) . Using a modified v ersio n of K ra uss and Weinheime r’s (1964) ref erential communica tion t a sk, th e se rese a rchers were a ble to e licit L 1 and L2 strat egic performance da ta f rom the same group of spe ake rs a nd to f ind evidence of si gnificant CHAPTE R 1: Literature re vi e w - 70 - similariti es be tween t h eir L1 a nd L2 re fe rential strategies . Partici pants des cribed the figures from the same perspective s, followed a simila r pattern of choice between hol isti c and segmental pe rspectives, an d reduced the complexity of their referential expres sions in a similar wa y when repeate d reference s were made to the same figure. Although diffe r ences were al so detec ted in the a mount of speech a nd the s tructural com plexity o f the r eferential expressions used t o perf orm both ve rsions of the tas ks, the resul ts obtained suggest that t here a re no intri nsic d ifferences in the nature of t he L 1 and L2 strate gies u sed. A simi lar c onclusion was a lso reache d by Bial ystok (1990), who compa r ed da ta obtained on L 2 CS use with the re sults of previous L1 research on children and adult NSs’ referential communi cation skills. From a n interactional a ppr oac h to the concept of CS, attempts have also bee n made t o c ompare L 1 and L 2 st rategy use –cf . Tarone a nd Yule (1989). For the purposes of this kind of research t he object of interest wa s the comparison of L2 learn ers’ strate gic behavior and that of t he NS of the TL . In order to el icit thi s kind of data, Tarone an d Yule (1989) designed a communicative task in which reference had to be made to a se t of real but unf amiliar objects expecte d to pose le xical difficulties to both L 1 and L2 spea kers. In line wi th thei r theoretical a pproach t o the study of CSs, Tarone and Yule’s analyses are focused on the form and structure of th e r eferential means of expressi on used to perform the tas k, rather than on t he c ognitive proce sses unde rly i ng them. Yet, they also reveal st riking similarities between L1 and L2 strate gies. The results obtained by all these different studies show that CSs not only consti tu te part of speakers’ L1 and L2 c ompet ence, they are also used alike i n L1 and L 2 com munication (Poul i sse et al ., 1990: 2). In other words, s peakers resort to t he sam e set of s tr a tegies when communicating in their L 1 a nd i n a L 2 . Thi s ki nd of evidence supports t he c l aims made in fa vor of an explana tion of CS use a ble to acc ount for L 1 as well as L 2 s trategic behav ior. With this objective in mind SLA researchers ha ve re lated the study of L2 CSs to the genera l study of L 1 communication a nd tried to e mbed these s tr ate gies into theoretica l m odels of c ommunication originally developed to e xplain non-strategic NL use. The f inal aim of this line of re s e arc h is to develop an all -encompassing the ory of CS use abl e to account for the se strategie s as an aspect of overal l communication and, in this way, de finitely remove t he study of CSs from its restricte d L 2 position. CHAPTE R 1: Literature re view - 71 - In s ec tion 1 .1.1 of this c ha pter, whe n disc ussing the c ontroversy e xisti ng around the issue of the definition of C Ss, we sa w how Færch an d Kaspe r (1983b) desc ribe strate gic b eha vior in te rms of Clark and Clark’s (1977) mo del of speech product ion, wherea s Poulisse et al . (1990) adopt Levelt’s (1989) mode l with a similar purpose. In Færch and K a sper’s a ppr oac h CSs a re a nalyzed a s plans. Strategies a r e c onsidered t o be a subtype of pla ns d evelope d by the spea ker to reach those communicative goals whic h present some kind of linguistic diffic ulty. They ar gue tha t these s tr a tegic plans work in a similar way as non-strategic on es; i .e. those de veloped for non-stra teg i c or non- problematic c ommunicative goals. On this basis they concl ude that “there is no compelling re ason imposed by re ality why a distinction shoul d be observ e d be tween plans and stra tegies” (Færch and Kasper , 1983b: 30). Poulisse e t al . (1990) a nd Pouli sse (1993, 19 97) t alk about processes, rather tha n plans. But they s imila rly be lieve that the processes involved in CS us e are the same as those leading to wha t is general ly ass umed t o be non-stra teg i c communication. The y recognize th at the procedure for encoding a m essage gets more complicated when lexica l di fficulties ari s e . They a rgue howe ver t hat t he na ture o f the proc ess in itself is not altere d and that therefore CSs shou ld be see n as just a nother way of enc oding messages (Poulisse et al., 1990: 66). These argument s are also in line with Bialystok’s (1990) cla im in fa vor of a more general app roach t o t he stu dy of CSs –cf. pres ent cha pt er, p. 28. This a uthor believes that no distin ction sho uld be made between strategic a nd non -strat egic language use since any atte mpt to reach a c ommunicative goal, whe ther i t in volves some kind of linguistic di fficulty or not, is strat egic. For t his pu rpose she propose s t o integrate the s tudy of CSs into a cogniti ve the ory of languag e acquisition and use ba sed on a dis tinctio n b e tween a nalysis of knowledge a nd cog nitive control proce sses –cf. also Kellerman and Bialystok (1997). Tarone and Y ule (1 989), approach ing the study of C Ss f rom a more interac tional perspecti ve, ado pt a diffe rent theoretical orientation t o e xplain t he s t rategic phenome na. Focusing the ir attentio n on st rategic linguistic res ources rathe r tha n cognitive processi ng, t hey trea t CSs as referential expressions a nd try to e xpla in t he m withi n the general f r ame work of re ferential comm unication. These au thors se e strategic compete nce as the speaker’s a bili ty to s elect the most effec tive refe rential e xpr es sion CHAPTE R 1: Literature re vi e w - 72 - availa ble to perf orm t heir com municative goals. D epending on the sp ea kers’ linguistic resourc es, their a ssessment of the interlocutor’s knowledge of the language a nd the conte xt of t he interaction, the l inguistic e xpression chosen will result in a nat ive or a non-native-li ke utterance. B oth kinds of utterances, however, ca n be s imilarly s een as the out come of t he speaker’s s trategic a ttempt to c ommunicate the ir message wi th the most effective possi bl e me ans. L a nguage us e, as a type of goal-related be havior, is therefore always strat egic, whe ther i t face s the spe aker with a lexical gap and a CS use to com pensate for it or not. CSs h ave also been analyzed withi n t he framework of the pattern the ory (Wagner, 1983), conversati on ana lysis (Wagner and Firth, 1997) or critical sociol ing uis ti cs (Ram pto n, 19 97). The resul ts obt ained by al l the se different studies suggest tha t CS uses can and should be c onsidered as spe cial cases of gene ral languag e use which, on th e other hand, can be seen a s a funda mental strategic a ctivity (Kasper and Kelle rm an, 19 97: 15). Therefo r e, not onl y can CSs be successf ul l y embedde d within models orig inall y de veloped to explain L1 non-strate gic communication. In fact, it seems that this is the way r esearch should t ake if a n overall c omprehens iv e understa nding of strategic communicat ion is aimed at. The comm unicative m odels re viewed so far a re m ostly c oncerned wit h the speaker a nd the speech product ion ph ase of the communi cative proc ess; therefore, they pay little or no attent ion at all to t he role of t he interlocutor. I n this cha pter we present a theoretica l m odel of c onv e rsation analysis tha t explains comm unication as t he res ult of the sp eak e r a nd the int erlocutor’s col laborative effort to ac hieve a j oint communicative goal. This m odel was orig i nally de vised b y C lark and W i lkes -Gibbs (19 86) in order to explain reference m aking as a j ointly ac hieved activity, i.e. to invest igate how spe akers and listeners coordina te the ir beliefs a nd behaviors i n order to re ach an agr eeme nt on reference . But it has proved to be suit able for t he gene ral analysis of conve rsational interacti on which, in these a uthors’ view, is alwa ys a collabo ratively c o-constructe d activity (Clark and Schaefe r, 1987, 1989; Wilkes-Gibbs, 1995, 1997). In 19 97 Kasper a nd Kellerman include d in their a lready me ntioned st ate of the art col lection of re s e arch on CSs a review of c ommunicati on studies carried out withi n the collaborative theory framew ork (Wi lkes-Gibbs, 19 97). They were thus suggesti ng CHAPTE R 1: Literature re view - 73 - the poss ibility for SL A re search on CSs to benefit f rom the ins ights off ered b y the collaborati ve analysis of L1 communic ation. One year a fterwards, Anderson (1998) mad e a f irst attemp t to apply the collaborati ve theory m odel, pr eviously rest ricted t o the study of NS conversation, to the ana lysis of N NS int eraction –cf . current c hapter, pp. 60-61. This m odel m ade it poss ible to look a t the role of bo th the speake r a nd the listener i n t he solution of th ose communicative problems that a rise i n FL inte raction due to interlocutors’ linguistic shortcomings. By looking at ‘both s ides of t he page’, i.e. spea ker and listener, Anderson mana ged to int egrate in one single f ramework of analysis two previously separate areas of inqui ry: stra te gi c c ommunicati on and ne gotiation of me aning. He was thus responding to Y ule a nd Tar one’s (1991) claim in favor o f a bilateral approac h to the study of comm unicative diffic ulti es in FL communi cation. The observati ons m ade by Anderson (1998) a bout i nteract ion a nd st r ate gi c language use, although rel atively l imite d in scope, suggest the suitabi li ty and feas ibility of f urther resear ch looking at CS us e fro m a collaborative perspective, paying a ttenti on to both speake r a nd i nterlocutor as co-constructors of m eaning. As expla in ed in the inroducti on, the collaborative theory mode l will a lso be adopt ed in our study f or the purposes of CS use analy sis. In f ollowing c hapters we will argue f or a n explanation of strate gic communication i n interactional contexts as the result of the joint collaborati on between the l earne r and their native or non-native interlocutor. Here w e introduc e the model and review the relevant lit erature that will constit ute th e ba ckground of our f uture ana lyses. 1.2.1. TH E COLLA BORATIVE THEORY As we have just mention ed, the collabo rat ive theory sees c ommunication a s a collaborati ve a ctivity, i nvolving ac tions by spea kers and i nterlocutors and their mu tual responsibi li ty in the understan ding of ea c h utterance. On this ba s i s, the c ollabora tiv e model wa s designed t o ac count f or how s peakers and interl ocutors coordinate t heir beliefs , and the conte nt and timing of their ac tions for the purpose of the co-cons truction of meaning. A s Cla rk and W ilkes-Gibbs (198 6: 2) cl early po int out, it i s relatively evident to assume that communi cation is a col laborative proc ess, but it i s s omethi ng diffe r ent to un derstand why an d how it works. The de s cript ion a nd c larification of t his CHAPTE R 1: Literature re vi e w - 74 - process is the final aim of the work c arried ou t within the framework of the collaborati ve t heory o f communicat ion –mos t notably Clark and W ilkes-Gibbs (1986), and Clark a nd Schaefer (1987, 1989). 17 The collaborative m odel appea rs as a reaction t o previous unilateral mod els o f conversati on ori gi na ted wit hin a variety of d iffe r e nt a pproaches to the study of communication: L in guistic s ( Heim, 1983), Philosophy (Austi n, 1962 ; Grice , 1957; Schiffe r, 1972; Searle, 19 69; Kamp, 1981; Lewis, 19 79; Stal nake r , 1978), or Psychology (Cl ark and Haviland, 197 7; Johns on-L a ird, 1983; van Dij k a nd Kintsch, 1983) –cf. Clark and Wil kes-Gibbs (1986: 1-3), and Clark and Schaefer (1989: 260- 262). All these implicitly or e xplicitly unilate r al accounts of di scourse e xplain speaking and understandi ng in conv e rsation a s autonomous activitie s. In these mo dels speakers are seen as senders, whose main responsibility is to issue comprehe nsible utterances, i.e. to produce the right utterance at the ri ght moment; and addressees a s receivers, who have to listen a nd unders tand wh at is said. T his view implies that spe akers assume that addresse es understand their utte rances, and that addressees assume t hey have c orrectly understood as intende d by the sp eaker. Howeve r, as Clark a nd W ilkes-G ibbs’s ( 1986) ana lyses reveal, this is an idea lization tha t does not always work. Addresse es do no t alwa ys understand as intende d, and speakers do not just utter words a nd sentences, they also re act to what t heir interlocutors say or do, changing t he c ourse of their utte rances and inv iting the m to c ollaborate in the me aning crea tion proces s by p roviding re pairs or completions of what they are saying. 18 In the col laborative approach both spea king a nd understandi ng ar e see n a s the mutual responsibilit y of all the interlo cutors. Speak ers and a ddressee s need to coordinate on what is meant and wha t is unde rstood so that s uccessful c ommunication i s ach iev e d. In thi s view c onversation , a s a ny ot her c oll ectiv e a ct, c an only suc ce ed if all the part ici pants work t og ether, perform ing a nd c oordinat ing t heir individual ac tions to ach iev e a com mon goal. The a im of the collaborative mode l is to i dentify a nd de scribe these actions, a s well a s t he procedures followed in orde r to coordinat e t he m in face -to- face interaction . 17 I n 19 86 Clar k an d Wilkes- Gi bbs desi gned a f irst version of the c o llab orative m odel a ble to ac count f o r how refer ence is m ade i n fac e-t o-face c onversati on . This model was r e v ise d in 198 7 a nd 1989 b y Clar k and Schaefer , who exte nded it s applica bilit y t o over all face -to - face oral c ommunication. 18 S ee C lar k (2 004) f o r a detaile d c omparative a cc ount o f unilateral and b ilater al mode ls o f communicati on. CHAPTE R 1: Literature re view - 75 - The st art ing point of the c oll aborat ive theory is the generally ag reed a ssumption that communication i s a common ground building ac tivity. W hen two or more p eople take part in a c onversati on they pres uppose they share a ce rtain am ount of knowle dge, beliefs a nd assumpt ions; eit her beca us e thes e are universal ly he ld i n the c ommunity to which the pa rticipants belong or be cause of their prior j oi nt personal experience s – communal versus personal co mmon ground (C lark and M arshall, 1 981; Cla rk, 1993, 1994). The se presuppos itions, t he sum of the p articipants’ pre suppos ed mu tua l knowledge, bel iefs a nd assum ptions, con stitute t heir com mon ground (St alna ker, 1978: 320). In the c ourse of interaction part icipants work to add to thi s initial common ground, i.e. they use language to try to build ne w m utual knowledge, belief s a nd assumptions out o f the ir i ndividual ones . Each c onversational move or cont ribution is produc ed a nd interpre ted in relation to the as sumed c urrent co mmon ground and results in t he creati on of new presuppositi ons and the destruction of old on es. Common ground is t hus bei ng constantl y ac cumulated a nd updated as the conversat ion pro ce e ds (Cla rk and Scha efer, 1989; Cla rk, 1994, 2004). These a ssumpti ons are general t o most models of co mmunica tion. Bu t whereas unilateral ap proaches sugg est t hat the main means to a dd to c ommon ground is the speaker’s ut terance of a comprehens ible sentence, the colla borative theory consid ers this to be ins ufficient. A s just m entioned, Cl a rk and Wil kes-Gibbs (1986 ) argue that speakers cannot just as sume that their ut terances are always c orrectly unde rstood by the addresse e and t heir c ont ent automatical ly added to their c ommon ground; ne ither can addresse es belie ve they have corre ctly inferred all t h e speakers’ communica tive intenti ons. F or com mon ground to be buil t and c onv e rsati on t o go on, interlocutors need evidence of th eir mutual unde rstanding. That is, s pea ker a nd a ddressee n eed to e stablish the m utual belief tha t what wa s meant by the speaker has in f act been correctly understood by the a ddressee . T his req uir es the j oi nt ac tion of al l t he pa rticipants in the conversati on a nd is accomplis hed by mea ns of wha t Clark a nd W ilkes-Gibbs (1986), and Clark a nd Schaefer (1987, 1989) ha v e ca lled the grounding process . 19 L et us say spea ker A w ants to make reference to x. This m ea ns A inte nds the identit y of this referent x to b ecome part of A and B’s common ground –B being the addresse e. For this purpose A ne ed s to present to B a referring expre ssion that, on the 19 Clar k and Schaefer (198 9) init iall y named i t the ‘c ontent groundi ng pr o c ess’ ; but for c onvenience it is usuall y referr ed to as the ‘groundi ng process’. CHAPTE R 1: Literature re vi e w - 82 - discourse but also how the y are going to do it. In m ost c ases t hey w i ll try to spe cify the conte nt they want to ground in one single and complete turn, a s we h ave see n in the previous exa mples. B ut one si ngl e turn c an a lso c ontain several different contribut ions and one single contribution may be presented i n several differe nt tur ns. On th is ba sis C lark and Sch aefer (1989) ha v e i dentified thr e e main patterns of contribut ions to di scourse: c ontributions by t urns , contri butions w i thin turns and contri but ions v ia s entence part s . Here we will ill ustrate how the c ollaborative model account s for the se different pa tterns of b ehavior and f or the refa sh i oning proce sses that are carrie d ou t within each of these contribu t ion ty pes every time a problem or difficu lty is encounte r ed. 1.2.2.2. Contributions b y turns Contributi ons by turn s oc cur, ac cording to C l ark and Schaefer (1989: 270), every time there is “a relevant ord erly cha nge in turns”. This me ans that the contribution i s initiated by a spe aker’s t urn in which t hey pres ent the content they wa nt to ground –u sually by mea ns of a full or elliptical sentence; a nd thi s is foll ow ed by a n a ddressee’s tu rn in which they indica te t heir state of unde rstanding of the s peaker’s pre sentati on. If the y believe th ey have hear d, n oticed a nd understood wha t the speaker mean t to add to their common ground well enoug h for their current purposes, t hey wil l accept i t. If, on the contrary, t hey ha ve t rouble i n un derstanding it , the y will use t heir turn to reje ct the presentat ion a nd initiate a side sequence that will lea d the ac ceptance phase into a series of c onse cutive t urns. In order to e xpress acceptan ce, the addressee needs to provide som e kind of positive evidence of t heir understanding. They ma y use several d ifferent ki nds o f procedures f or this purpose. T hey ma y asse rt t hei r accepta nce by m eans of ack nowledgments s uch as ‘yes’, ‘rig ht ’, ‘oka y’, ‘I s ee’ or ‘mh hm ’ –cf . e xample 1.1 above, p. 80; and, in f ace-to-face interaction, nonv e rbal s ignals such as he ad nods, smiles or rai s ed eyebrows. 22 They ca n demonst rate, i .e. with a paraphrase, or displ ay, i.e. with a verbatim rep etition, what they hav e unde rstood. But th e y may also presuppose acce pt a nce. The y do this by al lowing t he conversati on to c ontinue or by initiatin g t he mselves a rele vant next contribut ion. In the first case, their continued 22 See Clar k (2004) f or a stud y o f t he different effects that verbal and n onver bal form s of evidence have on the wa y the grounding proc ess is carrie d out. CHAPTE R 1: Literature re view - 83 - attention a nd t he pa ssing up o f the opport unity to indic ate t rouble or to initiate a re pair of a pos sible misunderstandi ng is consi dered as an acceptance of the last prese ntation. In the second c ase, the relevance of their utt erance s ugg es ts the ir correct understandi ng of the previous one –cf. example 1.2 in p. 81. Cla rk and Schaefer (1989: 27 2) argue that e very utterance that initiates a t urn belongs to the p resentation pha se of some c ontribution. This m eans that a n addressee’s answer to a sp ea ker pre sentation, whet her a n acknowledgment or a rele vant next turn , consti tu tes no t only an acc eptance b ut also a new pres entation that n eeds it self to b e accepted. The a cceptance phase be comes t hus in he rently recursive. T his re cursion stop s when the we akest form of evide nce is reac hed, i.e. when acceptance i s presupposed by mea ns of c ontinued attention and/ or the speaker’s i nitiation of a ne xt contribution (Clark and Schaefer, 1987: 36; 1989: 268). 23 Non-ac ceptance can a lso be expresse d through a variety of diffe r ent procedures. Addressees may directly a ssert their non-understandi ng of the whole or pa rt of the previous presentation. They m ay imply the ir la ck o f understanding by r equest ing a repeat or asking for more information, i.e. asking for clarification. T hey may repeat the presentat ion or pa rt of it with r ising in tonation, asking thus for confirma tion tha t the y have in fact correctly unde rstood the messa ge as intende d. The y may finally imply their non-acce ptance by means of quizzical looks or silences. These diffe rent f orms of non-understa nding evidence a lso need to be considered by t he speaker. In general, whe n a ddressees rejec t a p resentation a si de sequence is opene d a nd a refashioning proce ss ge ts i nitiated. Refashioning occurs a lso whe n speakers , as pointed out above , notice a problem in their ow n presentation or in their interlocutor’s under standing. Acco rding to Cla rk and Wil kes-Gibbs (1986), ref ashioning may be acc omplished i n three differe nt wa ys : by m eans of a repair , a n expans ion or a replacement . 