Microsoft Word - 6.Porter_7.1.docx EuroAmerican  Journal  of  Applied  Linguistics  and  Languages   Special  Issue,  Volume  7,  Issue  1,  February  2020,  78-­‐107   ISSN  2376-­‐905X     DOI  http://dx.doi.org/10.21283/2376905X.11.187   www.e-­‐journall.org     © Porter 2020. This work is licensed under a Creative Commons Attribution 4.0 International License.       Learning  French  sound/spelling  links     in  English  primary  school  classrooms         ALISON  PORTER   University  of  Southampton         Received  1  May  2019;  received  in  revised  form  12  July  2019;  accepted  30  September  2019   ABSTRACT EN Two classes (n = 45) of 9- to 11-year-old children in different English primary schools experienced weekly French phonics instruction as part of a foreign language (FL) learning programme. The underpinning rationale of the phonics instruction was that systematic and explicit FL phonics could contribute to the development of FL phonological decoding, operationalised as reading aloud individual word cards. The study explored progression in learning French sound/spelling links over this period with data from a sub-sample of students (n = 23) from both classes. Additional variables such as FL general proficiency and FL literacy, as well as L1 reading age, spelling age, and teacher assessed literacy levels were collected to explore relationships with the development of French phonological decoding. This study found that French sound/spelling links developed slowly but showed significant gains between mid- and post-test and that learning appeared resistant to attrition evidenced by delayed post-test scores. Successful phonological decoders were likely to be successful FL learners generally and proficient in L1 literacy. Key words: YOUNG FL LEARNERS, EARLY FL LITERACY INSTRUCTION, FL PHONICS ES Dos clases (n = 45) de niños de 9-11 años en diferentes escuelas primarias de inglés experimentaron semanalmente instrucción fónica francesa como parte de un programa de una lengua extranjera (LE). Este modelo pedagógico se implementó porque la instrucción fonética sistemática y explícita de una LE podría contribuir al desarrollo de la decodificación fonológica en la Le, esta última medida en la lectura en voz alta de tarjetas de una sola palabra. El estudio observó la progresión en el aprendizaje de sonido/ortografía con datos de una submuestra de estudiantes (n = 23) de ambas clases. Otras variables fueron consideradas para observar su correlación con el desarrollo de la decodificación fonológica en francés: el nivel de competencia lingüística y alfabetización en la LE, el nivel de lectura en la primera lengua, el nivel de ortografía y los niveles de alfabetización del profesor. Con este estudio se descubrió que la relación sonido-grafía se desarrolló despacio, pero con resultados positivos entre las pruebas durante y después de la instrucción. Además, el aprendizaje se resistió al fenómeno de atrición, como mostraron los resultados de las segundas pruebas después de la instrucción. Los que decodificaban los aspectos fonológicos con éxito tenían más probabilidades de aprender una LE con éxito y con una competencia avanzada en su lengua materna. Palabras clave: JÓVENES ESTUDIANTES DE LE, APRENDIZAJE TEMPRANO DE UNA LE, INSTRUCCIÓN FÓNICA EN LA LE IT Come parte di un programma per l’apprendimento di una lingua straniera (LS), due classi (n = 45) di bambini tra i 9 e gli 11 anni in due scuole primarie inglesi hanno sperimentato settimanalmente lezioni di francese con il metodo fonetico. L’ipotesi alla base del metodo è che l’istruzione fonetica sistematica ed esplicita in LS contribuisca allo sviluppo della decodificazione fonologica in LS espressa dalla lettura ad alta voce di schede con una singola parola. Lo studio analizza la progressione nell’apprendimento della corrispondenza fonema/grafema attraverso i dati di un sotto-campione di studenti (n = 23) ins entrambe le classi. Sono stati raccolti dati aggiuntivi relativi a variabili quali la competenza complessiva e l’alfabetizzazione in LS, insieme all’età di lettura in L1, l’età di scrittura e il livello di alfabetizzazione, stabilito dall’insegnante, al fine di analizzarne il rapporto con lo sviluppo della decodificazione fonologica in francese. Da questo studio emerge che le corrispondenze fonema/grafema in francese si sviluppano lentamente ma evidenziano progressi significativi tra il test in itinere e il post-test e che quanto appreso pare resistere all’attrito come dimostrato dai punteggi del post-test differito. Inoltre, è risultato che apprendenti abili nella decodificazione fonologica avessero un’alta probabilità di essere apprendenti di LS complessivamente competenti e con un alto livello di alfabetizzazione in L1. Parole chiave: GIOVANI APPRENDENTI DI LS, ALFABETIZZAZIONE IN UNA LS, METODO FONETICO IN UNA LS *  Alison  Porter,  University  of  Southampton                  A.Porter@soton.ac.uk LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 79   1.  Literature  review   1.1.  The  importance  of  FL  literacy  in  England       Foreign   language   study   is   now   compulsory   in   English   primary   schools   and   was   included   in   the   Primary   National   Curriculum   (DfE,   2013)   from   Key   Stage   2   (ages   7-­‐11).   The   curriculum   aims   to   develop   practical   communication   skills   in   one   language,   to   create   foundational   language   skills   as   preparation   for   ongoing   language   learning   and   to   develop   cultural   and   intercultural   awareness.   It   includes   FL   literacy   (reading  and  writing)  and  requires  children  to  “discover  and  develop  an  appreciation  of  a  range  of  writing”   (p.   193),   as  well   as   practical   skills   such   as:   understanding   and   responding   to  written,   authentic   text   (DfE,   2013).  There   are   specific   attainment   targets  which   relate   to   reading   for  meaning   and   reading   aloud  which   include  the  ability  to  link  sounds,  spellings  and  word  meaning.       The   importance   of   FL   literacy   as   a   component   of   the   development   of   practical   language   skills   and   cultural   awareness   is  widely   accepted   (Koutsompou,   2015).   However,   little   is   known   about   the   efficacy   of   particular  approaches   to   teaching   literacy   in  FL   school   settings   such  as   in   the  UK.  Research  has   found   that   reading   strategy   instruction   supported   reading   comprehension   and   FL   motivation   in   English   secondary   schools   (Macaro  &  Erler,  2008)  and  reading  comprehension   in  English  primary  schools   (Macaro  &  Mutton,   2009).  A  recent,   large-­‐scale  project  in  secondary  schools  examined  to  what  extent  three  different  pedagogic   approaches—phonics  instruction,  strategy  instruction,  or  text  experience  without  explicit  instruction—could   contribute  to  FL  (French)  linguistic  and  non-­‐linguistic  outcomes.  All  approaches  noted  statistically  significant   improvements   in  FL  reading  aloud,  FL  vocabulary,  and  FL  reading  comprehension  (Woore,  Graham,  Porter,   Courtney,  &  Savory,  2018).  However,   the  phonics   group  made  more  progress   at  FL   reading  aloud   than   the   other   conditions,   and   the   strategies   group   improved  most   for   self-­‐efficacy   in  French   reading   (Woore  et   al.,   2018).         Learning  FL   sound/spelling   links   in  English   schools   is   noted   to  develop   slowly   (Cable   et   al.,   2010;   Woore,   2011)   and   unsystematically   (Woore,   2011).   French   reading   aloud  was   particularly   problematic   for   primary   school   learners,  who   tended   to   be   influenced   by   knowledge   and   application   of   L1   sound/spelling   links   (Cable   et   al.,   2010;   Porter,   2014).   Similarly,   in   secondary   schools,  Woore   (2007)   found   that   after   18   phonics   lessons   only   the   two   strongest   L1   readers   made   successful,   unscaffolded   analogies   to   single   graphemes.  He  also  noted  that  the  weaker  L1  readers  tended  to  use  larger  syllabic  units  and  attributed  this  to   less  well-­‐developed  phonemic  awareness.  In  other  words,  a  period  of  focused  FL  phonics  instruction  did  not   appear  to  compensate  for  the  influence  of  L1  knowledge  and  skills.       As  FL   reading  and  writing  are   important   in   the  development  of  FL   skills   and  attitudes   towards  FL   learning,   it   is  essential   that  research  continues  to   investigate  the  effectiveness  of  particular  pedagogic   tools   and   the  process  of   learning   literacy   in  FL   classrooms.  Phonics   is  often  practised,   as   a  pedagogic   tool,   in  FL   settings  based  on   assumptions   grounded   in   the  popularity   of   phonics   instruction   in  L1  English   classrooms   and  evidence   concerning   its   contribution   to  word   recognition   skills   in   alphabetic   languages   (Adams,  1990;   Castles,   Rastle   &   Nation,   2018).   From   a   process   perspective,   whilst   L2   reading   has   been   researched   and   theorised   with   bilingual   and   multilingual   populations,   there   is   little   evidence   about   the   processes   or   outcomes   for   learning   to   read   in  FL   settings.  Again,   assumptions   cannot  be  made   that  pedagogic   tools   and   learning   processes   identified   in   L1   and   L2   settings  will   be   readily   transferable   to   beginner   FL   classrooms,   known  to   lack  the  rich  and  regular   input  which  supports   learning  in  L1  and  L2  settings  (Garton,  Copland  &   Burns,  2011;  Graham  et  al.,  2017).       1.2.  Learning  to  read  in  a  first  language  (L1):  the  importance  of  word  recognition       It  is  now  widely  accepted  that  rapid  and  accurate  word  recognition  is  a  fundamental  skill  in  reading   for  meaning  (Adams,  1990;  Oakhill  &  Beard,  1999;  Stanovich  &  Stanovich,  1999;  Vellutino  &  Fletcher,  2005).   Word  recognition  is  accessed  by  either  a)  decoding  spellings  into  sounds  then  retrieving  meaning  (sub-­‐lexical   route)  or  b)  retrieving  meaning  directly  from  spellings  (lexical  route)  (Castles,  Rastle,  &  Nation,  2018).       1.2.1.  The  process  of  developing  L1  word  recognition       Phonological   decoding,   involving   the   mapping   of   letters   or   groups   of   letters   onto   corresponding   sounds  (Ziegler,  Perry,  &  Zorzi,  2014),  is  essential  in  word  recognition.  Sight  word  recognition  is  believed  to   develop   in  stages  or  “overlapping  waves”  of  phonological  decoding,  starting  with  a  partial  alphabetic  phase   where  connections  are  made  between  a  few  sounds  and  their  corresponding  letters,  usually  initial  and  final   consonants  (Ehri,  2014,  p.  5;  Ehri,  1999).  This  development  is  accompanied  by  spelling  development  which   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 80   starts   with   written   words   generated   from   a   relatively   small   number   of   learned   letters   e.g.   giraffe   =   JRF   (Castles   et   al.,   2018).  Development   of   links   between   sound   and  print   is   known   as   “orthographic  mapping”   (Ehri,  2014:  5)  or  “orthographic  learning”  and  comprises  both  word-­‐specific  links  between  sound  and  print,   along   with   awareness   of   the   probabilistic   occurrence   of   letters   and   letter   combinations   in   a   particular   language  (Castles  et  al.,  2018,  p.  19).  Eventually  “spellings,  pronunciations  and  meanings  of  specific  words  in   memory”  are  bound  together  and  written  word  meaning  is  accessed  directly  in  memory  rather  than  indirectly   through  decoding  (Ehri,  1999,  pp.  88-­‐96;  Ehri  1992,  p.  108).       1.2.2.  L1  phonological  decoding  and  word  recognition       Phonological   decoding   is   also   believed   to   influence   how   representations   of   written   words   are   formed/stored   in   memory,   and   it   remains   accessible   even   in   skilled   readers   (Castles   et   al.,   2018).   Share   (1999)   suggests   that   phonological   decoding   is   “item-­‐based”   or   depends   on   frequency   of   exposure   to,   and   active   decoding   of,   written   forms   (p.   96).   In   other   words,   phonological   decoding   supports   the   learning   of   written  forms  of  words  throughout  all  stages  of  reading  development  (The  Self  Teaching  Hypothesis;  Share,   1999).  Phonological  decoding  of  embedded  non-­‐words  whilst  reading  connected  texts  was  found  to  support   children’s   word   recognition   and   spelling   (Share,   1999;   Ricketts,   Bishop,   Pimperton,   &   Nation,   2011).   The   process  of  actively  decoding  the  written  form  seems  to  lead  to  improved  “lexical  quality”  and  the  “precise”  yet   “flexible”   mental   storage   of   written   forms   (Castles   et   al.,   2018,   p.   21).   Kaefer   (2016)   too   posits   that   phonological  decoding  activates  explicit  awareness  of  orthographic  knowledge  (spelling,  word  recognition)  in   early  readers  (mean  age  5  years  10  months).       1.3.  Learning  to  read  in  a  first  language  (L1):  Skills  that  underpin  phonological  decoding       In   an   L1,   word   recognition   involves   “turning   listeners   into   readers”   (Stuart,   Masterson,   &   Dixon,   1999,  p.  100),  and  the  phonological  decoding  processes  outlined  previously  are  predicated  on  learned  skills   involving   “lower   level   verbal   processing   mechanisms”   (Koda,   1992,   pp.   52-­‐57)   such   as   phonological   awareness,   phonological   recoding,   and   verbal   working   memory   (McBride   Chang   &   Ho,   2005,   p.   119).   Phonological  awareness   is   “one  of   the  best  predictors  of   literacy  outcomes”  (Henbest  &  Apel,  2017,  p.  304)   and   involves   the   ability   to   segment   and   manipulate   spoken   words   (Verhoeven,   2011).   This   sensitivity   involves  skills  such  as  syllable  identification,  the  ability  to  manipulate  (delete/exchange)  phonemes  in  spoken   words,  and  sensitivity   to  onset-­‐rime,  or  sounds  before  and  after   the  nucleus  of  a  syllable  (Antony,  Lonigan,   Driscoll,  Phillips,  Burgess,  &  Cantor,  2002,  pp.  68,  87).  These  skills  are  deemed  essential  in  recognising  how   sounds   map   onto   their   respective   letters   (Ehri,   1999,   p.   102).   Morphological   awareness   or   the   ability   to   identify,   reflect   on,   and   manipulate   the   smallest   units   of   meaning   are   also   an   important   factor   in   word   recognition,   spelling,   and   reading   comprehension   (Apel,   Wilson-­‐Fowler,   Brimo,   &   Perrin,   2012;   Deacon   &   Kirby,  2004).     1.4.  Teaching  word  recognition:  The  role  of  phonics  in  L1  English  reading  instruction       Phonics   instruction   is   currently   practised   in  many   English   L1   school   education   contexts   (England,   USA,   Canada,   Australia,   New   Zealand)   as   a   means   of   developing   foundational   reading   skills   through   phonological   decoding   (Department  of  Education,   Science  &  Training,   2005;  National  Reading  Panel,   2000;   The  Rose  Review,  2006).  Whilst  phonics-­‐only  approaches  to  teaching  early  reading  remain  contentious  (Wyse   &   Styles,   2007),   there   is   evidence   that   teaching   approaches   which   combine   phonics   or   code   instruction   (sound/letter  links)  with  meaning-­‐focused  activities  (understanding  text)  resulted  in  faster  reading  progress   in  primary  school  children  (Adams,  1990;  Chall,  1967).  