Publications

Abstract

Mouthings and mouth gestures are omnipresent in Sign Language of the Netherlands (NGT). Mouthings in NGT commonly have their origin in spoken Dutch. We conducted a corpus study to explore how frequent mouthings in fact are in NGT, whether there is variation within and between signs in mouthings, and how frequent temporal reduction occurs in mouthings. Answers to these questions can help us classify mouthings as being specified in the sign lexicon or as being instances of code-blending. We investigated a sample of 20 frequently occurring signs. We found that each sign in the sample co-occurs frequently with a mouthing, usually that of a specific Dutch lexical item. On the other hand, signs show variation in the way they co-occur with mouthings and mouth gestures. By using a relatively large amount of natural data, we succeeded in gaining more insight into the way mouth actions are utilized in sign languages.

Abstract

In the present paper we investigate the effect of categorising raw behavioural data or computational model responses. In addition, the effect of averaging over stimuli from potentially different populations is assessed. To this end, we replicate studies on word learning and generalisation abilities using the ACORNS models. Our results show that discrete categories may obscure interesting phenomena in the continuous responses. For example, the finding that learning in the model saturates very early at a uniform high recognition accuracy only holds for categorical representations. Additionally, a large difference in the accuracy for individual words is obscured by averaging over all stimuli. Because different words behaved differently for different speakers, we could not identify a phonetic basis for the differences. Implications and new predictions for infant behaviour are discussed.

Abstract

In the present paper we investigate the effect of categorising raw behavioural data or computational model responses. In addition, the effect of averaging over stimuli from potentially different populations is assessed. To this end, we replicate studies on word learning and generalisation abilities using the ACORNS models. Our results show that discrete
categories may obscure interesting phenomena in the continuous
responses. For example, the finding that learning in the model saturates very early at a uniform high recognition accuracy only holds for categorical representations. Additionally, a large difference in the accuracy for individual words is obscured
by averaging over all stimuli. Because different words behaved
differently for different speakers, we could not identify a phonetic
basis for the differences. Implications and new predictions for
infant behaviour are discussed.

Abstract

Investigating human cognitive faculties such as language, attention, and memory most often relies on testing small and homogeneous groups of volunteers coming to research facilities where they are asked to participate in behavioral experiments. We show that this limitation and sampling bias can be overcome by using smartphone technology to collect data in cognitive science experiments from thousands of subjects from all over the world. This mass coordinated use of smartphones creates a novel and powerful scientific ‘‘instrument’’ that yields the data necessary to test universal theories of cognition. This increase in power represents a potential revolution in cognitive science

Abstract

This paper describes a computerized alternative to glottochronology for estimating elapsed time since parent languages diverged into daughter languages. The method, developed by the Automated Similarity Judgment Program (ASJP) consortium, is different from glottochronology in four major respects: (1) it is automated and thus is more objective, (2) it applies a uniform analytical approach to a single database of worldwide languages, (3) it is based on lexical similarity as determined from Levenshtein (edit) distances rather than on cognate percentages, and (4) it provides a formula for date calculation that mathematically recognizes the lexical heterogeneity of individual languages, including parent languages just before their breakup into daughter languages. Automated judgments of lexical similarity for groups of related languages are calibrated with historical, epigraphic, and archaeological divergence dates for 52 language groups. The discrepancies between estimated and calibration dates are found to be on average 29% as large as the estimated dates themselves, a figure that does not differ significantly among language families. As a resource for further research that may require dates of known level of accuracy, we offer a list of ASJP time depths for nearly all the world’s recognized language families and for many subfamilies.

Abstract

We describe the key features of the visual world paradigm and review the main research areas where it has been used. In our discussion we highlight that the paradigm provides information about the way language users integrate linguistic information with information derived from the visual environment. Therefore the paradigm is well suited to study one of the key issues of current cognitive psychology, namely the interplay between linguistic and visual information processing. However, conclusions about linguistic processing (e.g., about activation, competition, and timing of access of linguistic representations) in the absence of relevant visual information must be drawn with caution.

