Edited by: Jamie Reilly, Temple University, United States
Reviewed by: Harald C. Traue, University of Ulm, Germany; Fernando Marmolejo-Ramos, University of South Australia, Australia
This article was submitted to Theoretical and Philosophical Psychology, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
There are considerable gaps in our knowledge of how children develop abstract language. In this paper, we tested the Affective Embodiment Account, which proposes that emotional information is more essential for abstract than concrete conceptual development. We tested the recognition memory of 7- and 8-year-old children, as well as a group of adults, for abstract and concrete words which differed categorically in valence (negative, neutral, and positive). Word valence significantly interacted with concreteness in hit rates of both children and adults, such that effects of valence were only found in memory for abstract words. The pattern of valence effects differed for children and adults: children remembered negative words more accurately than neutral and positive words (a negativity effect), whereas adults remembered negative and positive words more accurately than neutral words (a negativity effect and a positivity effect). In addition, signal detection analysis revealed that children were better able to discriminate negative than positive words, regardless of concreteness. The findings suggest that the memory accuracy of 7- and 8-year-old children is influenced by emotional information, particularly for abstract words. The results are in agreement with the Affective Embodiment Account and with multimodal accounts of children’s lexical development.
Abstract words refer to concepts that are difficult to experience through the senses, like
It has been a challenge for some theories of lexical and conceptual knowledge to explain the acquisition of abstract concepts. There are several different theoretical views on how lexical knowledge is represented, and they can be placed on a spectrum based on the role of sensory and motor systems (
The strongly embodied accounts have the greatest difficulty explaining acquisition of abstract vocabulary; if word meanings cannot be experienced through the senses, how are they acquired? Amodal or unembodied accounts explain acquisition of abstract meaning in terms of linguistic representation: the meanings of abstract words are represented by association with already learned words (
Within the multimodal framework, it has been suggested that the emotional information associated with a word (i.e., word valence), provides a bootstrapping mechanism for children’s abstract vocabulary acquisition (
In adults, word valence affects lexical processing speed, with positive words processed faster than neutral words. Negative words are sometimes processed more slowly than neutral words (
Two previous developmental studies have found some support for predictions of the Affective Embodiment Account by testing the effects of word valence and concreteness on children’s lexical processing (
To further test predictions of the Affective Embodiment Account, in the present study, we shifted focus to a different cognitive process and investigated the effects of word valence and concreteness on children’s recognition memory. This allowed us to test whether children’s retrieval of abstract and concrete words are differentially influenced by emotional information. To our knowledge, the relationship between concreteness and emotional enhancement of memory has not yet been investigated, in either children or adults.
Emotional enhancement of memory is a phenomenon in which positive and negative emotional information is better remembered than neutral information. It has been demonstrated in adult studies using various stimuli, including pictures, stories, and words (
Many developmental studies have examined the effects of valence on memory using images or stories as stimuli (
Recently,
In short, there is limited research on the effects of word valence on children’s memory. Moreover, the existing research has used varying and inconsistent methods with mixed results, which calls for additional exploration. In the present study, we examined the effects of word valence and concreteness on adults’ and children’s recognition memory. We tested 7- and 8-year-old children, since this is the age of rapid abstract word acquisition, and compared the groups’ performance to each other. Since children have greater neural response and attentional bias toward negative stimuli (
The child participants were 45 7-year-olds (
We selected 120 monosyllabic words to achieve a factorial manipulation of valence (negative, positive, neutral) and concreteness (abstract, concrete). The negative words had valence ratings of 1–4, neutral words had valence ratings of 4.01–6.50, and the positive words had valence ratings of 6.51–9, based on the ratings from
Mean characteristics of word stimuli for each word type (standard deviations in parentheses).
Word type |
||||||||
Negative |
Neutral |
Positive |
||||||
Abstract | Concrete | Abstract | Concrete | Abstract | Concrete | Valence | Concreteness | |
Valence | 2.95 (0.60) | 3.31 (0.50) | 5.53 (0.60) | 5.38 (0.62) | 7.07 (0.50) | 7.04 (0.44) | <0.001 | 0.71 |
PLD | 1.23 (0.29) | 1.42 (0.30) | 1.26 (0.33) | 1.39 (0.30) | 1.42 (0.37) | 1.24 (0.25) | 1 | 0.34 |
Child spoken frequency | 47.95 (86.65) | 17.05 (37.06) | 44.70 (152.53) | 38.55 (146.50) | 26.16 (51.13) | 44.67 (174.59) | 0.94 | 0.82 |
Grade 2 print frequency | 34.95 (56.75) | 28.70 (29.34) | 26.15 (19.80) | 54.4 (155.37) | 54.47 (107.98) | 34.81 (43.42) | 0.80 | 0.90 |
Age of acquisition | 6.91 (1.83) | 6.83 (1.36) | 6.69 (1.31) | 7.08 (1.45) | 6.85 (1.27) | 6.70 (1.15) | 0.93 | 0.80 |
Imageability | 3.52 (0.77) | 5.29 (0.78) | 3.52 (1.00) | 5.33 (0.75) | 3.40 (0.55) | 5.39 (0.71) | 0.98 | <0.001 |
Concreteness | 2.82 (0.64) | 4.29 (0.37) | 2.85 (0.41) | 4.26 (0.48) | 2.61 (0.58) | 4.36 (0.40) | 0.96 | <0.001 |
Length | 4.30 (0.92) | 4.55 (0.83) | 4.45 (0.89) | 4.40 (0.68) | 4.47 (1.07) | 4.38 (0.97) | 1 | 0.92 |
For the recognition memory (retrieval) task, an additional 40 positive, 40 neutral, and 40 negative monosyllabic words were selected as foils. These were matched to the original words on valence, age of acquisition, and length in letters. All stimuli are available at
Procedures for child and adult participants were similar, except for minor differences in the filler and language tasks. The study consisted of encoding, filler, and retrieval phases.