Meta-­‐analyses  of  experimental  research  have   found   that   systematic   phonics   instruction   which   promotes   the   learning   of   sound/letter   links   in   targeted   and   sequenced   programmes,   had   a   statistically   significant   effect   on   reading   accuracy   (Ehri,   Nunes,   Stahl,   &   Willows,   2001;   Torgerson,   Brooks,   &   Hall,   2006),   reading   comprehension   (Ehri   et   al.,   2001),   and   spelling   (Ehri  et  al.,  2001).         Commercial   phonics   packages   such   as   Jolly   Phonics   (Lloyd  &  Wernham,   1992)   or   Read  Write   Inc.   Phonics   (Miskin,  2011)  promote  pedagogic   routines   involving  production  of   isolated  phonemes   to  segment   sounds  in  spoken  words,  picture  mnemonics  to  memorise  letters  and  exemplar  words  which  model  the  target   sound/spelling   links.   These   activities   are   supported   by   research   which   showed   that   articulatory   gestures   supported  memorisation  of  sound/spelling  links  and  word  reading  performance  when  coupled  with  practice   in   segmenting  words   (Boyer  &  Ehri,   2011)   and   that   letter  names   supported  by  picture  mnemonics   lead   to   faster   learning   of   sound/spelling   links   (Shmidman  &  Ehri,   2010).  However,   as   phonics   instruction   “instils”   LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 81   rather   than   “installs”   the   code   (Gough   &  Wren,   1999,   p.   75),   it   is   essential   that   phonics-­‐related   skills   are   operationalised   through   reading   aloud   which   involves   applying   sound/spelling   links   knowledge   (Hatcher,   Hulme,   &   Ellis,   1994).   Recall   that   this   is   likely   to   lead   to   fine-­‐quality   lexical   representations   stored   in   the   learner’s  memory  (Castles  et  al.,  2018).  It  is  further  argued  that  frequent  and  diverse  read  aloud  texts  support   word  recall  and  new  word  spelling  (Rosenthal  &  Ehri,  2011).         To   conclude,   systematic   phonics   instruction   supports   word   recognition   and   is   therefore   recommended  by  researchers  as  a  valuable  pedagogic  tool  in  L1  literacy  development  (Castles  et  al.,  2018).  It   is   also   important   to   note   that,   as   the   ultimate   goal   of   any   reading   instruction   programme   is   reading   comprehension,   secure   and   automatized   word   recognition   is   only   one   part   of   a   complex   developmental   process.   Other   factors   likely   to   influence   reading   comprehension   outcomes   include   oral   language,   parsing,   inferring,   comprehension  monitoring,  exposure   to  print,  morphological  awareness,   reading  motivation,  and   cognitive  resources  such  as  working  memory  (Castles  et  al.,  2018;  Perfetti  &  Stafura,  2014).       1.5.  Learning  to  read  in  an  L2   1.5.1.  A  repeated  process  –  shared,  underlying  skills  and  knowledge       Learning  to  read  in  a  second  or  additional  language  is  a  “repeated  process”  (Koda,  2008,  p.  74)  and   involves  a  constantly  active  “two  language  processing  system”  (Grabe  &  Stoller,  2002,  p.  35).  Research  in  bi-­‐ /multi-­‐lingual   settings  has  determined   that   some   foundational   L1  decoding   related   skills   are   evident   in  L2   literacy  development.  L1  phonological  awareness  supported  children’s  L2  reading  development  (Durgunoğlu,   Nagy,  &  Hancin-­‐Bhatt,  1993)  and  predicted  L2  word  reading  skills  (Lesaux  et  al.,  2008).  Opinions  vary  as  to   whether  these  L1  resources  are  available  in  the  early  stages  of  L2  oral  development  (Genesee,  Geva,  Dressler   &   Kamil,   2008,   p.   68)   or   whether   a   particular   level   of   L2   competence   is   required   (Interdependence   Hypothesis;  Cummins,  1979).         However,  it   is  important  to  note  that  differences  in  orthographies  and  syllabic  structures  as  well  as   participant  age  are  likely  to  play  a  role  in  cross-­‐linguistic  transfer  of  L1  literacy  skills  (Genesee  et  al.,  2008).   For  example,  Goswami  (2006)  posits   that  readers  of  L1  English   learn  to  map  sounds  with   letters   in  chunks   (Psycholinguistic  Grain  Size  Theory)  due   to   the   consistency  of  English  at   rime   level.  On   the  other  hand,  L1   readers  of  more  transparent  orthographies  such  Spanish,  with  consistent  1:1  mappings  between  sound  and   print,  will  have  recourse  first  to  decoding  at  the  individual  phoneme  level  (Landerl,  2006,  p.  518).         Interestingly,   in   FL   settings,   a   bi-­‐directional   influence   for   foundational   L1   literacy   skills   has   been   noted.  Clusters  of  L1  skills  such  as  spelling,  word  decoding,  phonological  awareness,  vocabulary,  and  reading   comprehension   influenced   FL   attainment   and   aptitude   in   high   school   (Sparks   &   Ganschow,   2012;   Sparks,   Patton,  Ganschow,  &  Humbach,  2009).  These  researchers  had  earlier  identified  that  similar  L1  skills  acted  as  a   foundation  for  L2  proficiency,  arguing  that  weak  phonological  processing  could  negatively  influence  both  L1   and  L2   learning   (Linguistic   Coding  Deficit  Hypothesis;   Sparks  &  Ganschow,  1991).  Research   in   younger  FL   classrooms  has  evidenced  bi-­‐directionality  of  phonological  awareness  influence,  finding  that  L2  phonological   awareness   can   also   support   the  development   of   L1  phonological   awareness   in   learners   aged  7-­‐8   (Murphy,   Macaro,  Alba,  &  Cipolla,  2014).     1.5.2.  A  different  process-­‐  new  sounds  and  new  mappings       Word   recognition   and   phonological   decoding   rests   upon   phonological   awareness   (deemed   transferable)   but   also   on   orthographic   mapping   or   the   development   of   links   between   sounds   and   their   respective   letters   (Ehri,   2014).   In   an   L1,   these   sounds   are   already   secure   and   just   need   to   be  mapped   to   letters.  However,  in  an  L2  or  FL,  sounds  and  letters  will  either  need  to  be  learned  “from  scratch”  (i.e.,  new  FL   sounds  linked  to  new  FL  letter/s)  or  re-­‐mapped  (i.e.  existing  letters/sounds  linked  to  different  counterparts).         In   terms   of   remapping,   Koda   (2007)   argues   for   a   process   where   L1   knowledge   and   skills   are   transformed  into  additional  L2  competencies  through  exposure  to  L2  print.  However,  this  seems  problematic   bearing  in  mind  that  L1  readers’  knowledge  and  decoding  skills  are  likely  to  have  transformed  from  effortful   to  automatized  through  the  process  of   learning  L1  reading  (Castles  et  al.,  2018;  Ehri,  2014;  1999).  The  task   for  learners  of  L2/FL  re-­‐mappings  will  likely  involve  adapting  automatized  L1  sound/spelling  mappings,  and   automatized  cognitive  language  learning  mechanisms  are  believed  to  be  particularly  resistant  to  adaptation   (Mitchell,  Myles,  &  Marsden,  2019).         Although   there   is   little   evidence   concerning   the   learning   of   novel   phonology   in   FL   settings,   L2   phonological  acquisition  has  been  extensively  investigated.  Whilst  recognising  important  differences  between   L2   and   FL   contexts   in   terms   of   input   quality,   extent,   and   intensity,   theories   relating   to   L2   phonology   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 82   acquisition  could  give  some  insight  into  the  process  of  learning  novel  sounds  when  an  L1  phonological  system   is  well  established.       Opinions  differ   regarding  ease  of  acquisition  of  L2  phonology.  For  some,  novel   sounds  are  deemed   harder   to   learn   either   due   to   entrenched   L1   phonological   representations   or   because   they   require   the   creation  of  new  L2  sound  categories  (Escudero  &  Boersma,  2004;  Lado,  1957).  Others  assert  that  dissimilar   sounds  are  easier  to   learn  as  they  are  more  likely  to  be  perceived  and  therefore  result   in  “new  L2  phonetic   categories”   (Speech  Learning  Model;   Flege,   1995).   Crucially,   L2   sounds  will   differ   from  monolingual   sound   acquisition  as  L2  sounds  will  interact  with  the  L1  and  presumably  subsequent,  additional  languages  over  an   individual’s  lifespan  (Flege,  1995;  Nimz,  2018).         For   example,   French   nasal   sounds,   for   example,   vowels   (often   novel   for   learners)   have   proved   problematic  showing  variation  across  advanced  learners  and  particular  types  of  nasal  vowels:    /ɔ̄/was   the  most  accurately  produced  but    /ɛ̄/  was  the  least  accurately  produced  and  tended  to  be  substituted  by    /ɑ̄/  (Adeline  Charlton,  2014;  Neveu  et  al.,  2010).  Foreign  language  instruction  seems  to  assist  the   development   of   novel   French   nasal   vowels   and   shifts   in   phonological   representations   occur   between   proficiency   levels   (Martinez,   2016).   Interestingly,   once   learners   are   literate   in   their   L1,   orthographic   knowledge  influences  the  perception  and  production  of  L2  phonology  (Bassetti  &  Atkinson,  2015;  Escudero  &   Wanrooij,  2010;  Nimz,  2018).  L1  orthography  is  indiscriminately  activated  when  written  cues  are  involved  in   sound  perception  and  this  has  an  ambivalent  effect  on  perception  (Escudero  &  Wanrooij,  2010).  It  should  also   be   noted   that   other   factors,   such   as   L1   syllable   structure,   L1   phonotactics,   and   individual   differences,   influence  L2  sound  perception,  particularly  of  vowels  (Tyler,  Best,  Faber,  &  Levitt,  2014).       1.6.  Teaching  FL/L2  reading        Bearing  in  mind  issues  such  as  transfer  of  L1  related  skills  and  competencies,  the  instructional  task   for   teachers   differs   from   that   of   teaching   children   to   read   in   an   L1.   Assumptions   should   not   be  made   that   pedagogic  tools  and  approaches  relevant  in  the  L1  will  automatically  have  the  same,  or  indeed  any,  effect  on   FL  development.       Clearly  when   learning   to   read   in  a  second  or  subsequent   language,   the   learner   is  not  a   tabula   rasa   (Porter,   2014).   Firstly,   a   learner   approaches   L1   reading   with   a   fairly   extensive   L1   spoken   vocabulary,   internalised  L1  grammar,  and  L1  phonological   inventory.  The  6-­‐7  year  old   (grade  2)  native  speaker  has  an   average  L1  root  word  vocabulary  of  5200  words,  rising  to  8400  words  by  age  11  (Biemiller  &  Slonim,  2001).   In  instructed  L2/FL  contexts,  the  starting  point  is  somewhat  different.  Often,  literacy  instruction  in  the  FL/L2   languages  classroom  is  initiated  after  the  development  of  oral  skills  and  it  is  suggested  that  L2  phonological   decoding   has   limited   effectiveness   unless   the   oral   representation   of   the   word   is   already   known   (Grabe   &   Stoller,   2002,   p.   42).   However,   in   FL   settings   where   instructional   time   is   limited,   waiting   for   the   right   or   optimum  time  to  introduce  decoding  and  focused  literacy  instruction  may  not  be  a  pedagogical  option.  On  the   other  hand,  it  appears  as  though  the  essential  foundational  literacy  skills  of  phonological  awareness  may  be   available  to  learners  if  they  are  already  developed  in  the  L1.  It  is  also  likely  that  increasing  cognitive  maturity   of  children  in  the  later  stages  of  primary  school  (when  FL  instruction  generally  commences)  may  assist  in  the   teaching  and  learning  of  FL  literacy.       To   summarise,  word   recognition   is   a  key   skill   in  developing  FL   literacy  and   is  underpinned  by   the   knowledge  and  application  of  FL  sound/spelling  links.  Research  indicates  that  teaching  FL  word  recognition   is  likely  to  be  a  shared  and  yet  different  process  to  its  development  in  the  L1.  Already-­‐literate  learners  could   bring  transferable  skills  to  the  task  of  FL  word  recognition,  such  as  phonological  awareness,  relative  cognitive   maturity,   and   shared   sound/spelling   links.   However,   some   L1   skills—sound   perception,   knowledge   of   FL   sound/spelling   links,   and  FL   oral   vocabulary—are   likely   to   complicate   the   learning  process.   It   is   therefore   important   that,   in   primary   FL   settings,   research   is   conducted   to   explore   the   potential   effectiveness   of   L1   literacy  pedagogic  practices  and  curricular  expectations.     2.  Method       The  study  set  out  to  investigate  whether/how  children  studying  French  as  a  new  FL  made  progress   at  reading  aloud  FL  word  cards  during  and  after  a  23-­‐week  teaching  programme  which  comprised  systematic   and  explicit   phonics   instruction   in  French   (research  question  1).  The   test  measures   also   aimed   to   contrast   how  successfully   the   children   read  aloud  known  (familiar)  words   that   they  had  practised  every  week  with   how   they   decoded   novel   (unfamiliar)   words   (research   question   2).   Lastly,   the   study   aimed   to   evaluate   LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 83   whether  there  were  any  relationships  between  L1  literacy  measures,  FL  proficiency  scores,  and  reading  aloud   performances  at  each  test  time  (research  question  3).  These  analyses  would  account  for  influential  factors  in   the  learning  process  and  therefore  possibly  explain  some  of  the  variation  in  FL  reading  aloud  outcomes.     -­‐ RQ1:   To   what   extent   do   learners   engaged   in   systematic   and   explicit   FL   phonics   instruction  make   progress  in  reading  aloud  familiar  and  unfamiliar  FL  words  at  mid-­‐,  post-­‐,  and  delayed  post-­‐test?   -­‐ RQ2:  Are  performances  on  reading  aloud  familiar  and  unfamiliar  word  cards  different  at  mid-­‐,  post-­‐,   and  delayed  post-­‐test?   -­‐ RQ3:  To  what  extent  are  L1  literacy  measures  and  FL  global  proficiency  scores  related  to  FL  reading   aloud  proficiency  at  each  test  time?     2.1.  Sample       This  study  took  place  in  two  English  primary  schools.  The  total  sample  (n  =  45)  was  tested  at  three   main   test   points:   pre-­‐,   post-­‐,   and   delayed   post-­‐test   on   reading   aloud   in   sentences.   The   focus   of   this   paper   concerns  a  sub-­‐sample  of  students  (n  =  23)  who  were  tested  on  all  the  main  measures  plus  word  card  reading   aloud  at  mid-­‐test  (approximately  10  teaching  weeks  after  the  phonics  instruction  commenced),  post-­‐test  (23   weeks  after),   and  delayed  post-­‐test   (7  weeks  after   the   teaching  programme   finished).  The  mean  age  of   the   sub-­‐sample  was   10.07   years   (minimum   9.10,  maximum   11.0)   and   the   children  were   taught   in  mixed-­‐year   classes  (England  Years  5  &  6;  US  Grades  4  &  5).  All  of  the  children  spoke  English  as  their  L1.  Although  some   had  exposure   to  other   languages  at  home  (e.g.  Nepalese,  Arabic,  Dutch),   these  were  not  spoken  as  primary   home   languages.   Both   schools   were   not   in   areas   of   social   deprivation   and   were   situated   in   small,   rural   villages.  The  FL  provision  in  both  settings  was  identical:  the  same  French  teacher  (the  researcher)  taught  one   50-­‐minute  FL  lesson  per  week.  This  teacher/researcher  had  worked  in  School  1  for  2  years  and  School  2  for  6   years  prior  to  the  research.     