Abstract
The age of acquisition (AoA) refers to the age at which an individual learns specific items or words. Research on word recognition has shown that items with lower AoA—those acquired earlier—can be processed more quickly and accurately. In the field of Japanese word recognition, a large-scale database that would enable mega-study approaches has not been well-established. In this study, we developed an AoA norm for over 5,000 Japanese words. A total of 1,345 adults rated the AoA of 5,736 words using a 7-point scale. These ratings demonstrated satisfactory reliability. Furthermore, when examining the correlation with lexical decision task performance, words with lower AoA showed shorter response times and higher accuracy than words with higher AoA. These findings indicate that the AoA rating database collected in this study serves as a valuable resource for research using Japanese words.
1 Introduction
The age of acquisition (AoA) refers to the age at which an individual learns specific items or words. Previous studies have demonstrated that items acquired early are processed more quickly and accurately than those acquired later. Rochford and Williams () conducted an experiment involving 32 adult patients with speech impairment and 120 children aged 2–11 years. Participants were asked to name the objects in the order in which they were presented. A significant correlation was observed between the correct response rates of children and adult patients. Specifically, the study revealed a negative correlation between the age at which 80% or more children correctly answered each item and adult patients' correct response rate for the task. Consequently, it was found that items acquired early, that is, those with a low AoA, could be named more accurately by participants. This phenomenon applies to various stimuli, including words, phrases, images, and faces (Elsherif et al., ).
This study focused on the AoA effect on word recognition. In word recognition research, the AoA effect has been observed in naming (Morrison and Ellis, ), lexical decision (Turner et al., ), and semantic decision tasks (Brysbaert et al., ), suggesting that the AoA effect is not limited to the speech system, which is crucial in naming tasks, and that the semantic system may also be involved.
Until the 2000s, many AoA studies employed an experimental approach to factorially compare items acquired early in life with those acquired later. However, the mega-study approach (Balota et al., ) gained prominence in subsequent decades. This approach involves collecting word attributes on a large scale and analyzing the attributes that influence cognitive performance in various tasks. Researchers have estimated the AoA of more than 40,000 English words (Brysbaert and Biemiller, ), while also collecting AoA ratings from Italian (Montefinese et al., ), Dutch (Brysbaert et al., ), German (Birchenough et al., ), Portuguese (Cameirão and Vicente, ), Spanish (Alonso et al., ), Chinese (Xu et al., ), and French (Ferrand et al., ). Łuniewska et al. () conducted a comparative analysis of AoA across 25 languages. They demonstrated the reliability of subjective ratings by establishing a correlation between them and both the previously collected AoA and quasi-objective AoA estimated using inventories designed to assess children's language development. In addition to subjective AoA ratings, objective AoA measures have also been collected, such as parental reports of the age at which children begin to use specific words (e.g., Frank et al., ) and estimations of AoA based on the age at which characters are designated for learning in school textbooks (Cai et al., ).
These AoA ratings also serve as control variables in word recognition research (e.g., Diveica et al., ; Pexman et al., ). When investigating the impact of a variable of interest on word recognition, excluding the influences of other variables is essential. As AoA is correlated with frequency, imageability, and concreteness (Zevin and Seidenberg, ), its impact must be controlled to examine these variables' roles.
The AoA effect has also been observed in Japanese. Havelka and Tomita () demonstrated that reaction times in naming tasks were shorter in low-AoA conditions than in high-AoA conditions. They observed that for Japanese, the AoA effect was greater for words written in kanji (ideographic characters) than those written in hiragana (phonetic characters). Yamazaki et al. () reported similar findings, demonstrating that when naming single kanji characters, both the age at which the word was learned to be spoken and the age at which it was learned to be written predicted naming speed.