Abstract

This study assessed, in a sample of 98 adult native speakers of Dutch, how their lexical skills and their speaking proficiency varied as a function of their age and level of education and profession (EP). Participants, categorized in terms of their age (18–35, 36–50, and 51–76 years old) and the level of their EP (low versus high), were tested on their lexical knowledge, lexical fluency, and lexical memory, and they performed four speaking tasks, differing in genre and formality. Speaking performance was rated in terms of communicative adequacy and in terms of number of words, number of T-units, words per T-unit, content words per T-unit, hesitations per T-unit, and grammatical errors per T-unit. Increasing age affected lexical knowledge positively but lexical fluency and memory negatively. High EP positively affected lexical knowledge and memory but EP did not affect lexical fluency. Communicative adequacy of the responses in the speaking tasks was positively affected by high EP but was not affected by age. It is concluded that, given the large variability in native speakers’ language knowledge and skills, studies investigating the question of whether second-language learners can reach native levels of proficiency, should take native-speaker variability into account.

Abstract

Disagreement exists about whether lexical selection in word production is a competitive process. Competition predicts semanticinterference from distractor words in immediate but not in delayed picture naming. In contrast, Janssen, Schirm, Mahon, and Caramazza (2008) obtained semanticinterference in delayed picture naming when participants had to decide between picture naming and oral reading depending on the distractor word’s colour. We report three experiments that examined the role of such taskdecisions. In a single-task situation requiring picture naming only (Experiment 1), we obtained semanticinterference in immediate but not in delayednaming. In a task-decision situation (Experiments 2 and 3), no semantic effects were obtained in immediate and delayed picture naming and word reading using either the materials of Experiment 1 or the materials of Janssen et al. (2008). We present an attentional account in which taskdecisions may hide or reveal semanticinterference from lexical competition depending on the amount of parallelism between task-decision and picture–word processing.

Abstract

Two groups of participants were trained to perceive an ambiguous sound [s/f] as either /s/ or /f/ based on lexical bias: One group heard the ambiguous fricative in /s/-final words, the other in /f/-final words. This kind of exposure leads to a recalibration of the /s/-/f/ contrast [e.g., 4]. In order to investigate when and how this recalibration emerges, test trials were interspersed among training and filler trials. The learning effect needed at least 10 clear training items to arise. Its emergence seemed to occur in a rather step-wise fashion. Learning did not improve much after it first appeared. It is likely, however, that the early test trials attracted participants' attention and therefore may have interfered with the learning process.

Abstract

Trilingual dictionary

Abstract

E. Dhooge and R. J. Hartsuiker (2010) reported experiments showing that picture naming takes longer with low- than high-frequency distractor words, replicating M. Miozzo and A. Caramazza (2003). In addition, they showed that this distractor-frequency effect disappears when distractors are masked or preexposed. These findings were taken to refute models like WEAVER++ (A. Roelofs, 2003) in which words are selected by competition. However, Dhooge and Hartsuiker do not take into account that according to this model, picture-word interference taps not only into word production but also into attentional processes. Here, the authors indicate that WEAVER++ contains an attentional mechanism that accounts for the distractor-frequency effect (A. Roelofs, 2005). Moreover, the authors demonstrate that the model accounts for the influence of masking and preexposure, and does so in a simpler way than the response exclusion through self-monitoring account advanced by Dhooge and Hartsuiker

Abstract

It has been argued that inhibition is a mechanism of attentional control in bilingual language performance. Evidence suggests that effects of inhibition are largest in the tail of a response time (RT) distribution in non-linguistic and monolingual performance domains. We examined this for bilingual performance by conducting delta-plot analyses of naming RTs. Dutch-English bilingual speakers named pictures using English while trying to ignore superimposed neutral Xs or Dutch distractor words that were semantically related, unrelated, or translations. The mean RTs revealed semantic, translation, and lexicality effects. The delta plots leveled off with increasing RT, more so when the mean distractor effect was smaller as compared with larger. This suggests that the influence of inhibition is largest toward the distribution tail, corresponding to what is observed in other performance domains. Moreover, the delta plots suggested that more inhibition was applied by high- than low-proficiency individuals in the unrelated than the other distractor conditions. These results support the view that inhibition is a domain-general mechanism that may be optionally engaged depending on the prevailing circumstances.

