Development of spelling skills in Spanish orthography: a longitudinal study

Francisca Serrano Universidad de Granada fserran@ugr.es

Gracia Jiménez Fernández Universidad de Granada [email protected]

Sylvia Defior Universidad de Granada [email protected]

Key words: spelling skills, spanish orthography, lexical effect, complexity effect, phoneme-grapheme correspondences

Abstract

The main aim of this research was to study the spelling skills development in the early stages of learning of Spanish children. Several studies show that the depth of the orthography affects spelling acquisition; it is supposed that spelling skills are easier to acquire in a transparent orthography than in a deep orthography. The Spanish written system is considered shallow, but has some complexities in the mapping of phonemes onto graphemes. Participants were 63 Spanish children in first grade of Primary Education. All the children are tested at three different Test Points (November, February and May) with four dictation tasks: Letter knowledge, simple and complex word spelling (6 items) and pseudoword spelling. The goals were to study the lexical and non-lexical spelling procedures according to the dual route models, on one hand, and to investigate the complexity effect in spelling. Syllabic structure influence and error analysis were also carried out. Results showed a high level of performance in all task and lexical effect, complexity effect and syllabic structure effect was found. The findings are discussed regarding Spanish written system transparency and the teaching methods used in the early stages of spelling acquisition.

Introduction

Reading and spelling acquisition is an instructional goal relevant enough to deserve the analysis of the conditions that could facilitate this acquisition. However, research in cognitive processes underlying spelling development has received less attention than those implied in reading skill development. In the last years, spelling studies have permitted to make progress in spelling development models as well as in the explanation of spelling disabilities (Brown & Ellis, 1994).

European languages vary significantly in orthographic complexity. According to the degree of consistency of the phoneme-grapheme correspondences, orthographies can be put on an orthographic depth continuum from transparent or shallow orthographies, with an almost perfect mapping of phonemes onto graphemes, to deep orthographies which contain orthographic inconsistencies and complexities, including multi-letter graphemes, context dependent rules, irregularities, and morphological effects. There is agreement that some European orthographies are relatively shallow (Finnish, Greek, Italian, Spanish, German) while others are deeper in the sense of containing more inconsistent correspondences as well as more morphological influences on spelling (Portuguese, French, Danish, English) (Seymour, Aro and Erskine, 2003). Cross-linguistic studies (Öney and Durgunoglu, 1994; Landerl, Wimmer and Frith, 1997; Seymour et al., 2003) have highlighted differences between the different orthographic systems. It is suggested that the results of the studies carried out with English speakers are not totally applicable to other languages. According to these studies, the development of literacy skills takes place differently depending on the characteristics of the writing system and the linguistic environment where the development of the learner takes place (Defior, 2004; Müller and Brady, 2001). Some recent studies also point out the influence of factors of educational nature as the reading method (Genard, Alegria, Leybaert, Mousty and Defior, 2004).

Spanish is a clear example of a shallow orthographic system in reading. Each grapheme has a clear and precise phonemic translation. The grapheme-phoneme correspondes (GPCRs) allows readers to determine the phoneme corresponding to each specific grapheme without ambiguity, and thus reading is controlled by a set of consistent rules. There are simple GPCRs concerning the five vowels (a, e, i, o, u) and 19 graphonemes (b, d, f, ch, j, k, l, ll, m, n, ñ, p, rr, qu, s, t, v, w, z). Each of these units maps to a single phoneme in reading. There is also one silent letter (h). There are some contextual GPCRs to establish the pronunciation of graphemes that may map to two or more phonemes (c, g, r, x, y), but these conversions are totally predictable and stable, depending on the grapheme that follows and/or their position in the word. For instance, the grapheme c is read /k/ when followed by the graphemes a, o, or u, as in casa (house) but it is read /q/ when followed by e or i as in cena (supper). The grapheme g is read /g/ when followed by the graphemes a, o, or u as in gato (cat) but it is read /x/ when followed by e or i as in gitano (gypsy). The grapheme r is read /r/ when it is not located in initial position as in cara (face), but it is read /R/ when located at the beginning of the word or following the letters n, l, or s as in rosa (rose) or Israel (Israel). The grapheme y takes the vowel value /i/ when isolated or located at the end of the word, as in rey (king), but it takes the consonant value /j/ in the rest of the cases, as in yema (yoke). The grapheme x is read /s/ when located at the beginning of the word, as in xilófono (xylophone), but it is read /ks/ in the remainder of cases, as in taxi (taxi) (Defior, Cary & Martos, 2002).