23 Clar k and Sc haefer (1989 ) e xplain t his proces s on t he basis of what t hey ha ve called t he stre ngt h of evidence p rinciple : “the partic ipants e xpect that, if e vidence e0 is ne eded f or accepti ng prese ntation u 0 , and e 1 for acce ptin g the pre sentati on o f e 0, then e 1 wil l be wea ker tha n e 0” ( Clar k a nd Schaef er, 1 989: 268). T his m eans tha t evide nce g e ts w ea ker a nd wea ker u ntil the w ea kest f orm o f e vid ence, c ontinued attention or initiati on of t he ne xt turn, i s reac hed. I n thi s w a y the acce pta nce proce ss wil l ne ver spin out indefinitel y . CHAPTE R 1: Literature re vi e w - 84 - 1.2.2.2.1. Re pairs Repairs, as u nderstood by Clark and Wilk es-Gibbs (1986: 21), a re a lways ca rried out by the speaker. They o ccur whe never contributors, detecti ng an e rror or a n ina dequacy in the pre s e ntation they are utteri ng, decide to change it midcourse. In ongoin g con versation, time pressure leads relatively of ten to the use of misc hosen or er roneous w ords and express ions w hich ne ed to be corrected a s so on as possible in order to pr event g roundi ng problem s. This is what is h a ppening in t he foll owing exa mple, where ‘his legs a re’ is repaired int o ‘hi s one le g is’ (Cla rk a nd Wilkes-Gibb s, 1986: 21). EXAMP LE 1.3: A: Um, next one is t he gu y , the p ers on with his hea d to the ri ght but his l egs are, his one leg i s kicked up to the l eft The revise d version of the prese ntation is what t he addressee is expected t o consider a nd accept. Theref ore repairs need to be ma de, not onl y a s soon as the prob lem is identifie d, but also i n an unam big uous way, making cle ar what part of the presentat ion is being revised and how. 1.2.2.2.2. Ex pansions Once a presentat ion has bee n utte red, the speake r or the add ressee ma y find it correct but not c lear e nough f or c urrent purposes. They are then expected t o ex pand it by adding new i nformation. Expansion s are directly add ed by speakers, without need of prompti ng from the addresse e, whenever t he pres entation proces s is initiated with a pro visional exp ression, i.e. an expre ssion use d under ti me pres sure that the spea ker knows cannot ac curately specif y the content they want to ground. T his explai ns why, in the f ollowing e xamp le, the spea ker has dec ided to e xpand their prese ntati on by a dding ‘the l eft side look s like a square’ (Clark and Wilkes-Gibbs, 1986: 21). EXAMP LE 1.4: A: Oka y, number 1 is the j ust kind o f bl o c k-like fi gure with t he jagged ri ght-han d side. The le ft side looks li k e a s quare. Expansion s c an a lso be reque sted by t he a ddressee when the y find they c annot accuratel y understand th e speaker’s pres entation as j ust uttere d. A sil ence where a reply is expec ted, a tentati v ely voice ‘um’, ‘uh huh’ or ‘ye s?’, a quizzical l ook, or a re peti t ion CHAPTE R 1: Literature re view - 85 - with rising intona tion o f some part of the prese nta tion are g en erally considered to b e a n addresse e’s request for an expansion. As ca n be s een in the f ollowing example (Cl a rk and Wilke s-Gibbs, 1986: 21-22), ea ch m ovement in the side se quence c onstitutes i tself a ne w presentation to be grounded. The addressee’s s ig nal of t rouble needs to be recognized by the sp eake r, a nd both the initial pre sentati on a nd its e xpansion ne ed to be finally unde rstood by the a ddressee and t hus indicate d t o the s peaker so t hat mutual acceptanc e can be reached and the conversation can g o on. EXAMP LE 1.5: A: Oka y, the next one is the rabbi t. B: Uh- A: That’ s asleep, you know, it looks li ke it’s got ears and he’ s pointin g down? B: Oka y . A third t ype of expa nsion occ urs w hen the addr esse es t hemselves e xpand the speaker’s p resentation. Usual ly thi s i s accomplishe d in t he f orm of a reque st for confirma tion or confirmation ch eck ut tered with ris ing intonation. In the following exa mple (Clark an d Wilkes-Gib bs, 1986: 22), the a ddressee does not consider th e speaker’s prese ntation cle ar enough f or com mon ground b uilding purpose s and deci des to sug gest a possible expansion. In turns 3 a nd 4 speaker and addre ssee give evidence of a mut ual acceptance of the initial pr ese ntation plus its expansion. EXAMP LE 1.6: A: Um, third one i s the gu y readi n g with, holding hi s book t o the left. B: Oka y , kind of standi ng up? A: Yeah. B: Oka y . 1.2.2.2.3. Re placements Finally, una cceptable presen tations can a lso be re fa shioned by me ans of a replac ement. Once the sp eake r’s presentat ion ha s finishe d, the add ressee ma y reject it and propose a substitu ting version to replace th e original one. T he addressee wil l normally utt er t his replaceme nt with rising int onation, ind ic a ti ng they are not full y sure they ha ve correc tly understood t he speake r’s ini tial presentation –cf. e xamp l e 1.7 ta ken f rom Cla rk a nd Wilkes-Gibb s (1986: 23). CHAPTE R 1: Literature re vi e w - 86 - EXAMP LE 1.7: A: Oka y, and the next one is the pers on that loo ks like t hey’re ca rr y i ng some thing and it’s stic king out to the left. It lo oks like a hat that’s upsi de down. B: The gu y t hat’s pointing t o the left again? A: Yeah, pointin g to t he left, t hat’s it! (laughs) B: Oka y . In the l ast two p atterns o f refas hion, interlocutors active ly coll aborate not only in the a cceptance of th e c ontribut ion but also in its presentatio n. T his i s act ually c o- construc ted by the joint action of the two interlocutors. Collaboration at this leve l account s a lso f or the co m municat ive suc cess o f what Clark a nd Schaefer (1989) have calle d contributions via s entence parts and will be analyz ed in m ore d etail i n the foll owing pag es of the present chapter. 1.2.2.3. Contributions within turns In eve ry day conversat ion, speake rs do not a lways gi ve the floor to t hei r i nte rlocutors every time t hey f inish the presentat ion of a new c ontributi on. Sometimes they take exte nded turns in which c ontributi ons f ollow one an other. C lark and Schaefer ( 1989) argue t hat, alt hough there i s no exchange in turns, there i s, h owever, a n ac ceptance process f or each co ntribution, i.e . fo r eac h se parate portion of the turn. On t his basis they have calle d this pattern of behavior cont ributions within tur ns. The speake r’s speech can a lways be interrupted by the addre ssee whenever they believe th ey may have not c orrectly heard or underst ood what is meant. If they pa ss up this op portunity to initia te a repair, t hei r cont inu ed a ttention can be c onsidered eno ugh evidence f or the inte rlo cutors to es tablish the mutual bel i ef that t he content c ontributed to discour se has been satisfactorily grounded. Quite often a ddressees use also continuers or assessme nts inserted during the speaker’s tur n to confirm they are following and understa nding wha t is being sa id. Continuers i nclude gestures s uch as head nods or sm iles, and expr e ssions s uch as ‘mm’, ‘mhm’ or ‘yes’ –c f. e xa mple 1.8, t a ken from Cl a rk and Sch aefer (1989: 280). Assessments a re signa ls s uch as ‘gosh’, ‘really? ’ or ‘oh’, that of fer a br ief j udgment of what has jus t been sa id. As we ha ve a lre ady see n t hese t wo types of acknowledgme nts can al so co nstitute an addressee’s turn, but in contributions within turns the y ar e spoken in th e bac kgr ound, us ually o verlapping wit h spe aker’s tal k. The a ddressee u tters them without ta king the floor from the speaker (Scheglof f, 1982). CHAPTE R 1: Literature re view - 87 - EXAMP LE 1.8: B: But yo u daren’t set s y nt hesis a gain yo u see, . you set anal y si s, and yo u can put the ans wers down, a nd y ou r assista nt *exa miners will work them ,* but if y ou given them a give n t hem A: * y e s quite, y e s, yes* B: a free hand on s ynthesis , and the y’d be ma rking all s orts of s tuff, becau se the y can’t do the stuff *the mselves, * I must watc h [continue s] A: *quite m* Cla rk a nd Schaefer’s (1989) ana ly se s e vidence that addressees us ually utter ack nowledgments when speake rs come to a n end of a certain thought. The y i ndicate the addresse e’s ac ceptance up t o that point and their rec ognition t hat t he speaker i s i n the middle of an ex t ende d unit of spee ch. O n the othe r ha nd, the speaker’s continuati on with the ir turn indicates t hey are acce pting t he addre ss ees’ a cknowledgments as sufficient evidence of their understa nding for the current purp oses. As Clark and Scha e fer (1989) als o point out ac knowledgments are used whe n interlocutors are quite confident of their mutual unde rstanding and are not expecting strong eviden ce. This e xplains tha t, d espite th e ir bre vity, they c an succeed in cr eati ng contribut ions from extended tu rns. 1.2.2.4. Contributions via se ntence parts All t he i nstan ce s of co ntr ibution s a nalyz ed so f ar a re ini tiated by a p resentation which includes, at lea st, o ne f ull or e lliptical s entence uttered b y the speaker with t he a im of specif ying the cont ent they want to c ontribute to dis course. Th ey pe rform ill ocutionary acts, such as as king questions , maki ng as sertions, requests, apologies, etc. –cf. Clark and Schaefer (1989: 282). B ut somet im es , initial presentati ons a re uttered as sentenc e consti tu ents which can only pe rform part s of the i ntended i llocutionary ac t and which, the speaker know s, cannot d irectly l ead to the addressee ’s under standing and a cceptance of their c ommunicative intentions . Quite o ften, speakers do so because they think the c ontent th e y want to a dd to common ground is too complex an d is therefore bett er presented in sev e ral sepa rat e turns – i nstallme nt c ontributions ; bec ause t hey are unsure of som e piece of informa tion and need the a ddressee’s confi rmation – trial constitue nts ; or because they nee d the addresse e’s he lp to complete the presen tation they ha ve j ust initiated – collaborativ e comple tions . CHAPTE R 1: Literature re vi e w - 88 - 1.2.2.4.1. Installm ent contributions Installment c ontributions ar e uttered in several e pisodes. The speakers divide the information they want to present in t o portions an d present e ach o f these portions a s a separate installment followed by a pa use. This pause indicat es they need confirmation of understa nding f rom their a ddressees. That is, each ins tallment of the contribution needs to be grounded, and only once the speaker and the a ddressee hav e e sta blished th e mutual belief tha t eac h e pisode has been c orrectly unde rstood and accepted can the contribut ion go on. When the last i nstallment is utte r ed, the whole presentation a lso nee ds t o be con sidered and mutuall y a ccepted a s an all-e ncompassing unit. Each no n-final installment is uttered with rising intonat ion, i ndicating that, although there is a pa use a nd t he f loor is given to t he addre ss ee, the presentation has not yet fini shed. The final installment is ut tered with falling intona tion to signal the end. In this way, al though the prese ntation is divided into episodes and uttered in different turns, it carries the intonation of a sin gle utterance. As Cla rk and Scha efer (1987) point out, insta llment pre s enta tions can be bot h considered as one or sev eral presentations. In the following example (Clar k and Schaefer, 1989: 284-5), the speaker divides th e desc ription of a wedding into four diffe r ent installments. A c ceptance is establishe d f or eac h of the f our di fferent pa rts of the description, which constitute them s e lves fou r different c ontributions w ith a presentat ion and a n acceptance ph ase. But the fin al a ssertion made by the addre ssee ‘it does sound, very nic e indee d’ grounds not only the last installment , but also the whole descri ption, i.e. a complex presentat ion made up o f f our embedded contributions . EXAMP LE 1.9: B: How how was t he weddin g – A: Oh it was it was reall y good, it was uh it was a l ovel y da y B: Yes A: And . it was a s uper place, . t o have it . o f cour se B: Yes – A: And we w e nt and sat on sat in an o r c hard, at Grantc hester, and ha d a huge tea afterward s (laughs . ) B: (lau ghs - -) A: Uh B: It does s ound, ver y nice i ndeed 1.2.2.4.2. Trial cons tituents Whe n s peakers are not sure whe ther their prese ntation is co rrect or c omprehensible for their interlocutors, the y may u tter it wit h a tr y m arker , i.e. rising intonation followed by CHAPTE R 1: Literature re view - 89 - a p ause (S ac ks and Schegloff, 1979). By a dding a try ma rker, th ey indicate t heir lack of confidenc e in what they have j ust sa id and recognize th e possibility f or a non- acceptanc e. Acc ording to Clark and W ilkes-Gibbs (19 86: 1 9), a se ntence uttered wit h f alling intonati on tends to project an explicit acceptance , i.e. the speaker expects the addres se e to accept it. But when it is mo di fi ed by a try marke r, it projects a verdict : accept or not accept. In the followi ng e xample (Cl ark and W ilkes-Gibbs, 1986: 18), the speaker is making an a ssertion, not a que stion. The final rising inton ation s imply indica tes t hat the speaker is not sure whether the presentation ca n be acc epted as j ust ut tered and is asking the addres see t o explicitly a ssert whet her t hey hav e or not under stood wh at the y mean. The c ontribution will not be fini shed until the addressee confirms thei r understanding. EXAMP LE 1.10: A: Um, the next one’s the pers on ice skating that has two a rms? Someti mes speake rs a re unsu re a bout j ust s ome pie ce of th e informat io n the y want to c ontribute t o discours e. In this ca se they need to m odify only so me part of t heir presentat ion with a try marker. The ma rked constitue nt, i.e. the t ri al constitue nt, ne eds to be accepted by the addressee before the prese nt ation ca n be completed. In the next e xample (Clark a nd Sc hae f e r, 1989 : 286), the speaker is trying t o asse rt t hat ‘A man called Annegra replie d, and…’, but is, howeve r, unsure abo ut this name . The s peaker presents it with rising into nation, i.e. as a trial const ituent, a nd pauses , waiting f or the addre ssee t o eit her accept it or rej ect it. The a ddr essee replaces ‘Annegra’ with t he correct name ‘ Allegra’, indicat ing they believe they ha v e correctl y inferred who the spea ker was ref e rring to de spite the erroneous na me. The speaker accepts the c orrection by repeati ng i t and goes on to com plete the prese ntat io n, which is thus accom plished in t hree different moves a nd co-construc ted by the two pa rticipants in the c onversation. EXAMP LE 1.11: A: So I wrote off t o . Bill, . u h who ((ha d)) pres umabl y disap peare d by this ti me, certainl y , a man calle d Anne gra? - B: Yeah, Allegra A: Allegra , uh replie d, . uh a nd I . put . t wo other people, who’d bee n in f o r . t he BBST j ob . with me [ continue s] CHAPTE R 1: Literature re vi e w - 90 - 1.2.2.4.3. Comple tions In the sta ndard patte rn of c ontributions speakers are assumed to be the o nes who present the content to b e grounded. They are therefore expe cted to utter an u tt erance that specif ies t his c ontent. But somet im es this utte ran ce i s only i nitiated b y the speak er and it is the ad dressee who actually comp letes it. In e xample 1.12 from Clark a nd Schaefer (1989: 28 6), the s pe a ker i nitiates the contribut ion and pauses. The intonation and the pause indicate t here is somet hing else t o come but the speaker is ha ving trouble to utte r it. Th e a ddressee , feeling they ma y know what the speaker is about to say, utters in the seco nd turn what th ey believe to be an appropriate final part for the pre s enta tion, i.e. completes the present ation. EXAMP LE 1.12: A: Um the problem i s a that you((‘ v e )) got to get planning c onsent - B: Before y ou start - A: Bef ore y ou start on that part, ye s Each com pletion ac ts a s eviden ce of t he addressee ’s a cceptance of what ha s bee n so far uttered and constitutes itself a new prese nt a tion that also needs to be grounded. On this ba sis, Cl ark a nd Schaefe r (1989) argue that i n c omple ti ons the re are two contributions . The addressee’s com pletion constit utes one contribution em bedded within t he presentat ion phase of an o verall cont rib ution that i ncludes th e comple tion a nd the spea ker’ s f ragment. Completio ns are particularl y significan t because they consti tute one of the cleare st example s of how mea ning in c onversation i s ac hi eved by the j oint action of more than one pa rticipan t. We see in s peech how speake r an d addressee colla borate t o ach iev e one single co ll ective goal, i .e. t o ground the ir be liefs abou t a meaning they are actively co-cons tructing. We finally ne ed to poi nt out that groundin g probl ems can also occur at any poin t during a contributi on v i a sentence parts: a n installment or a tri al constituent ma y not be correctly he ard, noticed or underst ood by t he inte rlocut or; or a com pletion ma y not be considered accepta bl e by the i nitial speaker. In a ny of these c a ses a ref ashi oning side sequence needs to be initiated. The non-accepta bl e pa rt of the c ontribution will be then repaired , exp a nded or replaced unt il a mutually acceptable version i s reached. CHAPTE R 1: Literature re view - 91 - 1.2.2.5. Failure s in the grounding proces s In all t he ex amples analyzed so f ar both the spe aker a nd the a ddressee a ct a nd collaborate a s expec ted a t both the presentat ion a nd t he acceptance phases. I t is precisel y thi s col l aborati on wha t mak es possible that, de spite the di fficultie s enc ountered in presenting a nd accept ing, i nterlocutors c an safely buil d ne w common ground; that is, communica tion of the intended content can be successfully ac hieved. If, however, at any poi nt during the contribution one of the participants doe s not collaborate as expected, grounding fa ilures may oc cur that c an pre vent the satisfact ory communication of the speaker’s message. Cla rk and Wilkes-Gibbs (198 6) have identified a t least three different types of non -collaborative beh avior that m ay lead to failed or unsuccessful c ommunication: ignoring , f eigni ng un derstanding a nd misunders tanding . 1.2.2.5.1. Ignoring Sometim es a ddressee s simply ignore the ir conversati onal partne rs’ pr esentations. The speaker ut ters a pre s e ntation, but the a ddressee does not pay at tention to it a nd goes on uttering a next t urn that is not relate d to t he previ ous one . There is no gr oundi ng process and the refor e no c ommon g rou nd is built. In t he f ollowing exa mple (Clark and Scha efer, 1989: 276), B a sks a question: ‘what wh at’s his name then Charlie’. An answer is the n expected from the interlocutor. Instea d, C utters a ne w prese ntation that is no t rele vant t o t he j ust uttered que stion. There is no evidence of C’s understa nding and no grounding. Th e qu estion ha s just been ignor e d. EXAMP LE 1.13: C: Well . I’ve got um . a bo y e x G ordonstoun - . B: I sa y C: Who sticks out like a s ore thumb B: What w hat’s hi s name the n C h ar lie - - C: And I’ve got . several fl ower pe ople B: Ooh uh tha- that ’s nice. Not onl y do address ees ignore spe aker s’ presentations, t hese may also ig nore addresse es’ ac ceptance moves. They ma y ignore ind ications of troubling i n understa nding in order to av oid a refa shioning proc ess . In ei th er case, the content of the CHAPTE R 1: Literature re vi e w - 98 - can a s s ist t heir L 2 le arning, an d on w hat aspe cts and types of inter action re searchers should foc us the ir a ttention i n orde r to enhance our und erstan ding of the language lea rning pr oce ss. Most of t he t heoretical and e mpirical work carried ou t s o fa r h as e mphasized the role of i nteraction as a prov ide r of input, arguing not only tha t in inte raction learners obtain t he input they need f or thei r IL system to deve lop, but also that th is input works better for acquisition than other f orms of l inguistic e vidence. This research a pproach derived originally f rom Kra shen’s c omprehensible input hy pothes is. Krashen c laimed that ac quisition would unco nsciously occu r through expos ure to compre hensible i nput, i.e. inp ut t hat is unde rstood by the le arner ( Krashen, 1980, 1982, 1985, 19 94). Although most sc holars now a gree tha t positive e vidence is ne cessary though ins ufficie nt f or SLA, K rashe n’s hypothe sis opened a ri c h line o f research t ryi ng t o find out how new input ca n be rendered comprehe nsible and in what wa ys interac tion may be a source of comprehe nsible input. In the e arly 80s Long proposed tha t, although input ma y be made comprehe nsible in a varie ty of different way s –e .g. modifie d in put in fore ig ne r ta lk, the most optimal cond ition for compr ehensible input to be obtai ned i s modified inte raction (L on g, 1980, 1981, 19 83a, 1983b, 1985). Modified int erac tion, u nderstood a s a modif i cation of the structure of the conve rsation versus th e modif ication of input occ urring in f oreigner t alk, occu rs through negotiati on of meaning . In negotiation of mea ni ng le arners and their interlocutors work to sol ve ant icipated, perceived or expe ri e nced diffic ulties in message c omprehensibility. Through the use of ne gotiation strate gies, suc h as comprehension checks, t opic shi f ts or cla rification reque sts, a nd repeating or re phrasin g the o rig i nal me ssage, i nterlocutors a chieve c omprehensibilit y (Pica, 1994). It was t hus hypothesize d t hat negotiatio n of m eaning lead s to L2 le arning because it prov i des learners with comprehe nsible input. Following t his a ssumption, ne got iati on of meaning beca me one of t he most active area s of research in the f ield of SLA durin g the 80s a nd most of t he 90s. Evi dence was f ound t hat negotiation l eads t o greate r com prehensibility of input a nd al so of how negotiation works a nd w hic h inte ractional c ontexts or factors f oster m ore n e gotiation of mea ni ng –c f. Pica (1994) f or a de taile d revie w of negotia tion of m eaning rese arch carried out during the 80s a nd the e arly 90s. However, a s Swain (2000) arg ues, it is not CHAPTE R 1: Literature re view - 99 - yet c lear to what e xtent this greater comprehensibility of inp ut ac hieved through negotiation may lead t o L2 learning 27 or in which way thi s may occ ur, i.e. what a r e th e mec hanisms or proc edure s t hat may acco unt for the tra nsformation of comprehe nsibl e input into L 2 knowle dge a nd use. In 1985 S wain f irst ca lled for a broade ning in t he foc us of research to pa y attention not only to input bu t a lso to ou tput. She argued tha t both input and out put ha ve a role to pl ay in L2 le arning and tha t, the refore, inte raction and neg otiat ion of me aning should be examined not only as a sourc e of input but a lso a s an opportunity to use the TL, i.e. an opportunit y for output. Swain’s a rgum e nts stemmed from re s earc h carried out with French immersi on student s revealing that, a fte r ye ars of compre he ns ible input exposure, the se lea rners’ productive use of language de viates not i c eably fr om t hat of the NS. Obse rvation of immersi on c lasses re ve aled t hat the se stude nts have l ittle opportunit ies to e ngage i n exte nded discourse and they are generally not pus hed by their tea c hers to produc e accurate as well as fluent T L . These insight s cha llenged Krashen’s hypothes is, suggesting that gr amma tical accuracy cannot always deve lop from input a lone. Swain suggested then t he need of pushed output for th e learner’s IL s ystem t o de velop towards the TL norm. Pushed output occ urs whe n the learner is “pushed toward t he d eli v ery of a message t hat is no t onl y conveyed, but that is conveyed p recisely, c oherently, and appropriately” (S w ain, 1985: 249). The basis of w ha t is now known as the comprehens ible output hypothesis li es in the belie f t hat outpu t producti on requi res deeper proce ssi ng than input comprehe nsion: “output may stimul ate learners to m ove from the s emantic, open-ended proces sing prevale nt in c omprehension t o the complete grammatical processing needed f or accurate production” (Swain, 2000: 99). In other words, while a m essage can be c omprehended with little syntac tic a naly sis , i.e. drawin g m ostly on sem antic know ledge, p roducing comprehe nsible output e ngag e s learners in morpho-sy ntac tic processing, a nd this is assumed t o foster acquisition. 