Phonological contribution during visual word recognition in child readers. An intermodal priming study in Grades 3 and 5
Abstract
This study investigated the phonological contribution during visual word recognition in child readers as a function of general reading expertise (third and fifth grades) and specific word exposure (frequent and less‐frequent words). An intermodal priming in lexical decision task was performed. Auditory primes (identical and unrelated) were used in order to directly activate phonological codes independently of orthographic processing. Overall, the results revealed a widespread phonological priming effect in both grades. There was a significant interaction between grade, priming condition and frequency, revealing that the impact of frequency on identity priming differed between grades. In third grade, the results indicated that the priming effect was greater for less‐frequent than for frequent words. In fifth grade, priming effects were similar for both frequent and less‐frequent words. These findings indicate that print and speech processing systems are interconnected in young readers. Moreover, phonological codes play an important role in word recognition throughout reading development.
Introduction
At the very first stage of reading acquisition, phonological processing plays a critical role in the translation of unfamiliar printed words into their spoken equivalents (phonological recoding). In skilled adult readers, phonological processing still influences word reading but had changed in nature and acts in a rapid and automatic fashion in parallel with orthographic processes (Rastle & Brysbaert, 2006). However, as reading skills develop, it is still unclear how phonological codes contribute to the word recognition process and whether the phonological contribution is dependent on visual word recognition system development. The present study investigated the contribution of phonological codes to word recognition in relation to reading experience. For that, a lexical decision task was used.
Learning to read fluently involves moving from an effortful phonological recoding process (by grapheme‐phoneme conversion) to an automatic process of recognition of words. According to the self‐teaching theory (Share 1995, 1999), at the very first stage, the successful phonological recoding of novel words provides the opportunity for learners to acquire the word‐specific orthographic knowledge necessary for fast and efficient visual word recognition. Several studies using lexical decision tasks or naming tasks have shown that with increasing reading experience, phonological recoding decreases (Backman, Bruck, Hebert, & Seidenberg, 1984; Coltheart, Laxon, Keating, & Pool, 1986; Seidenberg, Waters, Barnes, & Tanenhaus, 1984; Sprenger‐Charolles & Casalis, 1995; Waters, Seidenberg, & Bruck, 1984). For example, Schmalz, Marinus, and Castles (2013) investigated this issue by examining regularity effects in third and fifth graders who were asked to perform a go/no‐go lexical decision task on high and low frequency regular and irregular words. The children showed regularity effects only for low frequency words, indicating that children are using phonological decoding strategies to recognize unfamiliar words and third graders are already relying predominantly on an orthographic lexical strategy in their silent reading of familiar words. The increasing reliance on lexical orthographic processing as shown by sensitivity to word frequency (Burani, Marcolini, & Stella, 2002; Zoccolotti et al., 2005) is congruent with the results from masked priming experiments showing that children develop an automatic orthographic procedure to speed up lexical access (Acha & Perea, 2008; Castles, Davis, Cavalot, & Forster, 2007; Castles, Davis, & Letcher, 1999). Moreover, the reliance on an automatic orthographic procedure increases with general reading expertise or grade (Lété & Fayol, 2013; Ziegler, Bertrand, Lété, & Grainger, 2014) and with the number of exposures to words as assessed by frequency (Booth, Perfetti, & MacWhinney, 1999). Thus, grade and word frequency are reliable predictors of automatic orthographic processes. At this point, the question is whether phonological code is still contributing to visual word recognition as orthographic procedure is becoming more and more efficient.