However, Spanish orthography is slightly unpredictable in spelling, due to the existence of certain phonemes which can be transcribed with more than one grapheme, namely inconsistent graphonemes (Marín, Carrillo & Alegría, 1999). For example, /b/ could be spelled as b or v, as in vaso (glass) and base (base); /x/ could be spelled j or g with vowels e and i, as in jirafa (giraffe) and gitano (gypsy); or the case of letter h that have no sound, for instance, ola (wave) and hola (hello). The speller needs specific knowledge of the word (orthographic lexicon) for an adequate spelling of the words that contain inconsistent graphonemes (sounds that could be represented by more than one grapheme).

Spelling acquisition implies phonological skills, required for the application of the GPCRs when we write no familiar words and pseudowords. At the same time, the beginner speller needs lexical knowledge concerning the orthographic pattern of the words in the mentioned cases.

Some studies have proven that the features of any alphabetic system, mainly its degree of transparency, affect the development of its acquisition. These studies show that such features have a major influence in the time needed to acquire the code (Cossu, 1999; Wimmer, Landerl & Frith, 1999) and on the type of orthographic unit (phonemes, intrasyllabic units) relevant for such acquisition (Defior, 2004; Share & Levin, 1999). Beginning spellers in a shallow orthography may be helped by the phonological transparency of the language. It is expected that they would write better ‘simple graphonemes’ in which phoneme-grapheme correspondences are 1:1, than ‘complex graphonemes’ in which the correspondences are not biunivocal.

This study is part of a large longitudinal one in which the early development of cognitive processes involved in reading and spelling acquisition in a variety of orthographies is analysed. The main aim of this study was to focus on early spelling skill development through a longitudinal approach that consisted of three test points along the 1st grade Primary School, in the context of a shallow orthography as Spanish. The assessments were carried out in November (TP1), February (TP2), and May (TP3). As a secondary goal, lexical and non-lexical spelling procedures were investigated using a word and a pseudoword dictation task. Thirdly, complexity effect is studied contrasting simple versus complex words spelling. Finally, syllabic structure effect and type of spelling errors are analyzed.

Method

Participants

A total of sixty-three Spanish first graders from two public schools participated in the study. Their mean age was 76 months at the beginning of the study. The mean reading age was 7;5 years. The mean CI was 106.8 and the mean digit memory span percentile was 63.3. The teachers used a mixed method, in which the grapheme-phoneme correspondences and meaning was the main focus of the teaching-learning processes.

Materials

A set of dictation tasks was used. Each of the three test points included letter knowledge, simple and complex words and pseudowords.

A CI measure (Raven’s Coloured Progressive Matrices), a standardised measure of reading development (PEREL) and a verbal short memory task (Digit span, from WISC-R) were also administered in TP3.

Letter knowledge

It was tested with a set of twenty-seven letter included in the Spanish alphabet. Note that for letter knowledge, names were presented. Both spellings in upper and lower case spelling were admitted.

Simple word spelling

The task included 24 very familiar high frequency words. Half of the items were content words (nouns) and the other half were function words (conjunction, preposition and interjection). All the correspondences phoneme- grapheme were consistent (1 letter:1 sound). The words were two or three syllables length.

Complex word spelling

The task included 48 very familiar high frequency words. Half of the items were content words and the other half were function words. Each word contained one complex graphoneme (1:0, 2:1).

Pseudoword spelling

48 pseudowords containing different syllabic structures (CV, VC, CVC, CVCV, VCVC and CVCVC), with consistent phoneme-grapheme correspondences were built in order to analyse syllabic structure effect in children’s spelling development.

Procedure

A longitudinal design was used. The children were tested three times along the first Primary School year: The first test point (TP1) was in November, the second one (TP2) was in February and the last one (TP3) was in May. The dictation tasks consisted of asking the children to write the list of items. The children were given a sheet of paper with blanks to fill the words in. The words were presented orally in isolation.

Results

Table 1 shows mean percentage of correct responses in letter knowledge, word (complex and simple words) and pseudoword spelling. Separate analyses of variance (ANOVA) were conducted on accuracy in order to examine letter knowledge, lexicality effect, and complexity effect. Syllabic structure effect was analysed in pseudoword spelling. Qualitative analysis of error type was also conducted for complex words.