27 Ver y few studies have been able t o es tablish a direct c ausal l ink bet ween ne gotiation a nd acq uisition, some n oticeable although l imite d i n scope e xceptions a re Ellis, Tanaka a nd Yamaza ki (1994), and Macke y (199 5). CHAPTE R 1: Literature re vi e w - 100 - The com prehensible output hypothesi s was rev ise d in 19 95 by Swain herse l f, who su ggested then that output m ay pla y three ma in f unctions i n L2 l earning: a notici ng or conscio usness raising, a hypothesis testing and a metalinguistic functi on. Notici ng as a result of output oc curs when, in a n at tempt to produce T L , lea rners find out that th ey do not know how to expres s their mess age in a c oher e nt a nd a ccurate way. L e arne rs notice a hol e in their IL s ystem, i.e . become a ware of what the y ne ed to know but do not know yet about the TL –for evidenc e of this func tion of output see Swain and L a pkin (1995), Kowal a nd Swain (1997) and, in general, CS researc h. Consc ious of this problem learners try to solve it, e ngaging in cognitive processe s, such as ex ten ding L1 knowl edge to L2 c ontexts, ext ending L 2 know ledge to new L2 c ontexts, or f ormulati ng a nd testing hy pot heses. One way t o test a hy pothes is is to produce language. Eviden ce has bee n f ound that lea rners use output as a way to t ry out ne w langua ge forms a nd structures, i.e. t o see what works an d what do es not. The feedbac k obtained serves th em to confirm o r disconfirm their hypotheses, lea ding , whenever ne cessary, to a modif i cat ion of their initial output –empirical research, suc h as for instance Pica, Holliday, Lewis and Morgenthal er (1989), has dem onstrated that in negotiation of meaning learners modif y thei r out pu t in re sponse to their i nterlocutors ’ feedbac k. Finall y, output may serve a m etalinguis tic function, whe n learners consciously reflect on the form of their own TL use. Swain argues tha t th is reflection pla ys a maj or role in the proce ss of c o ntrol a nd interna lization of li ngui stic knowledge (Swain, 1997: 119). In the last decade, recognizing tha t linguistic accuracy c ould not be expected to develop onl y from com prehensible input exposure, researche rs ha ve broadene d considerabl y the focus of the ir approac h to interaction. Phenomena suc h as noticing , attenti on to form and feedback have thus become a major object o f study i n the field o f SLA. It is now g e nera lly ag r ee d that fo r i nput, particularly g ram ma tical input, to bec ome intake some noticing or c onscious a ttention to f orm must occur –c f. Schm idt and Frota (1986 ), Schmidt (1990, 1994, 1995), Gass (1990), or Gas s and Varo nis (1994). Learners ne ed to notice TL f orms in the input t hey rec eiv e (Ellis , 1994), gaps or diffe r ences between their T L use and that of the NS (Schmidt a nd Frota, 1986), a nd, as CHAPTE R 1: Literature re view - 101 - mentioned above, holes in their IL system. Resea rchers acknowledge in th is way that some ki nd of intervention or guidance may be necessary for grammat ical accuracy to develop out of i nteraction; th a t is, t o draw learners’ attention to form whil e communicating me aning. I n this line ne gative e vide nce , i.e. c orrective feedbac k offered by the in t erlocut or in response to a learner’s error, h as be en and continues to be the object of much em pirical res earch. Feedbac k obta ined through int erac tion d raws the lea rn e r’s attent ion t o form, while providin g t hem with T L models of thei r e rroneous IL utterances. It f acilitates modified output , i.e. th e lea rner’s modif i cation of t heir init i al output towa rds the TL norm, and subsequent L2 developm ent. Long’s input hypothesis was revised to acknowledge the functions not only of input, bu t als o of out put a nd feedback in L 2 le arning –cf. L ong (1996). Foc us on f orm models have bee n de vised and tested –cf. Lightbown an d Spada (19 90), Ly ste r a nd Ranta (1997), Doughty a nd W illiams (1998), Long and R obinson (1998), a nd L on g, Inagaki a nd Ortega (1998 ); an d, a s j ust menti one d, much researc h has a ccumulated on feedbac k –cf. Aljaafreh and L antolf ( 1994), Long et al . (1998), Mackey, Ga ss and McDonough (2000), and Morris (2002). Still, in all these theories and models interaction is primarily seen as a source of input a nd an opportunity f or output . Focu s on form and feedbac k are rele vant for L 2 acq uisition in so far a s t hey se rve to draw lea rners’ atte ntion to t he na ture of interacti onal input and/or output. This wi despread approach to the study of L2 inte raction ref lec ts a t heoretical orientation ba sed on the message model of communication (Donato, 1 994: 34; Swain and Lapkin, 1998: 320). F r om this perspective, comm unication is s een as m essage transmis sion and reception, as th e suc cessful s ending and receiving of linguistic tokens. Input and outp ut a re thus conc eived a s im ages of messages that are “transmitted as output from one source and rece ived as input elsewhere” (S wa in a nd Lapkin, 1998: 18). As Donato (1994), f ollowing Savignon (1991), point s out, in this view meaning is seen as a fixed and immutable construc t, to be s ent and re ceived. There is t herefore no pl ace for th e collaborative nature of the meaning c reation process we argued fo r in th e previous sec tion of the present chapter. Tarone (2003) argues als o tha t in i nput-output mod els of L 2 int erac t ion a nd lea rning (Van Patten, 1996; Gas s, 1997), a s well a s in f ocus on f orm (Doughty a nd CHAPTE R 1: Literature re vi e w - 102 - William s, 1998; L ong and R obinson, 1998; Long e t al ., 1998) or connec tionist models (Ellis, 2002), the l earner is seen as an automatic and logical language processor. L i ke a computer, the learner’s brain receives da ta, i.e. c omprehensible i nput or inp ut as feedbac k, analyzes it, incorporates it t o e xisting knowle dge, a nd uses it to generate output. I n thi s view, learners are assu med to be rather passive a nd t he social context is believed to play no significa nt role i n the language le arning process. Howev e r, Ta rone has provided evidenc e that in interaction learne rs manipulate the input they receive a nd stretc h their IL kno wledge, i.e . they a re ac tiv e and c reative; a nd the social context a nd the social relationship s exist ing bet ween the lea rner a nd their interlocutors affe ct L2 processi ng a nd learning in significa nt ways –cf . Tarone (2003). On thi s basis she argues that a n “adequate theory of SLA must inc lude the elements of soc ial context, sociopsychologica l process , and creat iv e choic e on the pa rt of the le arner” (Tarone, 2003: 110). Some more recent approaches to t he study of interaction have tried to a c count for t he s ocial factors inf luenci ng l anguage acquisition and t o t ake into c onsideration th e lea rn e r as both s ocial and ac tive in t his learning pr ocess . In this alternative a pproach to interacti on Tarone (2003) loc ates those t heorie s tha t try to expla in SLA providing sociol ing uistic models ( Preston, 2000, 200 2; Tar one, 2000 ), c onversation analysis models (Markee, 2000, 2004; B uckwalte r, 2001) and Vygotsk y an mo del s (L antolf, 2000; Swain, 2 000). The c oncept of collaborati ve dialogue (Swain, 1997, 2000) appea rs thus a s an alternati ve to the traditional study of in teraction foc used a lmost e xclusively on negotiation of mea ning exc hanges and feedback responses . Swain, in line with K ram sch (1995) a nd van Lier (2000) a mong others, attempts to go beyond the a na lysis of input and output in order “t o broa den the perspecti ve to one in which a ll social act ivity forms a part o f the l earning environment” (Swain, 2000: 99). In Swai n’s (1997, 2000) app roach to interact ion, lan guag e use is seen not only as communication but also as c ognitiv e activity, sociall y c onstructed through interacti on. That i s, out put ha s bot h a c omm uni cative and a c ognitive func tion. Swai n suggests tha t lea rners and the ir inte rlocutors use l anguage as a psychological tool t o solve proble ms, co-construct knowledge and thus aid L2 le arning. This means that language deve lopment does not only oc cur a s an i nternal mental process hidden in the CHAPTE R 1: Literature re view - 103 - hea ds of individuals as they proce ss rece iv e d input or f ee dback, it occ urs a lso thro ugh social inte r action and is av a ila ble in the soc ial and c ogni tive acti vity the l earner engages in wh en inte racting with other learners or wit h native spe aking i nterlocutors –c f. Donato (1994: 35). That is, instead of analyzing i nteract ion only a s comm unication, i. e. a s input reception and o utput producti on, Swain proposes to l ook a t it al so as cognitive and social act ivity . Ce rtainly, not all interact ion can be considered cog ni tive acti vity aiding L2 developme nt. Swai n distinguishe s betwee n the c onversational di a logue t ypical of negotiation of m eaning tasks , where la nguage is used f or communica tive purpose s, and collaborati ve dialogue, whe re the comm unicative and cogniti ve functions of la nguag e co-occur (Swain, 2000: 107). W ithout d enying the value of ne gotiation o f meaning studie s, Swai n claims that attenti on s hould a lso be pa id t o a different form of interacti on, i.e. c ollaborative dia logue, so t hat a mor e c omprehensive understandin g of how L2 learning occ urs can be achieve d. In t he f ollowing sec tion we will describe in detail the c oncept of c ollabora tiv e dialogue, illustrating it with exa mp les that will he lp t o c larify Dona to’s (1994) arguments and Swain’s (1997, 2000) t heoretical proposal. 1.3.2. COL LABORATIVE DI ALOGUE Swain’s propose d change in the a pproach to interacti on a nd SLA ha s its origin, a s already pointed out, in Vygotsky’s soc iocultura l theory of mind. The claim that dialogue can be seen a s social cognitive a ctivity lea ding to L 2 developme nt, i.e. that languag e and languag e knowledge are s o c i ally co-constructed th rough interac tion, is supporte d by the w ork of Vygotsk y (1978, 1987) and thos e s cholars who ha ve c ollaborated to articulate the sociocultural theory –cf . among others Leont’ev (1981), Wertsch (1985, 1991), Newma n, Griffin a nd Cole (1989), and Cole (1996). To understa nd the c on c ep t of collaborative dialogue and its rol e in L2 d evelopme nt we need therefore to consider first those Vy gotskian pr emises t hat underlie Swain’s theore tical fr amework. From a s ocioc ultural pe r spective, the origin of cognitive proc esses is sit ua ted in social act ivity. Socio cultural the orists argue that cognitive m ental fu nct ions s uch as reasoni ng, a ttention or vol untary memory a rise from the di alogical activities in which individuals participa te w hen engaged in so cial i nt er action. C ognitive pro cesses a re fi rst CHAPTE R 1: Literature re vi e w - 104 - socially co-c onstruc t ed and the n transf ormed into m ental act iv ity throu gh a gr adual process of int ernalization. Cognitive de velopment a rises thus as interpsy chologic al activity, a nd b ecomes intrapsychological through this internalization o r appropriation process. This means that t he origin of i ndividual kno wledge is di alogue, i .e. socially situa t ed interac tion occurring be twee n s p eakers. In other words, “hi gher psychological processe s unique t o hu mans can be acquir ed only through inte ractio n with ot hers, that is, through int erpsychological proces s e s that only late r wil l be gi n to be ca rried out indepe ndently by the i ndividual. When this happens, some of these processes los e their initial, e xternal form a nd are con verted into in t rapsychologica l processes” (Leont’ev, 1981: 56). In this view language pla ys a maj or role i n learning, ac ting as the me diating tool that makes possibl e the soc ial co-c onstruction of k nowledge through dialogue and th e transfor mation of s ocial cog nitive a ctivity into m ental one. For Vygotsky (1978, 1981), language const itutes a powerful semiotic tool. Just l ike a physica l tool, s uch as for insta nce a hammer, a llo w s us to carry out p hysica l activities w e would not be a ble to perform wi thout it, la nguage make s it possible to accompl ish mental acti vities we would not be able t o a ccomplish otherwi se –cf . Swain (2000: 103-10 4). Lan guage media tes our intera ction wi th the soci al env ironme nt a n d our physical and menta l act iv ities. L a ngua g e is the mea ns through which inte rpsychological ac tiv ity is c arried ou t a nd transfor med into intrapsychologica l functioning. In s um , S wa in’s app roa c h to t he natu re of the la nguage learning process is framed on the sociocultural theory belief that (1) cog niti ve p rocesses originate as so cial activity, ( 2) soci al co gnitive ac tivity is transform ed into m ental one through internalizati on, a nd (3) la nguage is the semiotic tool that mediates the c onstruction of social c ognitive proce ss es and t he internalization of t hese processes a nd t he kn owledg e they build. 28 Colla borative d ialogue, as de scribed in Swain (1997, 2000) and Swain and L a pkin (1998), a rises in face-to-face FL in teraction as c ollaborative pr oblem-solving activity. It occur s when le arners a nd their inte rlo c utors i dentify a proble m and work t o solve it in a j oint and collaborati ve way. For ins ta nce, in tryi ng t o produc e output the lea rn e r may not ice a h ol e in th e ir IL system, i.e . th a t th ey lack t he n ecessary TL 28 For a ful l acc ount of the soci oc ultural t heor y of min d an d how it has be en a p plied t o the ana l y s is o f SLA, see Lantolf an d Appel ( 1994), and Lantolf (2000). CHAPTE R 1: Literature re view - 105 - knowledge to express the me aning the y had in mind in an accurate and a ppropriate way, and try to seek fo r a solution to this pr oblem with the collaborative he lp of t heir interlocutor. With thi s ob jec tive in mind learner and in te rlocuto r e ngage in co gnitive processe s such a s f ormulati ng and testing hypot heses, extending L1 knowle dge to L2 conte xts or L2 knowle dge to new L2 co ntexts , off ering and assessing a lternatives, e tc. These cogniti ve proc esses a r e carrie d out through language use . We have alrea dy explaine d how, for instance, learners produce la nguage to see what works and what does not, i. e. to test thei r cu rrent hypotheses about TL struc ture –cf. c urrent cha pter, p. 100. L a ngua g e is thus used as a psychological tool to perform cognitive activity. T hat is, language media t es c ognit ive proce ssing. These cognitive activities a re also social acti on. They are not carried out by one individual al one. They a re t he jo int activity of a l l the interlocutors c ollaborating i n the problem-sol ving ta sk, i.e. the learner and their native or non- native speaki ng interlocutor or interlocutors. They are dialogi ca lly c onstituted in s ocially situated interacti on. This mea ns that the cognitive a ctivity le adi ng to the fi nal s olution of th e problem is interpsych olog i cal acti vity, socially c o-constructed through dia logue, i.e. through langua ge use. In this acti vity there is always a process of conscious reflection on la ngua ge . In collaborati ve dialogue, versus negotiation of meaning inte raction, attention i s foc used on b oth m eaning and form. L e arners and the ir int erlocutors work to co nvey me aning and to do so with an accurate and appropriate TL f orm. Language is here the vehicle a nd the object of reflection. Interl ocutors collaborate t o co-cons tru c t language, i.e. the language they nee d to solve t he linguistic problem enc ountered. The cognitive pr ocess es the y are e ngaged in result in ne w T L a nd t he modif ication of the learner’s original output. With the help of the interlocutor the learne r becomes able to use previously unavaila ble TL gr amm atical and/or lexical struct ures. In co llaborat ive di alo gue w e ca n thus s ee how pe rforma n ce outstrips c ompetence. Following Donato (1994), Swain (1997: 123) arg ue s tha t the linguistic change that occurs in col l abo rativ e dial ogue, a s lea rners’ modify thei r ini tial output, rep resents new TL knowledge. Through the proce ss of appropriat ion hypothesi zed i n the CHAPTE R 1: Literature re vi e w - 106 - sociocult ural t heory of mind, t he learne r i s e xpected t o inte rnalize this new la nguage and to become a ble to use it in the fut ure in a n individual and i ndependent wa y. That is, what is originally soc ially co-cons tructed l anguage and la nguage knowledge bec omes individual kno wledge, i .e. the learner become s “able to use the language of ot hers (a nd the m ental processe s that interaction has const ructed)” (Swain and L apkin, 1998: 321). 29 On this basi s Swain and L apkin support Don ato’s (199 4: 39) c laim that “the f ocus [in SLA] shoul d be … on o bserving t he construction of c o-knowl edge an d ho w thi s c o- construc tion process results in lin gui stic change a mong and wi thin i ndiv i dual s during joint act ivity ” (Swain and L apkin, 1998: 321). Cognitive problem-solving proces ses r e sult thu s in L2 learning. M ediati ng problem-sol ving a ctivity, language me diates al so the c onstruction of language a nd knowledge abou t la ngua ge . Langua ge (as a s emiotic t ool) mediate s language learni ng (knowledge buil ding) –cf. Swa in (2000: 104). Since the cognitive process es leading to knowle dge building ar e mediated by language and s ocially constituted through dia logue, they are di rectly observable in interacti on –cf. Donato and Lantolf (1990: 85) for a discus sion of this claim. Dial ogue is an ena ctment of cognitive act iv ity and in dialo gue we can fo llow the c og ni tive ste ps which f orm t he ba sis of language development. W hat norm a lly remai ns hidden in individually interna lized thought, i n j oint problem-solvin g ac tivity mani fests itself in dialogue (Swain a nd Lapkin, 1998: 321). That is, collabor ative dia logue is la nguage lea rning and i s also e vidence of la nguag e learning. Whe reas in negotia tio n of me aning research we see inpu t recept ion a nd output production as possible sources for L2 lea rning, in collaborative dialogue we percei ve actual L 2 learning ta king pl ace. In sum, collaborative dialogue i s “ knowledge-building dialogue. In th e ca se of our int erests in sec ond lan guag e l earning, it is dialogue that c onstructs linguistic knowledge. It is wha t allows perf ormance t o outstrip c ompe t ence. It is where la nguag e use a nd language learnin g ca n c o-occur. It i s language use mediatin g la nguage l earning. It is cognitive a ctivity a nd it is social activity” (Swa in, 2000: 97). 