The abovementioned studies employed experimental approaches and used specific controlled stimuli. They demonstrated AoA effects using rigorous experimental methods; however, their generalizability is limited owing to the small number of stimuli used. To determine whether similar AoA effects are observed across a broader range of Japanese words, a mega-study approach would be beneficial. However, AoA norms suitable for such an approach do not currently exist for Japanese. Only the Japanese AoA ratings from Nishimoto et al. ()—who collected the AoA, familiarity, and name agreements for 359 line drawings—are available publicly. In their study, the AoA ratings pertained to line drawings rather than words; thus, only concrete objects were used. The number of items rated was relatively small (359), which is another limitation for the mega-study approach. To facilitate the mega-study approach in Japanese word recognition research, collecting AoA ratings for significantly larger numbers of words is important. Therefore, this study aimed to develop a norm by collecting AoA ratings for over 5,000 Japanese words. To assess the validity and reliability of the collected data, we also examined the relationships between AoA in other languages, as well as the relationships between lexical and semantic variables in Japanese and performance on cognitive tasks. Based on previous research, we expected that items with lower AoA would show shorter reaction times and higher accuracy rates than items with higher AoA on lexical decision tasks.
2 Method
2.1 Participants
To collect AoA ratings from 40 participants per word, we recruited participants aged 18 years and older through Yahoo! Crowdsourcing (https://crowdsourcing.yahoo.co.jp/). Of the 1,376 respondents who completed the survey, 23 were excluded from the analysis owing to incorrect responses on the Directed Questions Scale (DQS; Maniaci and Rogge, ), while eight were excluded because they selected the same option for all rating items. Consequently, 1,345 participants were included in the analysis. All the participants reported that their primary language was Japanese. A total of 320 participants identified themselves as women, 1,012 as men, and 13 did not specify their gender. The average age was 50.70 years (SD = 12.04, range: 18–90 years). The participants' levels of education were as follows: 15 respondents had graduated from junior high school, 346 had graduated from high school, 854 had graduated from undergraduate university, 86 had completed a master's degree at a graduate school, 22 had completed a doctoral degree at a graduate school, and 22 did not respond.
2.2 Stimuli
We collected AoA ratings for 5,736 words, for which Ota and Mochizuki () collected data on lexical decision tasks. They adopted a mega-study approach and selected items with ratings or characteristics from wider databases to examine the relationships between lexical and semantic variables and lexical decision performance. By using these items, we can assess the relationship between AoA and various variables, as well as performance on lexical decision tasks on a large scale. The words used by Ota and Mochizuki () can be referred to by the following variables: word familiarity (Asahara, ; Fujita and Kobayashi, ), word frequency (Amano and Kondo, ; University of Tsukuba et al., ), word difficulty (Kajiwara et al., ), imageability (Sakuma et al., ), semantic orientations (valence; Takamura et al., ), abstractness (The Social Computing Laboratory and Nara Institute of Science and Technology, ), and body-object interaction (Mochizuki and Ota, ).
Following the procedure of Ota and Mochizuki (), the number of letters and morae (a unit of syllable weight) and the phonological and morphological neighborhood sizes were calculated as variables that can be calculated from the words. Subsequently, 5,736 words were randomly divided into 26 lists. Participants were randomly assigned to one of the 26 lists, and each participant rated the AoA for 220 or 221 words during the survey.
2.3 Procedure
The survey was conducted using Pavlovia Survey (Open Science Tools, Nottingham, UK). The participants accessed the survey from a crowdsourcing platform. Informed consent was obtained from all the participants after they received an explanation at the beginning of the survey. The participants were subsequently provided with detailed instructions based on the research of Cortese and Khanna () and Stadthagen-Gonzalez and Davis (). We modified the instructions to fit the survey format. Specific instructions and English translations are provided on the Open Science Framework (https://osf.io/fawmq/). In this survey, AoA was evaluated using a seven-point Likert scale. Participants were asked to rate the AoA using the following categories: 0–1 years old (1), 2–3 years old (2), 4–5 years old (3), 6–7 years old (4), 8–9 years old (5), 10–11 years old (6), and 12 years or older (7). For words that the participants did not know their AoA, they were instructed to select (7).1
Participants rated 10 practice items (恭しい[“respectful”], ばらつく[“scatter”], 下士官[“non-commissioned officer”], まさか[“impossible”], 疎ましい[“unwelcome”], みんな[“everyone”], 寝そべる[“sprawl”], チャート[“chart”], 騒ぐ[“be noisy”], 大きい[“large”]) before rating the target items. The practice items included words predicted to have a lower AoA (e.g., “large” and “everyone”) and words predicted to have a higher AoA (e.g., “respectful” and “unwelcome”). The participants familiarized themselves with the procedure using practice items and rated the target items. The participants rated 220 or 221 words in random order. In addition to the target items, one item was a DQS, and the participants were instructed to select the “10–11 years old” option.
