Age of acquisition effects on the decomposition of compound words

ABSTRACT

Age of acquisition (AoA) is a measure of learning experience and a strong predictor of lexical retrieval. The integrated account predicts that the AoA effect should be shown in the early processes of word recognition and the AoA effect should increase in tasks requiring greater semantic processing. The present study investigates the integrated account in compound words, which differ from monomorphemic words regarding ease of mapping and semantic processes in lexical retrieval. Forty-eight participants completed a compound lexeme segmentation (CLS) task, in which participants named either the head or modifier depending on the number above the compound word, to establish how semantics are involved in processing the head and the modifier. The results demonstrated that semantics influenced the naming of the modifier to a greater extent than the head, with the AoA effect being larger in the modifier than the head. This indicates that the AoA effect has multiple origins.

KEYWORDS:

• Age of acquisition
• compound lexeme segmentation
• compound word
• lexical retrieval
• morphology
• semantics

Data availability statement

The data and materials for all experiments are available at https://osf.io/g5b3m/.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Like English, compound words in German are right-headed.

2 Westfall et al. (2014) published a theoretical analysis of a mixed effects model and created a website (https://jakewestfall.shinyapps.io/two_factor_power/), which allows researchers to calculate the power of an experiment and the number of items/participants required for a well-powered experiment. We estimated the sample size using the following values: Effect size d: 0.063, residual variance: 0.61, participant intercept variance: 0.23, item intercept variance: 0.16, participant-by-item variance: 0, participant slope variance: 0, item slope variance: 0.

3 The patterns of the results did not change when alveolar was included or excluded as a fixed predictor.

4 Although the head was more frequent, early-acquired and shorter in letter length than the modifier of the compound word, suggesting that the former should be processed more quickly than the latter, the opposite was observed.

5 The different patterns that emerge for word frequency of and familiarity of the compound word may occur, as they do overlap in terms of being frequency measures. However, the database used in this study could drive the effect. The word frequency measure was extracted from SUBTLEX-UK database, while we extracted the familiarity of the compound word from Juhasz et al.’s (2015) database. Juhasz et al. asked participants to rate the familiarity of the compound word based on its meaning and frequency, thus the familiarity measure may not only be affected by subjective frequency but also meaning, while the word frequency is a measure of objective frequency. In addition, the correlation between these two measures was weak to moderate (r = . 33), thus supporting the argument that they may share a frequency-component but perhaps differ in terms of semantics, leading to the different pattern of findings.