29 I n the following secti on w e will see how this assum ption has bee n c onfirmed b y the r esult s of e mpirical research de monstrati ng tha t t he l angua ge str uctures c o-construc ted in social in teraction are later recognize d an d u sed b y t he learne r in i ndivi dual a ctivit y (Donat o, 199 4; LaPier re, 1994; S w ai n a nd Lapkin, 19 98). CHAPTE R 1: Literature re view - 107 - We will now exa mine some e xamples of c ollaborative dialogue that illustrate what we have jus t e xplained about the nature of this ki nd of interac tion a nd its role in the S L A proce ss. 1.3.2.1. Collaborativ e dialogue and L2 gra mmatical devel op ment The first excerpt of in terac tion w e are g oing t o discuss here comes from Kowal and Swain (1997), and has also be en include d in Swain (2000). Trying to iden ti fy how syntactic processing could be promoted in content-based immersion clas se s, K owa l a nd Swain engaged stude nts in a variety of ta sks specific ally designed to e ncourage f ocus on form while creating meaning. In thi s fi rst e xample we see how two Fre nch language lea rn e rs, ide ntified by the pse udonyms Rachel and Sophie, colla borate to rec onstruct in writing a text they hav e j ust heard. EXAMP LE 1.14: 1 Rachel: Cher[c hez] nou . .des nouveau x menaces. ( Lo ok u p new [as i n] ne w th re ats.) 2 Sophi e: Good one! 3 Rachel: Yeah, nou veaux, des nouveaux, de nouvea ux. Is it de s nouvea ux or de nouveaux? 4 Sophi e: Des nouvea ux or de s nouvelle s? 5 Rachel: Nou[ veaux], des n ou[veaux] , de nou[ veaux] . 6 Sophi e: It’s menace, un mena ce, u ne menace, un me nace, menac e ay a y a y ! [ exasperated]. 7 Rachel: J e v ais le pauser. (I’m going to p ut it on pau se [ie the ta pe-rec o r der] .) [ Th e y look up ‘ menace’ in the dicti onary. ] 8 Sophi e: C’est des nouvelle s! [tri umphantly ] . 9 Rachel: C’est fémi nin … des nou velles mena ces. The phrase included in the original text that Rache l and Sophie are trying to recreate is ‘de nou ve a ux problèm es’ –i n Englis h ‘new probl ems’. Instea d of produci ng a verbatim re petition, these two stude nts dec ide t o refor mulate the initial sentence and t o produce a phrase that is new f or t hem. R achel substitutes the ori ginal ‘prob lèmes’ with the synonym ‘men aces’ –i.e. ‘threats’; and in turn 2 Sophie co ngratulates he r on doing so: ‘Good one!’. Ac cording to Swain (2000: 101), they a re try i ng t o s tretch their IL a nd thus m ake th e task more challenging for them. Notice a lso that this is not a req uirement of the task . In Donato’s t erms, it is a sub-goal spontane ously genera ted by the learn ers themsel ves –about sub-goals s ee Don ato (1994: 43). In produci ng this new phras e, they co me to recognize a h ole in t heir IL s yste m . Rachel is not sure wheth e r the c orre ct pa rtitive i s ‘de’ or ‘des’, and Sophie realizes that CHAPTE R 1: Literature re vi e w - 114 - it is imperceptible” (S wa in and L a pkin, 1998: 321). In the following extract of interacti on Rick and Kim’s at tention focuses again on the TL w ord ‘oreil ler’. Rick is thus given an addi tional opport unity to check his c orrec t c omprehension o f the meaning of this word, to ref lect on it and to add it to his IL s ystem. In a post-test carried out two weeks afterwa rds Rick picks up the item ‘oreiller’ as t he c orrect word f or ‘pillow’ among f our differen t choices. T hat i s, Ric k shows t o have l earned the L2 le xical item ‘oreille r’ throug h colla borative dialogue. EXAMP LE 1.17: 1 Kim: Quelque c hose uh … est s ur l’… quel que chose e st s ur l’ oreiller. (S omething…i s on the…s omethin g is on the pill ow.) 2 Rick: Is that l’o reille r? (P ointing to somethi ng in t he picture.) (Is t hat the pillow?) 3 Kim: No, this is l’orei ller. (No, this is the pi llow.) 4 Rick: P illow ? 5 Kim: Yea h , pil low’s orei ller. The fina l exce rpt of interaction we a re going to discu ss c omes al so from the da t a collec ted in Swain a nd Lapkin’s ( 1998) stud y. Here the same t wo previous lea rners, Ric k and Kim, e ngage in a language-rela ted epi sode a s t hey c hoose be tween two competing French voc abulary items : ‘s’en aller’ and ‘march er’. In orde r t o expres s the notion of walking to s chool Rick proposes to u se the French verb ‘s’en alle r’, but Kim suggest s substituting it for ‘march er’. The two lear ners agree on t he second option, ‘marche r’, and they use this for m in their c ollabora tive writing of the s tory . Howeve r, the sec ond alternative is a di rect transfe r from the Eng lish verb ‘to walk’ and is not the cor rect Fr ench form . A French NS w ould have used in this context the verb ‘s’en aller’. EXAMP LE 1.18: 1 Kim: [ Elle voit un] gars (She sees a gu y ) 2 Rick: … g a rs, qui s’en va à l’éc o l e (gu y , who is going t o school) 3 Kim: qui mar che vers l’ é cole… marche (wh o is walking t owards sc hool, walkin g) Ric k’s initial utterance poses a problem, i.e. t he ne ed to choose betwee n ‘s ’en aller’ and ‘marc her’. Rick a nd Kim col laborate t o solve this probl e m agreeing to select the se cond f orm, ‘marche r’, a s the accurate mea ns of e xpression to convey thei r intended message. The initia l utterance is modified and a new TL e xpression is c o- CHAPTE R 1: Literature re view - 115 - construc ted with the collaborative eff ort of the two le arners. This modif ied output represe nts th e two learners’ current state of IL develop ment. But in thi s p rocess Rick and Kim rea ch a wrong soluti on. They co-constr uct wrong TL knowledge. This kind of interacti on gives rise t o an important pedagogical issue. As Swa in and L apkin (1998: 333) indicate, in c ollaborative tasks s ome kin d of teacher feedba ck is nee ded. Without it students ma y l earn the wrong grammati cal structure or le xical item. This mean s t hat, althoug h c ollaborative dia logue se ems to have an important learning potential to be profit ed from in the F L c lassroom , the teacher has also a signific ant role to pla y i n the process of L2 developme nt through le arner-learner interaction. 1.3.3. EM PIRICAL RESEARCH ON SLA F ROM A SO CIOCULTURAL PERSPECTIVE: RE SEARCH AND PE DAGOGICAL IMPLI CATIONS The value of c ollaborativ e dia logue for both the resea rch and the teaching p ractice is based on two major claims: firstl y, the s ocial c onstruction of language and la nguage knowledge i s mediated by langua ge, e nacted in lea rn er’s dialogue an d thus dire ctly observable for the researcher; and, s ec ondly, collective cogniti ve pro cesses a nd the linguistic knowledge these processes build are i nternalized by the lea rner a nd transfor med into individual me ntal re sources f or f uture i ndependent TL u se. Empirical research on SL A ca rried out f rom the pe rspective of the soci ocultural theory of mind has prov ided evidence t hat supports thes e two claims and Swain’s (1997, 2000) theoretica l perspective on interaction and S L A. The e xamples of c ollaborative dialogue so far di scussed s how tha t learners use language as a m ediating tool to perf orm problem-solving c ognitive proce sses a nd that, when these processes are sociall y co- constructed t hro ugh i nterac tion, learners’ dialogue consti tu tes an enactme nt of cogniti ve a c tivity. T hes e examples, as already mentione d, come f rom Donato (1994), Kowal an d Swain (1997) and Swain and Lapkin (1 998) –some of them have al so been included i n Swain’s (1997, 2000) theoretical approac he s to the conc ept of collaborative dialogue. H owever, the re are other re search projects carried ou t from a sociocultu ral perspective that also demonstrate that w h en language lea rn e rs sol ve problems “through dia logic interaction with anot her person”, this interacti on offers “ a window int o intra mental pro ce ssing ” (Goss, Yi ng-Hua a nd Lantolf, 1994: 267). CHAPTE R 1: Literature re vi e w - 116 - Goss et a l . (1994) examined lea rners’ g ra mma tical judgments in an ai m to identify what kind of cognitive proce sses they represe nt. Instea d of relying on introspective data or think aloud protoc ols, as had bee n do ne in m ost previous sim ilar research, these s c hola rs asked le arners’ dyads to deci de jointly on the gr ammaticality of a set of sentences. The dia logue they engaged in proved to be a n enactment of the ir cognitive proce ssing, which involved language mediated activities, such a s making translations or making ex plicit t heir meta knowledge through language use. That is, ana lyz i ng the d ialogic interacti on of learners in t his kind of collaborative prob lem- solving task, Goss et al . (1994) were able to observe the c ogni tive p roces ses that led these learne rs t o dec ide on t he gramm aticality of t he s entences proposed. Fur therm ore, they als o f ound similar respo nse patterns when th e ta sk was accomplished by dyads a nd by indi viduals working a lone. These res ults s uggest t hat t he me nt a l processes observable in j oint activity are sim ilar to those at work in indi vidual cognitive ac tivity. Goss et al .’s ( 1994) study doe s not onl y provide addi tional e vidence on collaborati ve di alogue as an enact ment of cognitive proce sses, it also directly sug gests that this kind of dialogue coul d an d should be used as a research t ool. It see ms t o offer a more direct source o f mental data than int rospectiv e techn iques. 32 In Swa in and L a pkin’s own word s col laborative d i alogue offe rs “a w i ndow in t o intra mental processi ng … more tra nsparent than the window provided through introspective tec hni ques” (Swain and L a pkin, 1998: 322). A simi lar c laim i s a lso ma de in Dona to a nd L a ntolf (1990), Donato (1994) and Swain (1997, 2000). These scholars do not deny the val ue of introspective techni ques. In fact, Swain and Lapkin e xplicitly a sk f or researc h combining t he a nalysis of collabo rativ e dialogue with retrospect ive da ta obtaine d through fol low-up interviews, a rg uing tha t thi s may provi de “ a more fine-grain ed und e rstanding of … mental p rocesses” (Swai n a nd L a pkin, 1998: 333). Collaborative dialogue i s ra the r posite d as an addi tional a nd compleme ntary too l for the study of m ental processes. The s e cond cla im conc erns the i s s ue of whet her the dialogically c onstitut ed cognitive processes evident in learne rs’ interacti on aid or fa vor L2 lea rning , i .e. whethe r 32 The pr oblems and limitation s of introspecti ve tec hnique s ha ve bee n widel y di scussed in t he fi eld of SLA resea rch. This issue will also be treat ed i n m ore detail in Chapter Thr ee of the pres ent s tudy. F or a discussion of the proble ms exi sting f o r obs ervin g mental pr ocesses relate d to S LA the re ader c an also see Donato an d Lantolf ( 1990). CHAPTE R 1: Literature re view - 117 - collaborati ve dialogue consti tutes an occasion for L 2 learni ng leading to future indepe ndent L2 p erformanc e. Evidence di r ec t ly supporting this c laim was fo und in at least four different e mpirical re search projects prev i ous t o Swai n’s (200 0) t heore tical account o f col l aborati ve dialogue and its role in SL A: Holunga (1994), La Pierre (1994), and the alr ea dy mentioned Donato (1994) and Swain and Lapkin (1998). As Sw a in (2000: 110) e xplains, t he rol e of language m ediati ng le arning is general ly accepted in o ther edu cational domains such as mathema tics, science or history –for some empirical research on this iss ue see Ta ly zina (1981) or Newman et al . (1989). However, the role of la nguage media ting language le arning seem s to be less well understood. She argues that maybe this is so beca use in L 2 de velopme nt language is both the vehicle and the product of learning. Its function a s a media ting too l is subseque ntly more di fficult to conceptualize and also less clear to observ e . In this line, Hol unga (1994) d esigned a study on the relationship between metac ognitiv e strategy instructi on and oral accuracy of ve rb f orms tha t e xplicitly tested the role of language me diating s trate gy use and it s effe cts on gramma tical developme nt. Holunga com pare d t h e performance o f three different groups of stude nts as they ca rried out a communicative task in pairs: a group who had not been instructed in st rate gy use ; a second group who h a d received instructi on on me tacognitive strategies, suc h as predicting, p lanning, m onitoring, a nd eva luating; and a third group who had rece iv ed the same instruction and wa s also asked to verbalize their strategy use a s they pe rformed the comm unicative task at hand. Analyzin g le arners’ interac tions during this task Holunga (1994) f ound that the dialogue of the sec ond group of stude nts wa s t ypical of ne gotiation of m eaning ta sks, i.e. i t was f ocused on meaning rather than on f orm. H owever, the talk of t he third group displa yed a tt entio n t o both content and f orm. These lea rners e xternalized hypothese s , tested them a nd rea ched rea soned solutions. They point ed out ea ch others’ errors, provi de d e xplanations for their i naccuracies and modified t heir out put whenever necessary. They did not onl y verbalize the verb form they neede d, but also why the y nee ded this form and not other . In gene ral, their dialogue reveal s th at they pred icted, planne d, m onitored and evaluate d their us e of verb fo rms. The se strategi c proces ses were c o- const ructed in interactio n and mediate d by l anguage. L an guage was the f ocus of their a ttention and the means by which the y performe d the task. CHAPTE R 1: Literature re vi e w - 118 - Holunga als o asked her students t o carry out a pre- and a post-tes t on verb forms. The results of these t e sts reveal that students w ho had received instruction on metac ognitiv e stra tegies ma de significant gains on the accuracy of verb forms whe n compa r ed to those who had no strategic instruction. But Holunga also found out that dyads who verbal ized their strategy use performed considerably bet ter and improved more than those who had not. On this ba sis she c oncluded that verbalization, i.e. exte rn a l speech, mediated not o nly t ask performance but also l a nguage learning. L e arners’ verba lization of s trategy use , i.e . the ac t of talki ng a bout what they w ere doing, a ide d t hem to co- c onstruc t c orrect grammat ical verb forms and to internalize t his co-cons tru cte d knowledge. Their dialogue was lang ua ge lea rning mediated by language use. We have alread y ex plained the main goal of Donato’s (1994) s tudy an d the methodol ogy e mployed to achieve it –cf. current c hapter, p. 109-110. This methodol ogy allowe d him to test actual le arning occ urring in sc affolded pe rformance w here both the L 1 a nd the L 2 a re used a s a cognitive too l to co-cons truct language and linguistic knowledge. Donato rela t ed l ea rn ers’ individual perf ormance i n an oral presentati on task to the s caffolding si tuations occu rring in a previous c ollaborative preparation of this presentat ion. He ide ntified 32 cases of scaffolding in t he pla nning session. 75 % of the language structures co-construc t ed in t hese s caffolds were used correc tly a nd indepe ndently in the presentation that t ook place one week af t erwa rds. This allowed him to conclude that, as s uggested in the sociocultural theory of mind theory, “L 2 developme nt i s b rought about on the social plane ”. T hat is, “ social interactions in th e classroom result in th e a ppropriation of linguistic know l edge by the individual” a nd therefore we can assert that “l earners c an e xert a developmental in f luence on eac h other’s inte rlang uage s ystem in observable ways” (Donato, 1994: 39). In LaPierre’s (1994) stud y, lea rners we re en gag e d in a dictogloss o r story reconst ru ction ta sk similar to the one used by Kowal and Swain (1997 ) –c f. present cha pter, p. 107. This researcher examined langua ge-r ela ted episodes that occurred as dyads of learner s tried to solve the problems e nco u ntere d in the writing of a previously hea rd story. L a Pierre’ s (1994) fina l aim was to te s t the hypothes is that when L 2 le arners nee d to talk a bout the language the y a re produ cing in order to pe rform a ta sk, thei r metatalk may be a source of L2 lea rning . W ith thi s obj ectiv e in mind, t he researche r CHAPTE R 1: Literature re view - 119 - designed a dyad-spe cific post-tes t that inc luded those language i tems tha t ha d bee n the object of previous language-rela t ed episodes . The results of t he post-te st, a dministered one week after the perfor manc e o f the f irst task, re veal that, for the 140 e pisodes wher e a corr ect solution to a l inguistic pr oblem had be en a ttained through c ollaborativ e dialogue, 80% of t he re lated po st-test items we re c orrect. Simila rly, whe n stud ents had co-cons tru cte d an inc orrect s olution t o the ir dif ficulties, t heir perf ormance in the p ost- test matche d t his inc orrect s olution in a pproximately 70% o f the items. That is, t he c o- construc ted knowledge, whethe r correct or incorrec t, tended to be retaine d by the lea rn e rs. In sum , conscious re flection on language through col laborative dia logue proved to be a n occasion for L2 learning. Swain a nd Lapkin’s (199 8) study is similar to L a Pierre’s (1994), although thes e two sc holars use d both pre- a nd pos t-tests. As al ready e xplaine d, they foc used t heir attention on bo th fo rm- ba sed a nd lexis-based language-relate d epis odes occurring a s dyads of lea rners performe d a j igsaw task where the y had to reconstruc t i n writing the story depi cte d in a set of pictures –f or a more detailed accoun t of the design o f t his study, conduct ed in a classroom cont ext, see current chapter, p. 112. They found, f irst, s triking diffe rences be tween dya ds in their p erformance of t he task. Whereas some students engaged r elatively often in language-related episode s , others seemed to focus a lmost exclusively on meaning and not to c are about grammatical or lexica l acc uracy. That i s, dif ferent pair s of s tudents interpreted the same task in different wa ys. This i s not su rprising , sin ce Coughlan and Duff (1994) ha d already shown that the “same task” is not the “same task” for diffe rent students. Swain and L apkin se lected the perf ormance of one pair of learne rs, Rick and Kim, to carry o ut t heir a nalyses. Thes e yiel ded evidenc e of di a logue as a n e nactme nt of mental proce sses a nd occasio ns f or L 2 lea rning . Ric k a nd Kim used la nguage as a mea ns of c ommunication a nd as a co gnitive t ool. When i n f ace of a problem, la nguage was us ed to solve it , i.e. to c o-construct the language re quired to express the ir inte nded mea ni ng with gramm atical and le xic a l accuracy. Through lan guag e use , both L 2 and L 1, they c onsciously si ngl ed out the L 2 a s an object of reflection and m anipulation, generat ed alte rn ative s and hypothese s, a ssessed or t ested t hem, and a pplied the re sulting knowledge to the solution of the problems e ncountere d. These socially co-cons tructed cognitive proc esses, mediated by language , were ma nif est in their dialogue. Thei r CHAPTE R 1: Literature re vi e w - 120 - outcome wa s co-constructed language a nd co-c onstructed language knowledge evident in their modificat ions of ini tially erroneous output, i.e. in linguistic change. In t he post- test the se learners showed to hav e retained th e correct a nd the i ncorrect kn owledge they had c o-constructed in language-rela ted ep isodes. Their c ol laborative dialog ue, i.e . their use of la nguage as communicati on and c ognitive ac tivity, served both as a tool to ai d L2 lea rning and as a window for the r esearchers to observe actual learning taking place . Taken all together, the re sults of these diffe rent research project s confirm that “la nguage learni ng occurs in c ollaborative di alogue, and t hat this external speech facilitates the a ppropriation of both st r ategic proce sses and l i nguistic knowledge” (Sw a in, 2000: 113 ). T his conclusion bears important research a nd ped agog ical implications. In order to a chieve a f ull und erstanding o f col laborative dialogue and i ts role in SLA i t seems t hat, as claimed b y Donat o (1994) and Swain a nd Lapkin (1998), “the focus [ in SLA] s hould be … on ob serving t he construction of c o-knowl edge a nd how this co -construction process results in linguistic change a mong and within indi viduals during join t ac tivity” (Dona to, 1994: 39 ). This means t hat interactionist research needs tasks that e ncourage these types o f processe s. That is, since in com municative t asks “atte ntion is principally focused on mea ning rather than form” (Nunan, 19 89: 10), more attention needs to be paid to c oll a borativ e t asks t hat, while still oriented to meaning communication, e ncourag e lea rners t o pay als o consci ous attention to l ang ua ge form in social interac tion (Swain, 1997: 119). Colla borative ta sks, as the ones used for the study of langua ge-related episodes and collabora tive d ialogue –cf. Kowal and Swa in (1997), or Swain and Lapkin (1998)– can be considered co mmunicative t as ks too, i n the se nse tha t they enga ge two interlocutors in mea ning c ommunication activities. These tasks, how ever, require also a joint final product: a writte n text or a n oral presentati on. The need to produce a final written or ora l text in the TL, wit h the highest p ossi ble degree of a ppropriateness a nd accuracy, is what e ncourages learne rs t o f ocus n ot only on meaning but al so on f orm, and to c ollaborate to create bot h meaning and TL structure s. From a pedagogical perspec t ive, it seems that th e use of collaborative tasks i n the FL c lassroom can encourage learne rs to e ngag e in a c ollaborative dia log u e that ha s proved to be pa rticularly hel pful f or thei r l ea rning of strategic proc esses, a nd CHAPTE R 1: Literature re view - 121 - grammatical and lexica l aspect s of the TL. In fa ct, the need to focus lea rn ers’ attenti on on both mea ning and f orm, and th e value o f collabor ative tas ks to push learn ers’ conscious reflec tion on the language the y use to comm unicate meaning a re now widely recognized among re searche rs interes t ed in pedagog ica l tasks –cf. Bygate , Ske han a nd Swain (2001). In rece nt years, foll owing Swain’s ( 2000) the oretical cla ims and proposal s, a considerable amount of e mpirical st udi e s have been conducte d on c ollaborative dialogue –cf . among others Swain (200 1), S wain and Lapkin ( 2001), L a pkin, Swai n and Smith (2002), L e e (2004) or Leese r (2004). Ne w ins ights are st ill requi re d in order to cla ri f y why, in the pe rformance of the same task, s ome stude nts engage m ore often in collaborati ve dialogue than others, why the y do so in ce rtain s ituations but not in others, or wh y this dia logue result s often, but not a lways, in indi vidual knowl edge. The literature here reviewed sugges ts the need of further work to give a nswer to t hes e still open i ssu es, but al so t he sui tability of more re se arch c arried out withi n this t heoretical framework. EMPIRIC AL STUD Y CHAPTE R 2: The object of st udy - 130 - be paid to the j oint act ions of spea ker a nd addressee, i .e. to strategic comm unicati on understood a s a mutual effort of two interlocutors to a gr ee on meaning. This explains t hat m ost previous res ea r c h, foc used on t he FL lea rner’s utte rances and mental processes, has provided ma ny insi ghts on the c ognitive a nd ling uisti c n ature of CSs . Nevertheless, as als o a rgued in the introduct ion, it has not been able t o accoun t for the role of C Ss in the strate gic communication process an d how c om munica tion of mea ni ng is ac tually achieved. We consi der, however, tha t an interactional definition of CSs s hould c ertainl y ack nowledge a more active role in the inte rlocutor’s part. On this basis and building on Tarone’s (1981) description, we have ela bor a ted our own working defi ni tion of CSs . This theoretical acc ount of the concep t of CS intends to clarif y the object of st udy of th e present investi gation and a lso t o s erve a s a guide for t he identi fi cation o f CS uses in FL interacti on, as a first step in the a nalysis and description of strategic communication. We consider that a CS is us ed in f ace-t o- face F L i nterac tion when: 1. The lea rner desires t o c ommunicate a n intended meaning to the interl ocutor. 