It  should  be  noted  that  the  mixed-­‐year  class   in  School  1,   in  particular  those  children  in  Year  6,  had   had   one   year   less   cumulative   instruction   than   the  Year   6   children   in   School   2.  No   children   had  previously   received   systematic   and   explicit   French   phonics   instruction,   and   FL   literacy   generally   had   not   been   given   special   or   sustained   emphasis.  Nonetheless,   there  was   only   one   significant   difference   in   FL  pre-­‐test   scores   between   schools,   which   concerned   FL   reading   aloud   scores.   Parents   and   guardians   were   approached   for   informed  consent,  and  the  children’s  verbal  consent  was  sought  in  order  to  conform  to  ethical  requirements   for   school-­‐based   teaching  and   learning   research.  This  was  an  action   research  study  so   there  was   flexibility   within   the   design   to   adapt   and   amend   teaching   practices.   However,   in   order   to   support   robustness   and   generalizability/transferability,   the  principles  underpinning  the  overarching  teaching  programme  (of  which   FL  phonics  instructed  formed  a  component)  and  those  principles  underpinning  the  specific  strand  of  phonics   instruction  are  outlined  below.       2.2.  Pedagogic  principles  for  literacy  instruction       This  small-­‐scale  study  aimed  to  build  a  phonics   instruction  programme  which  could  assist  primary   school   learners  of  French  as  a  FL,  with  extremely   limited   input,   to  start   to  develop  FL  sound/spelling   links.   The  teaching  programme  included  a  weekly  phonics  component  which  formed  part  of  lessons  which  aimed  to   develop   both   FL   oracy   and   literacy   skills.   The   phonics   instruction   specifically   had   three   overarching   principles  based  on  the  extant  literature  regarding  L1  phonics  instruction.     2.2.1.  Principle  1:  Focus  on  FL  sounds         First,   it   was   crucial   to   develop   distinctive   FL   sounds   (phonology)   as   phonics   instruction   involves   linking  known  sounds  to  letters.  This  task  required  the  development  of  distinctive  new  or  adapted  phonemes   such  as  nasal  vowels.  Rather  than  aiming  for  native-­‐like  pronunciation,  distinctive  FL  sounds  were  evidenced   by  sounds  which  were  clearly  differentiated  in  production  from  their  closest  L1  counterparts.         There   are   particular   French   phonemes  which   do   not   have   exact   counterparts   in   the   L1   (e.g.   nasal   vowels  /ɔ̄/,  /ɑ̄/,  /ɛ̄/,  /ɶ̄/).  Furthermore,  some  French  phonemes  have  similar  English  counterparts  and  could   therefore   be   difficult   to   discriminate   (e.g.   French   front   rounded   vowel   /y/   versus   the   similar   English   back   rounded   vowel   English   /u:/).   Bearing   in   mind   the   L2   phonological   development   literature,   the   pedagogic   programme  therefore  comprised  a  small  element  which  focused  on  FL  sound  perception  (discrimination)  and   production.       PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 84   2.2.2.  Principle  2:  Systematic  and  explicit  phonics  instruction       In  line  with  commercial  phonics  packages  (e.g.,  Read  Write  Inc.  or  Jolly  Phonics)  sounds  and  letters   were  taught  explicitly  and  in  a  planned  order  (Table  1).  Less  attention  was  given  to  phonological  awareness   such  as  the  ability  to  segment  spoken  words  into  sub-­‐lexical  units—individual  phonemes  and/or  onset-­‐rime   or  syllables—as  this  knowledge  was  assumed  to  transfer  from  the  L1  (e.g.,  Genesee  et  al.,  2008;  Lesaux  et  al.,   2008).    However,  children  were  occasionally  asked  to  segment  spoken  words   into  syllables  and  to  segment   and  blend  written  words.     Table 1 Target FL sound/spelling links Sound Letters Exemplar Gesture Visual Links with L1 Pre-lesson Letter names Alphabet Week 1 Final silent letters PETS,

, , , Week 2-5 /ɔ̄/ Cochon (pig) Finger on nose - snout Picture of pig Novel FL sound/existing L1 grapheme /y/ and <Û> Putois (ferret) Fingers pinching nose (smell) Picture of ferret Novel FL sound/ existing L1 and novel FL grapheme /j/ Billet (ticket) Taking a ticket Picture of train ticket Existing L1 sound/existing but different L1 grapheme Week 6 /ɑ̄/ and Enfant (child) Rocking a baby Picture of baby Novel FL sounds/existing L1 grapheme /ɛ̄/ , , Singe (monkey) Arms bent at sides Picture of monkey Novel FL sounds/existing L1 and novel FL graphemes Weeks 7-11 Practice all previous sound/spelling links Week 12 /e/ , , , Éléphant (elephant) Arm imitating trunk Picture of element Novel FL sound/existing L1 and novel FL graphemes /o/ , , <Ô>, Océan (ocean) Arm moving like a wave Picture of ocean Novel FL sound/existing L1 and novel FL graphemes Week 13-20 Practice all previous sound/spelling links       Sounds  and  their  corresponding  letters  were  also  presented  through  an  exemplar  word,  e.g.  cochon,   which  aimed  to  allow  for  associative  learning  and  the  possibility  that  this  could  be  used  through  analogy  to   decode   the   same   grapheme   in   unfamiliar  words.   It  was   anticipated   that   the   learning   of   FL   sound/spelling   links  would  be  facilitated  by  shared  links  with  English  but  that  learning  to  adept/amend  existing  mappings  or   LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 85   creating   entirely   novel   ones   would   be   more   problematic.   The   sounds   and   their   links   were   taught   on   a   cumulative  basis  so  that,  as  a  new  sound  was  introduced,  previous  sound/spelling  links  were  still  practised   regularly.     2.2.3.  Principle  3:  Applying  sound/spelling  links  through  reading  aloud  practice       It   was   also   deemed   necessary   to   offer   children   opportunities   to   practise   applying   these   sound/spelling   links   through   reading   and   writing   opportunities.   Weekly   spelling   activities   involving   core   vocabulary   (not   phonics   exemplar   words)   supported   the   children’s   learning   of   links   between   sounds   and   letter   in   French.   The   children   were   asked   during   a   sequence   of   tasks   designed   to   support   vocabulary   memorisation,  to  recall  a  word  orally  from  a  visual  cue  and  then  to  attempt  to  write  this  word  from  memory   (having  seen  the  written  form  several  times  previously).  Wherever  possible,  core  vocabulary  was  utilised  to   teach  phonics  sound/spelling  links  such  as  ondule  (slither)    =  /ɔ̄/and  fourrure  (fur)    =  /u/.  It  should   be   noted   that   these   words   were   chosen   to   support   broader,   communicative   language   use   rather   than   act   solely   as   practice   for   the   phonics   element   of   the   instructional   programme.   In   reading   aloud   opportunities   such   as   stories   and   pairwork,   peer-­‐assessed   reading   of   sentences   were   used   to   practise   phonological   decoding.   Again,   these   stories   were   mostly   focused   on   meaning-­‐related   activities   in   line   with   integrated   approaches  to  reading  but  comprised  occasional  decoding  opportunities  (Adams,  1990).       2.3.  Procedures       The  reading  aloud  test  battery  targeted  specific,  taught  FL  sound/spelling  links  as  well  as  a  broader   range   of   linguistic   skills   and   knowledge.   Due   to   the   lack   of   existing   resources   for   testing   phonological   decoding  in  primary  school  learners  of  French,  test  items  were  designed  by  the  teacher/researcher.  As  there   was  no  pilot  sample  with  similar  language  learning  experience  in  terms  of  age,  language,  and  extent  of  input,   these  tests  were  not  piloted.  The  broader  linguistic  FL  assessment  tasks  were  closely  aligned  with  classroom   activities   and   designed   to   test   the   constructs   of   FL   general   proficiency   and   FL   global   literacy.   These  were   administered   at   pre-­‐,   post-­‐,   and   delayed   post-­‐test.   The   post-­‐test   was   conducted   immediately   after   the   instructional   programme   ended,   and   the   delayed   post-­‐test   was   completed   seven  weeks   later.   The   general   proficiency  construct  comprised  an  oral  vocabulary  recognition  test  and  an  elicited  imitation  test  designed  to   measure   developing   interlanguage.   The   FL   global   literacy   tests   involved   reading   aloud   target   words   in   sentences  (with  illustrative  pictures  to  support  meaning)  and  reading  for  global  meaning  and  detail  of  a  short   text.  An  additional  reading  aloud  test  was  implemented  during  the  teaching  programme  in  order  to  capture   any  development  in  a)  applying  sound/spelling  links  through  decoding  familiar  exemplar  phonics  words  and   b)  applying   these   taught  sound/spelling   links   in  novel  contexts   (unfamiliar  words).  This   test  was  delivered   with  a  sub-­‐sample  of  pupils  (n  =  23)  at  mid-­‐test,  post-­‐test  and  delayed  post-­‐test.  The  whole  sample  (n  =  45)   took  the  FL  reading  aloud  word  card  test  at  post-­‐test  and  delayed  post-­‐test.       2.4.  Research  instruments   2.4.1.  FL  general  proficiency:  Receptive  vocabulary  and  elicited  imitation       This   variable   was   addressed   through   two   complementary,   general   proficiency   (non-­‐literacy)   measures.  The  FL  receptive  vocabulary  test  was  based  on  the  Peabody  Picture  Vocabulary  Test  (Dunn  &  Dunn,   2007)  and  the  ELIAS  British  Picture  Vocabulary  Scale  (Dunn,  Dunn,  Whetton,  &  Burley,  1997).  Lexical  items   which  had  been  taught  over  previous  years  where  chosen  and  individual  worksheets  were  completed.  From  a   spoken  cue  (provided  by  the  teacher),  the  children  had  to  circle  the  picture  which  correctly  represented  each   lexical   items.   Some   items   were   discrete   words   (e.g.,   le   père   –   father)   and   some   were   formulaic   utterances/constructions  (j’adore  faire  du  cheval  –  I  love  horseriding).  The  pictorial  cues  were  familiar  to  the   children  as  they  were  largely  derived  from  historic  teaching  materials  used  by  the  teacher.      The  FL   elicited   imitation   test  was   designed   to   show  developing   implicit   (subconscious)   L2   knowledge   (Erlam,  2006).  Learners  listened  to  utterances  incorporating  particular  syntactic  or  morphological  demands,   such  as   interrogatives  or  negation,  and  were  asked  to  repeat  them  after  a  short  gap.  Prior  to   listening,  they   were   guided   to   process   the   utterances   for   meaning   by   being   warned   that   they   would   have   to   answer   questions  about  the  utterances  after  they  had  heard  them.  All  the  stimuli  were  designed  to  be  long  enough  to   exceed  working  memory   capacity   and   therefore   could   only   be   reformulated   by   accessing  meaning   and   re-­‐ constructing  the  utterance  from  interlanguage.  For  example,  the  sentence  Xavier,   il  n’aime  pas  aller  à   l’école   (Xavier  does  not   like  going  to  school)  was  elicited  to  explore   learning  of  negation  expressed  through  dislike.   Elicited   imitation   is   deemed   an   efficient   way   of   accessing   spontaneous   oral   ability   in   young   learner   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 86   classrooms  where  there  are  often  practical  constraints  such  as  the  amount  of  available  test  time  (Campfield,   2015).   To   measure   performance,   reproductions   were   marked   by   target-­‐like   syllables.   Each   utterance,   regardless   of   length,   could   achieve   a  maximum   score   of   4   and   syllabic   errors  were   deducted   from   this   as   required  (Lonsdale,  Graham,  Kennington,  Johnson,  &  McGhee,  2008).  Therefore,  a  sentence  with  two  syllable   errors  (non-­‐target-­‐like  reproduction)  would  receive  a  mark  of  2.  Often  elicited  imitation  tests  with  younger   learners   are   marked   on   a   rating   scale   such   as:   3   =   mainly   target-­‐like   reproduction   to   0=unintelligible   production.  It  was  felt  that  a  syllabic  scoring  mechanism  might  give  a  better  indication  of  the  extent  of  FL  oral   processing   (reproduction   in   terms   of   sentence   length   and   complexity)   rather   than   an   arguably   more   subjective  rating  scale  which  focused  on  evaluation  of  overall  accuracy  of  sentences.  A  small  sample  of  post-­‐ test  scores  were  second-­‐marked  by  the  same  rater  with  a  reliability  score  of  67%.       2.4.2.  FL  literacy  proficiency:  Reading  aloud  (sentences)  and  reading  comprehension       For  the  FL  reading  aloud  (text)  test,  short  sentences  illustrated  with  pictures  were  read  aloud  by  each   child.  For  example,  Monsieur  Laurent  travaille  au  marché  (Mr  Laurent  works  on  the  market)  was  constructed   to  elicit  reproduction  of  specific,  underlined  sound/spelling  links.  Individual  words  (containing  instructed  FL   sound/spelling)  were  scored  for  target-­‐like  recoding,  and  a  clear  scoring  scheme  was  set  out  to  contribute  to   rater   reliability.   Recoding  was   strictly  marked   to   reflect   explicitly   taught   characteristics   of   FL   sounds   and   their  links  to  letters.  In  other  words,  recoding  of    was  only  deemed  target-­‐like  (scoring  1  =  correct)  if  it   reflected  nasality  /ɔ̃/.  Even  though  /ɒn/  would  have  been  largely  intelligible  to  a  sympathetic  speaker  of  the   language,  it  was  marked  incorrect  as  production/discrimination  of  novel  FL  phonology  was  an  integral  part  of   the  instructional  programme.  Second  marking  by  the  same  rater  yielded  a  reliability  score  of  83%.  This  test   was  conducted  on  a  one-­‐on-­‐one  basis,  during  lesson  time  in  a  private  setting  at  each  school.         The   FL   reading   comprehension   test   aligned   with   primary   FL   learning   outcomes   including:   “understand  the  main  points  and  some  detail  from  a  short  written  passage”  and  “reading  texts  for  enjoyment   or  information”  (DfES,  2005,  p.  57).  A  short  passage  with  11  (8%)  unfamiliar  words  was  designed  as  a  police   report   concerning   a   robbery.   Questions   were   presented   in   children’s   L1   and   designed   to   test   reading   for   understanding  of  the  main  points  and  some  detail.  Responses  were  designed  to  require  “limited  production”   (McKay,  2006,  pp.  106,  186)  in  that  children  could  identify  the  meaning  by  selecting  a  picture.  For  instance,   for   the  question   “What   item  was   stolen?”   children   could   choose   from  pictures   of   a  wallet,   a   bag   and   some   house  keys.  It  was  anticipated  that  this  would  mitigate  cognitive  loading  effects  for  children  with  weaker  L1   literacy.  Scoring  was  dichotomous,  and  this  test  was  conducted  using  individual  worksheets  in  a  whole  class   setting  without  conferring.  Children  were  allowed  approximately  15  minutes  to  complete  the  test.     2.4.3.  FL  phonics:  Read  aloud  word  card       This   test  was   designed   in   line  with   action   research   principles   and   cycles   of   research  where   initial   investigations   yield   new   or   more   refined   lines   of   inquiry.   The   teacher/researcher   deemed,   during   the   instructional  programme,   that   learning  and  applying  exemplar  sound/spelling   links  was  challenging  for  the   children.   She   felt   it  would  be   interesting   to  uncover  how  progress  was  unfolding  with   respect   to  decoding   exemplar  phonics  words  and  also  whether  this  knowledge  was  being  applied  to  novel  contexts.  