After the target items were rated, participants were asked to complete the ENDCOREs (Fujimoto and Daibo, )—a scale for measuring social skills—and the Plymouth Sensory Imagery Questionnaire (Fukui and Aoki, )—a scale for measuring sensory imagery—to measure their characteristics.2 The survey ended after the participants answered questions regarding their age, gender, first language, and highest level of education. After completing the survey, participants were given a reward code and received monetary compensation via the crowdsourcing platform.
3 Results
3.1 Data processing
As explained above, before summarizing the AoA ratings, we first excluded 31 participants who did not respond appropriately to the DQS and selected the same option for all target items. Means and standard deviations were calculated for each item. To calculate the reliability measures described below, we used each participant's individual rating values rather than the ratings' average.
3.2 Reliabilities
The participants' ratings were randomly divided into two groups, and the split-half reliability was calculated for 100 iterations, which resulted in a high reliability coefficient of M = 0.924, SD = 0.001. The mean absolute z-value, indicating the deviation of each participant's rating from the item-specific mean, was Mz = 0.805 and SDz = 0.269. The participant-sample correlation coefficient, which measures the degree of correlation between participants' ratings and the mean of their ratings on the assigned list, was Mr = 0.682 and SDr = 0.182, suggesting a satisfactory level of reliability.
3.3 Correlations with AoA ratings in other datasets
To validate our data, we calculated the correlations between the AoA ratings collected in this study and those from large-scale AoA datasets that provide English labels or translations. We first assigned what we considered to be the most representative English translations to the Japanese items. Subsequently, from six AoA datasets (see Table 1) that include English labels or corresponding English translations, we extracted only the items that matched those in the present dataset and calculated correlation coefficients. Additionally, although the number of items was relatively small, correlations were also calculated with AoA ratings collected from Japanese speakers (Nishimoto et al., ). As shown in Table 1, all ratings were only moderately positively correlated (r = 0.396–0.531). Among these, despite the limited number of items, the correlation between the AoA ratings from Nishimoto et al. () and those obtained in the present study was relatively high.
Table 1
| Datasets | Language | N. of items | N. of matched items | r | 95% CI | p |
|---|---|---|---|---|---|---|
| Cortese and Khanna () | English | 3,000 | 1,321 | 0.396 | 0.350, 0.441 | <0.001 |
| Cameirão and Vicente () | Portuguese | 1,994 | 1,255 | 0.428 | 0.382, 0.472 | <0.001 |
| Montefinese et al. () | Italian | 1,957 | 915 | 0.453 | 0.400, 0.503 | <0.001 |
| Birchenough et al. () | Germany | 3,259 | 2,082 | 0.489 | 0.455, 0.521 | <0.001 |
| Kuperman et al. () | English | 31,124 | 4,750 | 0.506 | 0.485, 0.527 | <0.001 |
| Brysbaert and Biemiller () | English | 43,991 | 4,587 | 0.514 | 0.492, 0.535 | <0.001 |
| Nishimoto et al. () | Japanese | 359 | 128 | 0.531 | 0.394, 0.645 | <0.001 |
Correlation coefficients between AoA ratings.
In Nishimoto et al. (), AoA ratings were collected for pictures and their concept names.
3.4 Relationships among AoA, psycholinguistic variables, and lexical decision task performance
Figure 1 shows the correlations with other lexical and semantic variables in Japanese, as well as with performance on lexical decision tasks (Ota and Mochizuki, ). The correlations between AoA and each psycholinguistic variable showed patterns consistent with previous research. To examine the effects of AoA and psycholinguistic variables on performance in the lexical decision task, multiple regression analyses were conducted (Table 2). Following Kuperman et al. (), who collected large-scale subjective AoA ratings, a baseline model was constructed including number of characters, number of morae, word frequency (log-transformed; University of Tsukuba et al., ), and orthographic Levenshtein distance (Yarkoni et al., ). When AoA was added to this model, AoA significantly predicted both response times and accuracy: words with lower AoA ratings were associated with shorter reaction times and higher accuracy (Model 2 in Table 2). Even when additional available psycholinguistic variables were included in the model to rule out their influence, the effect of AoA on response time persisted, such that lower AoA ratings were associated with faster responses. However, for accuracy, an unexpected pattern emerged: lower AoA ratings were associated with lower accuracy, contrary to the traditional AoA effect (Model 3 in Table 2).