Abstract

E. Dhooge and R. J. Hartsuiker (2010) reported experiments showing that picture naming takes longer with low- than high-frequency distractor words, replicating M. Miozzo and A. Caramazza (2003). In addition, they showed that this distractor-frequency effect disappears when distractors are masked or preexposed. These findings were taken to refute models like WEAVER++ (A. Roelofs, 2003) in which words are selected by competition. However, Dhooge and Hartsuiker do not take into account that according to this model, picture-word interference taps not only into word production but also into attentional processes. Here, the authors indicate that WEAVER++ contains an attentional mechanism that accounts for the distractor-frequency effect (A. Roelofs, 2005). Moreover, the authors demonstrate that the model accounts for the influence of masking and preexposure, and does so in a simpler way than the response exclusion through self-monitoring account advanced by Dhooge and Hartsuiker

Abstract

Background: A well-known test for measuring verbal adequacy (i.e., verbal effectiveness) in mildly impaired aphasic speakers is the Amsterdam-Nijmegen Everyday Language Test (ANELT; Blomert, Koster, & Kean, 1995). Aphasia therapy practitioners score verbal adequacy qualitatively when they administer the ANELT to their aphasic clients in clinical practice. Aims: The current study investigated whether the construct validity of the ANELT could be further improved by substituting the qualitative score by a quantitative one, which takes the number of essential information units into account. The new quantitative measure could have the following advantages: the ability to derive a quantitative score of verbal efficiency, as well as improved sensitivity to detect changes in functional communication over time. Methods & Procedures: The current study systematically compared a new quantitative measure of verbal effectiveness with the current ANELT Comprehensibility scale, which is based on qualitative judgements. A total of 30 speakers of Dutch participated: 20 non-aphasic speakers and 10 aphasic patients with predominantly expressive disturbances. Outcomes & Results: Although our findings need to be replicated in a larger group of aphasic speakers, the main results suggest that the new quantitative measure of verbal effectiveness is more sensitive to detect change in verbal effectiveness over time. What is more, it can be used to derive a measure of verbal efficiency. Conclusions: The fact that both verbal effectiveness and verbal efficiency can be reliably as well as validly measured in the ANELT is of relevance to clinicians. It allows them to obtain a more complete picture of aphasic speakers' functional communication skills.

Abstract

Computational models of reading have differed in terms of whether they propose a single route forming the mapping between orthography and phonology or whether there is a lexical/sublexical route distinction. A critical test of the architecture of the reading system is how it deals with multi-letter graphemes. Rastle and Coltheart (1998) found that the presence of digraphs in nonwords but not in words led to an increase in naming times, suggesting that nonwords were processed via a distinct sequential route to words. In contrast Pagliuca, Monaghan, and McIntosh (2008) implemented a single route model of reading and showed that under conditions of visual noise the presence of digraphs in words did have an effect on naming accuracy. In this study, we investigated whether such digraph effects could be found in both words and nonwords under conditions of visual noise. If so it would suggest that effects on words and nonwords are comparable. A single route connectionist model of reading showed greater accuracy for both words and nonwords containing digraphs. Experimental results showed participants were more accurate in recognising words if they contained digraphs. However contrary to model predictions they were less accurate in recognising nonwords containing digraphs compared to controls. We discuss the challenges faced by both theoretical perspectives in interpreting these findings and in light of a psycholinguistic grain size theory of reading.

Abstract

This research reports the results of eliciting pronouns in two syntactic environments: Object pronouns and quantitative er (Q-er). Thus another type of language is added to the literature on subject and object clitic acquisition in the Romance languages (Jakubowicz et al., 1998; Hamann et al., 1996). Quantitative er is a unique pronoun in the Germanic languages; it has the same distribution as partitive clitics in Romance. Q-er is an N'-anaphor and occurs obligatorily with headless noun phrases with a numeral or weak quantifier. Q-er is licensed only when the context offers an antecedent; it binds an empty position in the NP. Data from typically-developing children aged 5;0-6;0 show that object and Q-er pronouns are not acquired equally; it is proposed that this is due to their different syntax. The use of Q-er involves more sophisticated syntactic knowledge: Q-er occurs at the left edge of the VP and binds an empty position in the NP, whereas object pronouns are simply stand-ins for full NPs and occur in the same position. These Dutch data reveal that pronouns are not used as exclusively as object clitics are in the Romance languages (Varlakosta, in prep.).

Abstract

The present study explored whether language-nonselective access in bilinguals occurs across word classes in a sentence context. Dutch–English bilinguals were auditorily presented with English (L2) sentences while looking at a visual world. The sentences contained interlingual homophones from distinct lexical categories (e.g., the English verb spoke, which overlaps phonologically with the Dutch noun for ghost, spook). Eye movement recordings showed that depictions of referents of the Dutch (L1) nouns attracted more visual attention than unrelated distractor pictures in sentences containing homophones. This finding shows that native language objects are activated during second language verb processing despite the structural information provided by the sentence context. Research highlights We show that native language words are activated during second language sentence processing. We tested this in a visual world setting on homophones with a different word class across languages. Fixations show that processing second language verbs activated native language nouns.