2. The learner (and his /her inte rlo c utor) bel ieves the lex ic a l item des ir ed t o comm unicate this m eaning is unavailabl e or is not sha red with his/her interl ocutor. 3. The lea rner (and his/her interl ocutor) chooses t o (a) avoid, that is, t he l earner doe s not a ttempt to c ommunicate the int ended mea ni ng, or (b) try out an alternative means of expressio n to communica te the inte nd mea ni ng. The f irst tw o conditions esta blish CSs as problem-sol ving techni ques. CSs are used a s tools in order to overcom e the problem that ari ses whe n, in a n attem pt to communicate meaning in fac e-to-face oral inte raction, the TL l exical i t em o rig i nall y desire d to convey this meaning can not be used. Notice that , a s e xplained in the introduction c hap t er, in FL inte raction learners may face a w ide variety of c ommunicative difficult i es dir ec t ly resul tin g from their inade quate c ommand of the TL . T hey ge nerally ne ed m ore time than a NS in order to produce a n u tterance and they als o e ncounter r elatively fr equent c omprehensibilit y diffic ulti e s. To overc ome these problems t he y use produc tion or conversational CHAPTE R 2: The object of st udy - 131 - strate gies, su ch as fillers or repetit ions, that a llow them to gain time; and negotiation of mea ni ng strat egies, such a s c onfirmation checks or clari fication requests. Thes e diffe r ent phenome na occur, re latively of te n, in c ombinat ion wit h C Ss, but t hey c annot be considered as CSs in themselves. They have thus been left out of the definiti on o f CSs; although, in the analysis of strate gic interaction, we will see that, quite often, they are part of the overall strategic com munication of meaning process . The sec ond con di tion make s i t e xplicit, in line with Tarone (1981), that le xical diffic ulti e s may be t he re sult of a learner’s or a n i nterlocutor’s def icit in their IL system. Sometim es a l earner is unable to make use of a de sired word or express ion be cause it has not yet been a cquired. Tha t i s, the d esired TL item i s not ready f or u se because of a lea rn e r’s I L gap. On oth er occ asions, t his item is pa rt of the le arner’s IL s ystem but, because of the time pr essure of ongoing inte ract ion, it ca nnot be retrieved at a specifi c moment , i .e. it is u navailable because of a reca ll diffic ulty . In f ace -to- f ace i nterac tion, it may als o ha ppen that this T L ite m is a ccessible in the speaker’s IL syste m, but it cannot be used i n orde r to establish a n ag re eme nt on meaning with their interlocutor becau s e it is not yet pa rt of the latter’s k nowledge of t he TL , i .e. because it is not share d with the interlocutor . We b elieve that the se t hree different source s of dif ficulty pose t he same kind of problem and, subseque ntly, elicit th e same kind of strate gi c beha vio r. In h er desc ription of C S s Tarone makes ref e r e nce to li nguist i c a nd sociol ing uistic difficulties, but in the definition here proposed the scope of CSs ha s been limited to lexical problems. This doe s not mea n, howe ver, that we consider t hat CSs nee d to b e ne cessa ri ly rela ted to l exical difficult i es. L ike m ost previous researchers in the f ield we defend th a t CSs c an also be use d to deal wi th com municative problems resulting f rom phonological, s yntactical or pra gmatic def icits in t he FL l ea rner’s I L system. How ever, the pre sent invest igation is going to be re stric ted to the study of lexica l difficulties and, subseque ntly, the working de finition of CSs has been devised to meet thi s l exic al restriction. The s cope of the investigation needs to be somehow narrowed dow n. It i s not possible to deal, wit hin the l imits of thi s kind of re s ea r ch project, with eve ry sort of linguistic and s ociolinguistic difficu lty a lea rn er may face i n FL interac ti on. We hav e CHAPTE R 2: The object of st udy - 132 - dec ided to limit the scope o f our research by f ocu sing on lexical CSs, 33 in part, becaus e lexis see ms to play a cent ral role i n c ommunication. As Poulisse et al . po inted out “without knowledge of (at least a few of) the words of a language i t is impossible to communicate in it, however l arge one’s kno wle dge of the gramma r of tha t language may be” (Poulisse et al ., 1990: 12). Other rese archers have also argued that , in g ener al, lexica l CSs a re used m ore ofte n than oth e r kin ds of CSs, not only in FL interact ions bu t also in L 1 c ommunic ation –cf . Kellerman (19 91), Kasper a nd Kelle rman (1997) or Dörnyei and K ormos (1998). W e postulate, on these grounds, that their study ca n offer more relevant ins ights for our understandi ng of FL communi cation. On the other hand, Chapter O ne, section 1.1, has c learly shown that m ost previous resea rch adopted a similar focus. L ex ical CSs have receive d muc h mor e attention in the field of SLA r esearc h that any other form of CSs a nd, conse quently, their study is highly develope d. This pro j ect inte nds to build on these previous investigations. W e plan to study s trategic comm unication analyzing FL le arners’ interacti ons as we ll a s t heir retrospecti ve c om ments on these interac tions. Such a methodol ogical fra mewo rk s eems to benefit f ro m a focus on lexis. Some re s earc hers have fo und that learners are a ble to provide more e xtensive comments on the l exical problems t hey e ncounter and t he st rategies the y use to solve them t han on grammatical or phonol ogical difficulties (Gla hn, 1980 ; Poulisse et al ., 1990). T hese latter f orms of communicative d ifficult i es certainly de serve attention i n the field of S L A , but t aking into account the curre nt s tate of th e researc h we c onsider that a parti cular emphasi s on lexica l problems is more advanta geous for the specific purposes of the present study. In thi s workin g defi nition CSs are de sc rib ed as probl em-solving devic es, but they ar e als o c onc eptualized as c onscious strategies. I n this sense we disagree with Tarone (1981) who rejected the c onsciousne ss crite rion –cf. C hapter One, p. 25. We believe th at for a C S to be used, the lea rner, and possibly the interlocutor as well, need to “believe” , that is, t o be somehow a w a re, t hat a le xi ca l proble m has been e ncountered. In othe r words, we t hink, like Færch and Kaspe r (1983b), t hat pro blematic ity al ways implies a c ertain de gree of c onsciousness –cf. C hapter One, p. 27. A communicative problem is e xperienced w hen one or both of the i nt e rlo c utors become more or less 33 Note that t he e xpression ‘l exical CSs’ is used t o make ref erence to CSs used to solve le xical difficulties, n ot to the nature of the CS itsel f. CHAPTE R 2: The object of st udy - 133 - conscious –unde rstanding consci ousnes s as a gradua l phe nomenon– that a difficul ty has arise n in their attempt to reach a co mmunicative goal. In fact , as seen in Chapter One, sectio n 1.1.1, problematicity and consciousness have be en the t wo most com monly u sed c riteria in prev i ous de finitions of C Ss. Conscio usn es s has played a major role in CS res earc h. It implies tha t lea rn ers are awar e of problems an d c an, therefore, report on th e m an d on the strategies th ey use to solv e them. Retrospect ive c omments elicit ed fro m the learner on their FL pe rform ance have made it possible to dis t inguish CS uses from other relat ed IL phenom ena t hat may ha ve a sim ilar surface r ealization in interaction. N on-target-like utterances can reveal an automatic application of a n e rroneous I L rule o r t he us e of a C S. C ode switc hing, f or insta nce, occurs rela tively ofte n in FL lea rners’ spe ech. Sometimes it i s u s ed subconscious ly. It ca n also be emplo yed a s a consci ous social s trategy , f or i nstance to asse rt identity or to establis h a c loser rapport wit h the interlocut or. It may be fin ally used a s a CS to c ompensate for what the learner be lieves to be the unavailability of a TL lexica l item or s tructure. In orde r to distinguis h the se three different ki nds of phenome na the re searcher needs, relatively often, to rely on introspect ive d ata provided by the speakers t hemselves. As explained above, from an interactional approach problems arising in FL communication are mutuall y sha red probl ems. I n i nteraction spea ker and interlocutor share a c ommon goal, i.e. the successful c om m unication of meaning. Any lexi cal diffic ulty e ncountered in the a ttempt t o a chieve this goal is, s ubsequently, mutu ally shared. This does not mean, however, that both t he s peaker and the interlocutor need to be aware o f the problem and of t he use of a CS to solve it. The spea ker is al ways the one orig inal ly experienc ing the problem and th ey may raise their interl ocutor’s awareness on this proble m or not. That is, the interlocutor m ay c olla borate to e stablish a mutual agreem ent on mea ning wit hout being c onscious that t he TL struc tures u sed to convey this meani ng are not the spea ker’ s original ly inte nded ones. In Chap ter One, secti on 1.2., when re viewing the different t ypes of CSs availa ble, we c ould observ e t hat most of the m res ul t in erro neous or non-ta r get-like uses of the FL . The use, for ins tance, of a word c oi nage, a transfer or a wordy ext ended descri ption i nstead of t he appropriate and accurate T L te rm gene rally reveal that the lea rn e r is expe ri enc ing som e kind of l e xical difficul ty and is m ak ing use of an CHAPTE R 2: The object of st udy - 134 - alternati ve mean s of expression. Explici t statements, such a s ‘I don’t know how to sa y this’; di rect appe als for assistance, suc h a s ‘how do y ou say t his?’; o r problem indicators, s uc h as a n ov e ruse of pauses, f illers, re petitions, lengthening de vice s or hedges, like ‘a ki nd of’, als o act a s a signal fo r th e interloc utor suggesting tha t the speaker is c onfronting a linguistic proble m –c f. W agner and Fir th (1997). Some interacti onist sc holars, s uch as Wagner and Firth (1997) , argue tha t only when the lea rn e r signa ls to their interlocutor t hat a lexical diffi culty has bee n enc ountered by mea ns of one of th ese f lag gin g d evices a nd t he i nte rlo c utor responds t o it , ca n the researcher c onsider that a CS is be ing used: “an interactional approach de fi nes instances of talk a s CS if , and only if , the pa rticipants themse lves make i t public in the talk its e lf an enc oding-related problem a nd b y s o doing en gag e –indi vidual ly or c onjoi nt ly– in attempts to resolve the problem” (Wagner a nd Firth, 1997: 325-326). However, othe r re searchers have provided data sug gesting t hat, u nder certain conditio ns, speake rs ma y try t o c oncea l t heir lexical def icits, a nd tha t some learners, i n particular advance d level students, are sometimes a ble t o h ide the strateg i c natu re of their utterances from the i r interlocutors –cf. Fæ rch a nd K a sper (1984). On th is ba sis we believe, as m entione d a bove, t hat CSs ca n in ce rtain situa tions be used wi thout the interlocutor’s awa reness. This explains t he inclusio n o f t he interlocutor withi n parenthe ses in our working def inition of CSs: “the learner (and his/her int erlocutor) believes the l exical item de sired to com municate this m ea ning is un a vailable or is not shared wit h his/her interlocutor”. The third c riterion or condition in th is description of CSs draws a clea r distinction between two ma in types of str ategies: a voidance and achieveme n t. Following Tarone (1981) w e c onside r that, whe n a lexica l probl em occ urs i n FL communication, speake rs a nd inte rlocutors have t wo mai n option s: to c ir cumvent the lexica l de fi cit that caused this probl em by avoiding the meani ng they original ly intended to communicate, or t o compens ate f or the u na vailable TL lexic al item using a n alternati ve means of expressi on. Furthermore, we believe tha t, in face-to-face FL interaction, t his can be accomp lished by the speaker alone , or by the spea ker working wit h the inte rlo c utor; thus a gain the use of pa rentheses: “the learner (and his/her interlocut or) c hooses to avoid, tha t i s to sa y, the learner do es not at tempt to com municate the intended meaning, CHAPTE R 2: The object of st udy - 135 - or try out an al ternative mea ns of expr e ssion to c ommuni cate the intended mea ning”. In most previous research it has been assumed that the speak er originally expe riencing the problem i s the one who deve lops a nd pre s e nts a CS to solve it. We maintain, ho wever, that in interaction the interlocutor can a lso have a n active participation in t his process. The interl ocutor may infer, f rom the linguistic and soc ial context of the com municative exc hang e , the mea ning the learne r is t rying to convey and o ffer a CS to express it. It is also possible that an alter na tive mea ns of expres sion, such as a descript ion, becomes c o- construc ted by the learner and the interl ocutor; or tha t each of the two interacta nts offers diffe r ent but c omplem ent a ry CSs to conve y one si ngle meani ng. Finall y, we have omit ted from Ta ron e’s original d escription of CSs h er claim that “ the spea ker stops t ry ing a lter natives when i t seems clear to the s peaker that there i s shared mea ning”. This condition see ms t o im ply tha t a ll ac hievement CSs resu lt in the successful c ommunication of the orig inally int ended m essage. Previous rese arch has, however, re vealed that some CSs a re more effective t han others a nd th at an agreement on meaning is not always reac hed whe n the le xical structures desire d t o convey this mea ni ng are not available or not sha r e d –cf. C hapter One, pp. 61-62. The criteria established in thi s worki ng def inition of CSs d eli mit our object of study and will guide our investigation: t he kind o f data we are going to examine, the ana lyses we a r e going to conduct and the c onclusions we will draw . In Chapte r Four, we will analyze the samples of face-t o-face FL inte ra ction elicited f or the purposes of the research with the aim of identifying C S insta nces. An element o f the interaction will b e considered a CS and under the sc ope of our study if, and onl y if , i t f ulfills the t hree conditio ns of o ur working def inition . In t he following c hapt ers of the da ta an aly si s, Chapter Fi ve a nd Chapter Six, att ention will be paid not only t o the C S but also to the interlocutor’s respons e to t his strategy. T his mea ns that attenti on w ill be foc used on th e whole st ra tegic commu ni ca tion process and on the use of the previously i dent ified CSs as part of a mutual attempt of two interlocutor s to a gree on meanin g. In this wa y w e expect to be a ble to account for strate gi c communication of meaning in fa ce-to-fa ce FL interacti on as a c ollaborative a ctivi ty involving t he joint ac tions and e fforts of a ll the interacti onal parti cipa nts. CHAPTER 3 METHOD CH A PTER 3: Meth od - 139 - The current chapter is dev ot ed to the methodology e mployed for the collection of th e data tha t will be analy zed in the pre sent inv es tig ation. We describe here th e ex periment set up for th e elicitation of the data and the procedures followed. In the f irst secti on we talk about th e pa rticipants who c olla bor ated in our projec t. We explai n the m e thod appl i ed to sele ct our final s a mple o f subjects from all th e student s who volunteered to take pa rt in our experiments. In the next s ec ti on we discuss the data collec tion ins truments . T he first inv ol v es a spot-the-dif ference task spe cifica l ly de signed to elic it int eractional informati on that could be c onsider ed as repres entative as possible of natur ally occurring i nteractions , and at th e s ame time suitable for analytical purpose s, i .e. with a ce rtain degree of researcher’s con trol. Following most previous research on CSs, we i ntend to support the study of lea rners’ l anguage with ret rospective da ta, more specifically, with learne rs’ retrospect iv e comments on the part icular us es of la nguage bei ng an alyzed. For t his purpose we d eveloped a post-inte rview activity. The final de sign o f these two da ta collection tas ks, a s described in sec tion 3.2, is the result of a n init ial pilot experience . In the next part of the c hapter we explain the procedures follo w e d t o carry out both this piloti ng and the ma in exper im ent. W e discuss the problems e ncountered duri ng the first t rial experience, and how ce rtain modif i cations inc orporated int o the ori ginal desig n of the e xperiment all owed us to compe nsate for them in the actua l data c ollection proce ss. We c onclude this acc ount with a gene ral assessment of the expe riment and of the na tur e of the data obtained. In the fi nal secti on of this cha pter we de scribe t he system and p rocedures foll ow e d f or the tr ansc ription of the sam ples of inte raction elicited. We also e xplain how we decide d to present the lang uage dat a collecte d in this work. 3.1. SUBJECTS In t he intro duction t o the presen t study we pointed out our inte nt ion to analyze FL interacti ons involving English la nguage learners with two diffe rent proficiency le vels, working in d yads with othe r le arners of the same leve l or with NSs. We also expla in ed that t his work wa s to be con ducted i n t he context of the University of San tiago de Compost ela. Therefore, our n on-native s peaking participant s were, a t the moment of the data collection, unde rg r adua te a nd graduate s tudents of the English language a nd CH A PTE R 5: The collaborati ve cre ation of meaning - 243 - In this ch apte r strategic i nte r action be tween learners and th e ir nati ve or non- native speaking interlo cutors will be a nalyzed with the aim of finding an answer t o the fi rs t major quest ion which prompt ed our rese arch: how is com munication achieved through CS use i n face-t o-face FL interacti on ? The f inal objective of the present cha pter is t hus to describe a nd unders t and h ow languag e l earners a nd their interloc utors ma nage to reach succe ssful c ommunication of t heir messages despite their IL s hortcomings, i. e. when the desired T L item or structu re to c ommunicate an intended meaning is either unavaila ble or not shared by all the i nterlocutors. In the a pproach we take to the st udy of strate gic comm unication, CSs a re considered to belong to the inte raction t aking pla ce between t he learne r and their interlocutor, a nd defined as e lements of this interac tional exchange. In f ac e -to-face oral conversati on, spe akers and a ddressees share a common goal, the succes s ful communication of their me ssages. C omm unication of meaning, as we understa nd it, is alwa ys a jointly ach ieved col l aborat ive activity. For c ommunicati on to succeed speaker and addre ssees nee d t o agree on me anin g. The sp eaker has to prese nt a message and th e addresse e t o understand it, but the two of t hem also need to e stablish the m utual belief that wha t has be en said has been u nderstood as intended. T his requires the c ollaborative effort of all the participants , the coord ination of thei r individual actions and be lief s in the pu rsuit of one c ommon goal (Cla rk and Wilke s-Gibbs, 1986; Clark and Schaefer, 1987, 1989; W ilkes-Gibbs, 1997). Problems enc ountered along the communicative proce ss are also comm only shared probl ems and th eir sol ution the mutual r e sponsibility of s pe ak ers and addressees. Strategic c ommunication w ill thus be s een as “ a mutua l attempt of t wo interlocutors to agree on a me aning in situa tions whe re re quisite m ea ning str uctures do not se em to be shared” (Tarone, 1981: 288), and commu nication of meaning t hroug h CS use will be explaine d as a collaboratively c o-constructed activity. Our objec tiv e is t herefore to descri be and un derstand how i nte rlocutors collaborate in f ace -to- f ace oral interact ion to solve w hat we belie v e to be a shared problem. Th at is, we inten d to identify what actions are taken by both the s peaker and the addressee whe n the desired T L ite m to communicate an intended me anin g is u navailable, h ow the two of them coor dinat e t heir individual action s a nd beliefs in order to reach a f inal a gr ee ment on meaning, a nd t o CHAPTE R 5: The c ollaborati ve creati on of meanin g - 244 - what ex tent th e communicative out c ome of t he s trategic ex c hange re s ul ts f rom th e collaborati ve effort of all the interactiona l participants. In the pre vious stage of t he research, we ana lyzed the samples of L 2 spontaneous c onversational data e licited for the purpose s of our study and ide ntified th e occ urren ce of linguistic diffi culties a nd su bs equent CS use s. The results of these ana lyses provided som e ini tial evidenc e that strategic com munication canno t be f ully understood if attent ion is focused only on the FL lea rner ’s actio ns, as has be en d one in most previous res earch. W e have f ound that le arners m onitor t heir interlocutors’ reactions when they us e a C S and, if th ey detect any indic ation that the content of t heir message is not being correc tly u nderstood, they repa ir or substi tut e the initial CS –cf. exa mples 4.24 a nd 4.46 in Chapte r Four, pp. 207 and 229-230. Inte rlocutors col laborate with l earners in t he devel opm ent of alternative means of e xpression whe n communicative problems a rise: t hey repa ir s peakers’ utte rances proposing les s ambiguous –cf. e xample 4.46 line 24, in pp. 229-230; more ela borated –cf . example 4.46 li ne 3 0, in pp. 22 9-230; or a dditional CSs – cf . exam ple 4.36, in p. 218. Sometimes they even provide the desire d but u navailable TL lexic al item in orde r to fac ilitate th e successful communic ation of the message –cf. examples 4. 