TP 1 TP2 TP3 Letter Knowledge 80,94 86,38 94,74 (16.92) (12.76) (7.18) Word spelling Complex words 27,69 52,02 63,17 (26.08) (19.42) (17.24) Simple words 60,22 80,38 88,17 (29.78) (22.18) (13.79) Pseudoword spelling Total 50,13 62,5 75,07 (25.89) (19.66) (15.92) CV 78.23 90.73 91.94 (28.08) (12.99) (14.86) VC 54.03 70.97 81.05 (42.75) (28.34) (22.94) CVC 54.84 66.53 79.84 (35.88) (29.67) (20.67) CVCV 52.42 59.68 72.98 (38.86) (31.8) (26.51) VCVC 29.03 47.98 62.5 (33.33) (29.44) (27.81) CVCVC 32.26 39.11 62.1 (37.44) (27.77) (27.06)

Table 1. Mean percentage of correct responses (and standard deviation) as a function of task and TP.

Letter Knowledge

Results show that there is a schooling effect (F(2,116)= 42.87, p<.0001), showing that children’s letter knowledge improves from TP1 to TP2 (p<.001), and from TP2 to TP3 (p<.0001). Performance reaches a high level in TP3, near ceiling. However, some errors persisted in some letters from one TP to another, namely letter k (TP1: 57.14%; TP2: 47.62%; TP3: 20.97%), letter j (TP1: 49.21%; TP2: 23.81%; TP3: 12.9%), letter g (TP1: 44.44%; TP2: 33.33%; TP3: 14.52%), letter q (TP1: 34.92%; TP2: 28.57%; TP3: 22.58%) or, finally, letter h (TP1: 25.4%; TP2: 26.98%; TP3: 9.68%).

Lexicality effect

In order to evaluate the lexicality effect pseudoword spelling was compared to simple word spelling, since pseudowords were built without any complexity. Mean percentage of correct responses for word and pseudoword spelling in each TP are shown in Table 1. Analysis of variance (ANOVA) was conducted with Test point (TP1, TP2 and TP3) and Lexicality (word vs. pseudoword) as main factors. The results showed a significant main effect of Test point (F (2, 122)= 68.60; p<.0001). Children’s performance improves from TP1 to TP2 (p<.0001) and from TP2 to TP3 (p<.0001) both in word and pseudoword spelling. The main effect of Lexicality (F(1, 61)= 105.55; p< .0001) showed that children spell better words than pseudowords. A marginally significant interaction Test Point*Lexicality (F(2, 122)= 3.1; p< .05) was found (see Figure 1). Posthoc comparison shows that children’s performance improves in word through the three TPs (p<.0001). This is not the same in pseudoword spelling in which there was no improvement from TP2 to TP3.

100

90

80

70

60

50

40

30 TP1 20

percentage of correct responses TP2 10 TP3 0 word spelling pseudoword spelling LEXICALITY EFFECT

Figure 1. Mean percentage of correct responses as a function of Test Point and Lexicality.

Complexity effect

Mean percentages of correct responses for simple and complex word spelling in each TP are shown in Table 1. An ANOVA, including two factors, Test Point (TP1, TP2 and TP3) and Complexity (Simple vs. Complex words), was conducted. The main effect of Test Point was significant (F (2, 58)= 420.67; p< .0001), showing an improvement from TP1 to TP2 (p<.0001) and from TP2 to TP3 (p<.0002). As expected, the main effect of Complexity was significant (F (2, 116)= 106.92; p< .0001), with higher performance for simple than for complex words in all TPs. No significant interaction was found between Test Point and Complexity.

Syllabic structure in pseudoword spelling

Mean percentage of correct responses for each syllabic structure in each TP are presented in Table 1. Separate analysis of variance (ANOVA) was conducted on accuracy for monosyllabic structures (CV, VC, CVC) in each TP (TP1, TP2, TP3). There was a significant effect of Test Point (F(2,122)=36.28, p<.0001). Children’s performance improves from TP1 to TP2 (p<.0001), but there were no differences between TP2 to TP3. The results showed a significant main effect of Syllabic Structure (F (2,122)= 40.95, p<.0001). Posthoc comparison shows that performance in CV structure was significantly better than in VC structure (p<.0001) and performance in VC structure better than in CVC structure (p<.01).