With  this  in   mind,  illustrated  word  cards  were  designed  (arguably  more  as  a  formative  assessment  tool  than  as  a  research   instrument).  At  mid-­‐test,  due  to  time  constraints,  a  small  number  of  pupils  (n  =  23)  were  tested.  The  children   were   selected   based   on   their   FL   reading   aloud   performance   at   pre-­‐test   (sentence)   from   each   third   of   the   scoring  range   (Table  2).  The  children  were  each  asked   to   shuffle   the  cards  and   then  read  aloud  each  word   whilst  being  audio-­‐recorded.  Each  child  sat  alone  in  the  corridor  outside  the  classroom  at  mid-­‐test  whilst  the   teacher   taught   the  rest  of   the  class.  No   feedback  was  given  at  any  stage.  At  post-­‐test  and  delayed  post-­‐test,   this  word  card  measure  was   incorporated   into   the  overall   test  battery  and  delivered  on  a  one-­‐to-­‐one  basis   with   the   teacher/researcher   in   a   private   room.   Each  word   card   had   a   picture   for  meaning   (Figure   1).   For   every   targeted  FL  sound/spelling   link   there  was  a  card  with   the  phonics  exemplar  word  and  card  with   the   same  grapheme,  in  a  different  position,  in  an  unfamiliar  word.               LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 87   Table 2 L2 reading aloud pre-test, quartile breakdown Pre-test read aloud scoring (max 10) Participants Low (0-2) 8 Mid (3-4) 10 High (5-6) 5 Total participants 23     Cochon Ongle Figure 1. Example of exemplar phonics word card     2.4.4.  Underlying  L1  literacy  measures       In  line  with  national  data  collection  in  English  primary  schools  at  the  time,  this  study  had  access  to  a   range  of  standardised  L1  literacy  test  scores  (reported  on  a  national  level  by  school  and  child)  and  classroom-­‐ based  informal  teacher  assessments.  These  were  deemed  potentially  important  in  explaining  the  influence  of   L1   literacy  on  FL  reading  aloud  performances.  Children’s  chronological  reading  ages  were  collected  using  a   standardized   New   Group   Reading   Test   (GL   Assessment,   2018).   This   test   comprises   two   parts:   sentence   completion  (decoding  with  some  comprehension)  plus  passage  comprehension  where  children  are  tested  on   a  range  of  comprehension  skills.  The  Single  Word  Spelling  Test  (GL  Assessment,  2018)  was  also  administered   by   the   regular   class   teachers   to   determine   children’s   chronological   spelling   age   through   spelling  words   in   context  (sentences).  This  is  also  a  standardized,  commercially  produced  test.         In  addition  to  these  standardized  tests,  the  children’s  regular  class  teachers  (not  the  researcher  who   was   their   longstanding   yet   visiting   FL   teacher)   undertook   ongoing   formative   assessments   of   each   child’s   progression   in   all   curricular   subjects.   These   teacher-­‐centred   assessments,   backed   up   by   observation   and   other  classroom  data  such  as  examples  of  children’s  work,  were  then  linked  to  levels  of  attainment  which  at   the   time   formed   part   of   the   National   Curriculum.   Whilst   levels   have   now   been   eliminated   formally   from   government   documentation,   they   are   still   sometimes   referred   to   by   teachers   as   a  means   of   benchmarking   pupils’  performances  against   their  peers  and  previous  cohorts.  National  Curriculum   levels   ranged   from  1-­‐6   with  the  idea  that,  by  the  end  of  primary  school,  expected  attainment  for  children  would  reach  level  4.  Each   level  was  accompanied  by  descriptors  for  each  subject  and  had  sub  categories  (1c  was  the  bottom  sub-­‐level   within  level  1  whilst  1a  was  the  top  sub-­‐level).  To  facilitate  statistical  analysis,  these  levels  were  converted  to   a  continuous  scale  e.g.  level  1c  was  1  and  level  6a  was  18.  Numerical  levels  were  assigned  to  each  child  for  L1   reading  (reading  aloud,  comprehension)  and  L1  writing  (spelling,  complexity,  genre,  cohesion).         PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 88   3.  Results       A  Shapiro-­‐Wilk  test  (as  n  equalled  <50)  showed  that  assumptions  of  normality  were  violated  for  33%     of  the  pre-­‐  and  post-­‐test  FL  constructs.  Given  that  result,  and  bearing  in  mind  the  relatively  small  sample  size,   subsequent   statistical   analyses   were   conducted   using   non-­‐parametric   techniques.   Mann   Whitney   U-­‐Tests   were  conducted  to   identify  whether   there  were  any  differences  between  sub  samples   in  each  school   for  L1   and  FL  read  aloud  measures  at  pre-­‐,  post-­‐,  and  delayed  post-­‐test  times.  No  differences  were  found  between   school  or  gender.     3.1.  To  what  extent  do  learners  engaged  in  systematic  and  explicit  FL  phonics  instruction  make   progress  in  reading  aloud  familiar  and  unfamiliar  FL  words  at  mid-­‐,  post-­‐,  and  delayed  post-­‐ test?       Sections   3.1.1   and   3.1.2  will   report   on  whole-­‐sample   performances   for   reading   aloud   familiar   and   unfamiliar  word  cards,  during  and  immediately  after  the  instructional  programme  (at  mid-­‐test  and  post-­‐test).   Descriptive  statistics  will  portray  the  extent  and  nature  of  developing  of  FL  reading  aloud  scores  at  both  test   times.   Sections   3.1.3   and   3.1.4  will   then   use   statistical   techniques   to   establish  whether   any   improvements   were  statistically  significant  in  reading  a)  familiar  word  cards  and  b)  unfamiliar  word  cards.       3.1.1.  Reading  aloud  familiar  and  unfamiliar  word  cards  at  mid-­‐test       By  mid-­‐test,  which  was  conducted  10  weeks  after  the  FL  phonics  instruction  commenced,  descriptive   statistics  showed  that  progress  reading  familiar  word  cards  was  slow  overall  (mean  score  1.13;  median  score   1.00)  out  of  a  maximum  score  of  5  (Table  3).         Positive  skewness  and  kurtosis  values  indicated  that  scores  at  this  test  time  were  clustered  towards   the  lower  end  of  the  scoring  range  (Figure  2).  However,  the  range  of  scores  showed  that  whilst  one  child  (out   of  the  23  participants)  had  scored  close  to  full  marks,  others  had  made  no  progress  (seven  children  scored  0)   reading   familiar  words.  Despite  weekly   lessons  which   involved  practising  production  and  discrimination  of   FL  sounds,  FL  phonological  decoding   (spellings   to   sound),  and  reading  aloud  of   familiar   (exemplar)  words,   learners   had   made   little   progress   at   reading   aloud   the   practised   words   mid-­‐way   through   the   teaching   programme.         Table 3 Descriptive statistics, mid-test, reading aloud word cards (N = 23) Mid-Test Mean Median Mode SD Range Skewness Kurtosis Min Max Familiar words (max 5) 1.13 1.00 .00a 1.100 0 4 1.096 1.168 Unfamiliar words (max 5) .3042 .00 .00 .653 0 2 1.994 2.934 Note. a = multiple modes LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 89     Figure 2. Distributions of read aloud familiar words at mid-test   Unfamiliar  word  card  reading  scores  were  less  variable  (range  0  to  2).  Eighteen  children  scored  zero   whilst   two   children   scored   2  marks   at  mid-­‐test   points.   Almost   no   progress  was   evidenced   at  mid-­‐test   for   reading  unfamiliar  word  cards  (unfamiliar  mean  .3043;  median  .00)  and  again  skewness  and  kurtosis  showed   scores  weighted  toward  the   lower  end  of  the  range  (Figure  3).  Perhaps  unsurprisingly,  bearing  in  mind  the   limited  progress  for  familiar  (exemplar)  words,  the  children  had  also  made  scant  progress  at  applying  learned   sound/spelling  links  to  novel  contexts.       Figure 3. Distributions of read aloud unfamiliar words at mid-test     PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 90   3.1.2.  Reading  aloud  familiar  and  unfamiliar  word  cards  at  post-­‐test       By  post-­‐test,  at   the  end  of  FL  phonics   instruction  (23  weeks   in   total),  descriptive  statistics  showed   that   progress   reading   familiar   word   cards   had   improved   (mean   score   2.174;  median   score   2.00)   out   of   a   maximum  score  of  5  (Table  4).       Table 4 Descriptive statistics post-test reading aloud word cards (N = 23) Mid-Test Mean Median Mode SD Range Skewness Kurtosis Min Max Familiar words (max 5) 2.174 2.00 1.00 1.230 0 4.5 -.040 -.708 Unfamiliar words (max 5) .9130 1.00 .00 .950 0 3 .535 -.934   A  change  in  skewness  values  (now  negative)  indicated  that  scores  were  clustered  towards  the  mid  or   higher  end  of  the  score  range.  However,  positive  kurtosis  values  indicated  that  there  remained  a  heavy  tail  of   scores  in  the  lower  marks  (Figure  4).  Two  children  scored  4/4.5  and  two  children  scored  0  reading  familiar   words.  In  other  words,  the  range  was  still  wide  between  top  and  bottom  scores  but  the  lowest  scorers  were   less   numerous.   After   23   weeks,   the   sample   seemed   to   be   making   progress   in   reading   aloud   familiar   (exemplar)  words.     However,  scant  progress  was  evidenced  at  post-­‐test  for  reading  unfamiliar  word  cards  (mean  .9130;   median  1.00)   and   again   skewness   and  kurtosis   showed   scores   tended   towards   the   lower   end  of   the   range   (Figure  5).  Unfamiliar  word  card  reading  scores  showed  slightly  different  variability  (range  0  to  3  maximum)   with  10  children  scoring  zero  and  1  child  scoring  the  top  mark  of  3  at  post-­‐test.  23  weeks  after  the  weekly   phonics   instruction   commenced,   most   children   were   still   unable   to   consistently   apply   their   developing   knowledge  of  FL  sound/spelling  links  to  novel  contexts     Figure 4. Distributions of read aloud familiar word cards at post-test     LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 91   Figure 5. Distributions of read aloud unfamiliar word cards at post-test     3.1.3.  Reading  aloud  familiar  word  cards:  Differences  between  mid-­‐test  and  post-­‐test       In  this  and  the  following  section,  scores  between  mid-­‐  and  post-­‐test  are  compared  for  both  kinds  of   word  cards  separately  to  evaluate  whether  reported  improvements  were  of  statistical  significance.  Wilcoxon   Signed  Rank  Tests   (Table  5)  show  that   the  scores   for   familiar  reading  aloud  word  cards  between  mid-­‐  and   post-­‐test   were   significantly   different.   Target-­‐like   decoding   at   post-­‐test   was   significantly   higher   (median   2.000)  than  at  mid-­‐test  (median  1.000),  z  =  3.533,  p  =  <0.0001,  r  =  0.52,  moderate  effect  size.  So  by  the  end  of   the   teaching   programme,   there   was   a   significant   improvement   in   children’s   familiar   word   reading   aloud   scores.     Table 5 Mid-test To Post-Test Familiar FL reading aloud word cards – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Mid-test Post-test FL Word Card Familiar Reading Aloud 1.000 2.000 3.533 <.0001 19.245 0.52 Note. Max score familiar words 5; unfamiliar words 5.   3.1.4.  Reading  aloud  unfamiliar  word  cards:  Differences  between  mid-­‐test  and  post-­‐test       Wilcoxon  Signed  Rank  Tests  (Table  6)  show  that  unfamiliar  reading  aloud  word  cards  between  mid-­‐   and  post-­‐test  were  significantly  different.  Target-­‐like  decoding  at  post-­‐test  was  significantly  higher  (median   1.000)  than  at  mid-­‐test  (median  .000),  z  =  2.617,  p  =  .007,  r  =  0.40,  moderate  effect  size.  This  meant  that  the   children  became  significantly  better  at  reading  aloud  unfamiliar  word  cards  between  the  mid  and  end  point  of   the   teaching  programme.   It   is,  however,   important   to  note   that  medians  and  means   for  both  kinds  of  word   cards  at  mid-­‐  and  post-­‐test  were  still  below  the  mid-­‐point  of  the  scoring  range.     Table 6 Mid-test To Post-Test Unfamiliar FL reading aloud word cards – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Mid-test Post-test FL Word Card Reading Aloud .000 1.000 2.617 .007 10.937 0.40 Note. Max score familiar words 5; unfamiliar words 5. PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 92       Sections  3.1.5  will  now  turn  to  performance  at  delayed  post-­‐test  approximately  7  weeks  after  all  FL   instruction  had  finished.  This  will  evaluate  the  durability  of  any  learning  and  explore  whether  there  was  any   observed   attrition   related   to   word   reading   aloud   in   both   familiar   and   unfamiliar   contexts.   Descriptive   statistics  will   first   explore   the  nature  of   change   (Section  3.1.5),   and   inferential   statistics  will   then   compare   performances   between   post-­‐   and   delayed   post-­‐test   (familiar   and   unfamiliar   words   treated   separately)   to   evaluate  whether  any  observed  differences  were  statistically  significant  (Sections  3.1.6.  and  3.1.7.).     3.1.5.  Reading  aloud  familiar  and  unfamiliar  word  cards  at  delayed  post-­‐test       By  delayed  post-­‐test,  which  was  conducted  7  weeks  after  the  FL  phonics  instruction  ended  (with  no   French  language  input  in  between),  descriptive  statistics  show  a  dip  in  means  (to  1.8696  out  of  a  maximum  of   5)  but  not  in  medians  (Table  7).       Table 7 Descriptive statistics delayed post-test reading aloud word cards (N = 23) Mid-Test Mean Median Mode SD Range Skewness Kurtosis Min Max Familiar words (max 5) 1.8696 2.00 .00 1.440 0 4.5 .136 -1.093 Unfamiliar words (max 5) .6957 .00 .00 .974 0 3 1.331 .876   A   reversal   in  skewness  scores   (Table  7,  Figure  6)  was  also  noted  which   illustrated   that   scores  had   shifted   from  the  higher   range   (at  post-­‐test)  back   to   the   lower  range   (by  delayed  post-­‐test).  This   confirmed   that  reading  aloud  of   familiar  FL  word  cards  had  suffered  attrition  once  the  weekly  FL  instruction  (phonics   and  language)  had  ceased.     Figure 6. Distributions of familiar read aloud word cards at delayed post-test   For  unfamiliar  word  cards  the  mean  had  decreased  slightly,   from  an  already   low  point  at  post-­‐test.   Interestingly,   the  range  of  scores   for   familiar  and  unfamiliar  words  were   identical   to   those  at  post-­‐test  and   the  skewness  and  kurtosis  values  remained  positive  (Figure  7).  It  seemed  as  though  progress  in  reading  aloud   of   sound/spelling   links   in   unfamiliar   contexts   was   resistant   to   instruction   during   the   programme   and   therefore  scores  were  not  particularly  impacted  once  the  FL  instruction  finished.       LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 93   Figure 7. Distributions of unfamiliar read aloud word cards at delayed post-test     3.1.6.  Reading  aloud  familiar  word  cards:  differences  between  post-­‐test  and  delayed  post-­‐test   Wilcoxon  Signed  Rank  Tests  were  conducted  to  evaluate  the  extent  of  attrition  in  students’  scores  on   word  card  read  aloud  tests  (Table  8)  with  a  Bonferroni  correction  applied  for  repeated  tests  (p  =  .0125).  For   familiar  word   card   reading   between   post-­‐test   and   delayed   post-­‐test   (Table   8),   target-­‐like   decoding   clearly   was  not  significantly  lower  at  delayed  post-­‐test  (median  =  2.000)  than  familiar  word  card  reading  at  post-­‐test   (median  =  2.000),  z  =  -­‐1.613,  p  =  .107.  