13.2.1.1 Phonology

Dutch vowels are pronounced according to the phonological features of place of articulation (frontal, central, back) and height (high, middle, low) (See Table 13.1). At the same time, there is an opposition between long and short vowels: Long vowels tend to be tense and short vowels lax. And there are three diphthongs: frontal ει as in hei ‘heath’, central œy as in bui ‘shower’, and back ∧u as in dauw ‘dew’.

The system of Dutch consonants is relatively simple. Place and manner of articulation in addition to voicing are the main distinctive features (see Table 13.2). The voiced–voiceless opposition is distinctive for plosives but not so much for fricatives.

The vast majority of Dutch monosyllabic words follow a CVC pattern. Consonant clusters can occur in both initial and final position, with the following distribution possible – in principle – within a single word: CCCVCCCC or striktst (‘most strict’). When a liquid or nasal is followed by a non-homorganic consonant in postvocalic position, the word is pronounced with an unstressed schwa inserted between the two consonants. Thus melk ‘milk’ is pronounced as [mεlək], kalm ‘calm’ as [kαləm], but wens ‘wish’ as [wεns].

The main stress in polysyllabic words otherwise tends to be on the prefinal syllable. And with the exception of only the schwa, all vowels in Dutch can receive stress. In nominal and verbal compounds, primary stress is on the first part and secondary stress on the second part.

Sentence intonation in Dutch is rather flat. Unmarked Dutch declarative sentences have a low declining contour at the beginning, a high declining contour in the middle, and a final low declining contour at the end. Interrogative sentences have a high declining contour at the end.

13.2.1.2 Inflectional Morphology

The general rule for Dutch plural formation is adding -en to a singular word form, e.g., deur ‘door’ – deuren ‘doors’. However, many exceptions apply, e.g., jongen ‘boy’ – jongens ‘boys’. With respect to verbal morphology, Dutch can be characterized as a language with a restricted tense system and a very limited aspect system. Basically, there are contrasts among present, simple past, and perfect. The basic tense opposition is present versus past tense. Only the singular forms of verbs require differentiation for person (see van der Putten, Reference van der Putten1997). The first person form of the verb is verb stem + 0, while the second and third person forms are verb stem + t. For regular verbs, the past tense is formed by adding -de/-te to the verb stem. For so-called strong verbs, the past is formed by a vowel change, as in ik loop ‘I walk’ vs. ik liep ‘I walked’. The regular past participle is created with the addition of the prefix ge- and the suffix -d/-t, as in ge-droom-d ‘dreamed’ or ge-kus-t ‘kissed’. Most strong verbs take the suffix -en, as in ge-holp-en ‘helped’, to form the past participle. The auxiliaries for the perfect tense are hebben ‘have’ for transitive verbs and zijn ‘be’ for intransitive verbs at times. The auxiliary for the passive voice in Dutch is worden. Word order is the basic marker of syntactic roles in Dutch, which does not have case inflections (Geerts, Haeseryn, de Rooij & van den Toorn, Reference Geerts, Haeseryn, de Rooij and van den Toorn1984).

In Dutch words lacking an internal morphological structure, the main stress is generally placed on the prefinal syllable, which leaves the vowel in the final syllable unstressed. Depending on vowel length and syllable weight, however, the main stress may be shifted to the final syllable at times (see Kooij, 1990). According to the CELEX database, approximately 74 percent of Dutch bisyllabic words are pronounced with first syllable stress (Baayen, Piepenbrock & van Rijn, Reference Baayen, Piepenbrock and van Rijn1993). Stress assignment to the first syllable can thus be seen as regular and stress assignment to the final syllable as irregular. Prefixes are quite frequent in Dutch words (cf. Schreuder & Baayen, Reference Schreuder, Baayen and Feldman1995) but never stressed. For instance, the presence of a prefix in a bisyllabic verb form leads to a shift of stress from the first (i.e., prefinal) syllable to the second (i.e., final) syllable as in BŪKKEN (‘to bend’) and GEBŪKT (‘bent’, past participle), respectively. The unmarked case for stress assignment is thus violated by such prefixed words and therefore an interesting domain for study.