Figure 1
Table 2
| Variable | Response Time | Accuracy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 1 | Model 2 | Model 3 | |||||||
| β | SE | β | SE | β | SE | β | SE | β | SE | β | SE | |
| (Intercept) | 0.000 | 0.011 | 0.000 | 0.011 | 0.000 | 0.010 | 0.000*** | 0.013 | 0.000*** | 0.013 | 0.000*** | 0.012 |
| N. of letter | −0.182*** | 0.027 | −0.065** | 0.025 | −0.094*** | 0.024 | 0.020 | 0.030 | −0.026 | 0.030 | 0.024 | 0.029 |
| N. of mora | 0.083*** | 0.013 | 0.036** | 0.012 | 0.060*** | 0.012 | 0.042** | 0.014 | 0.061*** | 0.014 | 0.032* | 0.015 |
| Frequency | −0.466*** | 0.011 | −0.394*** | 0.011 | −0.335*** | 0.012 | 0.290*** | 0.013 | 0.262*** | 0.013 | 0.207*** | 0.015 |
| OLD | 0.274*** | 0.026 | 0.208*** | 0.024 | 0.259*** | 0.024 | −0.048 | 0.029 | −0.022 | 0.029 | −0.117*** | 0.028 |
| AoA | 0.315*** | 0.011 | 0.046** | 0.016 | −0.123*** | 0.013 | 0.139*** | 0.019 | ||||
| ONS | −0.146*** | 0.012 | 0.079*** | 0.014 | ||||||||
| PNS | 0.010 | 0.012 | 0.020 | 0.014 | ||||||||
| Difficulty | 0.063*** | 0.014 | 0.020 | 0.016 | ||||||||
| Familiarity | −0.161*** | 0.013 | 0.247*** | 0.015 | ||||||||
| Imageability | −0.285*** | 0.019 | 0.265*** | 0.023 | ||||||||
| Semantic orientation | −0.061*** | 0.010 | 0.052*** | 0.012 | ||||||||
| Abstractness | 0.014 | 0.013 | −0.075*** | 0.015 | ||||||||
| BOI | 0.084*** | 0.014 | −0.077*** | 0.017 | ||||||||
| 0.272 | 0.362*** | 0.438*** | 0.086 | 0.099*** | 0.195*** | |||||||
| ΔR2 | 0.090 | 0.077 | 0.014 | 0.096 | ||||||||
Multiple regression analyses predicting lexical decision task performance.
p < 0.05,
p < 0.01,
p < 0.001.
AoA, age of acquisition; BOI, body-object interaction; LDT, lexical decision task; OLD, orthographic Levenshtein distance (Yarkoni et al., ); ONS, orthographic neighborhood size; PNS, phonetic neighborhood size; RT, response time. Each variable was referenced from the following studies: difficulty is from Kajiwara et al. (), Frequency is from University of Tsukuba et al. (), Familiarity is from Fujita and Kobayashi (), Semantic orientation is from Takamura et al. (), Imageability is from Sakuma et al. (), Abstractness is from The Social Computing Laboratory and Nara Institute of Science and Technology (), and BOI is from Mochizuki and Ota (). The response time and accuracy data for the LDT for Japanese words were obtained from Ota and Mochizuki ().
4 Discussion
In this study, we collected AoA ratings for 5,736 Japanese words. The results suggest that the AoA ratings have a certain degree of validity and reliability, and that these are a valuable research resource for studies using Japanese words. This is the largest AoA norm available for Japanese. The strength of this dataset is that it enabled us to examine the relationship between the other 13 variables and the two cognitive performances. The AoA is known to be correlated with frequency, imageability, and concreteness. All items in this dataset were cross-referenced with these variables, allowing us to select stimuli based on the characteristics of multiple variables and examine each variable's influence.