To examine whether phonological coding still contributes to visual word recognition in developing reading after initial phases, it is necessary to distinguish among phonological processes described in the literature. First, phonological recoding is sublexical phonological processing involved at the first stages of learning to read and relying on a strict sequential conversion from graphemes to phonemes to activate the whole‐word phonological representations of orally known words. Second, in adult skilled readers, phonological processing refers to fast and automatic activation of phonology from letters processed in parallel. It might take place at both sublexical and lexical levels (see for example the bimodal interactive‐activation model; Grainger & Holcomb, 2010). Indeed, numerous studies show that phonological codes still play a role in expert reading and acts in parallel with orthographic processes (Brysbaert, 2001; Ferrand & Grainger, 1992, 1993, 1994, 1996; Grainger & Ferrand, 1994; Ziegler et al., 2000; for a discussion, see Rastle & Brysbaert, 2006). As mentioned, phonological activation occurs at both sublexical level, by phonological computation from letters, and lexical level, by retrieval of the phonological form from the orthographic lexical representation (for an in‐depth discussion see Frost, 1998).
Therefore, reliance on orthographic processes as reading skills develop does not necessarily mean that phonological activation vanishes after the initial phase of effortful phonological recoding. The masked phonological priming paradigm (Forster & Davis, 1984) has been especially useful in studying the automatic involvement of phonological code in expert reading. In this paradigm, a target word is preceded by a subliminal phonological prime that is phonologically similar to the target and is presented too briefly to be consciously perceived. Numerous studies have shown a facilitation of target recognition when the target word was preceded by a subliminal phonological prime. This demonstrates that phonological code activation by the prime occurs at a very early stage of word processing, at both sublexical and lexical levels.
In child readers, attempts to study automatic phonological activation during written familiar word recognition using masked phonological priming have produced inconsistent results. Davis, Castles, and Iakovidis (1998) failed to evidence activation of phonological code in word recognition in fourth grade in a lexical decision task. In contrast, the masked priming study of Booth et al. (1999) suggested a rapid and automatic involvement of phonological code during word recognition in both younger (mean of 8 years 6 months) and older (mean of 11 years 4 months) child readers. In fact, error rates from the writing task showed that older children benefited more from both orthographic and pseudohomophone priming than younger ones. The authors argued that orthographic and phonological codes were activated faster and more effectively in older readers compared to younger ones because of precise and redundant orthographic and phonological representations (Perfetti, 1992). More recently, Ziegler et al. (2014) tracked the development of phonological and orthographic processing in two masked sandwich priming experiments (Lupker & Davis, 2009), a pseudohomophone priming experiment and a transposed‐letter priming experiment, respectively. The masked sandwich priming, as defined by Lupker and Davis (2009), “involves briefly presenting the target itself prior to the presentation of each prime” (p. 618; e.g., place‐plasse‐PLACE). Ziegler and colleagues found a phonological priming effect that remained stable across grades, whereas the orthographic priming effect increased from first to fifth grade suggesting a greater reliance on orthographic procedure as children become better readers. Thus, these visual masked priming studies using pseudohomophone primes do not provide a clear picture about the development of the phonological contribution as reading experience increases.
An inherent problem in visual masked priming studies that investigate phonological involvement during word recognition in developing readers is that orthographic processing is still progressing. In fact, phonological code activation is entirely dependent on the efficiency of the prior automatic activation of sublexical orthographic representation by visual primes. Given that orthographic processing is not fully efficient in young learners, it is difficult to obtain masked phonological priming from print and thus to evaluate the contribution of phonological code per se during reading acquisition within this procedure. One way to overcome this problem is to activate phonological representations directly without going through orthographic code by using auditory primes as phonological primes. To the best of our knowledge, there has been only one study in children using an audio‐visual priming experiment to examine phonological involvement in accessing the written lexicon. Reitsma (1984) found that the auditory word‐fragment prime /b^/ facilitated the visual word processing of BUNNY in first and sixth grades. This indicates that written and speech processing systems are interconnected and that phonological information is involved in visual word processing. In studies examining lexical activation, audio‐visual priming effects were also found in expert readers (Connine, Blasko, & Titone, 1993; Friedrich, 2005; Friedrich, Kotz, & Friederici, 2004). Focusing on the strength of cross‐modal interaction, Holcomb, Anderson, & Grainger (2005) used an intermodal repetition priming study (e.g., pepper – PEPPER vs. window – PEPPER) using event‐related potentials, and found interactions between orthographic and phonological processing systems at the sublexical and lexical levels during word processing. They interpreted these effects in the framework of a bimodal interactive‐activation model (BIAM, Grainger & Holcomb, 2010).