13.2.1.3 Word Formation Processes

Word formation in Dutch is primarily accomplished via affixation and compounding (cf. Booij, Reference Booij1977). For affixation, the processes of declination, conjugation, and derivation can be distinguished. And for declination, there are more or less regular devices available to form the plural, possessive, comparative, and superlative.

Nominal compounds are not morphologically marked in Dutch. Nevertheless, additional elements not associated with the compounded words may be added and thus be considered a part of the compound structure itself. That is, a phonological element (-e-, -s-, -en-, -er-) is inserted between two compounded words as a sort of “morphonological glue” on many but not all occasions. And according to van den Toorn (Reference van den Toorn1982), phonological, syntactic, and semantic constraints can explain the occurrence of such binding phonemes.

13.2.2.1 Script and Punctuation

Dutch orthography uses the Latin alphabet. The twenty-six letters of the Roman alphabet and digraph combinations cover the thirty-five phonemes of which sixteen are vowels. Four consonants (‹c, q, x, y›) are used for only Latin-based loan words, such as centrum (‘center’), quasi, extra, yen. Excluding these foreign consonants, most consonants have a one-to-one correspondence to speech sounds. Exceptions are the stop consonants ‹b› and ‹d›, which represent voiced stops in word-initial position and unvoiced stops in word-final position. The same used to hold for the letter ‹g› representing the voiced velar fricative /y/ in word-initial position and the voiceless sound /x/ in word-final position but, nowadays, the latter tends to be used throughout. Other exceptions are the digraph ‹ch› for the voiceless velar fricative /x/, the digraph ‹ng›, and the ‹n› before ‹k› in word-final position representing the velar nasal /n/ (tong, ‘tongue’; rank, ‘slender’).

The conversion of vowel graphemes to phonemes is less straightforward than that for consonants. Table 13.3 shows the representation of Dutch vowels in simple CVC words. The single vowel graphemes ‹a›, ‹e›, ‹i›, ‹o›, and ‹u› can occur in initial or medial position representing the phonemes /a/, /e/, /i/, /o/, and /u/. In word-final position, however, these vowel graphemes represent long vowels. The ‹e› in word-final position represents the neutral schwa sound, as in de (‘the’). The ‹i› in word-final position occurs only in loan words such as ski, otherwise ‹ie› is used, as in knie (‘knee’). If the letter ‹u› occurs in penultimate position, it is followed by the letter ‹w› while the vowel pronunciation gets lengthened (ruw, ‘rough’). And if the vowel symbols ‹i› and ‹u› occur as the last elements of trigraph vowels, they represent the consonants /j/ and /w/, respectively as in baai ‘bay’ and dauw ‘dew’.

With regard to punctuation (cf. Donaldson, Reference Donaldson1997), Dutch declarative sentences start with a capital letter and end with a period. Interrogative sentences end with a question mark, exclamations with an exclamation mark. A semicolon can be used instead of a period at the end of a sentence and thereby connect two sentences. A colon is used to introduce a list, a quotation, or an illustrative example. A comma is used to distinguish coordinate or subordinate clauses within a sentence; may occur between adjectives referring to the same entity; and may be placed at the end of a quote. Quotation marks are used to indicate quoted material. Compound words may sometimes be hyphenated to prevent the pronunciation of two vowels as a single unit. The repeated parts of subsequent compound words may also be preceded by a hyphen. A diaeresis is placed on the first letter of a syllable to prevent the pronunciation of two vowels as a single syllable (e.g., kopieën ‘copies’). Finally, an apostrophe is used with the plural forms of singular roots that end with a short vowel (e.g., foto’s ‘photographs’) or in the case of letter omissions/contractions (e.g., ’s avonds ‘in the evening’).