In  other  words,  there  was  no  significant  attrition  between  post-­‐test  and   delayed  post-­‐test  for  children  reading  aloud  familiar  words.     Table 8 Post-Test to Delayed Post-Test Familiar FL reading aloud word cards – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Post-test Delayed post-test FL Word Card Reading Aloud 2.000 2.000 -1.613 .107 20.767 n/a Note. Max score familiar words 5; unfamiliar words 5.   3.1.7.  Reading  aloud  unfamiliar  word  cards:  differences  between  post-­‐test  and  delayed  post-­‐test       For   unfamiliar   word   card   reading   (Table   9)   between   post-­‐test   and   delayed   post-­‐test,   target-­‐like   decoding  was  not  significantly  lower  at  delayed  post-­‐test  (median  .000)  than  unfamiliar  word  card  reading  at   post-­‐test  (median  1.000),  z  =  -­‐1.155,  p  =  .248.  So,  there  was  also  no  significant  attrition  between  post-­‐test  and   delayed  post-­‐test  for  unfamiliar  word  card  reading.     Table 9 Post-Test to Delayed Post-Test Unfamiliar FL reading aloud word cards – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Post-test Delayed Post-test FL Word Card Reading Aloud 1.000 .000 -1.155 .248 12.124 n/a Note. Max score familiar words 5; unfamiliar words 5.   Overall,  then,  moderate  statistically-­‐significant  progress  was  made  for  familiar  words  between  mid-­‐   and  post-­‐test   and   for  unfamiliar  words  between  mid-­‐  and  post-­‐test.  Neither   familiar  nor  unfamiliar  words,   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 94   examined  separately,  showed  any  significant  difference  between  post-­‐  and  delayed  post-­‐test.  So  the  learning   of   sound/spelling   links   observed   in   both   familiar   and   unfamiliar   contexts   was   relatively   long-­‐lasting.   Nonetheless,  it  is  important  to  contextualise  the  nature  of  the  significant  differences  across  test  times  and  test   conditions.  The  shape  of  the  distribution  of  scores  for  familiar  word  cards  changed  between  mid-­‐  and  post-­‐ test  but  the  spread  of  scores  were  still  centred  around  the  lower  scoring  range.     3.2.  Are  performances  on  reading  aloud  familiar  and  unfamiliar  word  cards  different  at  mid-­‐,   post-­‐,  and  delayed  post-­‐test?       Now  the  analyses  will   focus  on  whether  word   familiarity  affects  FL  reading  aloud  performances  as   the  descriptive  data  examined  so  far  suggests  this  is  likely.  Sections  3.2.1  to  3.2.3  will  take  each  test-­‐time  in   isolation   and   explore  whether   FL   reading   aloud   outcomes   for   familiar   versus   unfamiliar  word   cards  were   statistically  different.  Alongside  comparison  of  group  medians,  a  more  detailed  examination  of  performances   will  be  covered  in  section  3.2.4  at  the  level  of  word  card.       3.2.1.  Reading  aloud  familiar  versus  unfamiliar  word  cards  at  mid-­‐test       Children’s   scores  were   compared,   using  Wilcoxon   Signed   Rank   Tests,   between   tasks   (familiar   and   unfamiliar  word  cards)  at  each   time  point.  This  was   to  discern  whether   there  was  a  difference  at  each   test   time   between   reading   familiar,   exemplar   words   (rehearsed,   whole   word   pronunciation)   compared   to   unfamiliar   words   (application   of   FL   decoding   and   therefore   FL   sound/spelling   link   knowledge).   Table   10   shows   that   mid-­‐test   familiar   word   card   decoding   was   significantly   higher   (median   =   1.000)   than   for   unfamiliar  word  card  reading  (median  =   .000),  z  =   -­‐3.368,  p  =  0.001,  r  =   -­‐0.49,  medium  effect  size.   In  other   words,   at   mid-­‐test   there   was   a   marked   difference   in   overall   performance   when   reading   aloud   familiar   compared  to  unfamiliar  words.   Table 10 Mid-Test Familiar versus Unfamiliar FL reading aloud word cards – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Familiar Unfamiliar FL Word Card Reading Aloud 1.000 .000 -3.368 .001 17.367 -0.49 Note. Max score familiar words 5; unfamiliar words 5.   3.2.2.  Reading  aloud  familiar  versus  unfamiliar  word  cards  at  post-­‐test   Similarly,   Table   11   shows   that   at   post-­‐test,   familiar   word   card   decoding   was   significantly   higher   (median  2.000)  than  unfamiliar  word  card  decoding  (median  1.000),  z  =  -­‐3.851,  p  =  <0.001,  r  =  -­‐0.57,  medium   effect   size.   By   post-­‐test   the   pattern   continued   that   familiar  word   learning   (reading   aloud)  was   better   than   unfamiliar  word  reading.       Table 11 Post-Test Familiar versus Unfamiliar FL reading aloud word cards – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Familiar Unfamiliar FL Word Card Reading Aloud 2.000 1.000 -3.851 <.0001 24.668 -0.57 Note. Max score familiar words 5; unfamiliar words 5.   3.2.3.  Reading  aloud  familiar  versus  unfamiliar  word  cards  at  delayed  post-­‐test   Finally,  Table  12  shows  that  at  delayed  post-­‐test,  familiar  word  card  decoding  was  still  significantly   higher  (median  2.000)  compared  to  unfamiliar  word  card  decoding  (median  .000),  z  =  -­‐3.521,  p  =  <0.001,  r  =  -­‐ 0.52   medium   effect   size.   This   demonstrates   that   at   all   test   points   familiar,   exemplar   words   were   more   successfully  decoded  than  unfamiliar  words.         LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 95   Table 12 Delayed Post-Test to Familiar versus Unfamiliar FL reading aloud word cards – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Familiar Unfamiliar FL Word Card Reading Aloud 2.000 0.000 -3.521 <.0001 22.864 -0.52 Note. Max score familiar words 5; unfamiliar words 5.   3.2.4.  Reading  aloud  by  word  card  at  each  test  time   Performances  across  familiar  and  unfamiliar  word  cards  at  all  test  times  were  also  examined  through   analysis   of   each   word   card.   A   breakdown   shows   that   overall   performances   were   centred   around   specific   sound/letter  links  (Figure  8).  Broadly  speaking,  novel  FL  sounds  linked  to  existing  L1  spellings  were  learned   best   in   familiar  words   (singe,  cochon,  enfant).  Those   sounds  which  were  either   a)   similar   to   an  existing  L1   sound  /y/  or  b)  an  extant  L1  sound  /j/  which  required  a  different  mapping  were  learned  slowest.  This  was   not  particularly  affected  by  order  of  presentation:  /j/  as  in  billet  was  recoded  less  successfully  than  /ɛ̄/  as  in   singe,   yet   /j/   had   been   presented   earlier   in   the   teaching   programme.   Sounds   that   were  most   successfully   decoded   in   post-­‐test   familiar   conditions   were   likely   to   show   relative   success   at   post-­‐test   unfamiliar   conditions,   although   singe/câlin   proved  an  exception.   It   is   important   to  note   that  grand   (unfamiliar)  didn’t   perform  as  differently  as  expected  although  this  is  likely  because  the  children  chose  to  use  this  word  as  part   of   their   extended   reading/writing  work   (the  adjective   formed  part  of   the   core   taught  vocabulary)   and   it   is   also  a  “false  friend”  (false  cognate)  in  English.         Figure 8. Performance at each test time for specific word cards/sound-spelling links (max score per word/test time 23)       3.3.   To   what   extent   do   L1   literacy   measures   and   FL   global   proficiency   scores   relate   to   FL   reading  aloud  proficiency  at  each  test  time?       This   research   question   seeks   to   identify   whether   particular   underlying   constructs,   known   to   influence  FL  reading,  could  explain  any  of   the  variance   in   this  study’s  FL  reading  outcomes.  Firstly,  Section   3.3.1   describes   the   nature   of   L1   proficiency   across   the   sample   and   then   explores   the   potential   for   relationships  between  L1  literacy  measures  and  FL  reading  aloud  performances  at  each  test  time.       PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 96       3.3.1.  L1  literacy  measures  and  FL  reading  aloud  performance       Descriptive   statistics   (Table  13)   show   the   range  of  L1   literacy  achievement  across   the   sub-­‐sample.   Mann  Whitney  U  tests  were  conducted  and  confirmed  that  there  were  no  significant  differences  in  L1  baseline   literacy  scores  between  School  1  and  School  2.     Table 13 Descriptive statistics for all L1 literacy-related measures Baseline L1 Scores Mean Median Mode SD Range Skewness Kurtosis Min Max NGRT L1 reading age 12.709 12.07 13.02 1.675 9.0 15.0 -.003 -.221 SWST L1 spelling age 10.517 10.07 10.02a 3.211 7.08 17.0 -1.118 5.115 NC Reading Levels (max 14) 9.96 10.07 10.00 2.383 6 14 2.383 -.319 NC Writing Levels (max 14) 9.35 10.00 10.00 1.991 5 13 -.189 -.325 Note. a = multiple modes exist       These  data  were  examined  using  Spearman’s  Rho  correlation  coefficients  to  establish  whether  there   were  relationships  between  the  individual  measures  of  L1  literacy  proficiency.  Strong  positive  relationships   were   found   between   both   National   Curriculum   reading   levels   (teacher-­‐assessed)   and   NGRT   standardised   reading  scores  rs  =  .809**  sig  <  .0001  (*  =  .05  alpha;  **  =  .001  alpha).  Moderate  relationships  were  also  found   between  National  Curriculum  writing   levels   (teacher-­‐assessed)   and  SWST  standardised   spelling   scores   rs   =   .556**  sig  =  .006.  NGRT  and  SWST  scores  also  correlated  moderately  rs  =  .580**  sig  =  .004.  Correlational  tests   also  showed  that  across  all  times  L1  literacy  measures  showed  positive  and  significant  relationships  with  FL   reading  aloud  outcomes  (FLRA  -­‐  Table  14).     Table 14 Relationships between L1 literacy-related outcomes and FL reading aloud at all test times Mid-Test FLRA (fam-unfam) Post-Test FLRA (fam-unfam) Delayed Post-Test FLRA (fam-unfam) rs sig rs sig rs sig L1 reading age NGRT .583** .003 .513** .012 .568** .005 L1 spelling age SWST .693** <.0001 .645** .001 .701** <.0001 NC Reading Levels .691** <.0001 .583** .004 .767** <.0001 NC Writing .516** .012 .614** .002 .713** <.0001       Whilst   the   influence  of  L1  reading  age  on  FL  reading  aloud  remained  significant  at  all   test   times,   it   diminished   slightly   after   the   teaching   programme,   evidenced   by   slightly   lower   rs   values.   Interestingly,   L1   spelling   and   teacher-­‐assessed   NC   levels   show   highest   rs   values,   and   therefore   strongest   relationships,   at   delayed   post-­‐test   although   with   the   exception   of   NC   writing,   these   did   not   result   in   different   values   of   statistical  significance.  Spelling  age  is  the  most  influential,  continuously  showing  rs  values  above  .6  (strong).     3.3.2.  Global  FL  scores  (literacy  and  general  proficiency)         Section   3.3.2   turns   to   examine   FL   global   performances   (collected   from   additional,   combined   FL   measures)   to   establish   whether   FL   word   card   reading   aloud   scores   at   each   test   time   were   related   to   underlying   FL   proficiency.   Firstly,   global   literacy   and   general   proficiency   scores   at   each   test   time   were   examined  to  determine  whether  FL  learning  occurred  during  the  teaching  cycle  and  if  any  attrition  occurred   between  post-­‐  and  delayed  post-­‐test  (Tables  15-­‐17).  These  composite  scores  were  derived  from  the  following   LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 97   datasets:  Global  FL   literacy  comprised  FL  reading  aloud  (target-­‐like  production  of  specific  sounds/word)   in   sentences  plus  FL  reading  comprehension  (reading  for  global  meaning  and  detail).  It  should  be  recalled  that   the  reading  aloud  task  was  different  from  the  FL  reading  aloud  word  card  task  reported  in  RQ1  and  RQ2.  FL   general   proficiency  was   derived   from   FL   receptive   vocabulary   (recognising  words   and   linking   to   pictorial   representations)  plus  FL  elicited  imitation  (sentence  reproduction).       At   pre-­‐test,   FL   global   literacy   and   general   proficiency   scores   showed  wide   ranges   in   achievement   across  the  sample  with  scores  for  spoken  language  measures  ranging  between  15-­‐57  marks  out  of  a  total  of   116  (receptive  vocabulary,  N  =  29;  elicited  imitation,  N  =  87).  For  FL  literacy,  scores  were  similarly  diverse   with   children’s   performances   ranging   from   1   to   9.5   out   of   a   total   of   18   (reading   aloud,   N   =   10;   reading   comprehension,  N  =  8  –  Table  15).  By  post-­‐test,  both  means  and  medians  had  risen,  but   it  should  be  noted   that   there  was   still   a   substantial   range   in  achievement   scores   in  both  FL  general  proficiency  and  global  FL   literacy  variables  (Table  16).  At  delayed  post-­‐test,  both  means  and  medians  had  risen  slightly  for  FL  general   proficiency  and  had  dropped  slightly   for  the  FL  global   literacy  scores  (Table  17).  Again,   the  range   in  scores   remained  wide  between  the  higher  and  lower  achievers.     Table 15 Descriptive statistics pre-test FL general proficiency and FL global literacy (N = 23) Pre-test Mean Median SD Range Skewness (S.E. Skewness) Kurtosis (S.E. Kurtosis) Min Max FL general proficiency (max 116) 36.609 41.000 9.981 15.0 57.0 -.822 (.481) .950 (.935) FL global literacy (max 18) 5.370 5.500 2.2270 1.0 9.5 .247 (.481) -.250 (.935)   Table 16 Descriptive statistics post-test FL general proficiency and FL global literacy (N = 23) Post-test Mean Median SD Range Skewness (S.E. Skewness) Kurtosis (S.E. Kurtosis) Min Max FL general proficiency (max 116) 45.391 46.000 10.4174 21.0 66.0 -.413 (.481) .346 (.935) FL global literacy (max 18) 6.957 6.500 2.4117 3.5 13.0 .482 (.481) .176 (.935)   Table 17 Descriptive statistics delayed post-test FL general proficiency and FL global literacy (N = 23) Post-test Mean Median SD Range Skewness (S.E. Skewness) Kurtosis (S.E. Kurtosis) Min Max FL general proficiency (max 116) 46.652 48.000 10.7602 26.0 67.0 -.304 (.481) -.379 (.481) FL global literacy (max 18) 6.500 6.500 2.5628 2.5 12.5 .433 (.481) -.130 (.935)       Learners’   progress   on   FL   general   proficiency   and   FL   literacy   constructs   will   now   be   explored   to   illustrate  broader  FL  learning  during  the  overall  teaching  programme  (Tables  18  and  19).  As  these  tests  also   comprised  two  literacy-­‐related  test  items,  it  is  interesting  to  note  progress  made  beyond  the  sub-­‐skills  of  FL   phonological  decoding  measured  through  the  FL  reading  aloud  word  card   test.  These  analyses  also  explore   whether  this  broader  FL  learning  was  durable  by  examining  scores  between  post-­‐  and  delayed  post-­‐test.       Table  18  shows  that  from  pre-­‐test  to  post-­‐test  (immediately  after  the  teaching  programme),  students   made  statistically  significant  progress  across  both  FL  combined  measures.  Global  general  proficiency  scores   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 98   (pre-­‐test  mdn   =   39.609;   post-­‐test  mdn   =   45.391,   z   =   2.930,   p   =   0.56;   r   =   0.43)   and   global   literacy   scores   increased  moderately   (literacy:   pre-­‐test  mdn   =   5.50;   post-­‐test  mdn   =   6.50;   z   =   2.930,   p   =   .003;   r   =   0.43).   Between  post-­‐test  and  delayed  post-­‐test,  medians  either   remained   the  same  or   increased  slightly  but  were   not  statistically  significant  (Table  19),  showing  that  the  FL  learning  was  relatively  durable.     