The BIAM (Grainger, Diependaele, Spinelli, Ferrand, & Farioli, 2003; Grainger & Ferrand, 1994; Grainger & Holcomb, 2009; McClelland & Rumelhart, 1981) is the only one model with written and spoken inputs and as such may account for intermodal priming effects. Interconnections between the print and speech processing systems are located at both the sublexical and lexical levels. This expert model describes the visual word recognition process using two routes. The written word activates sublexical orthographic representations from which the activation is spread to both orthographic lexical representations (orthographic route or procedure) and sublexical phonological representations, and then lexical phonological representations (phonological route or procedure). Connections between the orthographic and phonological representations are also described at the lexical level. Thus, the architecture of the BIAM accounts for the rapid involvement of phonological code during visual word recognition with phonological activation taking place at both sublexical and lexical levels. In contrast, there is a lack of precision at the level (sublexical and/or lexical) of phonological code activation.
The purpose of the present study was to examine the impact of the phonological contribution during the visual word recognition process as a function of reading experience. Two predictors of reading experience were used: general reading expertise (third and fifth grades) and word frequency (frequent and less frequent words). In fact, word recognition is increasingly efficient as general reading expertise increases (Ziegler et al., 2014) and word frequency increases (Booth et al., 1999). An intermodal priming in lexical decision task was used in which visual target words were preceded by auditory identical primes (e.g., /lyn/ ‐ LUNE [moon]) or unrelated primes (e.g., /bεt/ [animal] ‐ LUNE). In this way, the activation of phonological representations by the prime was direct and did not depend on prior orthographic processing, whose efficiency is highly variable among young children. The first hypothesis was that print processing and speech processing systems are interconnected in young readers, so priming effects are expected. Considering the role of reading experience in the contribution of phonological activation to word reading, two hypotheses may be contrasted. If the contribution of phonological code decreases as experience increases, we would expect priming condition to interact with grade, on the one hand, and word frequency, on the other hand. Given that word recognition is faster in fifth grade (vs. third grade) and for frequent words (vs. less frequent words), phonological code could contribute more in third grade than in fifth grade and in less frequent words than in frequent words. Moreover, given that third graders have been less exposed to words than fifth graders, the frequency effect should be amplified in younger children. Alternatively, if the contribution of phonological code is independent of the word recognition system development, and therefore stable throughout reading development, a constant priming effect whatever the grade or word frequency would be expected.
Method
Participants
Forty third graders (mean age = 8 years 9 months, SD = 4 months) and 37 fifth graders (mean age = 10 years 7 months, SD = 4 months) participated in the experiment. The mean reading age was 8 years 8 months (SD = 12 months) in third grade and 11 years 6 months (SD = 20 months) in fifth grade according to the standardized French reading test “l'Alouette” (Lefavrais, 1967), which is a text reading task. All subjects were French native speakers. They had normal or corrected‐to‐normal vision. According to their teachers, none of the children had a language impairment or learning difficulties. All participants had parental consent. Reading instruction at school was based on phonics.