13.2.2.2 Orthography

Dutch orthography is largely but not completely phonemic (Nunn, Reference Nunn1998). Monosyllabic words show a highly consistent mapping between letters and phonemes, but this is not always the case for longer, multisyllabic words. Dutch syllable structure is quite complex, in part because multiple consonants can occur in both onset and coda positions with specific constraints then applying. In words with postvocalic consonant clusters, for instance, the word may be pronounced with an unstressed schwa between the consonants but only when a liquid or a nasal precedes the other consonant. In longer words, numerous deviations from the one-to-one correspondence between letters and sounds occur. To start with, the vowel ‹e› can represent three different sounds in polysyllabic words: /ε/ in a closed stressed syllable; /e/ in an open stressed syllable; or /œ/ in an unstressed syllable. The Dutch word weggeven ‘give away’ encompasses all of these variants and is pronounced /wεggevœn/.

Furthermore, the phonological status of the schwa when it occurs as the central vowel in polysyllabic words is quite unclear. It occurs in unstressed syllables only where it can nevertheless be alternatively represented by the letters ‹e›, ‹ij›, or ‹i› as in bodem ‘bottom’, eerlijk ‘honest’, and handig ‘handy’, respectively. A schwa can also be inserted in sonorant consonant clusters in unstressed syllables, as in bever ‘beaver’. Yet another deviation concerns the written reduplication of vowels and consonants. Vowels that are normally spelled with reduplication (‹aa›, ‹ee›, ‹oo›, ‹uu›) are written with a single letter when occurring in an open syllable position; the plural of poot ‘leg’ is thus spelled poten. In contrast, a single letter consonant is reduplicated when it occurs in intervocalic position after a stressed short vowel; the plural of pot ‘jar’ is thus spelled potten.

With the standardization of Dutch orthography in the nineteenth century, four main principles or orthographic conventions were established (see also Bosman, de Graaff & Gijsel, Reference Bosman, de Graaff, Gijsel, Joshi and Aron2006). The first principle is that distinctive phonemes are represented by distinctive graphemes. Non-distinctive (sub-phonemic) variation in pronunciation is thus neglected. The second principle is that grammatical morphemes are written invariably, or in other words, the phonetic variation occurring in morphological equivalents should not give rise to different spellings. The singular of a plural noun like handen ‘hands’ in Dutch is thus written with a final -d even though a voiceless -t is pronounced for hand ‘hand’. The third principle requires that grammatical morphemes are written even when they are absent from the pronunciation of a word. By analogy to hij denk-t ‘he think-s’– which indicates the third person, singular, present tense – the spelling of the same form for another verb is hij bid-t ‘he pray-s’ but the final -t is not pronounced separately from the -d. When the relevant grammatical morpheme is added to a verb stem that already ends with the letter -t, it is not doubled, which shows autonomous rules also exist. The final principle is maintenance of etymological derivation in current spelling practices and thus differs from earlier orthographic stages with no current phonological impact. Examples of this principle are the alternative spellings of ει, as in eis ‘claim’ versus ijs ‘ice’. Viewed synchronically, these spelling differences have no significance for the morphophonological basis of Dutch orthography (see Kooij, Reference Kooij and Comrie1992).

From a theoretical point of view, Dutch orthography can be classified as deep synchronic: Conversion rules apply to phonemes, which take the morpheme as their main domain. Phoneme-to-grapheme conversion rules are applied on the basis of phonological context to Dutch morphemes. But Dutch uses partly different sets of phoneme-to-grapheme conversion rules for native versus non-native sublexicons (cf. Nunn, Reference Nunn1998).

Dutch orthography also calls upon autonomous rules that change the sequence of letters when morphemes are combined into words. Such rules are sensitive to the orthographic context but not to sublexicon differences, which can lead to non-isomorphemic written forms. An example is the Dutch rule of g-deletion from such graphemic sequences as ‹ngk›, allowing koning ‘king’ to become koninkje ‘little king’. Other such autonomous rules are as follows.