Notably, the correlation between the AoA ratings of other languages and the AoA ratings of this study was relatively small. The English translations of the items collected in the present study were assigned by the authors based on what is considered to be the most common or representative meanings. However, word meanings are not necessarily singular, and thus the corresponding items across datasets or languages may not always share identical meanings. This semantic discrepancy could have contributed to the relatively low correlation coefficients. Another potential explanation is that the AoA ratings may have been influenced by the orthographic form of each item. In general, Japanese speakers first learn hiragana and katakana, followed by kanji (Tsukada, ). Consequently, words that contain more hiragana and katakana characters may be rated as acquired earlier, whereas words with more kanji characters may be rated as acquired later. To investigate this, we calculated the proportion of kanji, hiragana, and katakana characters in each target item and examined their correlation with the AoA ratings. The results showed that the higher the proportion of kanji, the higher the AoA rating (r = 0.224, p < 0.001); the higher the proportion of hiragana and katakana, the lower the AoA rating (r = −0.311, p < 0.001; r = −0.053, p < 0.001). Although this was a preliminary analysis, the AoA ratings in the present study appeared to be influenced by the orthographic form in which the items were presented. This may partially explain the relatively low correlation between the AoA ratings in this study and those reported in alphabetic cultures.
The multiple regression analyses using models adopted in previous studies revealed the AoA effect: words with lower AoA ratings were associated with shorter response times and higher accuracy, consistent with findings from other languages. This suggests that AoA plays an important role in word recognition in Japanese as well. It is noteworthy, however, that in the model including multiple available psycholinguistic variables, AoA was a significant predictor of accuracy, but in the opposite direction from the typical AoA effect: higher AoA ratings were associated with higher accuracy (see Model 3 in Table 2). This unexpected finding suggests that, for Japanese words, the effect of AoA on accuracy may be confounded with other variables such as imageability, familiarity, and body–object interaction. Presently, we cannot posit any strong explanation for this result. One possible reason may be the relationship between orthographic forms and AoA. Specifically, as mentioned above, words with higher AoA ratings are more likely to include kanji characters. High AoA words may be more difficult (as shown in Figure 1), leading to greater processing demands; as a result, they may be processed more carefully, yielding longer reaction times but higher accuracy. Because the present study did not include pairs of words that differ only in orthographic form (i.e., no identical-meaning words with different script type), we were unable to directly test this possibility. However, Havelka and Tomita (), who demonstrated AoA effects experimentally in Japanese, found that AoA effects on reaction time were larger for words presented in kanji compared to that of the same words presented in kana. The influence of variables such as word length, frequency, and familiarity is known to differ between the processing of kanji and katakana words (Kusunose and Hino, ; Kusunose et al., ). Future research should further examine how AoA and orthographic forms interactively influence lexical processing, particularly with respect to accuracy.
Additionally, a further examination of the validity and reliability of AoA ratings is crucial. The current study employed a slightly different methodology to collect subjective AoA ratings in distinct categories, compared with the research of Cortese and Khanna () and Stadthagen-Gonzalez and Davis (). Therefore, it is important to note that directly comparing the ratings from their respective studies with the average AoA ratings from the present study is not appropriate. It has been pointed out that the AoA should not be rated using the Likert method, but rather by the age at which the target item was acquired (Kuperman et al., ). AoA ratings have been shown to differ depending on the individual's relationship with the child (Wikse Barrow et al., ) and on how they are asked, such as by rating the age acquired by the respondent themselves vs. the age acquired by the child in general (Łuniewska et al., ). Another effective approach is to estimate objective AoA based on materials included in textbooks and examining its relationship with relevant variables (Cai et al., ). Future research will not only examine vocabulary and concept acquisition in Japanese speakers using this dataset but also explore the development of more reliable and validated AoA ratings.
Furthermore, investigating the relationship of AoA with tasks such as picture naming and word naming, in which AoA effects have been well-documented, is also necessary. In the context of Japanese, the construction of datasets for psycholinguistic studies has not progressed sufficiently. To validate the subjective ratings of semantic variables, constructing databases related to behavioral outcomes, such as cognitive performance and written language data, will be necessary.