Analysis of error in complex words

A qualitative analysis of error was conducted in complex words in order to evaluate beginning speller’s orthographic knowledge. Some categories of error were established in order to characterise wrong children’s performance.

- Refusal. The child spells nothing, - Absence of CG. The complex grapheme is not spelled but the child tries to spell something. Ex.: ‘eipo’ instead of equipo (team), - Preserved CG. The complex grapheme is correctly spelled. The error is elsewhere. Ex.: ‘equibo’, - Erroneous CG but phonologically correct (EPC). The complex grapheme is not well spelled but it is phonologically correct. Ex.: ‘ekipo,' - Erroneous CG and phonologically incorrect (EPI). The complex grapheme is not well spelled it is and phonologically incorrect. Ex.: ‘ecipo’.

Total percentage of error was calculated in each TP, TP1 (71.3%), TP2 (48.01%) and TP3 (19.75%). Only one error for word was computed. Table 2 shows the distribution of the percentage of error in each error type as a function of Test Point. It is observed than refusals are less frequent in TP2 and TP3 than in TP1. The percentage of Absence of CG decreases through academic year. It may seem that children try to write as they advance in learning, instead that they are not sure of the correct form. In the same way, percentage of Preserved CG increases from TP1 to TP3, showing a schooling effect. Regarding EPC and EPI, a high percentage of these types of error is found in TP1 and TP2, and it decreases in TP3. Children spelling skills improve and at the end of the 1st year of Primary School they already reach a high level of mastery in the spelling of complex words.

Error Type TP 1 TP2 TP3 Refusal 21,96 8,26 4,94

Absence of CG 10,58 3,97 1

Preserved CG 4,63 6,42 3,24

EPC 16,01 16,99 6,02

EPI 18,12 12,37 4,55

Table 2. Distribution of the percentage of error in each error type as a function of Test Point.

Discussion

This study tried to examine the beginning spelling skills in Spanish children in the fist year of formal literacy instruction. A first important result is the high performance achieved in all tasks, letter knowledge, word and pseudoword spelling, at the end of the first school year. It is found a clear progression since such early stages of spelling development, which is observable, even after two months of formal instruction.

These high results are especially obvious in letter knowledge task in which children reach nearly ceiling at the end of the school year. However, some errors persist as an evidence of gaps in children spelling skills. However, all these errors appear in letters that correspond to graphonemes that have particularities in Spanish orthography. For instance, letter k is rarely used in Spanish orthography as it appears only in very few foreign words. Moreover, it is related to other graphonemes, as the sound /k/ could be written k, c or qu. For these reasons, children in this study tended to write ‘ca’ instead of ‘k’ when they were dictated /ka/. Something similar happened in the case of the letter q, that has also the sound /k/. On the other side, letters j, g and h are related to inconsistent graphonemes, and children tended to confuse them.

Another important result is the appearance of an early Lexicality effect. From the beginning of the year, words are written better than pseudowords, and this tendency maintains through the academic year. This might indicate that Spanish children use lexical procedure as well as non-lexical one in their spelling activities. This finding contradicts stages theories of spelling development (Ehri, 1986, Gentry, 1982, Henderson, 1985) showing that children use multiple strategies from the onset of the experience with print (Treiman and Bourassa, 2000).

A complexity effect is also found in spelling, as simple words are better spelled than complex words in all TPs. This might indicate that there is still partial understanding of the orthographic conventions used in the Spanish writing system at this moment of learning. Spanish children rely on phonology and that allows simple words spelling. This idea agrees with other studies in transparent orthographies and it is support by the analysis of error, since Spanish children tend to produce correct or phonemically acceptable (high EPC errors) word spelling. However, Spanish children do not have important problems in ‘complex graphonemes’ considered in this study. For instance, Spanish digraphs are taught from the beginning of the learning and they are frequently found in words that are part of children’s vocabulary and books. This is the opposite pattern founded in studies in deep orthographies. The transparency of Spanish language as well as the familiarity with these graphonemes could help children in their spelling performance. Moreover, the predominant use of the phonological procedure reinforces the mentioned idea that spelling development consists on the use of multiple strategies from the beginning although there are some that are predominant depending on different points in time (Treiman and Cassar, 1997).