13.3.1.1 Phonological Development and Phonological Awareness

For Dutch phonology, five manners of articulation have been identified (Rietveld & van Heuven, Reference Rietveld and van Heuven1997): plosives (b, d, k, p, t), fricatives (f, g, s, v, z), liquids (l, r), nasals (m, n), and glides (h, j, w). Manner of articulation has been found to influence the difficulty of the items used to assess phonological awareness.

Position has also been found to influence phonological awareness with the isolation of initial phoneme easiest, followed by final phoneme and middle phoneme (de Graaff, Hasselman, Bosman & Verhoeven, Reference de Graaff, Hasselman, Bosman and Verhoeven2008). In a follow-up study, an interaction was found between type of task and the influence of phoneme position (de Graaff, Hasselman, Verhoeven & Bosman, Reference de Graaff, Hasselman, Verhoeven and Bosman2011). Children performed better on both phoneme isolation and phoneme segmentation tasks for phonemes in initial position, but better on phoneme blending for phonemes in word-final position.

In two experiments, Schreuder and van Bon (Reference Schreuder and van Bon1989) examined the effects of Dutch word properties such as length, consonant-vowel structure, syllabic structure, and meaning on phonemic segmentation. This was done for a group of fifty first graders who could segment shorter words better than longer words. Phonemic segmentation was also found to be easier for them between onset and rime than within an onset constituent. The type of phoneme has also been shown to affect children’s phonemic processing. Geudens and Sandra (Reference Geudens and Sandra2003) found differences in the segmentation of CV and VC syllables as a function of the sonority or vowel-likeness of the component consonants: sonorants (liquids and nasals) cohered more strongly with the preceding vowels than obstruents (plosives and fricatives) and were thus harder for the children to manipulate (i.e., segment). In a similar vein, Geudens, Sandra, and Van den Broeck (Reference Geudens, Sandra and Van den Broeck2004) indicated that the cohesion between phonemes interacts with the sonority of consonants. And in an experimental study with 5-year-old preliterate children, Wagensveld, van Alphen, Segers, and Verhoeven (Reference Wagensveld, van Alphen, Segers and Verhoeven2012) indeed found Dutch rhyming skill (which requires determination of the phonemic structure of words) to be influenced by global similarities between words; phonologically related pairs (bel – bal) were more difficult to rhyme/segment than phonologically unrelated pairs (sok – bal). In other research, the same effect was found for both literate children and adults (Wagensveld, Segers, van Alphen & Verhoeven, Reference Wagensveld, Segers, van Alphen and Verhoeven2013). Moreover, in two additional studies, neural evidence of a global similarity effect for rhyme processing was found in both children (Wagensveld, van Alphen, Segers, Hagoort & Verhoeven, Reference Wagensveld, van Alphen, Segers, Hagoort and Verhoeven2013) and adults (Wagensveld, Segers, van Alphen, Hagoort & Verhoeven, 2012).

When Vloedgraven and Verhoeven (Reference Vloedgraven and Verhoeven2007) examined the development of phonological awareness in relation to reading skills for 172 kindergartners and 173 first graders, they found performance on four different tasks to reflect a single underlying ability. The tasks ranged in difficulty with rhyming proving easiest; phoneme segmentation most difficult; and phoneme blending and phoneme identification occurring in between. In addition, strong growth in phonological awareness was observed between kindergarten and first grade. In a follow-up study (Vloedgraven & Verhoeven, Reference Vloedgraven and Verhoeven2009), the phonological awareness of children from kindergarten through fourth grade was examined in greater detail but again found to be unidimensional across tasks and also grades with the items measuring rhyming, phoneme identification, and phoneme blending proving easier than the items measuring phoneme segmentation and phoneme deletion.