Statements
Data availability statement
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 at: https://osf.io/fawmq/.
Ethics statement
The studies involving humans were approved by the Research Ethics Committee of the College of Humanities and Sciences at Nihon University (Approval No.: 06-48). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.
Author contributions
MM: Methodology, Conceptualization, Resources, Writing – review & editing, Project administration, Investigation, Data curation, Software, Funding acquisition, Writing – original draft, Formal analysis, Visualization. NO: Conceptualization, Writing – review & editing.
Funding
The author(s) declare that financial support was received for the research and/or publication of this article. This study was supported by JSPS KAKENHI Grant Number JP24K15684.
Acknowledgments
We would like to express our gratitude to all the participants who participated in the survey.
Conflict of interest
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.
Generative AI statement
The author(s) declare that no Gen AI was used in the creation of this manuscript.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Footnotes
1.^Note that these categories slightly differ from that used by Cortese and Khanna () and Stadthagen-Gonzalez and Davis ().
2.^These items were collected in order to analyze them together with data collected in a different study.
References
1
AlonsoM. A.FernandezA.DíezE. (2015). Subjective age-of-acquisition norms for 7,039 Spanish words. Behav. Res. Methods47, 268–274. 10.3758/s13428-014-0454-2
2
AmanoS.KondoK. (2000). NTT Database Series. Lexical Properties of Japanese, No. 7. Frequency.Tokyo: Sanseido.
3
AsaharaM. (2020). Word familiarity rate and register type estimation using a Bayesian linear mixed model. J. Nat. Lang. Process. 27, 133–150. 10.5715/jnlp.27.133
4
BalotaD. A.YapM. J.HutchisonK. A.CorteseM. J. (2012). “Megastudies: what do millions (or so) of trials tell us about lexical processing,” in Visual Word Recognition: Models and Methods, Orthography and Phonology, ed. AdelmanJ. S. (London: Psychology Press), 90–115.
5
BirchenoughJ. M. H.DaviesR.ConnellyV. (2017). Rated age-of-acquisition norms for over 3,200 German words. Behav. Res. Methods49, 484–501. 10.3758/s13428-016-0718-0
6
BrysbaertM.BiemillerA. (2017). Test-based age-of-acquisition norms for 44 thousand English word meanings. Behav. Res. Methods49, 1520–1523. 10.3758/s13428-016-0811-4
7
BrysbaertM.StevensM.De DeyneS.VoorspoelsW.StormsG. (2014). Norms of age of acquisition and concreteness for 30,000 Dutch words. Acta Psychol. 150, 80–84. 10.1016/j.actpsy.2014.04.010
8
BrysbaertM.Van WijnendaeleI.De DeyneS. (2000). Age-of-acquisition effects in semantic processing tasks. Acta Psychol. 104, 215–226. 10.1016/S0001-6918(00)00021-4
9
CaiZ. G.HuangS.XuZ.ZhaoN. (2022). Objective ages of acquisition for 3300+ simplified Chinese characters. Behav. Res. Methods54, 311–323. 10.3758/s13428-021-01626-1
10
CameirãoM. L.VicenteS. G. (2010). Age-of-acquisition norms for a set of 1,749 Portuguese words. Behav. Res. Methods42, 474–480. 10.3758/BRM.42.2.474
11
CorteseM. J.KhannaM. M. (2008). Age of acquisition ratings for 3,000 monosyllabic words. Behav. Res. Methods40, 791–794. 10.3758/BRM.40.3.791
12
DiveicaV.PexmanP. M.BinneyR. J. (2023). Quantifying social semantics: an inclusive definition of socialness and ratings for 8388 English words. Behav. Res. Methods55, 461–473. 10.3758/s13428-022-01810-x
13
ElsherifM. M.PreeceE.CatlingJ. C. (2023). Age-of-acquisition effects: a literature review. J. Exp. Psychol. Learn. Mem. Cogn. 49, 812–847. 10.1037/xlm0001215
14
FerrandL.BoninP.MéotA.AugustinovaM.NewB.PallierC.et al. (2008). Age-of-acquisition and subjective frequency estimates for all generally known monosyllabic French words and their relation with other psycholinguistic variables. Behav. Res. Methods40, 1049–1054. 10.3758/BRM.40.4.1049
15
FrankM. C.BraginskyM.YurovskyD.MarchmanV. A. (2017). Wordbank: an open repository for developmental vocabulary data. J. Child Lang.44, 677–694. 10.1017/S0305000916000209
16
FujimotoM.DaiboI. (2007). ENDCORE: a hierarchical structure theory of communication skills. Jpn. J. Pers. 15, 347–361. 10.2132/personality.15.347
17
FujitaS.KobayashiT. (2020). “Resurvey of word familiarity and comparison with past data,” in Proceedings of the Twenty-Sixth Annual Meeting of the Association for Natural Language Processing (Kyoto: Association for Natural Language Processing), 1037–1040.