34, 4.35 and 4.46 lines 22 and 42, in pp. 216, 217-21 8 a nd 229-230. E vidence was a lso fo und tha t, all thr ough this process, speakers and addressees exploit the combina tory possibilitie s of different C S types and th e contextual condit ions of the co-c onstructed ong o ing int erac tion. On thi s basis w e a rg ued t hat, in orde r t o un dersta nd how co mmunicatio n of m eaning is ach iev e d, attention ne eds to be pa id not only to the CS used by the FL lea r ner, but to the whole strategic i nteractional exchange ta king pla ce betw e en th e learner and the interlocutor. With this objective in mind we have propos ed the concept of C S epi sode a s a n alternati ve framework of a nalysis t o the traditi onal s tudy of C Ss a s le arners’ indepe ndent and isolated s trategic utterances. C S e pisodes have been defi ned as segments of i nteraction in which learne rs a nd their interlocutors work to rea ch an agreem ent on m eaning whe n t he TL ite m or s tructure de sired to express this mea ning is not rea dily a vailable. They may i nvolve o ne CS or a c ombination of diffe rent verbal and/or nonverba l CSs to fulfill one single communicative func tion. The s e a re ut tered by the learner, the interlocutor or both of the m working to gether alo ng a non- predetermined CH A PTE R 5: The collaborati ve cre ation of meaning - 245 - and unpredictable number of c onversational t urns. CS episode s a re thus always interacti onally and c ollaboratively co nstructed by the lea rner and their interlocutor –cf. Chapter Four, pp. 239-24 0. In order to acc omplish the objective s proposed in the introduc tion a nd thus g ive answer t o those qu estions whic h prom pted our rese arch, t he basic unit of our ana lyses will not be the C S, bu t t he C S e pisode. CSs, as des cribed in Cha pter Two, will be considered a s elements of the strategic e xchange taking place withi n the limit s of the CS episode, u nderstood as a hi gher order a nd more enc omp a ssing uni t of a nalys is, a ble to cap tur e the interaciona l nature of strategic communicatio n. In line with the approach we have taken to the s tudy and concept ualization of strate gic c ommunication, in the pre sent cha pter CS e pisodes w ill be analyzed within the theoretica l f ramework of the collaborative theory of c ommunication; a s de s cribe d in Cla rk and Wilkes-Gibbs (1986), Clark a nd Schaefer (1987, 1989), and W ilkes-Gibbs (1997). 53 The c ollaborative t heory expl ains c omm unication as a common ground bu ildi ng activity. People us e la nguage in order t o a dd new c ontent t o the common ground they share with the ir inte rlo c utors, i.e. to create new mu tual knowledge, b eliefs a nd assumpt ions out of their individual one s. Whereas most traditional accounts of communication e xplain thi s proce ss a s the re sult of the speaker is suing t he ri ght utterance at the right mome nt and the listener understandi ng it, the collaborative the ory sees it as a c ol laborati ve proce ss. The understanding of each utterance is the s hared responsibi li ty of the s peaker and the addressee and it can be succes s f ully achieved onl y if both of them work together to reach t his common goal. As al ready explained in se ction 1. 2 of C hapter One , spea kers c ontribute to t his process by present ing an ut terance w hi ch, they be lieve, specifies the co ntent they want to add to com mon groun d. But nei the r s peakers nor addre ssee s can j ust assume this utterance is always understood as intended. For c ommo n g round to be safely built, they nee d to establi sh the mutua l be lief tha t what is meant by the s peaker i s a ctually understood by the a ddressee well en ough f or the ir curre nt purpose s. This m utual agreem ent on me aning is achieved through a grou nding proces s in whi ch t he a ddresse e accepts the speake r’s presentation prov idi ng some kind of evide nce of the ir 53 For a de tailed revie w o f the colla borative theor y of c ommunicati on and the r esearc h c arri ed o ut within this frame work, see secti on 1. 2 in Chapter One. CHAPTE R 5: The c ollaborati ve creati on of meanin g - 246 - understa nding, and the spe aker recognizes a nd acc epts this ev i dence. If addr e ssees believe they have not been able to understand the speak er’s presentation, the y are expected to i ndicate their trouble and i nitiate a side sequence. The initi al prese ntation is then refashioned until a new versi on is achieved tha t can be c orre ctly understood a nd accepted by t he a ll t he i n terlocutors. In t his vi e w a ny communicative act involves the specif ication an d unders tandi ng of c ontent, plus its grounding, i.e. the speaker a nd addresse e’s building of the mutu al belief that the conte nt has been c orre ctly understood and satis factorily a dded to the ir c ommon ground (Clar k and Sch aefer, 1987: 20 , 1989: 262). The colla bor ative m ode l descri bes how this process is c arried out along a conversati onal unit c alled the contrib uti on (Clark a nd W ilkes-Gibbs, 1986: 35), and explains c onversation a s a com bin ati on of sequ entia l a nd hierarchical a rrang e me nts of contribut ions. T hese ha ve been defined as “a st retch of talk in which t he pa rticipants specif y a nd ground the c ontent of a cohe r ent pi ece of infor mation” (Clark and S c haefer, 1987: 20). The e xtent o f the unit th us d ef ined matches th e one d e limite d by the concept of C S episode , understood a s “ any segme nt of the interactio n in whic h lea r ners a nd their interlocutors e ngage in the s trategic communica t ion of a message, i.e. work t ogether to reach a mutual agreement on a me ani ng when the desired TL means of e xpress ion to convey this me aning i s not avail able or sha red” –cf . Chapter Four, pp. 239 -240. CS episodes wil l thus be here s een and exa mined as a particular form of conversati onal contribut ions, i.e. as a st r etch of F L talk in wh ich le arners and their int erlocutors specify and ground a c onte nt or mea ning fo r which t he desired TL me ans of expression is not availa ble. A con tribution, as described by the c ollaborative model of c om munication, involves a lways t wo identif iable phases : presentation and acceptance . In th e presentat ion phase the s peaker s pecifies the cont ent o f t he m es sa ge a nd i n the acceptanc e phase speaker and interlo cutor work to ground it. Any act, verbal or nonve rbal, by whic h a conversati onal participant i nt ends to mea n somet hing to an addressee i s cons idered a pr ese ntation (Clark and Schae fer, 1989 : 266). A CS, a n utterance by whi ch the FL lea rner inte nds to convey a meaning to their interlocutor, is t herefore a presentat ion. T hat is, we bel ieve that lea rner s use CSs in CH A PTE R 5: The collaborati ve cre ation of meaning - 247 - order to prese nt the content they wa nt to be added t o the ir and their interl oc ut ors’ share d common gr ound. In Chap ter O ne, section 1.2, we als o illustrated how in L1 non-stra tegic contribut ions spea ker s are expe cted to utter the prese nt ation that c an be st spe cify t heir message, so that their interlocutors can directly understand it. In CS episodes, however, the presentation c a nnot be uttered as initially intended by the spea ker, be cause the desire d TL ite m or structure is una vailable. This ne eds to dev e lop a n a lt ernative mea ns of expression in order to compe ns a te for the lack of wha t the y b elieve to be t he m ost ade quate or eff icient TL form. Strate gi c presentations , therefore, involve a lways a n extra amo unt of effort in terms of bot h production and c omprehension. In the previous chapter we a nalyz ed these a lt ernat ive means of expression. That is, we p aid a ttention to the stra tegic techniques language lea rners use for th e presentat ion of t heir messages whe n le xi cal difficulties arise in FL c ommunication. In this way we identif ied the d ifferent types of CSs ava ilable as well a s their combi natory possibil ities withi n the conte xt of ongoing oral interaction. In t he prese nt c hapter attention will be focused on the grounding of the content presented by mea ns of these CSs. W e will ana lyze CS episode s in or der to i dentif y and ca tegorize the procedu re s used in FL inte raction to a ssure that the C S utte red by the spea ker is unde rstood as intended by the addre ssee. As al ready point ed out, the aim is to understand how agreem ent on mea ning is c ollaboratively re ached, paying att ention to t he j oint a ctions and effort of speaker and addre ss ee all throug h the CS episode . Work done within th e framework of the c ollaborative theory of c ommunica tion has id entified the procedures a vailable in face-to-face ora l i nteraction f or mutual agreem ent on meani ng and subse quent s uccessful co mmunicati on of the m essage to be established. T hese grounding proce dures a re us ed by the addre ssee in orde r t o provide evidence o f their s tate of understa nding of the preceding prese ntation a nd by the speake r in order to acc ept this evidenc e. Tec hniques are al so available f or the c ollaborative solution of the und e rstanding proble ms that may arise during the communi cative process. Following the resea rch carried out on L1 c ommunication, we i ntend to identify how these different procedures work in FL strate gic interaction. We will f irst pa y atte ntion t o the devic es a ddres see s use in order to pr ov ide evidence of their sa tisfactory u ndersta nding of the s peaker’s co mmunicative i ntention s. CHAPTE R 5: The c ollaborati ve creati on of meanin g - 248 - That is, ho w addre ssees an d spea kers e stablish a greement on meanin g when the conten t presented throu gh a CS is succes sfully interpreted by the inte rlocutor. In t he sec ond section o f the c hapter t hose CS epi sodes in whic h understandi ng diffic ulti e s h ave be en en cou ntered wi ll be analyzed. The commu nicative model predicts a joi nt collabo ration between s pe a ker and addressee in the refashioning o f t hose presentat ions that cannot b e directly unde rstood a s in itially utte red. W hen a CS is no t understood by the addresse e, thi s a nd the learne r are expected to r eformulate i t until a version is achieve d that ca n be m utually accepted. The aim o f our study i s to id e ntify how this proce ss is carried out within the limits of the C S episode. In the final section we wil l a nalyz e those CS episodes that ha ve resulted i n unsucces s ful com munication. That i s, thos e strategic int eracti ons i n w hi ch, be cause of the learner’s use of an avoida nce CS or de spite th eir use of a n achieveme nt strategy , final a greement on meaning has not be en re ached. We will try to ide ntify the cau ses of these failures i n com munication and to de termine to what ex t ent they ar e due to th e nature of the CSs employ e d or t o unsolved problem s occurring during t he g rounding process. 5.1. PROVIDING EVI DENCE OF CS UN DERSTANDING FOR GROU NDING PURPOSES IN CS EPISODES In Chapte r O ne, se ction 1. 2, w e saw that speakers and addressees adapt their utte rance s and ac tions to each other a nd to the con textual circum stances of their ongoing interacti on. Content can thus be presented and grounded in an unpredictable a nd unlimited num ber of ways . In Cha pter Four we al so saw that CSs can not only be of diffe r ent types, the y can als o be uttered a lone or in c ombinati on with other CSs, in one single turn or in seve ral subseque nt turns, by the l earner or thei r interlocutor, or even co- construc ted by t he j oint action of both of t hem. W e should therefore expec t gr oundin g devices t o vary de pending on t he na ture o f the strate gic presentation an d the purpose s and demands of the specific intera ctional context to which it belongs. A deta iled qualitative analysis of CS epis odes rev ea ls that, w henever the interlocutors be lieve they have correctl y heard and unde rstood what the FL learne r is try in g to mean with a j us t ut tered CS , t hey can use a varie ty of difference procedures in order t o p rovide e vidence of this unders t anding. They may a ssert the ir unde rstandi ng CH A PTE R 5: The collaborati ve cre ation of meaning - 249 - and acceptanc e of the speaker’s previous pre sentation providing a n aff irmative response such as ‘yea h’, ‘okay’ o r ‘uh huh ’. The se kinds of c onfirmatory devic es a re kno wn i n the colla borative theory of communicati on as acknowledgme nts , c ontinue rs or backchannel responses ( Cla rk a nd Wilkes-Gib bs , 1986: 8; Cla rk and Schaefe r, 1989: 280-283; C lark, 199 4: 994 and 997). The y may also show wha t it is that t hey have understood. For t his pu rpos e the y some times us e ve rbatim re pet itions of all or pa rt of the previous s trategic presentation. In other c ircumstances t hey para phr ase what the speaker h a s just said, pr oviding an alternative CS or what they be lieved to be the lea rn e r’s inte nded but una vailable T L it em. R epetitio ns a re usually ref erred to as display s and pa raphrases as demonstrations in the c ollaborative the ory research (Clark and Schaefe r, 19 89: 267). W e have finall y found t hat a ddressee s may a lso pre suppose their acceptance of the sp eaker’s strategic presentation, initiating a releva nt n ext contribut ion or all owing the speaker t o proceed with the conversat ion by showin g continued at tention (Clark and Wilkes -Gibbs, 1986: 7 and 9; Clark and Schae fer, 1989: 267 and 271-272; Clark, 1994: 993-994; W ilkes-Gibbs, 1995: 246). All these d ifferent kind s of grounding p roced ures have been widely d ocumente d in L1 c ommunicati on studies. In fac t, Cla rk and Schaefe r identified t hese five diff erent categories of grounding device s a s the most common proce dures used by a ddressees in order to provide e v idence of un derstanding of a previous s peaker’s utterance (Cla rk and Schaefer, 1989: 267). He re we wil l a nalyze a nd de scribe how thes e strategies work i n FL CS epis odes. That is, how speakers and a ddressees use t his evidence of understa nding to bui ld the mu tual bel i e f that a j ust uttered C S has been int erpreted by the addressee as in tende d by the s peaker and thus reach an agreement on the mea ning they are try i ng to communic ate. 5.1.1. ACKNO WLEDGM ENTS AS A G ROUNDING PROCEDURE IN CS EPISODE S Taking int o account t he app roach a dopted to the conceptualiza tion and study of CSs, and the specific purpose s of o ur re search proj ect, we spec ified in Chapter Four th e criteria that a n ele ment of interaction needs to fulfi ll in order to be considered a C S a nd within the scope of our anal yses –cf . Chapter Four, p. 178. Followi ng the working definiti on thus proposed, we consi der that in the next extract of interaction a CHAPTE R 5: The c ollaborati ve creati on of meanin g - 250 - combination of two different CSs is us ed when the learner, A, desire s to comm unicat e the meaning ‘rolled up’ to his interlocutor, B , and f inds that the lexical item des ired to ach iev e t his goal is unavailable . 54 That is, th e learn e r i ntends to add the meaning ‘ro lled up’ to the common ground he sha res wit h his interlocu tor. For this purpose he needs to present a n utterance that ca n specif y thi s content, but the l exical ex pression that h e desire s t o use in this presentation, i . e . ‘rolled up’, is not yet part of his IL syste m. The lear ner deci des to attem pt a n alte rn ate mea ns to c ommunicate this meani ng and present s a combination of an approx imation and a nonverbal s tr a tegy. Instead of the intended but unavailab le T L e xpression ‘rolle d up’ he uses the IL it em ‘up’, which he believes to share e nough se mantic fea tures w ith the origina lly i nte nded one to be her e correctly understood by his interlocutor. He illustr ates the meanin g of this oral strategi c expressi on with an enacting nonverbal C S: mimicking the ac tion of rolling up a sleeve . EXAMP LE 5.1: 55 ADV-NS R A: rolled up REFERE NT: to r oll up/ father rolls up his sleeve INTERLANG UAGE DATA ( 406-408): 1 A: he remains wi th the:: tch {with the:: (0.5) sleev e (0.5) of 2 { A’s LH mimi cs rolling u p her right 3 sleeve } 4 A: the tshirt (1. 8) up? 5 B: {{uhuh}} } 6 {{ B nods }} 7 A: a::nd an:d no w the man with glasses, (0.5) seems to be very: 8 happy NATIVE LANG UAGE D ATA: A: el padre empieza a a remangarse en plan c omo si no sa le pues que le va a pegar o q ue le quiere pegar NATIVE LANG UAGE D ATA: A: la manga remangad a también el pa- el padr e el de bigo te SENDER’S RETR OSPECTI VE CO MMENTS ( 406-408): I: aquí? A: remangada sí es q ue no no sabía como deci rlo RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: raising up the sleeve s of hi s shirt 54 T he interl ocutor i dentifie d as A in the transcr ipt is a lways the learner takin g the sender role i n the picture- sto r y spot- the-differe nce task. Th e receive r in this tas k, whether another lear ner o r a native speakin g interl ocutor, will be identif ied as B. For the purp oses of our ana l y s is, t he int erlo cutor init iating the strate gic presentati on of t he message a nd t he CS epi sode will be r eferr ed to as ‘t he spea ker’ and their partner as ‘the addre ssee’. In o r der t o f acilitate ref erence w e will a ls o treat the speaker as a m ale a nd the addressee a s a f e male, i ndepende ntly of t heir act ual sex as specifie d in ta bles 3.1 an d 3. 2 –cf. Chapter Three, pp. 142-144, a nd in Ap pendix B. 2 and A ppendix C. 55 Like in Chapter Four, we include f or eac h example inform ati on on t he proficienc y le vel o f the learner, A, an d t he inter locut o r , B, if he/ she is al so a learner an d, if n ot, in dication that he/she is an E nglish NS. We als o specif y t h e int ended mean ing of t he CS, the tar get r eferent that elicite d the CS as prese nted in table 3.4 –cf. Cha pter Three , p. 1 52, a nd the line s of the tr anscri p t of the whole inte raction t o w hich the excerpt in the exa mple be lon gs. Finall y we incl ude, whe never r elevant, NL baseli ne data, re trospective data and t he results of the diff erence ide ntification task. CH A PTE R 5: The collaborati ve cre ation of meaning - 251 - The repetitions, drawls, pauses and paus e f illers act as a s ignal for the a ddr essee evidenc ing that the learner is experiencing some ki nd of l ing uistic problem in the utterance of his presentation. Th is is also modified by wha t Sacks and Sch egloff (1979) have called a try marker , i.e . it is uttered with f inal risin g i ntonation. W ith t his t ry marker the speaker indicates that he is not sure a bout the und erstandability of his presentat ion and that, for mut ual agreem ent on me aning t o be reac hed, he needs a n explicit confirmation of unde rstandi ng from his inte rloc ut or (Clark a nd W ilkes-Gibbs, 1986: 18-19). 