Interestingly, de Jong and van der Leij (Reference de Jong and van der Leij1999) documented changing relations between phonological awareness and the reading abilities of children over time. Children’s kindergarten phonological awareness did not relate to their later reading skills, while their phonological awareness in first grade did. Phonological awareness in first grade was a strong predictor of their later reading skill, but this predictive power subsided for phonological awareness in subsequent grades. This effect was tentatively explained in terms of the transparent nature of Dutch orthography. However, a study of phonological awareness and alphanumeric naming speed as predictors of the word decoding skills of Dutch versus English school-aged children showed no differences in the prediction of word decoding speed and word decoding accuracy in the two languages (Patel, Snowling & de Jong, Reference Patel, Snowling and de Jong2004). When de Jong and Olson (Reference de Jong and Olson2004) explored the extent to which the emergence of letter knowledge can be predicted by children’s phonological working memory and rapid automatized naming (RAN), however, phonological working memory was found to be a strong predictor and the effects of RAN much smaller. A life-span perspective on the development of continuous naming speed of numbers, letters, colors, and pictures was given by van den Bos, Zijlstra, and lutje Spelberg (Reference van den Bos, Zijlstra and lutje Spelberg2002), whereas its impact on word reading fluency was evidenced in van den Bos, Zijlstra, and Van den Broeck (Reference van den Bos, Zijlstra and Van den Broeck2003).

13.3.1.2 Morphological Development and Morphological Awareness

When Wexler, Schaeffer, and Bol (Reference Wexler, Schaeffer and Bol2004) used extensive data to test cross-linguistic models of the development of verb agreement in children, they found that Dutch-speaking children produced more root infinitives than English-speaking children. They also found specific tense and agreement errors to occur in Dutch but not in English.

In other research, morphological awareness was found to be differentially related to the word recognition and spelling skills of Dutch first- and sixth-grade children (Rispens, McBride-Chang & Reitsma, Reference Rispens, McBride-Chang and Reitsma2008). In first grade, only children’s awareness of nominal inflectional morphology related to their word decoding. In sixth grade, the children’s awareness of derivational morphology contributed to both their word reading and spelling, while their awareness of verbal inflectional morphology related to only their spelling. Knowledge of lexical compounding did not relate to the children’s reading or spelling skills in either of the two grades.

Children have been found to be sensitive to morphological segments, even in pseudowords, from an early age on. Verhoeven, Schreuder, and Baayen (Reference Verhoeven, Schreuder and Baayen2006) examined the extent of sensitivity to morphological boundaries in reading Dutch bisyllabic pseudowords and, more specifically, children’s interpretation of the initial syllable as a content morpheme, a prefix, or a random string. The children’s pronunciations and assignment of stress for the pseudowords were both found to depend on word type, which shows them to be able to identify morpheme boundaries and prefixes at an early age.

In a follow-up study, Verhoeven, Schreuder, and Haarman (Reference Verhoeven, Schreuder and Haarman2006) provided further evidence for children’s morphological sensitivity by having both children and adults read Dutch bisyllabic words with the first syllable being either a real prefix, a phonological prefix (i.e., same sound pattern as a prefix), or a pseudo-prefix (i.e., sound pattern deviant from a prefix). The results showed that both children and adults retrieve words with phonological prefixes more quickly and more accurately than words with a pseudo-prefix. These data show that both beginning and advanced readers of Dutch successfully apply conversion rules to identify not only elementary grapheme–phoneme correspondences but also correspondences for larger orthographic units.

13.3.2.1 Word Decoding Development

In a longitudinal study, the acquisition of Dutch word decoding throughout the elementary school grades was examined by Verhoeven and van Leeuwe (Reference Verhoeven and van Leeuwe2009, Reference Verhoeven and van Leeuwe2011) for words that varied in orthographic transparency (cf. Nunn, Reference Nunn1998): (1) regular CVC words, (2) complex monosyllabic words with consonant clusters in prevocalic and postvocalic positions, and (3) polysyllabic words (see Figure 13.1). The word decoding of beginning readers of Dutch started out slow and laborious with a fair number of mistakes but quickly reached a level of virtual mastery by the beginning of second grade (see Figure 13.2). The growth of word decoding skill in Dutch appears to be largely a matter of increased speed. And the exponential increases in the learning curves that level off later parallel the logistic learning functions featured in neural networks representing the reading process under supervised learning conditions.