18
FukuiH.AokiS. (2022). The development of the Japanese version of the Plymouth Sensory Imagery Questionnaire (Psi-Q). Jpn. J. Pers. 31, 87–99. 10.2132/personality.31.2.2
19
HavelkaJ.TomitaI. (2006). Age of acquisition in naming Japanese words. Vis. Cogn. 13, 981–991. 10.1080/13506280544000156
20
KajiwaraT.NishiharaD.KodairaT.KomachiM. (2020). Language resources for Japanese lexical simplification. J. Nat. Lang. Process. 27, 801–824. 10.5715/jnlp.27.801
21
KupermanV.Stadthagen-GonzalezH.BrysbaertM. (2012). Age-of-acquisition ratings for 30,000 English words. Behav. Res. Methods44, 978–990. 10.3758/s13428-012-0210-4
22
KusunoseY.HinoY. (2017). “Differences in word frequency and orthographic familiarity effects for Japanese Kanji versus Katakana words,” in Proceedings of the 81st Annual Convention of the Japanese Psychological Association (Tokyo: Japanese Psychological Association), 684. 10.4992/pacjpa.81.0_2C-060
23
KusunoseY.YoshiharaM.IdaXue, K.IjuinJ.HinoM. Y. (2014). Word length effects for kana and kanji words in lexical decision tasks. Jpn. J. Cogn. Psychol. 11, 105–115. 10.5265/jcogpsy.11.105
24
ŁuniewskaM.HamanE.Armon-LotemS.EtenkowskiB.SouthwoodF.AndelkovićD.et al. (2016). Ratings of age of acquisition of 299 words across 25 languages: is there a cross-linguistic order of words?Behav. Res. Methods48, 1154–1177. 10.3758/s13428-015-0636-6
25
ManiaciM. R.RoggeR. D. (2014). Caring about carelessness: participant inattention and its effects on research. J. Res. Pers. 48, 61–83. 10.1016/j.jrp.2013.09.008
26
MochizukiM.OtaN. (2024). “Collection of body-object interaction ratings for 5,736 Japanese words,” in Proceedings of the 21st Conference of the Japanese Society for Cognitive Psychology (Fukuoka: Japanese Society for Cognitive Psychology), 4–10. 10.14875/cogpsy.2024.0_96
27
MontefineseM.VinsonD.ViglioccoG.AmbrosiniE. (2019). Italian age of acquisition norms for a large set of words (ItAoA). Front. Psychol. 10:278. 10.3389/fpsyg.2019.00278
28
MorrisonC. M.EllisA. W. (1995). Roles of word frequency and age of acquisition in word naming and lexical decision. J. Exp. Psychol. Learn. Mem. Cogn. 21, 116–133. 10.1037/0278-7393.21.1.116
29
NishimotoT.MiyawakiK.UedaT.UneY.TakahashiM. (2005). Japanese normative set of 359 pictures. Behav. Res. Methods37, 398–416. 10.3758/BF03192709
30
OtaN.MochizukiM. (2025). JALEX: Japanese version of lexical decision database. Front. Lang. Sci. 3:1506509. 10.3389/flang.2024.1506509
31
PexmanP. M.MurakiE.SidhuD. M.SiakalukP. D.YapM. J. (2019). Quantifying sensorimotor experience: body-object interaction ratings for more than 9,000 English words. Behav. Res. Methods51, 453–466. 10.3758/s13428-018-1171-z
32
RochfordG.WilliamsM. (1962). Studies in the development and breakdown of the use of names: I. The relationship between nominal dysphasia and the acquisition of vocabulary in childhood. J. Neurology, Neurosurg. Psychiatry25, 222–227.