Abstract

This paper investigates the effects of novel words on a cognitively plausible computational model of word learning. The model is first familiarized with a set of words, achieving high recognition scores and subsequently offered novel words for training. We show that the model is able to recognize the novel words as different from the previously seen words, based on a measure of novelty that we introduce. We then propose a procedure analogous to novelty preference in infants. Results from simulations of word learning show that adding this procedure to our model speeds up training and helps the model attain higher recognition rates.

Abstract

Understanding speech usually seems easy, but it can become noticeably harder when the speaker has a foreign accent. This is because foreign accents add considerable variation to speech. Research on foreign-accented speech shows that participants are able to adapt quickly to this type of variation. Less is known, however, about longer-term maintenance of adaptation. The current study focused on long-term adaptation by exposing native listeners to foreign-accented speech on Day 1, and testing them on comprehension of the accent one day later. Comprehension was thus not tested immediately, but only after a 24 hour period. On Day 1, native Dutch listeners listened to the speech of a Hebrew learner of Dutch while performing a phoneme monitoring task that did not depend on the talker’s accent. In particular, shortening of the long vowel /i/ into /ɪ/ (e.g., lief [li:f], ‘sweet’, pronounced as [lɪf]) was examined. These mispronunciations did not create lexical ambiguities in Dutch. On Day 2, listeners participated in a cross-modal priming task to test their comprehension of the accent. The results will be contrasted with results from an experiment without delayed testing and related to accounts of how listeners maintain adaptation to foreign-accented speech.

Abstract

Foreign-accented speech is often perceived as more difficult to understand than native speech. What causes this potential difficulty, however, remains unknown. In the present study, we compared acoustic similarity and accent ratings of American-accented Dutch with a cross-modal priming task designed to measure online speech processing. We focused on two Dutch diphthongs: ui and ij. Though both diphthongs deviated from standard Dutch to varying degrees and perceptually varied in accent strength, native Dutch listeners recognized words containing the diphthongs easily. Thus, not all foreign-accented speech hinders comprehension, and acoustic similarity and perceived accentedness are not always predictive of processing difficulties.

Abstract

We present the first study on the acquisition of the Dutch quantitative pronoun er in sentences such as de vrouw draagt er drie ‘the woman is carrying three.’ There is a large literature on Dutch children’s interpretation of pronouns and a few recent production studies, all specifically looking at 3rd person singular pronouns and the so-called Delay of Principle B effect (Coopmans & Philip, 1996; Koster, 1993; Spenader, Smits and Hendriks, 2009). However, no one has studied children’s use of quantitative er. Dutch is the only Germanic language with such a pronoun.

Abstract

1. Introduction External input is necessary to acquire language. Consequently, the comprehension of various constituents of language, such as lexical items or syntactic and semantic structures should emerge at the same time as or even precede their production. However, in the case of pronouns this general assumption does not seem to hold. On the contrary, while children at the age of four use pronouns and reflexives appropriately during production (de Villiers, et al. 2006), a number of comprehension studies across different languages found chance performance in pronoun trials up to the age of seven, which co-occurs with a high level of accuracy in reflexive trials (for an overview see e.g. Conroy, et al. 2009; Elbourne 2005).

Abstract

In the present paper we show that a general-purpose word learning model can simulate several important findings from recent experiments in language acquisition. Both the addition of background noise and varying the speaker have been found to influence infants’ performance during word recognition experiments. We were able to replicate this behaviour in our artificial word learning agent. We use the results to discuss both advantages and limitations of computational models of language acquisition.