Figure 13.1 Word decoding accuracy proportions for words varying in orthographic transparency over the six primary grades. f = fall; s = summer.

Figure 13.2 Growth of word decoding efficiency over the six primary grades. f = fall; s = summer.

Furthermore, the three measures of word decoding show considerable commonalities and strong longitudinal interrelationships over the years. Some unique variance also characterized the three word decoding skills, which shows that specific orthographic complexities have to be learned over the years to become a fully proficient reader of Dutch. Further evidence for the unidimensionality of Dutch word decoding comes from the results of a study by Keuning and Verhoeven (Reference Keuning and Verhoeven2007) showing the means of item response theory analysis for a representative set of Dutch word decoding items administered to children throughout the elementary school grades to reflect a single underlying ability.

In other research, Verhoeven and van Leeuwe (Reference Verhoeven and van Leeuwe2011) investigated the roles of gender and linguistic diversity in the development of children’s word decoding. Small but significant effects were observed: Girls generally did better than boys – mainly on regular CVC word patterns – and children learning to read Dutch as a first language generally did better than children learning to read Dutch as a second language – but mainly on polysyllabic words. From a structural point of view, the development of word decoding was nevertheless highly comparable for the different subgroups.

In a range of experimental studies, the role of morphology in the development of Dutch word decoding was investigated and found to be greatly complicated by the understanding of graphotactic rules for which the reader must convert a phonological representation on the basis of spelling adaptation rules. Verhoeven, Schreuder, and Baayen (Reference Verhoeven, Schreuder and Baayen2006) examined the learning of autonomous graphotactic rules, indicating that the contrast between a long versus short vowel is expressed by the alternation between single versus double consonant letters in open syllables, but by single versus double vowel letters in closed syllables (see Section 13.2.2.2). Both children and adults were found to be significantly less accurate and slower when they had to recognize plural word forms that had undergone a vowel change as a result of the pluralization as opposed to no such change. Graphotactic rules greatly complicate early Dutch word identification and continue to do this, even in adult reading.

Verhoeven and Schreuder (Reference Verhoeven and Schreuder2011) examined the extent to which beginning and advanced readers, including dyslexic readers of Dutch, call upon morphological access units to read polymorphemic words and found a complicated picture. In lexical decision experiments, the influence of the frequency of the singular root form of a word on the reading of the plural form of the word was investigated. For adults, both the speed and the accuracy of the lexical decision process was largely determined by the frequency of the plural word form. The frequency of the singular root form also played a role only for low-frequency plural word forms. For children, the frequencies of both the singular and the plural word forms played a role in their lexical decision making. In a related study, de Zeeuw, Schreuder, and Verhoeven (Reference de Zeeuw, Schreuder and Verhoeven2013) investigated the extent to which the reading of compound words was determined by the frequency of the compound as a whole and the frequency of the constituent parts. For children learning Dutch as a first language but also those learning Dutch as a second language, both frequencies affected the speed and accuracy of their responding. The results of these two studies taken together make it clear that the characteristics of constituent morphemes affect the reading of polymorphemic words in conjunction with the level of reading skill and experience but also word and morpheme frequency.

13.3.2.2 Word Spelling Development

Keuning and Verhoeven (Reference Keuning and Verhoeven2008) examined the dimensional structure of spelling development throughout the elementary grades. Factor analysis of the results showed that spelling can be conceptualized as a unidimensional ability for second grade through to sixth grade. When the rate of developmental change was examined in conjunction with gender, ethnicity, and word reading skill, a different pattern of results presented itself. The children’s spelling ability was indeed found to systematically increase from the beginning of second grade to the end of sixth grade. However, the children also showed a tendency to master specific types of spelling problems during different – partially overlapping – periods in their development. Further structural analyses showed the children’s spelling ability to nevertheless be highly consistent. And while word reading skill proved an important predictor of spelling growth, neither gender nor ethnicity showed an influence.