As far as syllabic structure is concerned, results show that the best performance is in open syllables (CV). This can be explained because this is the most frequent structure in Spanish and it is mainly used in the initial literacy teaching.

To summarise, taking into account all these results, it may be said that Spanish children reach a high level of spelling skills from early stages of development and in the first year of formal teaching of this ability, as opposite to what happens in deep orthographies (Seymour et al., 2003). They might take advantage of the shallow nature of Spanish orthography, which helps them in the spelling mastering. However, to master the spelling skills children need to incorporate other knowledges as morphosyntactical and semantic ones, in the way to build the orthographic lexicon up.

References

Brown, G.D.A. & Ellis, N.C. (Eds.). (1994). Handbook of spelling. Theory, process and intervention. Chichester: John Wiley.

Cossu, G. (1999). The adquisition of Italian orthography. In M. Harris & G. Hatano (Eds.), Learning to read and write: A cross-linguistic perspective (pp. 10-33). Cambridge: Cambridge University Press.

Defior, S. (2004) Phonological Awareness and learning to read: A crosslinguistic perspective. In P. Bryant & T. Nunes (Eds.) Handbook on children’s literacy. (pp. 163-174). Amsterdam: Kluver.

Defior, S., Cary, L. & Martos, F. (2002). Differences in reading acquisition development in two shallow orthographies: Portuguese and Spanish. Applied Psycholinguistics, 23, 135–148.

Ehri, L. C. (1986). Sources of difficulty in learning to spell and read. Advances in Developmental & Behavioural Paediatrics, 7, 121-195.

Frith, U. (1985). Beneath the surface of developmental dyslexia. In K. Patterson, M. Colheart & J. Marshall (Eds.), Surface dyslexia. (pp. 301-330). London: LEA.

Genard, N., Alegria, J., Leybaert, J., Mousty, P. and Defior, S. (2004). La adquisición de la lectura y la escritura en francés (Reading and writing acquisition in French). Paper presented in Simposium ‘Procesos fonológicos y lectura: un enfoque translingüístico’ II Congreso Hispano-Portugués de Psicología. Lisbon.

Gentry, J. R. (1982). Analysis of Developmental spelling in GNYS AT WORK. The Reading Teacher, 36, 192-200.

Henderson, E.H. & Beers, J.W. (Eds.). (1980). Developmental & Cognitive Aspects of Learning to Spell: A Reflection of Word Knowledge. Newark, DL: International Reading Association.

Landerl, K., Wimmer, H. & Frith, U. (1997) The impact of orthographic consistency on dyslexia: a German English comparison. Cognition, 63, 315-334.

Marín, J., Carrillo, M. & Alegría, J. (1999). El proceso de aprendizaje de la escritura en español y francés: Un estudio comparativo (Learning to spell in Spanish and French: A comparative study). Paper presented in IV Simposio de Psicolingüística (IV Symposium in Psycholinguistic), Miraflores de la Sierra (Madrid).

Müller, K., & Brady, S. (2001). Correlates of early reading performance in a transparent orthography. Reading and Writing: An Interdisciplinary Journal, 14, 757-799.

Öney, B. & Durgunoglu, A. (1997). Beginning to read in Turkish: A phonologically transparent orthography. Applied Psycholinguistics, 18, 1-7. Seymour, P.H.K., Aro, M. & Erskine, J.M. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94, 143-174.

Share, D. & Levin, I. (1999). Learning to read and write in Hebrew. In M. Harris & G. Hatano (Eds.), Learning to read and write: A cross-linguistic perspective (pp. 89-111). Cambridge: Cambridge University Press.

Treiman, R. & Cassar, M. (1997) Spelling acquisition in English. In C. Perfetti, L. Rieben & M. Fayol (Eds.), Learning to spell: Research, theory and practice across languages (pp. 61-80). Mahwah, NJ: Lawrence Erlbaum Associates.

Treiman, R., & Bourassa, D. (2000). The development of Spelling Skills. Topics in Language Disorders, 20, 1-18.

Wimmer, H., Landerl, K. & Frith, U. (1999). Learning to read German: normal and impaired acquisition. In M. Harris & G. Hatano (Eds.), Learning to read and write: A cross-linguistic perspective (pp. 34-50). Cambridge: Cambridge University Press.