Abstract

Speakers often use spatial metaphors to talk about musical pitch (e.g., a low note, a high soprano). Previous experiments suggest that English speakers also think about pitches as high or low in space, even when theyʼre not using language or musical notation (Casasanto, 2010). Do metaphors in language merely reflect pre-existing associations between space and pitch, or might language also shape these non-linguistic metaphorical mappings? To investigate the role of language in pitch tepresentation, we conducted a pair of non-linguistic spacepitch interference experiments in speakers of two languages that use different spatial metaphors. Dutch speakers usually describe pitches as ʻhighʼ (hoog) and ʻlowʼ (laag). Farsi speakers, however, often describe high-frequency pitches as ʻthinʼ (naazok) and low-frequency pitches as ʻthickʼ (koloft). Do Dutch and Farsi speakers mentally represent pitch differently? To find out, we asked participants to reproduce musical pitches that they heard in the presence of irrelevant spatial information (i.e., lines that varied either in height or in thickness). For the Height Interference experiment, horizontal lines bisected a vertical reference line at one of nine different locations. For the Thickness Interference experiment, a vertical line appeared in the middle of the screen in one of nine thicknesses. In each experiment, the nine different lines were crossed with nine different pitches ranging from C4 to G#4 in semitone increments, to produce 81 distinct trials. If Dutch and Farsi speakers mentally represent pitch the way they talk about it, using different kinds of spatial representations, they should show contrasting patterns of cross-dimensional interference: Dutch speakersʼ pitch estimates should be more strongly affected by irrelevant height information, and Farsi speakersʼ by irrelevant thickness information. As predicted, Dutch speakersʼ pitch estimates were significantly modulated by spatial height but not by thickness. Conversely, Farsi speakersʼ pitch estimates were modulated by spatial thickness but not by height (2x2 ANOVA on normalized slopes of the effect of space on pitch: F(1,71)=17,15 p<.001). To determine whether language plays a causal role in shaping pitch representations, we conducted a training experiment. Native Dutch speakers learned to use Farsi-like metaphors, describing pitch relationships in terms of thickness (e.g., a cello sounds ʻthickerʼ than a flute). After training, Dutch speakers showed a significant effect of Thickness interference in the non-linguistic pitch reproduction task, similar to native Farsi speakers: on average, pitches accompanied by thicker lines were reproduced as lower in pitch (effect of thickness on pitch: r=-.22, p=.002). By conducting psychophysical tasks, we tested the ʻWhorfianʼ question without using words. Yet, results also inform theories of metaphorical language processing. According to psycholinguistic theories (e.g., Bowdle & Gentner, 2005), highly conventional metaphors are processed without any active mapping from the source to the target domain (e.g., from space to pitch). Our data, however, suggest that when people use verbal metaphors they activate a corresponding non-linguistic mapping from either height or thickness to pitch, strengthening this association at the expense of competing associations. As a result, people who use different metaphors in their native languages form correspondingly different representations of musical pitch. Casasanto, D. (2010). Space for Thinking. In Language, Cognition and Space: State of the art and new directions. V. Evans & P. Chilton (Eds.), 453-478, London: Equinox Publishing. Bowdle, B. & Gentner, D. (2005). The career of metaphor. Psychological Review, 112, 193-216.

Abstract

Researchers in human language processing and acquisition are making an increasing use of computational models. Computer simulations provide a valuable platform to reproduce hypothesised learning mechanisms that are otherwise very difficult, if not impossible, to verify on human subjects. However, computational models come with problems and risks. It is difficult to (automatically) extract essential information about the developing internal representations from a set of simulation runs, and often researchers limit themselves to analysing learning curves based on empirical recognition accuracy through time. The associated risk is to erroneously deem a specific learning behaviour as generalisable to human learners, while it could also be a mere consequence (artifact) of the implementation of the artificial learner or of the input coding scheme. In this paper a set of simulation runs taken from the ACORNS project is investigated. First a look `inside the box' of the learner is provided by employing novel quantitative methods for analysing changing structures in large data sets. Then, the obtained findings are discussed in the perspective of their ecological validity in the field of child language acquisition.

Abstract

Previous studies on Cantonese Chinese showed that rising
question intonation contours on low-toned words lead to
frequent misperceptions of the tones. Here we explored the
processing consequences of this interaction between tone and
intonation by comparing the processing and identification of
monosyllabic critical words at the end of questions and
statements, using a tone identification task, and ERPs as an
online measure of speech comprehension. Experiment 1
yielded higher error rates for the identification of low tones at
the end of questions and a larger N400-P600 pattern, reflecting
processing difficulty and reanalysis, compared to other
conditions. In Experiment 2, we investigated the effect of
immediate lexical context on the tone by intonation interaction.
Increasing contextual constraints led to a reduction in errors
and the disappearance of the P600 effect. These results
indicate that there is an immediate interaction between tone,
intonation, and context in online speech comprehension. The
difference in performance and activation patterns between the
two experiments highlights the significance of context in
understanding a tone language, like Cantonese-Chinese.