Materials
Two sets of 40 written words of four to eight letters in length were selected from Manulex (Lété, Sprenger‐Charolles, & Colé, 2004). Words were monosyllabic or bisyllabic (syllabic number mean = 1.6; SD = 0.5). One set consisted of frequent words (frequency mean = 192; SD = 91) and the other consisted of less frequent words (frequency mean = 42; SD = 14). Word frequencies were calculated by adding all forms with the same orthographic string. For example, the frequency of the noun “(le) rouge” [(the) red] was added to the frequencies of other orthographically identical forms (e.g., the adjective “rouge”). Numbers of letters (letter mean of frequent words = 5.8; SD = 1; letter mean of less frequent words = 5.8; SD = 1), syllables (syllabic mean of frequent words = 1.6; SD = 0.5; letter mean of less frequent words = 1.6; SD = 0.5) and phonemes (phoneme mean of frequent words = 4.4; SD = 1; phoneme mean of less frequent words = 4.3; SD = 0.9) were controlled. Each target word was paired with two types of spoken word prime: an identical prime (e.g., frequent words: /lyn/ ‐ LUNE [moon]; less frequent words: /zon/ ‐ ZONE [area]) and an unrelated prime (e.g., /bεt/ [animal] ‐ LUNE; /lam/ [blade] ‐ ZONE). Unrelated primes were words that did not share phonemes in the same position as the target. In the unrelated priming condition, frequent word primes (frequency mean = 195, SD = 89) were associated with frequent word targets while less frequent word primes (frequency mean = 42, SD = 13) were associated with less frequent word targets. Two sets of 40 written pseudowords were created from two sets of words (matched in frequency and letter, syllable and phoneme numbers with word targets) by replacing one letter in each word with another letter. Each pseudoword was matched with two spoken word primes: a related prime, which was the base word of the pseudoword target, and an unrelated prime. Two versions of the experiment were assembled from these items. Each target appeared only once in each version but across lists, targets appeared in each condition. Each participant processed only one version. The presentation order of the items was randomized. A full list of primes and target words can be found in the Appendix.
Procedure
Children were assessed individually at their school in a quiet room. They were tested in a single session lasting about 20 min. They were seated on a chair in front of a DELL computer using the E‐prime software and wore headphones. Participants performed a practice session of 15 trials before the experiment. Each trial consisted of the following sequence: first, a fixation cross in the middle of the screen was presented for 500 ms; second, it was followed immediately by the spoken word prime for a mean duration of 398 ms (SD = 75). Then, the prime was followed by a short inter stimuli interval (ISI) of 50 ms. Then, the target remained on the computer screen until the participant's response or it disappeared after 3000 ms. All the targets were followed by a 500 ms blank screen marking the end of the trial. A short pause was introduced after each series of 20 items. Participants were instructed to ignore the spoken word and perform a lexical decision task on print. They were asked to decide whether the letter sequence presented in uppercase letters was a word or not, responding as quickly and as accurately as possible. They indicated their responses by pressing one of two response buttons of the E‐prime SRbox. Latency was measured from target onset until the participant's response.
Results
Items and participants producing more than 30% errors were excluded from the analysis. This criterion led to the exclusion of one item in third grade (DRAP [sheet]) and one participant in third grade and in fifth grade. Correct response rates were 94.28% in third grade and 97.47% in fifth grade for words and 90.13% in third grade and 96.05% in fifth grade for pseudowords. Reaction times (RTs) were measured from target onset to response onset. RTs below 300 ms were discarded (less than 0.01%). The mean correct RTs and error rates are summarized in Table 1.
| Grade 3 | Grade 5 | |||
|---|---|---|---|---|
| RTs (SD) | Err | RTs (SD) | Err | |
| Less frequent word target | ||||
| ID condition | 957 (184) | 2.17 | 741 (159) | 0.95 |
| UR condition | 1350 (307) | 13.85 | 978 (148) | 5.95 |
| Frequent word target | ||||
| ID condition | 943 (198) | 0.63 | 683 (128) | 0.68 |
| UR condition | 1230 (282) | 6.25 | 888 (153) | 2.57 |
| Priming effect | ||||
| Less frequent word | 393 ms | 237 ms | ||
| Frequent word | 287 ms | 205 ms | ||
As Ziegler et al. (2014) advocate the use of inverse RTs in developmental studies in order to normalize the latency distributions (Ratcliff, 1993), we performed analyses with inverse RTs and using linear mixed‐effects models (Baayen, Davidson, & Bates, 2008). The error data were analysed using logit mixed‐effects models (Jaeger, 2008). Participants and items were treated as random factors. Grade (third and fifth), Frequency (frequent and less frequent) and Priming condition (Identical and Unrelated) were treated as fixed factors.