Table 18 Pre-Test to Post-Test FL general proficiency and FL global literacy scores – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Pre-test Post-test FL General Proficiency 39.609 45.391 3.771 <.0001 30.763 0.56 FL Global Literacy 5.50 6.50 2.930 .003 28.666 0.43 Note. Max score FL general proficiency 116; FL global literacy 18.   Table 19 Post-Test to Delayed Post-Test FL general proficiency and FL global literacy scores – Wilcoxon Signed Rank Tests (N = 23) Median z p S.E. Effect Size (Cohen’s d) Pre-test Post-test FL General Proficiency 45.391 46.652 1.123 .261 30.708 0.17 FL Global Literacy 6.50 6.50 -1.051 .293 26.646 -0.16 Note. Max score FL general proficiency 116; FL global literacy 18.   In  Table  19,  comparison  of  FL  global  literacy  scores  using  a  Wilcoxon  Signed  Rank  Test  shows  that  there   was  no  significant  difference  in  scores  between  post-­‐test  (Mdn  =  6.50)  and  delayed  post-­‐test  (Mdn  =  6.50),  z  =   -­‐1.051,  p  =   .293,  thereby  showing  that  the  progression  made  in  FL  literacy  during  the  teaching  intervention   appeared   relatively   long-­‐term.   However,   general   proficiency   scores   increased   slightly   (though   not   significantly)  between  post-­‐test   (Mdn  =  45.391)  and  delayed  post-­‐test   (Mdn  =  46.652),   z  =  1.123,  p  =   .261,   small   effect   size   r   =  0.17).   Section  3.3.3  will   turn   to   global   FL  proficiencies   (general   proficiency   and  global   literacy)  to  establish  whether  these  abilities  were  related  to  the  FL  word  card  reading  aloud  scores  at  each   test  time.       3.3.3.  Global  FL  and  FL  reading  aloud  scores   The  relationships  between  FL  general  proficiency  and  FL  global  literacy  with  FL  reading  aloud  (word   card)  outcomes  at  all  test  times  were  explored  using  Spearman’s  Rho  correlations  (Table  20).  All  pre-­‐test  FL   outcomes   (with   the   exception   of   FL   global   literacy   at   mid-­‐test   FLRA)   showed   statistically   significant   relationships  with  word   reading   outcomes   at  mid-­‐,   post-­‐,   and   delayed   post-­‐test   times.   The   influence   of   FL   outcomes  increased  in  strength  immediately  after  the  teaching  programme  (at  post-­‐test)  and  also  at  delayed   post-­‐test.  General  proficiency  and  literacy-­‐based  FL  outcomes  showed,  broadly  speaking,  the  same  strength  of   influence  at  post-­‐test,  evidenced  by  similar  rs  values.       Table 20 Relationships between combined FL outcomes and FL reading aloud word card (familiar and unfamiliar) at all test times Mid-Test FLRA (fam-unfam) Post-Test FLRA (fam-unfam) Delayed Post-Test FLRA (fam-unfam) rs sig rs sig rs sig Pre-test FL General Proficiency .480* .020 .486* .019 .568** .005 Pre-test FL Global Literacy .379 .074 .561** .005 .570* .005 Post-test FL General Proficiency .643** .001 .552** .006 .626** .001 Post-test FL Global Literacy .612** .002 .605** .002 .690** <.0001 Delayed Post-test FL General proficiency .501* .015 .535** .009 .562** .005 Delayed Post-test FL Global Literacy .583** .004 .786** <.0001 .781** <.0001 LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 99   Whilst  it  seems  likely  from  extant  literature  that  these  FL  and  L1  variables  could  be  predictors  of  FL   outcomes   including   FL   reading   aloud   outcomes,   the   sample   size   (n   =   23)   precluded   statistical   analysis   to   identify  such  predictors  (Field,  2014).     4.  Discussion   4.1.  To  what  extent  do  learners  engaged  in  systematic  and  explicit  FL  phonics  instruction  make   progress   in  reading  aloud   familiar  and  unfamiliar  FL  words  at  mid-­‐,  post-­‐  and  delayed  post-­‐ test?       For   familiar  word   card   reading   aloud,   the   children  made   slow  but   statistically   significant   progress   both  during  and  immediately  after  the  FL  phonics  teaching  programme.  The  slow  progression  is  evidenced  by   a   low  mean   2.174/median   2.00   at   post-­‐test   out   of   a   total   score   of   5.   Unfamiliar  words   and   application   of   sound/spelling   links  showed  similar  patterns.  Ten  weeks   into   the  phonics   instruction  programme  this  sub-­‐ sample   (n  =  23)  had  made  almost  no  progress   in  applying   learned  FL  sound/spelling   links   to  novel  words.   The   children’s   progression   (for   unfamiliar   word   cards)   between   mid-­‐   and   post-­‐test   was   slow   (mean   .9130/median  1.00  at  post-­‐test)  but  statistically  significant  with  a  moderate  effect  size  (rs  =  .40),  likely  due  to   a   zero-­‐starting   point.   These   findings   resonate   with   prior   research   in   a   similar   context   which   found   that   sound/spelling   links,   particularly   those   taught   in   French   (an   irregular   orthography)   developed   slowly   and   that  children’s  reproductions  tended  to  reflect  L1  sound/spelling  links  (Cable  et  al.,  2010).  An  investigation  of   phonics  instruction  conducted  in  English  secondary  schools  (children  aged  11-­‐13)  found  that,  after  16  weeks   of   systematic  and   targeted  weekly   sessions,   some  participants   still   failed   to   successfully  decode  written  FL   non-­‐words   through  a   sound-­‐alike   test   (Woore  et   al.,   2018).   It   seems   likely   that  development  of  FL   reading   aloud  skills   in  both  familiar  and  unfamiliar  contexts   is  a   long-­‐term  pedagogic  endeavour  and,  consequently,   progress  may  have  been  further  hindered  by  the  relatively  low  amounts  of  weekly  instructional  time,  which  is   known  to  affect  primary  FL  outcomes  (Graham  et  al.,  2017).       Like  other  primary  and  secondary  school  studies  (Cable  et  al.,  2010;  Woore  et  al.,  2018),  this  study   found   that   the   range  of  performances  was  diverse  at   all   test   times.  Almost  one   third  of   the   sample   (n  =  7)   scored  zero  for  reading  familiar  words  at  mid-­‐test,  ten  weeks  after  the  phonics  instruction  commenced  and,   after  23  weeks  of   learning,   there  were  still   children  who  were  unable   to   recognise  and  pronounce   familiar   words   and   to   decode   unfamiliar   words.   Correlational   analyses   showed   that   during   and   after   the   teaching   programme   both   L1   reading   proficiency   and   FL   proficiencies   were   related   to   achievement   at   FL   reading   aloud.   It   could   be   possible   that   the   learners   who   apparently   made   no   progress   following   the   phonics   instruction  programme  had   lower  L1   literacy  skills.   It  should  also  be  noted  that,   in   this  sample,  L1  spelling   ages   showed   considerable   range   (7.08   years   to   17.0   years).   This   measure   was   found   to   have   strong   relationships  with  FL  outcomes   in   this   study  and  L1  spelling   is  also  known   to  be   influenced  by  L1  phonics   instruction  (Castles  et  al.,  2018),  therefore  L1  spelling  ranges  could  have  explained  some  of  the  range  in  FL   reading  aloud  scores.  Perhaps  the  phonics  pedagogy  should  have  focused  more  systematically  on  practising   writing  both  exemplar  and  other  words  incorporating  the  targeted  sound/spelling  links.  On  a  broader  scale,  it   should  be  noted  that  diversity   in  FL  outcomes  is  a  known  phenomenon  in  FL  education  for  young  learners.   Other  studies  across  Europe  have  found  that  FL  oracy  and  literacy  scores  vary  considerably  amongst  learners   of  the  same  ages;  even  within  the  same  classes  (Cable  et  al.,  2010;  Enever  et  al.  2011;  Graham  et  al.,  2017).         Finally,   the   kinds   of   sound/spelling   links   taught   also   appeared   to  matter   to   some   extent.   The   best   performances  at  post-­‐test  for  familiar  words  centred  around  novel  FL  sounds  (e.g.,  nasal  vowels)  which  had   to   be   linked   to   existing   L1   spellings.   Those   involving   extant   L1   sounds   which   required   new  mappings   to   different  FL   spellings   seemed  more  problematic.  This  perhaps   could  be   explained  by   the  distinctiveness  of   novel  FL  sounds  which  are  easier   to  perceive  and  therefore   link  and/or  remap  to  their  respective  spellings   (Flege,  1995).  It  should  also  be  noted  that  the  phonics  instruction  which  involved  both  gestures  and  pictures   worked  particularly  well  as  clues  for  the  articulatory  properties  of  nasality.  For  example  the  nasal  vowel  /ɔ̃/   was  linked  to  the  exemplar  word  coch  (pig).  This  was  accompanied  by  a  picture  of  a  pig  and  whole  class   enactment   of   an   action  which   involved   pushing   the   tip   of   the   nose   to   resemble   a   snout  whilst   saying   the   sound.  This,  of  course,  drew  attention  to  the  nasality  and  potentially  acted  as  a  retrieval  cue  for  articulatory   properties.   However,   no   phonetic   analysis   of   sound   perception   or   production   was   involved   in   the   study,   therefore  this  explanation  should  be  treated  with  caution.     PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 100   4.2.  Are  performances  on  reading  aloud  familiar  and  unfamiliar  word  cards  different  at  mid-­‐,   post-­‐  and  delayed  post-­‐test?       Similar  patterns   for   familiar  and  unfamiliar  words   showed   that   learners  made  slow  but   significant   progress  at  reading  both  word  types  during  the  teaching.  Comparison  of  familiar  versus  unfamiliar  word  card   performance  at  each  test  time  also  showed  consistently  significant  differences.  In  other  words,  just  10  weeks   after  the  phonics  programme  started,  familiar  words  were  more  successfully  decoded  than  unfamiliar  words   (p  =  .001;  rs  =  0.49  moderate).  This  preference  was  replicated  at  post-­‐test  (p  <  .0001;  rs  =  -­‐.057)  and  delayed   post-­‐test  (p  <  .0001;  rs  =  -­‐.052).  At  each  time  point  learners  were  significantly  better  at  reading  familiar  words   compared  to  unfamiliar  words.       These  results  show  that  enhanced  practice  and  exposure  is  key  in  phonics  instruction  in  developing   reading   aloud   of   familiar   and   unfamiliar  words.   It   demonstrates   the   importance   of   familiarity  with  words   (item-­‐based  learning)  to  ensure  accurate  decoding,  in  line  with  L1  literature  (Share,  1999)  and  frequency  of   presentation  (Share,  1999).  This  could  also  be  linked  to  active  decoding  and  the  development  of  better  quality   lexical  representations  (spelling,  sound  and  meaning)  in  memory  (Castles  et  al.,  2018;  Ehri,  2014).  However,   there   are   a   number   of   complex   issues   surrounding   the   difference   in   performance   between   known   and   unknown   words.   Perhaps   the   multimodal   (enhanced)   nature   of   presentation   of   phonics   exemplar   words   (written   words,   pictures,   gestures   plus  meaning)   could   have   boosted   the  memorisation   of   familiar   words   such  as  sge  (monkey)  (Boyer  &  Ehri,  2011).         Due   to   frequency  of  presentation  and   rehearsal,   it   is  possible   that   these  exemplar  words  were  not   actually  phonologically  decoded  in  a  conscious  manner  and  could  have  been  linked  as  whole  word  forms  to   their  corresponding  phonological  representation.  It  is  also  feasible  that  the  children  had  memorised  the  oral   form   of   the  word   and   could   have   generated   this   from   the   picture   on   the  word   card   but   it   is,   nonetheless   interesting,  that  when  the  visual  was  presented  alongside  the  written  form,  some  learners  were  still  unable  to   suppress  the  corresponding  L1  mappings  and  produced  a  non-­‐target-­‐like  form  for  the  familiar  word  after  23   weeks  of  phonics   instruction.  Again,   this   seems   to  suggest   that   learning  FL  sound/spelling   links   takes   time   and   plenty   of   practise.   Historic   pedagogic   advice   and   anecdotal   teacher   opinion   contends   that   FL   reading   should  be  avoided  until   reliable  phonological  representations  of  words  have  been  established  so   that   these   will  counteract  L1-­‐based  mispronunciations  (Hurrell,  1999).  This  study   instead  suggests   that  print-­‐induced   L1   mispronunciations   will   result   regardless   of   timing   for   L1   literate   learners.   It   seems   possible   that   increasing   success   rates   at   post-­‐test   show   that   early   opportunities   to   tackle   FL   reading   can   yield   eventual   target-­‐like   decoding.   However,   as   section   5.1   notes,   it   is   beyond   the   remit   of   this   study   to   attribute   any   outcomes  directly   to   the   efficacy   of   FL  phonics   instruction   and  pedagogic   recommendations  would   benefit   from  further  research.       Evidently  the  phonological  decoding  of  unfamiliar  words  is  the  bona  fide  indicator  of  application  of   conscious   sound/spelling   link   knowledge.   However,   particular   sound/spelling   links   which   were   more   successfully  realised  in  familiar  forms  performed  better  in  unfamiliar  contexts  too,  and  this  seems  to  indicate   that  there  could  be  a  link  between  memorisation  of  exemplars  and  conscious  phonological  decoding.  Whilst   not   a   focus   of   analysis,   learners   tended   to   make   L1   related   errors   when   phonological   decoding   such   as   substituting   ON   =   /ɒn/   for   ON   =   /ɔ̃/.   This   seems   to   show   that   a   degree   of   automatized   L1   phonological   decoding  was   indeed  activated,  even   for  conscious  or  effortful  decoding  of  unfamiliar  words  and  successful   performance   could   have   indicated   whether   children   were   able   to   override   these   automatized   L1   sound/spelling   link  knowledge.   It   appears   that,  whilst  many  of   them  were  making  progress   in   this   regard,   reading   in   an   additional   language   is   a   dual   (or   multiple)   language   affair,   and   overriding   competing   L1   knowledge  is  challenging  (Cable  et  al.,  2010;  Grabe  &  Stoller,  2002;  Koda,  2008;  Woore,  2009).  Perhaps  the   learners   could   have   benefited   from   more   opportunities   to   apply   practised   sound/spelling   links   in   novel   contexts,   thereby   strengthening   recall   of   FL   sound/spelling   link   knowledge.   The   phonics   instruction   programme  could  have  included  more  opportunities  to  practise  reading  a  range  of  unknown  words.     4.3.   To   what   extent   do   L1   literacy   measures   and   FL   global   proficiency   scores   relate   to   FL   reading  aloud  proficiency  at  each  test  time?       Statistical  analyses  showed  that  all  L1  literacy  measures  (both  standardised  test  scores  and  teacher   assessments)  correlated  strongly  and  positively  at  all  test  times  with  FL  word  reading  aloud  test  scores.  This   indicated,   as   expected   from   the   literature,   that   successful   L1   readers   and   writers   were   likely   to   be   more   successful   at   FL   reading   aloud.   Wider   research   has   found   that   young   learners   FL   literacy   and   indeed   FL   LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 101   outcomes   generally   are   likely   to   be   influenced   by   L1   literacy   (Courtney   et   al.,   2017;   Sparks   &   Ganschow,   2012).  The  relationships  between  L1  literacy  and  FL  word  card  reading  aloud  decreased  very  slightly  at  post-­‐ test,  then  increased  again  at  delayed  post-­‐test.  