56 The use of the a pproximation and nonverbal strategies has a llowed the speaker to pre sent t he mean ing he wa nts to contri bute to di scourse, but t his prese ntation a lon e can not account f or the s uccessful c ommunication of t he messag e. For c ommunicati on to be achieved agreement on meaning between the learner and his i nterlocutor needs to b e established, i. e. the content here presented al so needs to be grounded. The CS episode e nters thus t he ac ceptance phase, in which the add ressee is expected to p rovide some ki nd of evidence of her state of understanding and he r subseque nt acceptance or non-acce ptance of the CSs prese nted by the speaker. Here, the addresse e believes she ha s been a ble to identif y the m eani ng t he learner int ends to contribute to their sha red com mon ground. She indica tes her underst anding of the approximatio n and nonverba l strategies offering a n asse rting ‘uhuh ’ a nd a head nod. Within th e c oll aborat ive theo ry frame work these are consi dered to be ac knowledgments (Clark and Scha e fer, 1989: 267), i.e. aff irmative verbal or n onverbal s ignals offered by the ad dressee in order to assert thei r a cceptance of the speaker’s previous prese ntation. For a greement on me aning to be reache d the learner also ne eds t o accept these ack nowledgments as enoug h e vidence o f the addr essee ’s sa t isfact ory underst andin g of his intended message. In line s 7 a nd 8 of the t ranscript we ca n see how he initiates a new co ntribution: ‘a: :nd a n:d now the man wi th glass es, seems to be very: ha ppy’. B y allowing the conversation to proceed to the next topic the speaker is acc epting the evidence of und erstanding of fere d by t he addressee , i.e. the spea ker is ac cepting the addresse e’s accepta nce. Mutua l ac ceptance is thus e stablished. The j oint a nd coordinated actions of speaker and a ddressee have all owed them to establis h the mu tual belief tha t wha t the speaker meant with his initi al pre sentation –with the approximation 56 The use of tr y markers a nd their r ol e in the groun ding process have been anal y ze d in de tail in Chapter One, pp. 88-8 9. CHAPTE R 5: The c ollaborati ve creati on of meanin g - 258 - pp. 88-89, tria l c onstituents a re elem ents of a presentation that a r e m odified w ith a t ry marker, i. e. with rising into nation f ollowed by a pause. Some times spea kers a re unsure about the understandability of jus t some part of the pre s entation the y are utteri ng. To prevent f inal mis understanding proble ms of the whol e prese nt ation they utter this portion with rising int onation. Wi th this t ry marker they are as king the ir addressee s fo r an expl icit acceptance o f the trial c onstituent. Only a f ter the a ddressee has provided t his acceptanc e does the speaker cont inue with his turn an d complete his presenta tion. On this basis we ca n a r gue that when a CS is uttered as a t ri al constituent the speaker is i ndicati ng that they are not a s c ertain of the understa ndability of thi s portion of their presentati on as they a re of the rest, a nd that the y consi der ne cessary to ground the co ntent of the CS before thei r overall encompassing presentati on can be com pleted. In the CS e pisode presented in examp l e 5. 4 l earner A uses an appro xi mati on strate gy to pre sent the m eaning ‘rolled up’, illustrate d by a n e nacti ng non verbal strate gy : rolling t he sle ev e s of his s hirt. Notice that the a pproximation CS has re sulted in a ctual c orrect T L le xis: ‘ ro::lled up?’. But n either the spea ker nor the addre ss ee are aware of this fact. They use an I L e xpression tha t the y believe c an convey a s imila r mea ni ng to the ori ginally i ntende d one but c annot be a ccurately used in this conte xt. This IL e xpressi on i s uttered with rising intonation and f ollowed by a short pa use, i.e. is uttered as a trial constituent. EXAMP LE 5.4: ADV-ADV MM: rolle d up REFERE NT: rolle d up sleeve s INTERLANG UAGE DATA ( 701-703): 1 A: well (0.3) you know at the: the {the s leeves were ro::lled up? 2 { A’s R H rolls up h er LA 3 sleev e } 4 (0.5) 5 B: yeah. 6 (0.5) 7 A: for the small man} NATIVE LANG UAGE D ATA: B: las mangas de la camisa SENDER AND RECE IVER’ S RETR OSPECT IVE CO MMENTS ( 553-56 1): 59 A: no sé como se dic e remangarse rolled up? I: sí to roll up o s í e ti entendiches non? B: si I: polo que te dicía ? ou pola mímica? polas dúas cosas? A: yo creo que fue u na combinación heh B: eu pen- si I: e ti sabes dicir que as mangas remangadas 59 A f irst ref erence t o the mea ning ‘rolled up’, previous to the one here ana lyzed, was made in l ine s 553- 561 of the tr anscript of the wh ole intera ction, see Appendix B.2. CH A PTE R 5: The collaborati ve cre ation of meaning - 259 - B: rolled up non? o sea eu non sei se hai ou tra forma de RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: rolled slee ves The addressee provides an a cknowledgment, ‘yeah’, i n orde r to indicate that she believes she has c orrectly unde rstood the content t he learner i s trying to ad d to their shared common ground. Af ter the a cknowledgment th e spea ker continues with his presentat ion: ‘for the sma ll man’. The m eaning ‘rolled up’ h as t hus be en grounded. Speake r and addre sse e ha ve s uccessfully re ached a mutu al a gr ee ment on this mea ning through their col laborative joint actio ns. In this excerpt of interaction we c an thus distinguish t wo diffe rent co ntributions. There is a n all-enc om passing cont ribution that incl udes the whole presentation uttere d by the speaker, ‘well (0.3) you k now a t the: the the sleeves were ro::lled up? for the smal l man ’, and its a ccep tance in subsequent tur ns. There is also an embedded contribut ion, t he CS epis ode. T his is ini tiated wit h the speaker’s presentati on of the approximatio n a nd nonverbal CSs. It is accompli shed with the a gr ee ment on the mea ni ng ‘roll ed up’ t hat is rea ched when the addressee a nd the speake r e stablish their mutual acce pt ance of the trial constit uent. The a nalysis of our dat a reveals that the use of verbal and nonverbal ack nowledgments in order to p rov ide evidence o f understandi ng of a pre viously ut tered CS is an extremely frequent g rou nding proce dure in FL strategic inte raction. In Appendix C the reader can find mul tiple examples of CS episodes where ack nowledgments are used f or grounding purp ose s in both N NS -NS a n d NNS- NNS interacti ons, involving eit her advanced or intermediate proficiency le vel l ea rners. That is, t his grounding proce dure is fr equently use d in CS epi sodes by bot h native and non- native s peaking inte rlocutors in depen dently of the learners’ l evel of proficiency in th e TL. All these d if ferent e xa mple s support the a nalyses here c arried out and the conc lusions dra wn. Utt ered at the end or in the m iddle of t urns, a cknowledgments are offere d by the addressee in order t o i ndicate t hat they be lieve they have correctly interpre ted th e mea ning of the CS presented by the speaker. Once recognized and accepted by th e speaker, t hey allow this and th e ir i nterlocutor to reach an agr ee ment on the meaning t hey are trying to com munic ate throu gh CS use . Succe ssful communicati on CHAPTE R 5: The c ollaborati ve creati on of meanin g - 260 - of the message is thus c ol labo rat ively achieved through the joint a ction of the learner and their nati ve or non-native spea king interlocutor. 5.1.2. REP ETITIO N AS A GROUNDING PROCEDURE IN CS EPISODES Another common procedure for a ddressees t o indica te that they bel i eve the y have noticed, heard and understood what has j ust been said i s t he verbatim repetition of all or part of th e sp ea ker’s previous presentation . This kind of repetiti v e behav ior is usually known a s display in the collabora tiv e theory of communica tion literature (Clark a nd Schaefer, 1989: 2 67). In t he f ollowing e xtract o f inte raction, example 5.5, we c an see how the addres se e re pe a ts th e C S uttered by the spea k er as a means to c ollaborate in the grounding of the content just presente d. In line 1 o f th e following trans cript the spe a ker, here B, 60 desires to mak e reference to a ‘knoc ker’ . The TL lexi cal item de sired to present this mea ning is unavaila ble in h is IL s ys tem and he c hooses to mak e use of a ci rcumlocution strategy: ‘something to knoc k ? ’. This circumlocut ion a lso involves an all-purpose word: ‘something’. The prese ntation i s at t he sa me time modif ied by a tr y ma rker, i.e. uttere d with rising intona tion and a f ollowing pause . EXAMP LE 5.5: INT- INT FC: knoc k er REFERE NT: knocker/ door knocker INTERLANG UAGE DATA ( 417-418): 1 B: tch a:::nd (0. 5) something to knock? 2 (1.4) 3 A: yeah something :: to kn- to knock. SENDER AND RECE IVER’ S RETR OSPECT IVE CO MMENTS ( 390-39 9): B: y tampoco sabíamo s decir esto I: el knocker B: ah! se dice knock er! I: sí A: knocker I: que tú en un mome nto llegaste a decir kno cker B: sí! A: sí pero no no sab ía I: tu dijste knocker pero como no B: sí sí A: dije knocker 60 Because of the desi gn of the data collect ion ins trument –c f. Chapter Three , p. 15 5, in t he excer pts of interaction we are a nal y zi ng the sender i n the pic ture-st ory spot-the-d ifferenc e t ask is usua lly the participant w ho prese nts the first CS and i nitiates t he CS episode. That i s, A is generall y the speaker. However, un der cert ain ci rcumsta nces, the r ecei v er , B, ma y be t he one wh o pre sents t he CS an d takes the role of s peaker i n the CS ep isode. T he se nder , A, a ct s t hen a s t he a ddressee o f the CS. Thi s c hange in roles needs to be ta k e n into account in the anal ysis of this CS episode, as well as in the CS epis odes presented i n examples 6.6, 6.1 0 and 6.1 9. CH A PTE R 5: The collaborati ve cre ation of meaning - 261 - B: sí lo dijiste I: pero como como ya no tal pues entonces vo lvisteis a e ntonces no sé tú cuando le o íste decir knocker B: yo decía sí I: o sea lo de knock er tú lo reconociste? B: sí supe que estab a diciendo pero digo yo no sé si ser á yo si I: o sea no sabías s i era la palabra o si B: claro no sabía si se lo había inventado o si cuadraba realmente A: uhuh! es que me l o había inventado knocke r knocking knocker heh heh B: heh heh heh A: está bien RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: to knock , the thing to knock In line 3 the addres see a sserts her unde rstanding and acce ptance of the speak er’s strate gic presentation with an ac knowledgment, ‘yeah’, and the re petition of th e C S : ‘something:: to kn- to knoc k.’. Af ter t hi s e vi dence of unde rstanding th e conversati on continues. Speaker and addres see allow t he conversation to proce ed to the ne xt topic , which, as already poi nted out, i ndicates that they co nsider th e CS e pisode c ompleted and it s content satisfactorily grounded. The addressee has combined h ere two diff erent kinds of proce dur es for the purpose of c ommon ground bu ilding: t he us e of a n ac knowledgment an d the verba tim repetition of the CS initial ly uttered by the speake r. This acti on a nd the subsequent continuation of the c onversation h ave a llowed spe aker a nd addressee to es tablish the mutual belief that the mea ning ‘knocker’ h as been understood to a criteri on sufficien t for their current purposes . By repeating the sp eaker’s words the addressee has participated a s e xpected in the grounding proce ss a nd in the collaborati ve c o- construc tion of meaning. This use of repetition in F L stra te gi c inte r actions had a lready be en doc um ente d in Anderson (1998) –cf. Chapte r One, p. 60. Wi thin the field of S L A resea rch, repetiti on responses or echoing what s om eone has j ust sai d ha d be en ofte n s tudied as tim e gaining strate gies –c f . Dörnyei a nd Scot t (1997); or conf irmation checks –cf. Long (1980 ). But Anderson show ed in his w ork t hat other-r epe tition can also be used for gr ounding purposes, i.e. t o in di cate t o the speaker that the content just prese nted h as bee n c orrectly understood a nd grounded (Anderson, 19 98: 35-37). The res ults of our a nalyses c onfirm Ande rson’s conc lusions abou t t he use of repetition. A lthough in C S episode s this gr ounding procedure is not as common as ack nowledgments, w e have f ound tha t b oth native and non-nati ve speaking inte rlocutors CHAPTE R 5: The c ollaborati ve creati on of meanin g - 262 - resort to ot her-repetit ion i n order to provide evide nce of what they bel ieve to be their correct underst anding of a learner’s pre vious CS. Again, the rea der ca n f ind in Appendix C more examples of CS e pisodes e videncing this kind of be ha vior. The one here analyzed ill ustrates the use of repetition for grounding purposes in a n interactional exc hang e between two intermediate level languag e learners. A re presentative example of th is strategy in inte raction involving advance d level s tudents may be found in ADV- ADV SO: 146-149. Nat ive sp eaki ng interlocut ors resorted als o to thi s s trategy in thei r conversati ons with advanced level learners, see, f or ins tance, ADV-NS IM: 259-264, and with interm ediate level speakers, see INT-NS AE: 68-70. 61 5.1.3. INIT IATING A RELEVANT NEXT CONTRIBUTIO N AS A GROUNDING PRO CEDURE IN CS EPI SODES By offe ring acknowledgment s and repetitions a ddressees asse rt their understa nding and acceptanc e o f the speaker’s previous p res entation but, as explained above, accepta nc e can al so be presu pposed. In the ana ly si s of ou r da ta we have f ound that addre ssees somet i mes respond to spe akers’ strategic pre s e ntation s initiating a r ele vant next contribut ion. That is, they o ffer a n utterance which c onstitutes the pre sentation of a ne w conte nt a nd the init iation of t he next c ontribution, but tha t i s bui lt on the speak er’s previous prese ntation. A relevant next ut t erance indicates through it s structure and/ or its conte nt the understa nding of what ha s been previously sa id, a nd is c onsidered and treated for grounding pur poses a s an implicit indication of acce ptance. Following Sche gloff and Sac ks (1973), Cla rk and Scha ef er sugg e st tha t many contribut ions follow this pattern of acceptance beca use they a re organized in a djacency pairs (C lark and Scha efer, 1989: 270-2 71). The y are co nstituted of two ordered utterances or pai r parts and, on the basis of what is know n as the conditi onal relevance property , t he first one projects the s tru c ture and conte nt of the second one. Thus, if the speaker presents a question, it is conditi onally relevant f or the a ddressee to re spond wit h an a nswer. A request is expec ted to be followed by a c ompliance or ref usal, a proposal by an a cceptance or rejection, a greeting by a greeti ng a nd so on. 61 As expl ained i n C hapter Thre e, p. 1 45, f or identificati on pu rposes, w e inclu de, f or each g i ven exa mple, informati on c oncerning the i nterl o c utors taking part in t he exchange –pr oficienc y le vel i n the TL a nd the initials of the ir names. We a lso indica te the li nes of the tra nscript of the whole in teracti on to which eac h CS epis ode bel ongs. The r e ader can thus ea sil y ide ntif y in App en dix B.2 a nd Appen dix C the transcripti on and a nal y si s of eac h of the CS e pisodes here m entioned. CH A PTE R 5: The collaborati ve cre ation of meaning - 263 - The C S e pisode i n example 5. 6 is initiated by B present ing a quest ion for A to be answered. W ithin this que stion the s peaker intends t o ma ke re fe rence to a ‘knocker’, but t his T L ite m is unavailable i n his IL system. In order to prese nt this m eaning a nd thus compe nsate f or his ignorance of the word ‘k nocker’, he decides to develop a n alternati ve mean s of e xpression. In line s 1-4 o f the transcript B pre sents a n all-purpose word, ‘t hing’, as pa rt of a wordy e xtended desc ription of the i ntended object of reference : ‘the:: thing to kn ock’. T his c ircumlocution CS is a ccompa nied by an ena cti ng nonverbal strategy that il lustrates the oral description: mim icking knocking on a doo r. EXAMP LE 5.6: INT- INT C L: knocker REFERE NT: knocker/ door knocker INTERLANG UAGE DATA ( 628-629): 1 B: you: you still watching the:, (1.2) {t he:: thing t o knock. 2 { B ’s RH mimics knocking on 3 a door } 4 B: (1.0) on the d oor?=} 5 A: {=no!} because the door is completely open, 6 { A shakes her head } NATIVE LANG UAGE D ATA: A: está simplemente para llamar a la puerta el bueno no sé el el éste pa llamar a la puerta heh NATIVE LANG UAGE D ATA: A: bueno y el chisme este para llamar que si go sin saber como se llama heh el llam ador vamos el que cuando no hay timb re A: sí el RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: llamador (t oc toc) The ut terance by the addres see, here A, of a relevant a nd expectable answer t o the first part of the ad jace ncy pa ir indicates that s he has understood what the speake r mea nt : ‘no! because the door is c ompletely open’. Fol lowing Clark a nd Sc ha efer (1989: 271), we can say that A is providing here e vi dence of her understanding at thre e diffe r ent levels. She is pass ing up the opp ortunity to a sk for a repair which, as we have already see n, is cons idered as an impl icit acceptance of the pres entation. By initiati ng an answer she is recognizing that a questi on has been asked and, through the c ontent of this answer, di splay ing her corre ct unde rstanding of this question. In order for the two int erlocut ors to be able to establish the ir mutual be lief on the correct interpre tation o f the C S and the question, this ans wer a lso needs to be acc epted by the spea ker. Furthermore, with this r espo ns e t he add ressee is also presenting ne w conte nt, i.e. initiating a rele vant next c ontribution. The utterance in 5- 6 i s pa rt of th e acceptanc e phase of t he CS e pisode, but a lso of the pre sent a ti on phase of a sec ond CHAPTE R 5: The c ollaborati ve creati on of meanin g - 264 - embe dded c ontribution. Its c onte nt needs theref ore to be grounded. Here spea ker a nd addresse e move on to th e ne xt t opic. Allowing the conversation to continue they accept each other’s presentati ons and establish t he mutual belief t hat t he question i n 1-4 has bee n sa tisfactorily a sked and answered, and t he cont ent c ont ribute d to discourse t hus adde d to their share d common ground. The c ommunicat ion of the meaning ‘knocke r’ has bee n collaboratively ach ieved through the s peaker and the a ddressee’s c oordina tion of their individual actions and beliefs. In the ne xt CS epi sode the ad dressee’s accept ance utterance is re lated to the speaker’s strategic presentation in a different way , but it c an a lso be consi dered as the initiatio n of a relevant next contri bution. The speaker, A, desires t o communicate the mea ni ngs ‘painte d’ a nd ‘wal lpapered ’ to the addres see, B. But in utt ering his presentat ion he finds that the TL i tems desired for this purpose a re unavailable in h is IL system. He pre sents a n alternative te rm of reference , ‘decorated’, whi ch is be lieved to share enough sem antic f eatures wi th the intended one s to be c orre ctly interpreted by his addresse e. He is thus making use of an approximation CS. EXAMP LE 5.7: ADV-ADV LT: painte d, wallpa pered REFERE NT: wallpa per/flowe red wallpa per INTERLANG UAGE DATA ( 201-202): 1 A: a::n:d behi::n :d him, (1.8) there’s a: : wall, (1.4 ) e::h 2 decorated with (.) flowers, or 3 B: {mine is plain .} 4 { B shakes her head } NATIVE LANG UAGE D ATA: A: eh hay se ve como una pared eso con B: sí A: estampada de flor es B: en la mía no tien e ni flores ni tiene el suelo con SENDER AND RECE IVER’ S RETR OSPECT IVE CO MMENTS ( 201-20 2): I: ahí por ejemplo e n lo de a wall decorated with flower s? ... no pensaste que fuer a empapelada? A: quise decir un no sabía otra forma de dec irlo I: entonces cuando d ices lo de decorada te r efieres a A: eso I: empapelado? A: o con flores pint adas I: y tú qué entiende s? simplemente? B: eh con flores pue s empapelada o sea pero sí con A: bueno es porque e s papel heh B: bueno pues ya enc ontramos la diferencia y punto o sea no I: y tú sabrías deci r empapelado? B: eh papered o algo así? I: wall wallpaper B: wallpaper RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: wall decor ated with f lowers CH A PTE R 5: The collaborati ve cre ation of meaning - 265 - The add r essee believes she ha s un derstood t he speaker’s message well eno ugh for t heir current purposes. As e xplained in Chapte r O ne , pp. 78-79, i nterlocutors ar e not alwa ys seeking perfec t understa nding o f eac h othe r’s prese ntations. Clark and Wilk es - Gibbs (1986) su ggest tha t pa rticipants i n a conversati on follow a pr in ciple of mutual respons ibility , which means that the y “try to establi sh, roughly by the initiati on of each new c ontribution, the mutua l be l ief t hat the listeners ha ve unde rstood wh at the speaker mea nt in the last utterance to a crit erion s uf ficient f or curren t p urpose s ” (Cla rk a nd Wilkes-Gibb s, 1986: 33). Depe nding on the circumsta nces of their i nteraction and t heir goals in each cont ribution, the c riterion of understanding ma y be hi gher or lower. At the same time, they are trying to minimi ze the ir collaborati ve ef fort and to avoid extra steps in bot h the presentati on and the accep tance phase of their contribut ions. Spe aker and addres see will some t imes agree t o t olerate s ome lack of understa nding or u ncertai nty in order t o minimize t heir joi nt effort, a s long as t his unce rt a inty doe s not interfere with the princi ple of mutual responsibility. 