While the spelling errors of younger or poor spellers are usually less consistent than those of older or good spellers (Bosman, Reference Bosman1994), there is no evidence that phonology plays a less important role for younger/poorer spellers than for older/better spellers (Bosman, Vonk & van Zwam, Reference Bosman, Vonk and van Zwam2006). Formal spelling instruction has proved necessary to help children advance their spelling abilities (Cordewener, Reference Cordewener2014), and visual dictation has been found to be an excellent device for doing this: The speller carefully studies the word to be spelled, covers the word, spells it from memory, and then checks the spelling with the original word and makes the necessary correction (van Leerdam, Bosma & van Orden, Reference van Leerdam, Bosman, van Orden, Reitsma and Verhoeven1998). Another procedure shown to be particularly effective for the memorization of words is overpronunciation (i.e., the regularization of the spelling of words by reading the word according to prototypical grapheme-to-phoneme relationships aloud). This approach is particularly effective for the memorization of loan words (Bosman, van Hell & Verhoeven, Reference Bosman, van Hell and Verhoeven2006; Hilte & Reitsma, Reference Hilte and Reitsma2006). Finally, explicit rule instruction is obviously effective for the acquisition of rule-based spelling conventions. When both Hilte and Reitsma (Reference Hilte and Reitsma2011) and Kemper, Verhoeven, and Bosman (Reference Kemper, Verhoeven and Bosman2012) compared an implicit instruction condition in which spellers were asked to simply practice spelling words with an explicit instruction condition in which the spellers were presented the rule together with examples, explicit instruction was found to be more effective than implicit instruction for good as opposed to poor spellers but only for words with a clear morphological rule (Kemper et al., Reference Kemper, Verhoeven and Bosman2012). The finding of significant differences between the instructional conditions for good as opposed to poor spellers suggests that a fundamental threshold degree of spelling awareness may be needed to benefit from explicit spelling instruction.

13.3.2.3 Reading and Spelling Difficulties

For many children, learning to read is a process that proceeds without major difficulties. However, about 10 percent of the children experience serious difficulties with word decoding and spelling including about 4 percent who are dyslexic (Blomert, Reference Blomert2005). Behavioral studies have shown that dyslexic readers have no problems in processing digraphs as perceptual units (Marinus & de Jong, Reference Marinus and de Jong2008), although their orthographic representations have been found to be less specified and less redundant as compared to typical readers (de Jong & Messbauer, Reference de Jong and Messbauer2011; van den Bos, Reference van den Bos2008; van den Bos et al., Reference van den Bos, Zijlstra and lutje Spelberg2002; van den Bos et al., Reference van den Bos, Zijlstra and Van den Broeck2003; Marinus & de Jong, Reference Marinus and de Jong2011) and that this is especially true for reading nonwords (Van den Broeck & Geudens, Reference Van den Broeck and Geudens2012; Van den Broeck, Geudens & van den Bos, Reference Van den Broeck, Geudens and van den Bos2010). Furthermore, their reading and spelling problems have been found to be associated with problems in phonological awareness, phonological memory, and rapid naming (de Bree, Wijnen & Gerrits, Reference de Bree, Wijnen and Gerrits2010; Gijsel, Bosman & Verhoeven, Reference Gijsel, Bosman and Verhoeven2006). In several studies of the role of phonological abilities in the development of word decoding in Dutch, de Jong and van der Leij (Reference de Jong and van der Leij2003) found manifestations of phonological problems as concomitant reading problems. Poor readers, for example, also showed rapid naming impairments from kindergarten through sixth grade. Their phonological impairments manifested themselves in first grade but tended to disappear by the end of elementary school – depending on task demands, which led the authors to conclude that the different manifestations of phonological awareness and phonological deficits may lead to distinct developmental pathways.