33
SakumaN.IjuinM.FushimiT.TastumiI.TanakaM.AmanoS.et al. (2005). NTT Database Series. Lexical Properties of Japanese, No. 8. Imageability. Tokyo: Sanseido.
34
Stadthagen-GonzalezH.DavisC. J. (2006). The Bristol norms for age of acquisition, imageability, and familiarity. Behav. Res. Methods38, 598–605. 10.3758/BF03193891
35
TakamuraH.InuiT.OkumuraM. (2005). “Extracting semantic orientations of words using spin model,” in Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics (ACL'05), eds. KnightK.NgH. T.OflazerK. (Ann Arbor, MI: Association for Computational Linguistics), 133–140. 10.3115/1219840.1219857
36
The Social Computing Laboratory and Nara Institute of Science and Technology (2021). Data from: AWD-J: Abstractness of Word Database for Japanese Common Words. Available online at: https://sociocom.naist.jp/awd-j/ (accessed January 1, 2025).
37
TsukadaY. (2007). A study of invented spelling and developing orthographic concepts in Japanese. L1 Edu. Stud. Lang. Lit. 7, 5–29. 10.17239/L1ESLL-2007.07.03.11
38
TurnerJ. E.ValentineT.EllisA. W. (1998). Contrasting effects of age of acquisition and word frequency on auditory and visual lexical decision. Mem. Cognit. 26, 1282–1291. 10.3758/BF03201200
39
University of Tsukuba, National Institute for Japanese Language and Linguistics, Lago Institute of Language. (2013). NINJAL-LWP for TWC. University of Tsukuba, National Institute for Japanese Language and Linguistics, and Lago Institute of Language. Available online at: http://corpus.tsukuba.ac.jp (accessed January 1, 2025).
40
Wikse BarrowC.Nilsson BjörkenstamK.StrömbergssonS. (2019). Subjective ratings of age-of-acquisition: exploring issues of validity and rater reliability. J. Child Lang.46, 199–213. 10.1017/S0305000918000363
41
XuX.LiJ.GuoS. (2021). Age of acquisition ratings for 19,716 simplified Chinese words. Behav. Res. Methods53, 558–573. 10.3758/s13428-020-01455-8
42
YamazakiM.EllisA. W.MorrisonC. M.Lambon RalphM. A. (1997). Two age of acquisition effects in the reading of Japanese Kanji. Br. J. Psychol. 88, 407–421. 10.1111/j.2044-8295.1997.tb02648.x
43
YarkoniT.BalotaD.YapM. (2008). Moving beyond Coltheart's N: a new measure of orthographic similarity. Psychon. Bull. Rev. 15, 971–979. 10.3758/PBR.15.5.971
44
ZevinJ. D.SeidenbergM. S. (2004). Age-of-acquisition effects in reading aloud: tests of cumulative frequency and frequency trajectory. Mem. Cognit. 32, 31–38. 10.3758/BF03195818
Summary
Keywords
age of acquisition, Japanese, lexicon, word norms, word recognition
Citation
Mochizuki M and Ota N (2025) Collection of age of acquisition ratings for over 5,000 Japanese words. Front. Lang. Sci. 4:1605224. doi: 10.3389/flang.2025.1605224
Received
03 April 2025
Accepted
14 May 2025
Published
10 June 2025
Volume
4 - 2025
Edited by
Zhenguang Cai, The Chinese University of Hong Kong, China
Reviewed by
Mikihiro Tanaka, Ritsumeikan University, Japan
Zebo Xu, The Chinese University of Hong Kong, China
Updates
Copyright
© 2025 Mochizuki and Ota.
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.
*Correspondence: Masaya Mochizuki mochizuki.masaya@nihon-u.ac.jp
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.