Analyses were carried out in R version 2.14.2 (R Development Core Team, 2013) using the lmer4 package (version 0.999999‐0, Bates, Maechler, & Bolker, 2012). A random structure approach was used by including all random effects according to the following principle: when the factor was between‐participant (Grade) or item (Frequency), a random intercept was used and when the factor was within‐participant (Frequency and Priming condition) or item (Grade and Priming condition), a by‐participant or item random slope was used (i.e., intercepts and slopes for participants and items and slopes for fixed effects). The same principles were applied to interactions. By‐participant or item random slopes were used for any interactions where all factors comprising the interaction were within‐participant or item (Barr, Levy, Scheepers, & Tily, 2013). We looked at main effects and interaction in a forward direction (Barr et al., 2013). The log‐likelihood ratio test (χ2) was used to test significance.
Model comparison on inverse RTs showed that adding of each factor (Grade, Priming condition or Frequency) significantly improved the goodness of model fit reflecting an effect of Grade, fifth graders responded faster than third graders (822 ms vs. 1120 ms; χ2 (1) = 30.28, p < .001), an effect of Priming condition, responses were faster in identity condition than in unrelated condition (831 ms vs. 1112 ms; χ2 (1) = 155.33, p < .001) and an effect of Frequency, responses were faster for frequent words than less frequent words (936 ms vs. 1007 ms; χ2 (1) = 9.15, p = .002). Moreover, both the Grade × Priming condition two‐way interaction and the Grade × Frequency two‐way interaction significantly improved the goodness of model fit (respectively, χ2 (1) = 11.18, p < .001; χ2 (1) = 5.47, p = .019), while the Frequency × Priming condition two‐way interaction yielded only a marginal improvement of the model fit (χ2 (1) = 3.50, p = .061). Interestingly, the Grade × Priming condition × Frequency three‐way interaction significantly improved the goodness of model fit (χ2 (4) = 21.16, p < .001).
Data analyses were carried out by Grade. For third graders, model comparison on inverse RTs showed that adding of each factor (Priming condition or Frequency) significantly improved the goodness of model fit indicating that there was an effect of Priming condition, responses were faster in identity condition as compared to unrelated condition (950 ms vs. 1290 ms; χ2 (1) = 96.20, p < .001) and an effect of Frequency, responses were faster for frequent words than less frequent words (1085 ms vs. 1154 ms; χ2 (1) = 7.27, p = .006). Finally, the Priming condition × Frequency two‐way interaction significantly improved the goodness of model fit (χ2 (1) = 6.54, p = .011) indicating that priming effect was stronger for less frequent words than frequent words (393 ms vs. 287 ms). For fifth graders, model comparison on inverse RTs showed that adding of each factor (Priming condition or Frequency) significantly improved the goodness of model fit indicating that there was an effect of Priming condition, responses were faster in identity condition as compared to unrelated condition (712 ms vs. 933 ms; χ2 (1) = 66.90, p < .001) and an effect of Frequency, responses were faster for frequent words than less frequent words (785 ms vs. 860 ms; χ2 (1) = 16.42, p < .001). In contrast, the Priming condition × Frequency two‐way interaction did not improve the goodness of model fit (χ2 (1) = 0.39, p = .53).
We used the same analysis procedure for accuracy than for RTs. Model comparison on accuracy showed an effect of Grade (χ2 (1) = 9.49, p = .002), an effect of Priming condition (χ2 (1) = 47.09, p < .001) and an effect of Frequency (χ2 (1) = 17.09, p < .001). In contrast, the Grade × Priming condition, the Grade × Frequency and the Frequency × Priming condition two‐way interactions and also Grade × Priming condition × Frequency three‐way interaction did not improve the goodness of model fit (χ2 (1) = 0.73, p = .39; χ2 (1) = 0.56, p = .56; χ2 (1) = 0.40, p = .53 and χ2 (4) = 2.71, p = .61, respectively).