During the encoding phase, participants were presented with 10 practice trials followed by 120 words and 120 non-words. Participants were not told that their memory for the words would be tested in the subsequent retrieval session. Participants sat beside the experimenter, in front of a computer screen. Both wore headphones. Participants were instructed to distinguish real and fake words as if they were “word detectives” via button press. On each trial, a white fixation cross was presented on a black background. An audio recording was presented via headphones. After the participant pressed one of the two buttons on the response box, the fixation cross was replaced with a letter “B,” and the experimenter pressed a key to proceed to the next trial. The order of word presentation was randomized and stimuli were presented using E-prime (
After the encoding phase, participants completed the filler task at an adjacent table, consisting of five manual dexterity subscales of the Bruininks-Oseretsky Test of Motor Proficiency (BOT-2;
In the retrieval phase, participants returned to the computer and completed the recognition memory task. They were told that they were playing a memory game and were asked to distinguish words heard in the encoding task (i.e., old) from words they did not hear in the encoding task (i.e., new) via button press. We presented 240 trials, in a random order, including 120 old and 120 new words.
For child participants, the Peabody Picture Vocabulary Test (PPVT-4;
One common way of analyzing recognition memory data is to use a signal detection paradigm (
We removed three words that had accuracy <50 in the encoding task:
Children’s and adults’ hit rates, false alarm rates, d’ scores, and Criterion C values for words as a function of valence and concreteness. Horizontal bars indicate medians; diamonds indicate means. Colored boxes correspond to the range between the first and third quartiles (i.e., the 25th and 75th percentiles). Whiskers indicate the lowest and highest values up to 1.5 times the interquartile range from the first and third quartiles, respectively. Dots represent all other values beyond the whiskers.
The three-way interaction was not significant [
No interactions reached statistical significance (
No interactions reached statistical significance (
No interactions reached statistical significance (
To facilitate comparison with children’s responses, we removed the same three items that were removed from those analyses. Data preparation and analyses were the same as in the child analyses, but without age as a factor.
The analysis revealed a significant interaction of valence and concreteness [
The interaction between valence and concreteness was not significant [
The interaction between valence and concreteness was not significant [
The analysis revealed a significant interaction [
The purpose of the present study was to investigate emotional enhancement of memory and its differential effect for abstract and concrete words, in a developmental context. We tested predictions derived from the Affective Embodiment Account, that emotional enhancement might be stronger for abstract words than for concrete words.
We measured several aspects of children’s and adults’ memory performance. One of these provided support for the Affective Embodiment Account: we found interactions of valence and concreteness in both children’s and adults’ hit rates (correct identification of “old” items), such that effects of valence were significant for abstract but not for concrete word memory. This supported the assumptions that for abstract words emotional information is important for grounding word meanings, and that for concrete words, which have more grounding in sensorimotor networks, valence is less influential.
We presented child and adult participants with the same stimuli and found both similarities and differences in memory effects. In particular, the nature of the valence effects for abstract word memory differed for children and adults. While children showed more accurate hit rates for negative (vs. neutral and positive) abstract words, adults showed higher hit rates for negative
Adults’ valence effects showed the usual
The differences in the nature of the valence effects we observed for children and adults may be due to developmental differences in emotional processing. Children undergo significant development in their ability to understand, regulate, and change emotional responses (for a review, see
In contrast to the results described for hits, our other measures of children’s memory performance showed main effects of valence and concreteness separately, without an interaction. Specifically, children showed better discrimination in their memory for negative than for positive words. They also made more false alarms (judging that new items were “old”) for positive words than for negative and neutral words. When words were of positive valence, children were more likely to incorrectly respond that they had seen those words before. Our findings are reasonably consistent with previous observations of valence effects in children’s word recall (
For both children and adults, word concreteness influenced memory performance. Both groups produced more false alarms for abstract than for concrete words, and they also showed more accurate discrimination for concrete than abstract words. This recognition memory advantage for concrete words is consistent with previous literature (e.g.,
One potential limitation of the present study was the relatively low memory accuracy rates. The average hit rates for both children and adults were in the mid 60s, suggesting that the recognition memory task was quite challenging. We presented participants with 120 words at encoding in order to test retrieval with a good number of items of each word type. The recognition memory task was unexpected; we did not warn participants that their memory would be tested because we wanted to assess incidental effects on memory. These methodological choices likely contributed to making the recognition memory task challenging for all participants and may have contributed to the different results we observed across hit rates and discrimination scores. Another limitation is created by the fact that we conducted the study in English. As such, our results cannot necessarily be generalized to other languages. In addition, the study was conducted with a WEIRD (Western, Educated, Industrialized, Rich, and Democratic) sample. Therefore, the findings cannot be applied to other populations and may not generalize to human conceptual development more broadly (
In conclusion, this study, for the first time, investigated the link between emotional enhancement of memory and word concreteness. The results showed that emotional valence plays an important role in lexical memory of both children and adults, particularly for abstract words. The results thus provide insights about how children acquire the meanings of abstract words. In keeping with the Affective Embodiment Account, and with multimodal accounts of lexical semantics, our findings point to emotion as a factor that shapes children’s word representations and helps them process and understand abstract concepts.
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below:
The studies involving human participants were reviewed and approved by University of Calgary Conjoint Faculties Research Ethics Board. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.
JK, DS, and PP conceptualized the study and jointly wrote the manuscript. JK tested the participants. JK and DS analyzed the data. All authors contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Cohen’s