Age of acquisition, not word frequency, affects object naming, not object recognition

Abstract

Word frequency is widely believed to affect object naming speed, despite several studies in which it has been reported that frequency effects may be redundant upon age of acquisition. We report, first, a reanalysis of data from the study by Oldfield and Wingfield (1965), which is standardly cited as evidence for a word frequency effect in object naming; then we report two new experiments. The reanalysis of Oldfield and Wingfield shows that age of acquisition is the major determinant of naming speed, and that frequency plays no independent role when its correlation with other variables is taken into account. In Experiment 1, age of acquisition and phoneme length proved to be the primary determinants of object naming speed. Frequency, prototypicality, and imageability had no independent effect. In Experiment 2, subjects classified objects into two semantic categories (natural or man-made). Prototypicality and semantic category were the only variables to have a significant effect on reaction time, with no effect of age of acquisition, frequency, imageability, or word length. We conclude that age of acquisition, not word frequency, affects the retrieval and/or execution of object names, not the process of object recognition. The locus of this effect is discussed, along with the possibility that words learned in early childhood may be more resistant to the effects of brain injury in at least some adult aphasics than words learned somewhat later.Treatments of word retrieval in general, and object naming in particular, commonly propose that word frequency is an important determinant of the speed and accuracy of lexical access in speech production (e.g., Monsell, Doyle, & Haggard, 1989;. It is widely accepted that the less frequent a word is in the English language, the slower one will be to name a pictorial representation of the object. The standard reference given in support of this assertion is Oldfield and Wingfield (1965), who reported an experiment in which 12 subjects named 26 line drawings of different objects. Oldfield and Wingfield found an inverse relationship between naming latency and frequency, such that the lower the frequency of a particular item (i.e., the less common its occurrence in English), the longer it took to name it. There are a number of reasons to be cautious about accepting the results of that study.First, Oldfield and Wingfield (1965) failed to consider variables that correlate highly with frequency, such as length, imageability, and the age at which a word is acquired (age of acquisition). Second, they used the Thorndike-Lorge (1944) frequency count. This is an American word count drawn from school books and other similar reading material, and it may not have been a good index of the true frequency of word use by Oldfield and Wingfield's subjects (9 of the 12 of whom were under-

Age of acquisition, word frequency, and the role of phonology in the lexical decision task

“…Logie (1981a, 1981b) found no evidence for an AoA effect on visual or auditory word recognition thresholds, which suggests that observed AoA effects in naming and lexical decision must result from processes operating after lexical access. Morrison et al (1992) found that AoA affected picture naming latencies, but not performance on a living versus a nonliving categorization task on the same pictures, which suggests that the AoA effect in naming arises from postsemantic processes. The most likely locus of the AoA effect is, therefore, at the stage of retrieval of lexical phonology.…”

“…Currently, there is no firm evidence of AoA effects in any lexical processing task that does not involve the retrieval of phonology. Admittedly, there have been only few attempts to look for possible AoA effects in tasks in which phonological variables have no major influence; for example, Morrison et al (1992) found no effect of AoA in the semantic categorization ofpictures, even when a clear AoA effect on the naming latencies to the same pictures was observed. However,additional studies ofthe possible effects of AoA in other lexical processing tasks will be required in order to investigate this further.…”

Comparing Electrophysiological Correlates of Word Production in Immediate and Delayed Naming Through the Analysis of Word Age of Acquisition Effects

“…As explained in the Introduction, two possible loci of AoA effects in picture naming have been suggested: lexical-semantic (Belke et al, 2005;Johnson & Barry, 2005) and lexical phonological encoding processes (Chalard & Bonin, 2006;Morrison and Ellis, 1995;Morrison et al, 1992). Different amplitudes between early-and late-acquired words were observed in four distinct time periods.…”

“…However, the reliably of lexical frequency effects has been challenged since effects of word AoA have also been reported (e.g., Bonin, Fayol, & Chalard, 2001;Bonin, Chalard, Méot, & Fayol, 2002;Chalard, Bonin, Méot, Boyer, & Fayol, 2003;Morrison, Ellis, & Quinlam, 1992;Morrison & Ellis, 1995). Especially, when those variables are operationalised following the recommendations made by Zevin and Seidenberg (2002) 1 , reliable AoA effects were reported on conceptually driven spoken production (as in picture naming paradigms, Bonin et al, 2004).…”

“…This means that early-acquired word forms are encoded faster than late-acquired ones. This 6 hypothesis is based on the fact that AoA effects were observed neither in lexical-semantic tasks (Chalard & Bonin, 2006;Morrison et al, 1992) nor in delayed production tasks (Morrison & Ellis, 1995), therefore excluding lexical-semantic encoding levels or motor processes. However, AoA effects have been recently reported with semantic tasks (Belke, Brysbaert, Meyer, & Ghyselinck, 2005;Johnson & Barry, 2005), supporting the alternative hypothesis of a lexical-semantic locus of AoA effects.…”