Abstract

Daily conversations contain many repetitions of identical and similar word forms. For bilinguals, the words can even come from the same or different languages. How do such repetitions affect the human word recognition system? The Bilingual Interactive Activation Plus (BIA+) model provides a theoretical and computational framework for understanding word recognition and word repetition in bilinguals. The model assumes that both phenomena involve a language non-selective process that is sensitive to the task context. By means of computer simulations, the model can specify both qualitatively and quantitatively how bilingual lexical processing in one language is affected by the other language. Our review discusses how BIA+ handles cross-linguistic repetition and masked orthographic priming data from two key empirical studies. We show that BIA+ can account for repetition priming effects within- and between-languages through the manipulation of resting-level activations of targets and neighbors (words sharing all but one letter with the target). The model also predicts cross-linguistic performance on within- and between-trial orthographic priming without appealing to conscious strategies or task schema competition as an explanation. At the end of the paper, we briefly evaluate the model and indicate future developments.

Abstract

1. Introduction Despite a large literature on Dutch children’s pronoun interpretation, relatively little is known about their production. In this study we elicited pronouns in two syntactic environments: object pronouns and quantitative er (Q-er). The goal was to see how different types of pronouns develop, in particular, whether acquisition depends on their different syntactic properties. Our Dutch data add another type of language to the acquisition literature on object clitics in the Romance languages. Moreover, we present another angle on this discussion by comparing object pronouns and Q-er.

Abstract

This paper presents an analysis of phoneme durations of emotional speech in two languages: Dutch and Korean. The analyzed corpus of emotional speech has been specifically developed for the purpose of cross-linguistic comparison, and is more balanced than any similar corpus available so far: a) it contains expressions by both Dutch and Korean actors and is based on judgments by both Dutch and Korean listeners; b) the same elicitation technique and recording procedure were used for recordings of both languages; and c) the phonetics of the carrier phrase were constructed to be permissible in both languages. The carefully controlled phonetic content of the carrier phrase allows for analysis of the role of specific phonetic features, such as phoneme duration, in emotional expression in Dutch and Korean. In this study the mutual effect of language and emotion on phoneme duration is presented.

Abstract

In this paper we investigate a computational model of word learning, that is embedded in a cognitively and ecologically plausible framework. Multi-modal stimuli from four different speakers form a varied source of experience. The model incorporates active learning, attention to a communicative setting and clarity of the visual scene. The model's ability to learn associations between speech utterances and visual concepts is evaluated during training to investigate the influence of active learning under conditions of uncertain input. The results show the importance of shared attention in word learning and the model's robustness against noise.

Abstract

In foreign-accented speech, listeners have to handle noticeable deviations from the standard pronunciation of a target language. Three cross-modal priming experiments investigated how short- and long-term experiences with a foreign accent influence word recognition by native listeners. In experiment 1, German-accented words were presented to Dutch listeners who had either extensive or limited prior experience with German-accented Dutch. Accented words either contained a diphthong substitution that deviated acoustically quite largely from the canonical form (huis [hys], "house", pronounced as [hoys]), or that deviated acoustically to a lesser extent (lijst [lst], "list", pronounced as [lst]). The mispronunciations never created lexical ambiguity in Dutch. While long-term experience facilitated word recognition for both types of substitutions, limited experience facilitated recognition only of words with acoustically smaller deviations. In experiment 2, Dutch listeners with limited experience listened to the German speaker for 4 min before participating in the cross-modal priming experiment. The results showed that speaker-specific learning effects for acoustically large deviations can be obtained already after a brief exposure, as long as the exposure contains evidence of the deviations. Experiment 3 investigates whether these short-term adaptation effects for foreign-accented speech are speaker-independent.

Abstract

The modality learning effect, according to Mayer (2001), proposes that learning is enhanced when information is presented in both the visual and auditory domain (e.g., pictures and spoken information), compared to presenting information solely in the visual channel (e.g., pictures and written text). Most of the evidence for this effect comes from adults in a laboratory setting. Therefore, we tested the modality effect with 80 children in the highest grade of elementary school, in a naturalistic setting. In a between-subjects design children either saw representational pictures with speech or representational pictures with text. Retention and transfer knowledge was tested at three moments: immediately after the intervention, one day after, and after one week. The present study did not find any evidence for a modality effect in children when the lesson is learner-paced. Instead, we found a reversed modality effect directly after the intervention for retention. A reversed modality effect was also found for the transfer questions one day later. This effect was robust, even when controlling for individual differences.