Discussion
The present study aimed to evaluate the phonological contribution during visual word recognition in the course of learning to read. To shed light on this issue, we compared the effect of phonological code activation in two groups of child readers differing in general reading expertise (third and fifth grades). In addition, word target frequency was manipulated to investigate the phonological contribution as a function of word exposure. We used an intermodal priming paradigm with auditory primes and written targets in a lexical decision task. In this way, phonological code activation did not depend on prior orthographic processing, whose efficiency is highly variable among young children.
Overall, the results showed that visual word recognition in third and fifth grades was widely facilitated by auditory primes. This indicates that print processing and speech processing systems are interconnected in young readers and the phonological code involved in word recognition is the speech phonological code (Holcomb et al., 2005 in expert readers; Reitsma, 1984 in children). This result also suggests that phonological code continues to play an important role in word recognition throughout reading development.
The main interest of the present research was in the development of the phonological priming effect as a function of reading experience. We observed a three‐way interaction between group, target frequency and priming condition indicating that third graders and fifth graders displayed different patterns of results. Indeed, the two‐way interaction between target frequency and priming condition was significantly greater in third graders than in fifth graders. The stronger reliance on phonological code for processing less frequent words in third grade suggests that when the visual word recognition process is less efficient (Ziegler et al., 2014), it relies more on phonological codes. As regards children from fifth grade, the results suggest that phonological code could continue to play a role after years of reading instruction (Booth et al., 1999; Ziegler et al., 2014) even when lexical access is highly efficient (i.e., frequent targets). In addition, contrary to in third grade, in fifth graders the phonological code contribution is less affected by word recognition system development. Thus, interestingly, it seems that the phonological contribution does not decrease with the increasing efficiency of word recognition, suggesting that phonological involvement in word recognition is general and non‐optional.
The results of this intermodal priming study can readily be accommodated by the BIAM (Grainger et al., 2003; Grainger & Ferrand, 1994; Grainger & Holcomb, 2009; McClelland & Rumelhart, 1981). In fact, the prediction made by the BIAM that print and speech processing systems are interconnected is also validated in young children. Moreover, as proposed by this model, orthographic processes and phonological processes interact during visual word recognition; the visual word process benefitted from prior activation by the auditory prime. This means that visual word recognition involves phonological code as predicted by the model. Note that the model does not allow the level (sublexical or lexical) at which connections between orthographic and phonological codes interact to be specified. Indeed, according to BIAM, phonological code activation by an orthographic input or a spoken input occurs at both sublexical and lexical levels during visual word recognition.
From a methodological perspective, the present study showed that the intermodal priming paradigm, in which primes are auditory, is relevant for examining phonological involvement in word recognition in reading development. In fact, auditory primes activate phonological code without orthographic code involvement. In this way, the activation of phonological code is not dependent on orthographic processing efficiency, contrary to the visual pseudohomophone priming paradigm. The intermodal priming paradigm therefore provides a novel way to investigate the contribution of phonological code to word recognition. However, we have to acknowledge some limitations of this paradigm. First, the phonological code activation by auditory primes might take place at both sublexical and lexical levels. Thus, we can only consider the phonological code activation as a whole without distinguishing between these two levels when the visual word recognition process is examined. Second, given that the auditory primes are not masked, we cannot confirm that phonological code is involved in an automatic way during visual word recognition. Finally, the visual target word was preceded by an auditory prime, either identical or unrelated. This design cannot distinguish between a facilitatory effect of the identical prime and an inhibitory effect of the unrelated prime. Further investigations could include a control condition of no prime.
In summary, the present study aimed to investigate phonological code involvement in visual word recognition in developing readers. A key aspect of our study is that phonological coding was directly activated by using auditory primes in an intermodal priming paradigm. Hence, phonological code activation by the prime did not depend on prior orthographic processing, which evolves with development. The results clearly indicate that print and speech processing systems are interconnected in child readers. Moreover, they show that phonological code continues to exert an influence on visual word recognition throughout reading development.
Acknowledgements
This study was made possible by a PhD Grant awarded to KS from the Ministry of Research of France. The authors cheerfully thank all the persons without whom this study would not have been possible; in particular the children, their parents and school staffs. They wish to express their gratitude to the Paul Fort and Frédéric Chopin schools in Villeneuve d'Ascq and Henri Matisse school in Tressin.