This  was  despite  the  differences  between  post-­‐test  scores  and   delayed   post-­‐test   scores   not   being   statistically   significant.   It   seems   possible   that   this   could   relate   to   a   particularly   low   start   for   FL   reading   aloud   scores   which   increased   immediately   following   the   FL   phonics   instruction  programme.   It   could  also   show   that   some  of   the   lower  achieving  L1   literacy   learners   increased   their  FL  reading  aloud  skills  as  their  L1  literacy  skills  were  improving  too.  This  was  not  evidenced  as  the  L1   data   collection  was   focused  on   the  beginning  of   the   school   year,   however,   it   is   likely   that   L1   literacy   skills   would   be   boosted   markedly   during   the   last   two   years   of   primary   school   when   pupils   are   preparing   for   national  standardised  literacy  and  numeracy  tests.         Both  L1  reading  age  and  L1  spelling  age  were  influential  at  all  test  times,  although  spelling  age  was   marginally  more  important,  tending  to  demonstrate  strong  rather  than  moderate  relationships.  This  could  be   because  the  L1  reading  age  test  examines  word  recognition  and  reading  comprehension  by  asking  student  to   select   the   correct   written  word   from   a   range   of   options   to   complete   a   gapped   sentence,   whilst   the   SWST   spelling   test   involves   typing   or   writing   a   target   word   in   a   gapped   sentence   from   an   auditory   cue   (e.g.   orthographic  decoding).  Arguably  the  SWST  would  act  as  a  more  direct  indicator  of  L1  relationships  between   sounds   and   letters.   It   could   also   be   relevant   to   note   that   the   SWST   scores   showed   greater   range   than   the   NGRT  scores  and   this  perhaps   influenced   the   statistical   analyses.  This   could  also   reflect   links  with  phonics   knowledge  and  spelling  evidenced  in  the  L1  literature  (Rosenthal  &  Ehri,  2011).       Interestingly,   FL  outcomes  were   generally   related   to  FL   reading   aloud  word   card   scores   at   all   test   times,  with  the  exception  of  FL  global  literacy  scores  at  mid-­‐test.  So,  if  learners  were  high  achievers  at  either   FL   literacy   or   FL   global   proficiency,   then   they   tended   to   score   higher   for   FL  word   card   reading   aloud.   As   discussed  in  the  results  section,  the  sample  size  was  not  large  enough  to  explore  these  variables  as  predictors,   so  it  is  impossible  to  discern  whether  these  FL  proficiencies  are  likely  to  have  determined  performance.  It  is   also   important   to   note   that   relationships   between   FL   outcomes   and   FL   reading   aloud   could   have   been   mediated  by   other   factors   such   as   learner   self-­‐efficacy   and  motivation,   as   earlier   young   learner   FL   studies   have  noted  links  between  attitudinal  factors  and  learning  outcomes  (Courtney,  2017;  Graham  et  al.,  2017).       5.  Conclusion       To   summarise,   over   the   course   of   a   23-­‐week   programme   of   systematic   weekly   phonics   activities   where  French  sound/spelling   links  were   taught  and  practised,   learners  made  steady  progress.  The  reading   aloud   process   showed   great   variation   inter-­‐participant,   but   the   overall   learning,  whilst   slow,   tended   to   be   durable  and  was  evident  at  least  seven  weeks  after  the  instructional  programme  finished.  The  influence  of  L1   literacy-­‐related  skills,  particularly  L1  spelling  age,  was   important   for  FL  reading  outcomes.   In  other  words,   skilled   L1   readers   had   a   clear   advantage   for   reading   aloud   FL  words.   For   all   learners,   familiar   FL   phonics   exemplar  words  were  more  accurately  read  aloud  than  unfamiliar  words,  but  after  23  weeks  of   instruction,   progress  applying  learned  sound/spelling  links  to  novel  (unfamiliar)  contexts    started  to  emerge.  This  study   shows   that   learning   to   read   aloud   FL   words   is   a   slow   process   but   that   children   do   make   progress.   On   a   broader  scale,  this  study  confirmed  that  young,  beginner  FL  learners  are  able  to  make  meaningful  progress  in   limited  input  classroom  settings.  It  was  also  apparent  that  whilst  L1  literacy  skills  have  facilitative  role  in  FL   outcomes,  the  process  of  re-­‐mapping  (and/or  suppressing)  L1  sound/spelling  links  is  a  complex  and  arduous   one.     5.1.  Limitations       It  is  important  to  recall  that  this  was  not  a  direct  evaluation  of  the  phonics  instructional  programme   alone,   as   the   children  were   engaged   in   a   rich   programme   of   oracy   and   literacy   activities,   which   has   been   documented   in   the   supplementary   materials.   This   is   in   line   with   wider   literacy-­‐related   research   which   evidences  the  importance  of  both  phonics   instruction  and  meaning-­‐based  text  activities  to  support   learning   (Adams,  1990).  To  evaluate  the  direct  implications  of  the  phonics  instruction  package,  one  school  would  have   had  to  have  only  the  phonics  instructional  programme.  As  little  is  known  about  best  practice  for  FL  phonics   instruction   or   its   effectiveness,   this   study   undertook   to   investigate   the   possible   outcomes   and   learning   processes  in  a  limited  input,  limited  timeframe  FL  setting.  It  is  also  important  to  consider  that  teacher  factors   such  as   linguistic  ability,   language  specialist   training,  and  teaching  time  are   likely  to  underpin  FL  outcomes   (Graham  et  al.,  2017)  and  could  have  influenced  progression  in  this  context.  Arguably,  the  teacher/researcher   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 102   stance   meant   that   this   study   was   implemented   by   someone   who   was   heavily   invested   in   the   teaching/learning  process.       5.2.  A  future  research  agenda       As   it  cannot  be  assumed  that   the  benefits   for  phonics   in  English  L1   literacy   learning  can  simply  be   transferred   to   any   FL   context,   future   research   should   aim   to   investigate   the   effectiveness   of   FL   phonics   in   primary  schools  in  a  controlled  experiment.  It  would  also  be  useful  to  examine  the  learning  of  sound/spelling   links  of  different  foreign  languages  in  young  learner  instructed  settings  as  well  as  the  teaching  and  learning  of   FL  literacy  for  learners  with  different  L1  (non-­‐alphabetic)  backgrounds  such  as  Chinese.       Funding   This   study  was   funded   through   an  ESRC  University   of   Southampton  Doctoral   Training  Centre   +3  Research   Studentship.     Acknowledgements   Figure  1  Images:  Image  of  a  pig  produced  by  https://commons.wikimedia.org/wiki/User:Joshua_Lutz  is  in  the   public  domain.     Image  of  a  fingernail  (https://commons.wikimedia.org/wiki/File:Ongle_du_doigt_de_la_main_-­‐ _Fingernail_hand.jpg)  produced  by  Bierenard  is  licensed  under  CC-­‐SA  4.0  International   (https://creativecommons.org/licenses/by-­‐sa/4.0/deed.en).                       References   Adams,  Marilyn   J.   (1990).  Beginning   to   read:  Thinking  and   learning  about   print.   Cambridge,  Massachusetts:   MIT  Press.   Antony,  Jason  L,  Lonigan,  Christopher  J.,  Burgess,  Stephen  R.,  Driscoll,  Kimberley,  Phillips,  Beth  M.,  &  Cantor,   Brenlee  G.  (2002).  Structure  of  pre-­‐school  phonological  sensitivity:  Overlapping  sensitivity  to  rhyme,   words,  syllables  and  phonemes.  Journal  of  Experimental  Child  Psychology,  85,  65-­‐92.     Apel,   Kenn,   Wilson-­‐Fowler,   Elizabeth   B.,   Brimo,   Danielle,   &   Perrin,   Nancy   A.   (2012).   Metalinguistic   contributions  to  reading  and  spelling  in  second  and  third  grade  students.  Read  Write,  25,  1283-­‐1305.     Bassetti,   Bene,  &  Atkinson,  Nathan.   (2015).   Effects   of   orthographic   forms   on  pronunciation   in   experienced   instructed  second  language  learners.  Applied  Psycholinguistics,  36(1),  67-­‐91.   Biemiller,   Andrew,   &   Slonim,   Naomi   (2001).   Estimating   root   word   vocabulary   growth   in   normative   and   advantaged   populations:   Evidence   for   a   common   sequence   of   vocabulary   acquisition.   Journal   of   Educational  Psychology,  93(3),  498-­‐520.   Boyer,   Nancy,   &   Ehri,   Linnea   (2011).   Contribution   of   phonemic   segmentation   instruction  with   letters   and   articulation  pictures   to  word  reading  and  spelling   in  beginners.  Scientific  Studies  of  Reading,  15(5),   440-­‐470.   Cable,   Carrie,   Driscoll,   Patricia,   Mitchell,   Rosamond,   Sing,   Sue,   Cremin,   Teresa,   Earl,   Justine,   Eyres,   Ian,   Holmes,   Bernardette,   Martin,   Cynthia,   &   Heins,   Barbara.   (2010).   Primary   modern   languages:   A   longitudinal  study  of  language  learning  at  Key  Stage  2.  Nottingham,  United  Kingdom:  Department  for   Children,  Schools  &  Families.   Campfield,  Dorotea   (2015).   Lexical   difficulty   –   using   elicited   imitation   to   study   child   L2.  Language  Testing,   34(2),  1-­‐12.     Castles,   Anne,   Rastle,   Kathleen,   &  Nation,   Kate   (2018).   Ending   the   reading  wars:   Reading   acquisition   from   novice  to  expert.  Psychological  Science  in  the  Public  Interest,  19,  5-­‐51.   Chall,  Jeanne  (1967).  Learning  to  read:  The  great  debate.  New  York,  New  York:  McGraw  Hill.   Charlton,  Adeline  (2014).  The  second  language  acquisition  of  French  nasal  vowels  by  advanced  English  learners.   Second  Language  Acquisition  of  Phonology  2012  workshop,  University  of  York,  2012.  Retrieved  from   http://www-­‐users.york.ac.uk/~sh581/POSTER-­‐ELOL%20Charlton.pdf.   LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 103   Courtney,   Louise   (2017).   Transition   in   modern   foreign   languages:   a   longitudinal   study   of   motivation   for   language  learning  and  second  language  proficiency.  Oxford  Review  of  Education,  43(4),  462-­‐481.   Courtney,   Louise,   Graham,   Suzanne,  Marinis,   Theo,   &   Tonkyn,   Allan   (2017).   Individual   differences   in   early   language  earning:  A  study  of  English  learners  of  French.  Applied  Linguistics,  38(6),  824-­‐847.   Cummins,   Jim   (1979).   Linguistic   interdependence   and   the   educational   development   of   bilingual   children.   Review  of  Educational  Research,  49,  222–251.     Deacon,  S.  Hélène,  &  Kirby,   John  R.   (2004).  Morphological  awareness:   Just  more  phonological?  The  roles  of   morphological   awareness   and   phonological   awareness   in   reading   development.   Applied   Psycholinguistics,  25,  223-­‐238.   Department   of   Education,   Science   &   Training   (2005).   Teaching   reading:   Report   and   recommendations.   National  Inquiry  into  the  Teaching  of  Literacy.   DfE   (2013).  The   national   curriculum   in   England.   Key   Stages   1   and   2   Framework   Document.   Retrieved   from   https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/fil e/425601/PRIMARY_national_curriculum.pdf.   DfES  (2005).  The  key  stage  2  framework  for  languages.  Nottingham,  United  Kingdom:  DfES  Publications.     Dunn,  Lloyd  M.,  &  Dunn,  Leota  M.  (2007).  PPVT-­‐4  Pearson  Picture  Vocabulary  Test.  Minneapolis,  Minnesota:   Pearson  Assessments.   Dunn,   Lloyd   M.,   Dunn,   Leota   M.,   Whetton,   Chris,   &   Burley,   Juliet   (1997).   British   Picture   Vocabulary   Scale.   Windsor,  United  Kingdom:  NFER.   Durgunoğlu,   Aydin   Y.,   Nagy,   William   B.,   &   Hancin-­‐Bhatt,   Barbara   J.   (1993).   Cross-­‐language   transfer   of   phonological  awareness.  Journal  of  Educational  Psychology,  85,  453-­‐465.   Ehri,  Linnea,  C.  (2014).  Orthographic  mapping  in  the  acquisition  of  sight  word  reading,  spelling  memory  and   vocabulary  learning.  Scientific  Studies  of  Reading,  18(1),  5-­‐21.   Ehri,   Linnea,   C.   (1999).   Phases   of   development   in   learning   to   read   words.   In   Jane   Oakhill   &   Roger   Beard   (Eds.),   Reading   development   and   the   teaching   of   reading   (pp.   79-­‐108).   Oxford,   United   Kingdom:   Blackwell.   Ehri,   Linnea,   C.   (1992).   Reconceptualising   the   development   of   sight   word   reading   and   its   relationship   to   recoding.  In  Phillip  B.  Gough,  Linnea  C.  Ehri,  &  Rebeca  Treiman  (Eds.),  Reading  acquisition  (pp.  107-­‐ 143).  Hillsdale,  New  Jersey:  Lawrence  Erlbaum.   Ehri,  Linnea  C.,  Nunes,  Simone  R.,  Stahl,  Steven  A.,  &  Willows,  Dale  M.  (2001).  Systematic  phonics  instruction   helps   students   learn   to   read:  Evidence   from   the  National  Reading  Panel’s  meta-­‐analysis.   Journal   of   Direct  Instruction,  2(2),  121-­‐166.   Enever,   Janet,   Krikhaar,   Evelien,   Lindgren,   Eva,   Lundberg,   Gun,   Mihaljević   Djigunović,   Jelena,   &   Muñoz,   Carmen   (2011).   ELLiE   –   Early   Language   Learning   in   Europe.   London,   United   Kingdom:   British   Council.   Erlam,  Rosemary   (2006).  Elicited   imitation  as   a  measure  of   L2   implicit   knowledge:   an   empirical   validation   study.  Applied  Linguistics,  27,  464-­‐491.   Escudero,   Paola,   &   Boersma,   Paul   (2004).   Bridging   the   gap   between   L2   speech   perception   research   and   phonological  theory.  Studies  in  Second  Language  Acquisition,  26,  551–585.   Escudero,   Paola   &  Wanrooij,   Karin   (2010).   The   effect   of   L1   orthography   on   non-­‐native   vowel   perception.   Language  and  Speech,  53(3),  343-­‐365.   Flege,   Jim  E.   (1995).  Second   language  speech   learning.  Theory,   findings,  and  problems.   In  Winifred  Strange   (Ed.),   Speech   perception   and   linguistic   experience.   Issues   in   cross-­‐language   research   (pp.   233-­‐277).   Baltimore,  Maryland:  York  Press.   Garton,  Susan,  Copland,  Fiona,  &  Burns,  Ann  (2011).  Investigating  global  practices  in  teaching  young  learners.   London,  United  Kingdom:  British  Council.   Genesee,   Fred,   Geva,   Esther,   Dressler,   Cheryl,   &   Kamil,   Michael   L.   (2008).   Cross-­‐linguistic   relationships   in   second   language   learners.   In   Diane   A.   August   &   Timothy   Shanahan   (Eds.),  Developing   reading   and   writing  in  second  language  learners  (pp.  63-­‐93).  New  York/Abingdon:  Routledge.   GL   Assessment.   (2018).   New   Group   Reading   Test   (NGRT).   Retrieved   from   https://www.gl-­‐ assessment.co.uk/products/new-­‐group-­‐reading-­‐test-­‐ngrt/).   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 104   GL   Assessment.   (2018).   Single   Word   Spelling   Test   (SWST).   Retrieved   from   https://www.gl-­‐ assessment.co.uk/products/single-­‐word-­‐spelling-­‐test-­‐swst/).   Goswami,  Usha  (2006).  Orthography,  phonology  and  reading  development:  A  cross-­‐linguistic  perspective.  In   R.  Malatesha  Joshi  &  P.  G.  Aaron  (Eds.),  Handbook  of  orthography  and  literacy  (pp.  463-­‐480).  Mahwah,   New  Jersey:  Laurence  Erlbaum.   Gough  Phillip  B.,  &  Wren,   Sebastian   (1999).   Constructing  meaning:  The   role   of   decoding.   In   Jane  Oakhill  &   Roger   Beard   (Eds.),   Reading   development   and   the   teaching   of   reading   (pp.   59-­‐78).   Oxford,   United   Kingdom:  Blackwell.   Grabe,   William,   &   Stoller,   Frederika   L.   (2002).   Teaching   and   researching   reading.   Harlow,   Essex:   Pearson   Education.   