More recently, neurocognitive studies have provided additional support for the hypothesis of a phonological deficit underlying developmental Dutch dyslexia. First, it is generally concluded that letter–speech sound pairs develop into unique audiovisual objects that need to be processed in a unique way in order to enable fluent reading (Blomert, Reference Blomert2011; Blomert & Froyen, Reference Blomert and Froyen2010). Deficits in the integration of letters and speech sounds and letter processing have been demonstrated in beginning dyslexic readers (Blau et al., Reference Blau, Reithler, van Atteveldt, Seitz, Gerretsen, Goebel and Blomert2010; Blomert, Reference Blomert2011) and interpreted as a major cause of later problems in reading fluency (Blomert, Reference Blomert2011; Bonte & Blomert, Reference Bonte and Blomert2004). A similar orthographic-phonological deficit was also evidenced in children at familiar risk for dyslexia (Blomert & Willems, Reference Blomert and Willems2010). Furthermore, Noordenbos, Segers, Wagensveld, and Verhoeven (Reference Noordenbos, Segers, Wagensveld and Verhoeven2013) showed that preliterate children at risk for dyslexia have difficulty in distinguishing rhyming word pairs from word pairs with phonological overlap. In a parallel study (Noordenbos, Segers, Serniclaes, Mitterer & Verhoeven, Reference Noordenbos, Segers, Serniclaes, Mitterer and Verhoeven2012), it was shown that children at risk tend to have problems with categorical perception of speech sounds. They tended to make a shift in their speech perception abilities from an allophonic mode of perception in kindergarten to a phonemic mode of perception in first grade.

13.3.3.1 Predictors of Reading Comprehension

In the simple view of reading, it is assumed that reading comprehension is the product of word decoding and listening comprehension (Hoover & Gough, Reference Hoover and Gough1990). It is further claimed that listening comprehension – or the linguistic processes involved in the comprehension of oral language – strongly constrains reading comprehension. In keeping with the simple view of reading, Droop and Verhoeven (Reference Droop and Verhoeven2003) found the development of reading comprehension in the intermediate grades of elementary school to be more influenced by top-down oral language comprehension than by bottom-up word decoding for both first and second language learners. Oral language skills turned out to be more prominent in the explanation of the variation in the reading comprehension for the first language learners than for the second language learners, however.

When word decoding and listening comprehension were more recently examined in relation to the reading comprehension of children throughout elementary school (Verhoeven & van Leeuwe, Reference Verhoeven and van Leeuwe2012), the longitudinal data showed that the two components accounted for all the systematic variance in both first and second language learners. With the progression of grade, the impact of word decoding on reading comprehension decreased while the impact of listening comprehension increased to similar extents for the two groups. However, the reciprocity of the relationship between listening comprehension and reading comprehension tended to be less prominent for the group of second language learners: Improved reading comprehension did not correlate with improved listening comprehension to the same extent for the two groups of learners.

13.3.3.2 Word-Level Effects in Text Comprehension

When Verhoeven, van Leeuwe, and Vermeer (Reference Verhoeven, van Leeuwe and Vermeer2011) longitudinally examined the associations between vocabulary growth and reading development for a representative sample of Dutch elementary school children, they found knowledge of word forms and word meanings to be significant predictors of reading comprehension over time. Significant progress was seen for all measures over time: basic and advanced vocabulary, word decoding, and reading comprehension. The stability of the vocabulary measures was high, moreover, and beginning vocabulary predicted both early word decoding and reading comprehension as the so-called lexical restructuring hypothesis predicts. Starting in second grade, word decoding also predicted children’s later vocabulary development, and finally a reciprocal relationship between advanced vocabulary and reading comprehension was found. The clear prediction of reading comprehension by children’s knowledge of word forms and word meanings provides support for the lexical quality hypothesis and the assumption that word-level knowledge drives early text comprehension.

In an analysis of the structural relations between word decoding, vocabulary, listening comprehension, and the reading comprehension of a representative sample of Dutch children (Verhoeven & van Leeuwe, Reference Verhoeven and van Leeuwe2008), it became clear that both the quantity and the quality of word representations were important for adequate word identification and reading development to take place. The data showed that a rich vocabulary along with a high level of listening comprehension help children integrate words into text.