Appendix : Material
| Targets Less frequent | Primes PH | UR | Targets Frequent | Primes PH | UR |
|---|---|---|---|---|---|
| AIGLE | aigle | fuite | ACTION | action | espace |
| ARMOIRE | armoire | victime | AFFAIRE | affaire | chevaux |
| AUTOMNE | automne | échelle | ARGENT | argent | demain |
| BILLE | bille | panne | ARTICLE | article | semaine |
| BILLET | billet | cochon | CAMPAGNE | campagne | vêtement |
| CAILLOU | caillou | enquête | CHASSE | chasse | langue |
| CHARBON | charbon | janvier | CHEMIN | chemin | bateau |
| CHARIOT | chariot | ficelle | CHEVEUX | cheveux | instant |
| CHATTE | chatte | mousse | DAME | dame | robe |
| COFFRE | coffre | risque | DIABLE | diable | photo |
| COURONNE | couronne | baguette | FORCE | force | sable |
| COUSIN | cousin | vivant | FRÈRE | frère | suite |
| CRÈME | crème | score | FRUIT | fruit | oncle |
| CULOTTE | culotte | mission | GÂTEAU | gâteau | accord |
| DOUZE | douze | singe | IMAGE | image | élève |
| DRAP | drap | mine | INDIEN | indien | épaule |
| FEMELLE | femelle | chagrin | JAMBE | jambe | corps |
| GAMIN | gamin | ruban | LISTE | liste | carte |
| GARE | gare | type | LUMIÈRE | lumière | journal |
| HUILE | huile | pente | LUNE | lune | bête |
| JEUDI | jeudi | savon | MADAME | madame | étoile |
| LÉGENDE | légende | carotte | MALADE | malade | besoin |
| MAIRIE | mairie | aspect | MÉDECIN | médecin | conseil |
| MOUCHE | mouche | paille | MINUTE | minute | papier |
| PELAGE | pelage | soupir | MONDE | monde | fille |
| PLUMAGE | plumage | vitrine | NOMBRE | nombre | jardin |
| POIRE | poire | stade | NUAGE | nuage | triste |
| RADIS | radis | coton | OBJET | objet | avion |
| REQUIN | requin | buffet | PASSAGE | passage | endroit |
| ROBOT | robot | lundi | PLUIE | pluie | herbe |
| SÉJOUR | séjour | casque | POCHE | poche | bande |
| SOIRÉE | soirée | cabine | RÉPONSE | réponse | famille |
| SOLDAT | soldat | fourmi | SECRET | secret | voleur |
| SOURCE | source | vendre | SERVICE | service | théâtre |
| TIMIDE | timide | sévère | TABLEAU | tableau | journée |
| TORTUE | tortue | église | TRAIN | train | femme |
| VASE | vase | code | TROU | trou | chef |
| VIOLON | violon | samedi | VENTRE | ventre | course |
| VOISINE | voisine | facteur | VITESSE | vitesse | pluriel |
| ZONE | zone | lame | VOILE | voile | plume |
Biographies
Karinne Sauval is a PhD in Cognitive Psychology at the University of Lille, France. She is interested by reading acquisition in children and particularly the development of the relationship between orthographic and phonological codes.
Séverine Casalis is a professor of cognitive psychology at the University of Lille, France. Her research interests include reading acquisition, developmental language and reading disorders and second language learning.
Laetitia Perre is an associate professor in the Department of Psychology at the University of Lille in France where she investigates the cognitive and neural mechanisms of language processing. Issues she is especially interested in include the relationship between reading and speech in normal and atypical development.
Number of times cited: 1
- Sanne W. van der Kleij, Margriet A. Groen, Eliane Segers and Ludo Verhoeven, Enhanced semantic involvement during word recognition in children with dyslexia, Journal of Experimental Child Psychology, 10.1016/j.jecp.2018.09.006, 178, (15-29), (2019).