Graham,   Suzanne,   Courtney,   Louise,   Marinis,   Theo,   &   Tonkyn,   Allan   (2017).   Early   language   learning:   The   impact  of  teaching  and  teacher  factors.  Language  Learning,  67(4),  1-­‐37.     Hatcher,  Peter  J.,  Hulme,  Charles,  &  Ellis,  Andrew  W.  (1994).  Ameliorating  early  reading  failure  by  integrating   the   teaching   of   reading   and   phonological   skills:   The   phonological   linkage   hypothesis.   Child   Development,  65(1),  41-­‐57.   Henbest,  Victoria  S.,  &  Apel,  Kenn  (2017).  Effective  word  reading  instruction:  What  does  the  evidence  tell  us?   Communication  Disorders  Quarterly,  39(1),  303-­‐311.   Hurrell,   Alison   (1999).   The   four   language   skills.   In   Patricia   Driscoll   &   David   Frost   (Eds.),   The   teaching   of   modern  foreign  languages  in  primary  school  (pp.  67-­‐87).  London,  United  Kingdom:  Routledge.   Kaefer,   Tanya   (2016).   Integrating   orthographic   and   phonological   knowledge   in   early   readers:   Implicit   and   explicit  knowledge.  Child  Development  Research.  https://doi.org/10.1155/2016/6036129.     Koda,  Keiko  (1992).The  effects  of   lower-­‐level  processing  skills  on  FL  reading  performance:   Implications   for   instruction.  Modern  Language  Journal,  76(4),  502-­‐512.   Koda,  Keiko  (2007).  Reading  and  language  learning:  Cross-­‐linguistic  constraints  on  second-­‐language  reading   development.  Language  Learning,  57(s1),  1-­‐44.   Koda,  Keiko  (2008).  Impacts  of  prior  literacy  experience  on  second-­‐language  learning  to  read.  In  Keiko  Koda   &  Annette  Zehler  (Eds.),  Learning  to  read  across   languages:  Cross-­‐linguistic  relationships   in   first  and   second-­‐language  literacy  development  (pp.  68-­‐96).  New  York/Abingdon:  Routledge.   Koutsompou,   Violetta-­‐Irene   (2015).   The   use   of   literature   in   the   language   classroom.   Methods   and   Aims.   International  Journal  of  Information  and  Education  Technology,  5(1),  74-­‐79.   Lado,   Robert   (1957).   Linguistics   across   cultures:   Applied   linguistics   and   language   teachers.   Ann   Arbor,   Michigan:  University  of  Michigan  Press.     Landerl,  Karin.  (2006).  Reading  acquisition  in  different  orthographies:  Evidence  from  direct  comparison.  In  R.   Malatesha   Joshi   &   P.   G.   Aaron   (Eds.),   The   handbook   of   orthography   and   literacy   (pp.   513-­‐530).   Mahwah,  New  Jersey:  Lawrence  Erlbaum.   Lesaux,  Nonie  K.,   Geva,   Esther,  Koda,  Keiko,   Siegel,   Linda   S.,  &   Shanahan,   Timothy   (2008).  Development   of   literacy   in   second   language   learners.   In   Diane   A.   August   &   Timothy   Shanahan   (Eds.),   Developing   reading  and  writing  in  second  language  learner  (pp.  27-­‐60).  New  York/Abingdon:  Routledge.   Lloyd,   Janet,  &  Wernham,  Susan  (1992).   Jolly  phonics:  A  handbook  for  teaching  reading,  writing  and  spelling.   Essex,  United  Kingdom:  Jolly  Learning.   Lonsdale,  Deryle,  Graham,  C.  Ray,  Kennington,  Casey,  Johnson,  Aaron,  &  McGhee,  Jeremiah  (2008).  Scoring  an   oral  language  test  using  automated  speech  recognition.  In  Nicoletta  Calzolari,  Khalid  Choukri,  Bente   Maegaard,   Joseph  Mariani,   Jan  Odijk,  Stelios  Piperidis,  &  Daniel  Tapias  (Eds.),  Proceedings  of   the  6th   International   Conference   on   Language   Resources   and   Evaluation   (pp.   1604-­‐1610).   Marrakech,   Morocco:  European  Language  Resources  Association.   Macaro,   Ernesto,   &   Erler,   Lynne   (2008).   Decoding   ability   in   French   as   a   foreign   language   and   language   learning  motivation.  The  Modern  Language  Journal,  95(4),  496-­‐517.   Macaro,  Ernesto,  &  Mutton,  Trevor   (2009).  Developing  reading  achievement   in  primary   learners  of  French:   Inferencing  strategies  versus  exposure  to  graded  readers.  Language  Learning  Journal,  37,  165-­‐182.   Martinez,  Márquez   (2016).  The   acquisition   of   French   nasal   vowels:   From   first   language   allophony   to   second   language  phonological  contrast.  Unpublished  PhD  dissertation.  Indiana  University.     LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 105   McBride   Chang,   Catherine,   &   Ho,   Connie   Suk-­‐Han   (2005).   Predictors   of   beginning   reading   in   Chinese   and   English:  A  two  year   longitudinal  study  of  Chinese  kindergartners.  Scientific  Studies  of  Reading,  9(2),   117-­‐144.   McKay,  Penny  (2006).  Assessing  Young  Learners.  Cambridge,  United  Kingdom:  Cambridge  University  Press.   Miskin,  Ruth  (2011).  Read  Write  Inc  phonics  handbook.  Oxford,  United  Kingdom:  Oxford  University  Press.   Mitchell,  Rosamond,  Myles,  Florence,  &  Marsden,  Emma  (2019).  Second   language   learning  theories.  London,   United  Kingdom:  Hodder  Arnold.   Murphy,  Victoria,  Macaro,  Ernesto,  Alba,  Sonia,  &  Cipolla,  Claudia  (2015).  The  influence  of  learning  a  second   language   in   primary   school   on   developing   first   language   literacy   skills.   Applied   Psycholinguistics,   36(5),  1133-­‐1153.   National   Reading   Panel   (2000).   Teaching   children   to   read:   An   evidence-­‐based   assessment   of   the   scientific   research   Literature   on   reading   and   its   implications   for   reading   instruction.   Washington,   District   of   Columbia:  National  Reading  Panel.   Neveu,   Frank,  Muni-­‐Toké,   Valelia,   Klinger,   Thomas,   Durand,   Jacques,  Mondada,   Lorenza,   &   Prévost,   Sophie   (2010).   Evaluation   des   voyelles   nasales   en   français   L2   en   production:   de   la   nécessité   d’un   corpus   multi-­‐tâches.   In   Franck   Neveu,   Valelia   Muni   Toke,   Thomas   Klingler,   Jacques   Durand,   Lorenz   Mondada,   &   Sophie   Prévost   (Eds.),   Proceedings   of   the   Congrès   Mondial   de   Linguistique   Française.   Paris,  France:  EDP  Sciences.  https://doi.org/10.1051/cmlf/2010119.     Nimz,   Katharina   (2018).   Sound   perception   and   production   in   a   foreign   language:   Does   orthography  matter?   Potsdam,  Germany:  Universitätsverlag  Potsdam.   Oakhill,   Jane,   &   Beard,   Roger   (1999).   Reading   development   and   the   teaching   of   reading.   Oxford,   United   Kingdom:  Blackwell  Publishing.   Perfetti,  Charles,  &  Stafura,  Joseph  (2014).  Word  knowledge  in  a  theory  of  reading  comprehension.  Scientific   Studies  of  Reading,  18(1),  22-­‐37.   Porter,  Alison  (2014).  An  early  start  to  French  literacy:  Learning  the  spoken  and  written  word  simultaneously  in   English   primary   schools.   Unpublished   PhD   Thesis.   Southampton,   United   Kingdom:   University   of   Southampton.   Ricketts,   Jessie,  Bishop,  Dorothy  V.,   Pimperton,  Hannah,  &  Nation,  Kate   (2011).  The   role   of   self-­‐teaching   in   learning  orthographic  and  semantic  aspects  of  new  words.  Scientific  Studies  of  Reading,  15(1),  47-­‐70.   Rosenthal,   Julie,  &   Ehri,   Linnea,   C.   (2011).   Pronouncing   new  words   aloud   during   the   silent   reading   of   text   enhances  fifth  graders’  memory  for  vocabulary  words  and  their  spellings.  Read  Write,  24,  921-­‐950.   Share,   David   L.   (1999).   Phonological   recoding   and   orthographic   learning:   A   direct   test   of   the   self-­‐teaching   hypothesis.  Journal  of  Experimental  Child  Psychology,  72,  95-­‐129.   Shmidman,  Adina,  &  Ehri,  Linnea,  C.  (2010).  Embedded  picture  mnemonics  to  learn  letters.  Scientific  Studies   of  Reading,  14(2),  159-­‐182.   Sparks,   Richard   L.  &  Ganschow,   Leonore   (1991).   Foreign   language   learning  differences:  Affective   or   native   language  aptitude  differences.  The  Modern  Language  Journal,  75(1),  3-­‐16.   Sparks,   Richard   L.,   Patton,   Jon,   Ganschow,   Leonore,   &   Humbach,   Nancy   (2009).   Long-­‐term   crosslinguistic   transfer  of  skills  from  L1  to  L2.  Language  Learning,  59(1),  203-­‐243.   Sparks,  Richard  L.,  Patton,  Jon,  &  Ganschow,  Leonore  (2012).  Profiles  of  more  and  less  successful  L2  learners:   A  cluster  analysis  study.  Learning  and  Individual  Differences,  22(4),  463-­‐472.   Stanovich,  Keith  E.,  &  Stanovich,  Paula,  J.  (1999).  How  research  might  inform  the  debate  about  early  reading   acquisition.   In   Jane  Oakhill  &  Roger  Beard   (Eds.),  Reading  development  and   the   teaching  of   reading   (pp.  12-­‐41).  Oxford,  United  Kingdom:  Blackwell  Publishing.   Stuart,   Morag,   Masterson,   Jackie,   &   Dixon,   Maureen.   (1999).   Learning   to   read   words   turns   listeners   into   readers.   How   children   accomplish   this   transition.   In   Jane   Oakhill   &   Roger   Beard   (Eds.),   Reading   development   and   the   teaching   of   reading   (pp.   109-­‐130).   Oxford,   United   Kingdom:   Blackwell   Publishing.   Rose,   Jim   (2006).   Independent   review   of   the   teaching   of   early   reading.   Nottingham,   United   Kingdom:   Department  of  Education  and  Skills.   PORTER   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 106   Torgerson,  Carole  J.,  Brooks,  Greg,  &  Hall,  Jill  (2006).  A  systematic  review  of  the  research  literature  on  the  use  of   phonics  in  the  teaching  of  reading  and  spelling.  DfES  Research  Report  RR711.   Tyler,   Michael   D.,   Best,   Catherine   T.,   Faber,   Alice,   &   Levitt,   Andrea   G.   (2014).   Perceptual   assimilation   and   discrimination  of  non-­‐native  vowel  contrasts.  Phonetica,  71(1),  4–21.   Vellutino,  Frank  R.  &  Fletcher,  Jack  M.  (2005).  Development  dyslexia.  In  Margaret  J.  Snowling  &  Charles  Hulme   (Eds.),  The  science  of  reading:  A  handbook  (pp.  362-­‐378).  Oxford,  United  Kingdom:  Blackwell.   Verhoeven,  Ludo  (2011).  Second  language  reading.  In  Michael  L.  Kamil,  David,  Pearson,  Elizabeth  B.  Moje,  &   Peter  Afflerbach  (Eds.),  Handbook  of  reading  research  (pp.  661-­‐683).  New  York/London:  Routledge.   Woore,  Robert  (2009).  Beginners’  progress  in  decoding  L2  French:  Some  longitudinal  evidence  from  English   modern  foreign  languages  classrooms.  Language  Learning  Journal,  37(1),  3-­‐18.   Woore,  Robert  (2011).  Investigating  and  developing  beginner  learners’  decoding  proficiency  in  second  language   French:  An  evaluation  of  two  programmes  of   instruction.  Unpublished  PhD  thesis.  Oxford:  University   of  Oxford.   Woore,  Robert,  Graham,  Suzanne,  Porter,  Alison,  Courtney,  Louise  &  Savory,  Claire  (2018).  Foreign  language   education   unlocking   reading   (FLEUR):   A   study   into   the   teaching   of   reading   to   beginner   learners   of   French  in  secondary  school.  Oxford,  United  Kingdom:  University  of  Oxford.   Wyse,   Dominic   &   Styles,   Morag   (2007).   Synthetic   phonics   and   the   teaching   of   reading:   The   debate   surrounding  England’s  Rose  Report.  Literacy,  41(1),  37-­‐42.           LEARNING  FRENCH  SOUND/SPELLING  LINKS   E-­‐JournALL  7(1)  (2020),  pp.  78-­‐107 107     Alison Porter, University of Southampton A.Porter@soton.ac.uk EN Alison Porter is a lecturer in Applied Linguistics at the University of Southampton. Her primary research area is young language learning in instructed settings. She is particularly interested in the teaching and learning of foreign language literacy and has participated in teaching and learning research in both early (ages 5-7) and late (ages 9-11) primary school classrooms. She also taught foreign languages in English primary schools for ten years. Most recently, Dr. Porter collaborated on a large-scale teaching intervention study in early secondary school classrooms (ages 11-13) which aimed to evaluate the effects of different types of reading instruction: phonics and strategy use. In 2017, Dr. Porter established a Southampton University Primary Languages (SUPL) Research and Practice partnership with Emeritus Professor Ros Mitchell. SUPL currently collaborates with 20 local primary schools to co-construct, implement, and evaluate foreign language literacy practices in classrooms. Participating teachers are concurrently trained in classroom research methods, and several small- scale investigations have already resulted. Dr. Porter is also a member of the Research in Primary Language network (RiPL), founded by the University of Essex, which engages with practitioners on a local level as well as national stakeholders in a bid to inform policy and sustainability issues relating to primary foreign language provision. ES Alison Porter es una profesora de Lingüística Aplicada en la Universidad de Southampton. Su área de investigación es principalmente el aprendizaje de la lengua en niños en un contexto educativo. Se ocupa sobretodo de enseñanza y aprendizaje de alfabetización de una lengua extranjera y ha contribuido en investigaciones sobre la enseñanza y el aprendizaje tanto en niños de 5-7 años como en los de 9-11 años. Impartió clases de lenguas extranjeras en escuelas primarias durante diez años. Más recientemente, la profesora Porter ha colaborado a un estudio a gran escala sobre la intervención de la enseñanza en las escuelas secundaria (11-13 años de edad) que tenía como objetivo evaluar los efectos producidos por diferentes tipos de enseñanza de lectura: instrucción fónica y uso estratégico. En 2017, la profesora Porter estableció la asociación Southampton University Primary Languages (SUPL) Research and Practice junto con el Profesor émerito Ros Mitchell. Hoy en día SUPL colabora con 20 escuelas primarias locales para desarrollar juntos, implementar y evaluar las practicas de alfabetización de lenguas extranjeras en clase. Los docentes que participan son formados al mismo tiempo en métodos de investigación desarrollando ya numerosas investigaciones a pequeña escala. La profesora Porter es también miembro de Research in Primary Language network (RiPL), fundado por la University of Essex, que se relaciona con tanto con profesionales a nivel local como con stakeholders a nivel nacional con el intento de informar sobre cuestiones de política y sostenibilidad relacionadas con la oferta de la lengua extranjera en la escuela primaria. IT Alison Porter è Ricercatrice in Linguistica applicata presso l’Università di Southampton. La sua area di ricerca principale è l’apprendimento linguistico dei bambini in contesto di istruzione formale. È particolarmente interessata all’insegnamento e apprendimento della letto-scrittura nella lingua straniera e ha partecipato a ricerche sull’insegnamento e apprendimento in classi iniziali (5-7 anni) ed avanzate (9-11) nella scuola primaria. Ha inoltre insegnato lingua straniera nella scuola primaria inglese per dieci anni. Ultimamente, Porter ha collaborato a un ampio studio sull’intervento didattico in classi iniziali della scuola secondaria (11-13 anni) finalizzato alla valutazione degli effetti di tipi diversi di insegnamento della lettura: il metodo fonetico e l’uso di strategie. Nel 2017, Porter ha stabilito una collaborazione dell’Università di Southampton con il professore emerito Ros Mitchell per Southampton University Primary Languages (SUPL) Research and Practice (SUPL). Attualmente, SUPL collabora con 20 scuole locali per la creazione, implementazione e valutazione di pratiche di alfabetizzazione per la classe. Gli insegnanti coinvolti vengono al contempo istruiti sui metodi di ricerca per la classe e ne sono già derivati diversi studi su piccola scala. Porter è anche membro della Research in Primary Language network (RiPL) fondata dall’Università dell’Essex, che si confronta con gli operatori del settore a livello locale e con le parti interessate a livello nazionale nel tentativo di orientare questioni di politiche e sostenibilità relative all’insegnamento delle lingue straniere nella scuola primaria.