62 In the task we de signed for the col lection of our da ta , descri bed i n Chapte r Three, section 3.2.1, spe aker a nd a ddressee’s final c ommunicative goal is to iden t ify th e diffe r ences e xisting betwe en the ir two versions of the same pi c ture - story. In the previous excerpt of interaction, which belongs to the de scription of picture eight, their aim is to i dentify whe ther the w all of t he house depi cted in t heir pictures is som ehow diffe r ent in th e ir two versions of the story. Wher eas f or the rece iver, learner B, thi s wall is plai n, for t he sender, A, i t has a pa ttern with fl owers. This is the re le va nt piec e of information t hey need t o c ommunicate. It is t herefore not significant whethe r these flowers are ac tually painted on t he w all or part of a wall paper. Whe n B identif i es this dif ference she considers she has understood A’s presentat ion well enough f or their current purpose s, i.e. we ll e nough for the succes s f ul performance of the c om munica tiv e ta sk at hand. She inte rrupts t he s pea ke r mi d-turn in order to indicate not only t hat she belie ves s he h a s already understoo d th e message, but also that she does not need a more e xtensive desc ription. 63 The relevant content and timing of B’s ut t erance, ‘mine is plain’, shows her unders t anding of the speaker’s 62 The princ iples of mutual res ponsibilit y a nd le ast c ollaborati ve e ffort have bee n acc ounted for in detail in Chapter O ne, pp. 78-80. 63 About the r ole o f interrupti ons in gr ounding processes, a lso di scusse d in C hapter One, p. 86, see Clark and Wilkes- Gibbs (19 86: 23-24). CHAPTE R 5: The c ollaborati ve creati on of meanin g - 266 - presentat ion, includin g the CS. The C S e pisode is closed here. Speaker and addres se e move to a ne w topic. Allowing the conversa tion to continue the speaker indicate s that h e accepts t he a ddressee’s in terrupting u tt eranc e as evide nce of unde rstanding of the previous pres entation and a lso as the p resentation of a new cont ribution. In this way speaker and a ddressee a gree that what the spea k er m eant wit h the approx imation C S has bee n und e rstood by the addressee well e nough for thei r current purposes. The a ddressee’s prese ntation of a rele vant ne xt utte rance ha s made possible the successful grounding of the co ntribution. T his uttera nce does not c onstitute the sec ond part of an a djacency pair, as we saw for th e CS episode in e xample 5.6, but is rele vant in terms of conte nt a nd timing. A greement on mea ning a nd succe ssful communication of the orig inally int ende d message have once again be en achie ved through the collaborati ve effort of both speaker and ad dressee. We are going to anal yz e here one more example of a ddr e ssees ’ initiation o f a relevan t next c ontribution as a groundin g procedure in C S epis odes. The CS epis ode in exa mple 5.7 is th e result of the speaker a nd the addre ssee’s ig nora nce of the T L item ‘rolled up’. In li nes 1-2 the learner presents this meaning using an a lternativ e approximate term of reference , t he word ‘upwa rds’, a nd a pointing nonverbal strategy. The lengthe ning , t he pa use and the pa use filler ‘eh’ s erve as e vi dence for the a ddresse e that the le arner has encountered a lexical difficulty and is not using the orig i nally intended lexis. This presentation is al so marked by rising intonat ion fo llowed by a pause. As we e xpl aine d in the analysis of examples 5.1 a nd 5.5, this i ndicates the speaker’s uncertainty abo ut the understa ndability of h is prese ntation and the nee d of an addresse e’s explicit confir mati on of understandi ng . The s peaker is thus checking f or his addresse e’s compre hension. EXAMP LE 5.8: ADV-ADV LT: rolled up REFERE NT: to r oll up/ father rolls up his sleeve/ r olled up s leeves INTERLANG UAGE DATA ( 378-380): 1 A: and he’s {stil l with the slee:ve e::h (2.5) upward s?} 2 { A’s LH points to RA sleeve o ver her elbo w } 3 (2.0) 4 B: ah! okay, mine has no sleeve upwards. but as it th e is it as 5 it’s the same, so it’s not. (2.8) i th ink that’s a ll. NATIVE LANG UAGE D ATA: A: eh después en el siguiente eh ya se está remangando l a manga de la camiseta B: no el mío so- apa rte que sólo se le ve es tá en la mis ma actitud que en la en la v ez anterior y sólo se le ve ese braz o SENDER AND RECE IVER’ S RETR OSPECT IVE CO MMENTS ( 378-38 0): CH A PTE R 5: The collaborati ve cre ation of meaning - 267 - A: ahí ella no sé si me entendió o no B: sí le entendí que tenía la manga subida I: mhm remangada? B: sí I: sin embargo no lo apuntaste aquí B: no pero sí ... es que hay cosas por ejemp lo the badge que me preguntaste al pr incipio si me acordaba o no y no me dijiste la palabra pero desp ués ahora ahora me acord é y claro es badge RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: only one arm not holdin g his s leeve The addressee pr ovides evidence of her unde rstanding a nd a cceptance of th e speaker’s presen tation i n the form of acknowle dgments: ‘ah!’ and ‘okay’. But she a lso present s a relevant ne xt ut te rance: ‘mine has no sle eve upwards’. B’s turn i n lines 4-5 consti tu tes a ne w presentati on and it a lso indicates in it s content the a cceptance of the previous utterance. Furthermore , in this n ew pres entation the addressee r epea ts th e lea rn e r’s a pproximation st rategy: ‘slee ve upwa rds ’ . She is thus indicating not only that she bel ieves she has u nderstood the meaning the l ea rn er is trying to c ommunicate , but also that she considers this C S a s an a greed alternative term of re f erence to compensate for the unavailable TL item ‘rolled up’. Three diffe rent procedures are here com bin ed for the purposes of c o mmon ground buildin g. The addres see provides ac knowledgments, initiates a re le vant next contribut ion a nd repeats the CS initia lly prese nted b y the spea ker. After the se different forms of evidence have been p rovided, the speaker a llows the add ressee to cont inue with t he conversation. He is thus indicatin g his a cce ptance of the addressee’s acceptanc e, a nd collaborating to e stablish the necessary mutua l agreement on me aning for comm unication to succeed and new com mon gr ound to be built. The examp l es here analyzed ill ustrate how the in itiation of a rel e vant next contribut ion ha s be en use d f or groundin g p urpose s in NNS-NNS inte rac ti ons, involving both inte rmediate a nd a d vance d le vel learners. In Appen dix C t he re ader can also fi nd exa mples of the us e of this groundin g p roc edure by na tive spe aking i nterlocu tors in their conve rsations wit h interm ediate level students, see, for in stan ce , ADV-NS RA: 384-394; as well as with a dvanced level ones, see INT-N S PM: 246-248. The CS episodes in Appendix C indicate that this grounding procedure has been re latively frequentl y use d in the fo ur di ff erent interactional conditions here analyzed: intermediate -intermediate learner, advanced-a dvanced learner, intermediate l earner-NS and advance d learner-NS. CHAPTE R 5: The c ollaborati ve creati on of meanin g - 274 - 6 B: oh! the letter s! {(the letterbox)} 7 { A opens her AA, B is not looking } NATIVE LANG UAGE D ATA: A: están delante de una puerta con un buzón SENDER AND RECE IVER’ S RETR OSPECT IVE CO MMENTS ( 146-14 9): A: the letterbox por que no me salía B: eu puxen mailbox RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: no mailbo x Acc ording to the c omments e licited in the retrospective interview, these CSs are used to compensate f or the l ea rn er’s inability to r e trieve the TL lexical it em ‘letterbox’. This item is, howe ver, part of the addressee’s IL s ys tem . She i s a ble to rec ognize th e communicative proble m enc ountered by the speake r and to c oll aborate in i ts soluti on provi din g the originally de sired TL word. In line s 6-7 of the previous tra nscript the addressee indicates he r a cceptance of the speaker’s strategic pr esentation providin g a n acknowled gment, ‘oh!’, a repetition of the last part of t his pre se ntati on, ‘the letters’, a nd a dem onstration of wha t it is that she has understood, ‘the letterbox’. W ith t his re sponse the a ddressee ind icat es that sh e believes she has co rrectly under stood t he s pea ker’s inte nded meani ng. B ut she a lso indicates tha t she has rec ognized the communicative p roblem encounte red by the speaker and tha t she be lieves t here is a more ac curate f orm of e xpression availa ble f or the prese nt ati on o f t he intended meaning. She collaborate s in t he communicative process not only providi ng evidenc e of her understandi ng , but also rephrasing the speaker’s pre sentation to offer wha t sh e believ es to be a correct native-like TL form. By allowi ng the c onversa tion t o continue the speaker indicates that he accepts the evidence of und erstandi ng pro vided by t he add ressee. This mea ns that he h a s a lso recognized ‘l etterbox’ as a T L l exical item to express h is originally intend ed me ssage. Agreement on both meaning and form has thus been col laboratively reached, through the joi nt and co ordinated ac tions of a ll the in terlocutors ta king p art in this interacti ona l exc hang e . In the CS epi sode illustrated in the next example agreement on what is believed to be the c orrect T L lexical item to c ommunicate the originally inte nded me ssage is a lso reached . In this partic ular case a CS is uttered by the speaker when he finds out that the desire d word to c ommunicate the re feren t ‘waistcoat’ is availa ble in his I L system but is CH A PTE R 5: The collaborati ve cre ation of meaning - 275 - not sh ared with his interlocutor. T hat is, a CS is used because of the addressee’s , not the speaker’s, lac k of knowledge of the TL f orm ‘waist coat’. 65 EXAMP LE 5.12: ADV -A DV SO: waistcoat REFERE NT: vest/slee veless jumper INTERLANG UAGE DATA ( 58-62): 1 A: e::h (0.4) he has two buttons, e:h in bueno (‘well ’), (1.3) 2 {e::h he wears a:::} {a waistcoat? (4. 4)} {and he has two 3 { A’s RI points to her chest } 4 { A’s RI outlines an armhole o n her LA 5 shoulder ×2, A holds the ge sture } 6 { A’s RI points to 7 where t he buttons 8 would b e on her 9 chest × 2 } 10 A: buttons, on it ,} (2.6) a:n::d mm 11 (1.2) 12 B: a waistcoat!? (0.6) what’s that? 13 A: heh heh heh °c haleco (‘vest’)° heh 14 B: ah! (1.0) (heh )okay NATIVE LANG UAGE D ATA: A: mm un chaleco con dos botones B: hm A: eh tiene dos bols illos también en el chal eco SENDER AND RECE IVER’ S RETR OSPECT IVE CO MMENTS ( 58-62): A: estaba casi segur a ... pero como me puso cara de ... hombre yo yo es que lo dije as í toda la vida pero no B: no porque é que e u non sabía como se dicí a como se di chaleco en inglés I: mm entón cando el a di que di waistcoat ti que pensach es que era? B: eu o primeiro que pensei supuxen que que sería un un chaleco porque coat waist coat I: heh B: saqueino máis ou menos polas súas palabra s e tal pero eso ó final acabamos dicindo a palabra en gale- en c astellano A: sí en castellano ... B: eu non me acordab a de nada RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: no ‘chalec o’ In line s 2 a nd 3-5 of the transc ript, t he speake r presents th e mea ning ‘wai stcoat’. His presentation cannot howeve r be acce pted by the addressee bec ause she canno t understa nd this word. In order to solve thi s probl em and thus col laborate in the successful c ommunication of the message, t he spe ake r refashio ns his i nitial pre sentation substitu ting the item ‘waistcoat’ with a code switching st rategy : ‘cha leco’. This CS allows the addressee t o i de ntify t he spea ker ’s intende d m ea ni ng. In 14 she offe rs two ack nowledgments t hat indicate not only tha t s he ha s be en a ble t o unde rstand the 65 As c learl y stated i n C hapter Tw o, p. 131, we c onsider that CSs are use d not onl y when the s peaker is unable t o use an unknown TL w o r d, but al so whe n the y cannot use a known word be cause t his i s n ot shared with their inte rlocutor. CHAPTE R 5: The c ollaborati ve creati on of meanin g - 276 - message, but a lso that she recognizes now t he term ‘wais tcoat’ as acc urate TL le xis: ‘ah!’ and ‘okay’. The conversat ion c ontinues evidenc ing the mut ual agreem ent on both mea ni ng and form collaboratively reac hed between speaker a nd a ddressee . 66 The use of this f orm of demonstration as a grounding proc edure is mor e frequent in interacti ons i nvolv i ng native spea king int erlocuto rs. S om e represent at i ve e xamples that can be found in Appendix C are t he fol lowing: ADV-NS C S 389-3 92, ADV-NS IM: 322-325, INT-NS AE: 234-237 and INT-NS SL: 146 -149. This does not mean, however, tha t learners are not a ble to offer corrective T L le xis to t heir interlocut ors. But , as we wi ll s ee in the f ollowing p ages, whe n t he inte rmediate or advanc ed level stude nt corrects ano th er stude nt, the y usuall y modify t heir c orrection wi th a try marke r. This kind of acc eptance move will be anal yzed in the next section of the present chapter. 5.2. PROVIDING EVIDENC E O F TRO UBLE IN UNDER STANDIN G FOR GROUNDING PURPOSES IN CS EP ISODES In a ll the CS e pisodes so f ar analyzed the ad dr essee has bee n a ble to recognize the lea rn e r’s pres entation as a strategic attempt to compen s a te f or an u navailable TL ite m, to identify the meani ng this intends to c onvey an d to indicate it. But CSs c annot always be interpre ted by the interlocutor a s intended by the speake r. L e ss effe ctive CSs can lead to understa nding problems on t he part of t he add re ssee, who m ay not be a ble to identify the m essage the learner is trying to con vey or to unde rstand it to a criterion suff icient for their c urrent purposes. In face -to-face interaction interloc utors c an use a v ariety of dif ferent procedures to col laborate in the solution of these unders tanding difficul ties. Following t he work carried out within the framework of the col laborative theory of communic ation, we have identified two main categories o f procedur es t ha t addre ssees c an use to i ndicate their troubles in the interpre tation of a speake r’s CS: c larification re quests a nd con firmati on checks . The use of o ne of the se t echniques initiates a side sequence wit hin the CS episode in which the initial p resentation i s refashioned by the spea ker, the addressee or the two of t hem working to gether. This pro cess e xtends until a mutually a cc eptabl e 66 In the prese nt c hapter the main f ocus of our a nal y ses is the meani ng creati on process in strate gic communicati on, i.e. how lea rners a nd their interl ocut o r s collab o r ate to a gree on mea ning whe n the desire d TL lexical items or struc tures are n o t a vailable. Learners a nd a ddressees’ c ollabor ati on for the purposes of TL f orm c onstructi on an d the im plications tha t this proce ss ma y have on t he develop ment of t he lear ner’s IL system wi ll be anal y ze d in detail i n Chapter Six. CH A PTE R 5: The collaborati ve cre ation of meaning - 277 - version, whic h all ows them t o establis h a m ut ua l a gr ee ment on m eaning, can be reached . The c ontributions thus em erging c learly illus trate how the fi nal s uccess of strate gic communication depends not so much on the initi al strateg y pres ented by the FL lea rner, a s on the colla borative effort of thi s and t heir inte rlo c utor all al ong the CS episode. The collaborative model al so pre dicts tha t problem s ca n o cc ur wit hin the acceptanc e phase of a CS episode , even when the addre ssee believes and indicates they have c orrectly unde rstood t he strategy presente d b y t he learner. The e vidence of understa nding provided by the add ressee ma y not be rec ognized b y the spea ker, it may be cons idered insufficie nt or i nadequate , or even reveal an a ctually misundersta nding of the message . This kind of troubl e c an a lso be col laborativel y solved with the speake r’s initiatio n of wha t Cla rk a nd Wilkes- Gibbs have calle d a follow-up (Cla rk and W ilkes- Gibbs, 1986: 24). In this s ection w e will anal yz e t he r ole of all thes e dif ferent grounding procedures i n FL s trate gic interaction. We will thus t ry to iden tify and und erstan d how the collaborative e ffort of sp e aker and a ddressee makes possible the solution o f understa nding pr oblems oc curring withi n CS ep isodes , a nd t he s ubsequent succe ssfu l communication of the originally intende d message. 5.2.1. CONFIRM ATION CHE CKS AS E VIDENCE OF TROUBLE IN UNDERSTANDING FOR G ROUNDING PURPOSE S IN CS EPISOD ES The use of confirmation check s i n FL interaction, either in NNS- NS or in NNS-NNS communicative e xchanges, has be en wid ely doc ume nted within t he negotiati on of mea ni ng literature –cf. Long (1980, 1983a), Varonis and Gass (1985), Pica , Young and Doughty (1987), and Pi ca (1994 ) for some of the most represen tative research in this field. Confirmation c hecks can be d efined a s move s “i mmediately following an utterance by t he interlocutor whic h a re d e signed t o e licit c onfirmation t hat the ut teranc e has been c orrectly heard or understood by th e li ste ner” –adapted from Long (1983a: 137). Initia lly confi rmation chec ks we re studied as NSs’ re sponse s t o NNS s’ utterances a nd considered to in volve alwa ys repetition, with rising i ntonation, of a ll or part of the speaker’s precedi ng utterance ( L ong, 1983a : 137). Subse quent research has CHAPTE R 5: The c ollaborati ve creati on of meanin g - 278 - found that NNSs als o us e this negotia tion of meani ng s trate gy when they have diffic ulti e s to understand thei r native as well as non-native speakin g int erlocutors. The concept of conf irmation check has also been widened t o i nclude any mo ve f rom the addresse e, repe titive or not, that seeks to confirm heari ng and understanding of a previous uttera nce. In t he analysis of CS e pisodes we have noticed that, in order to check for confirma tion, addressees re peat sometimes all or part of th e speaker’s precedi ng strate gic utterance with rising intonat ion. Furthermore , the y can a lso paraphrase what they think may be the meaning of the CS. For this purpose , they may us e an a lternative strate gy wi th rising intona tion or what they be lieve may be t he c orrect TL f orm of the speaker’s non-na tive-like s trate gic utterance. In CS episode s a re quest for conf irmation can thus also contain implicit c orrective fee dback and initiate a proce ss o f simul taneous negotiation of both m eaning an d f orm. All these diffe rent uses of c onfirmatio n checks will be h e re ana lyzed and exemplifie d. Notice also th a t c onfirmation ch ec ks are considere d to pro jec t a positive verdict. This mea ns that they are usually answ erable by a simp le confi rmation, an ack nowledgment such as ‘yeah’ or a head nod, and require l it tle effort f rom the interlocutor in te rms of production. On this basis we can a rg ue t hat a confirmation check is gen erally use d by the add ressee w hen they ha ve a cert ain confi dence i n the poss ibl e accuracy of t heir interpretation o f the speaker’s preceding CS. Unde rstanding needs to be seen as a cont inuum, ra the r tha n a yes or no issue. Wha t an addressee indica tes in their ac ceptance move is their level of unde rstandi ng and whethe r they consi der it sufficien t for their and their interlocut ors’ common purposes. In most of t he C S e pisodes we wi ll a nalyze he re t he c onfirmation check p rov ide d by the addresse e reveals that in fact they ha ve corre ctly u nderstood the s peaker’s CS. The y j ust do not consider t heir level of understandi ng high enough for the sa tisfactory grounding of the me ssage in this interactional co ntext. 5.2.1.1. Repetition as e vidence of trouble in unde rstanding f or groundin g purposes in C S episodes In the CS epis ode presented in exa mple 5. 13, with t he re petition of t he CS j ust uttered by the spea ker, the addressee indicates s he is having som e trouble to under stand th e CH A PTE R 5: The collaborati ve cre ation of meaning - 279 - lea rn e r’s intende d message. She needs some further evidenc e in order to be able to accept this strat egic presentation. The s peaker, A, uses in the pre s e ntation of t he c ontent he wan ts to contribute to discourse the IL ite m ‘d iamonds’ –cf . l ine 1 of the t ranscript. Th is is a n approxim ation CS to c ompens ate for the u navaila bility of the originally intended TL lexi cal i te m ‘rhombus’. EXAMP LE 5.13: ADV -A DV SO: rh ombus REFERE NT: chec kered fl oor/ (tiles) INTERLANG UAGE DATA ( 193-197): 1 A: a:n:::d there a::re e:::h (2.0) diamon ds? {on the floor, (1.5) 2 { A waves her RH palm 3 down ×n } 4 A: so:me black an d some white,} 5 (0.8) 6 B: diamonds? 7 (0.4) 8 A: yeah, (0.7) e: :h (0.4) {°rombos (‘rhom buses’).° he h=} 9 { A’s RI draws a rhombus sha pe in the 10 air } 11 B: =oh! (2.2) dia monds, sí (‘yes’). NATIVE LANG UAGE D ATA: A: el suelo pues tie ne rombos unos blancos y otros negro s SENDER AND RECE IVER’ S RETR OSPECT IVE CO MMENTS ( 193-19 7): B: eu aí entendín qu e había diamantes e dixe n onde están os diamantes porque nonos vía hai diamantes de verdade? e despois claro fixeime din me de conta da forma ... I: mhm y diamond lo dices? A: como rombo pude decir cuadrados pero me salió eso he h I: cuando ella no te entiende lo de diamonds A: pude haberle dich o otra cosa lo dije en e spañol porqu e no porque no lo pensé no sé RECEI V ER’S REFE RENT IDE NTIFIC ATION TASK: nothing pa inted in t he flo or The a ddressee has heard wha t t he speaker has said and recognized the expressi on he has used, bu t she is not sure about the e xact meaning he i s trying t o convey. She res ponds t o the spea ker’s CS with a co nfirmation check t hat indi ca tes her state of understanding: ‘diamonds? ’. Notice that i n th is confirmation check the addr e ssee does not repeat all the precedi ng prese ntation, but only t hat part she i s not s ure about, i .e. t he CS. She i s thus presupposing she has understood e very t hing but the te rm ‘d iamonds ’. Accord ing t o Cla rk and Schaefe r (1987: 22) , the addre ssee’s ac ceptance mo ve shows she is in diffe r ent s t ate s of understandin g for different cons tituents of t he presentation. Furthermore, by signaling exa ctl y whe re the t r ouble i s, she is c ollaborating with he r [Document text truncated for crawler view.]