|
Article Excerpt
Conventionally, object naming is separated into three distinct levels: a perceptual level, a semantic level and a lexical level. There are a number of variables that affect the speed at which a picture of an object progresses through these stages of processing. At the level of perceptual recognition, these include image agreement (Barry, Morrison, & Ellis, 1997), visual complexity (Ellis & Morrison, 1998), and structural similarity (Humphreys, Riddoch, & Quinlan, 1988). At the semantic level these include imageability (Strain, Patterson, & Seidenberg, 1995) and at the level of lexical access, frequency (e.g. Humphreys et al., 1988; Jescheniak & Levelt, 1994; Oldfield & Wingfield, 1965). However, a further variable, age of acquisition (AoA) is also believed to have an impact on the process of identifying objects although the precise locus of its effect is still uncertain. The effect of AoA within the levels of processing involved in object perception, comprehension and naming are addressed within this study.
The time required to name objects whose name was acquired early in life is shorter than the time required to name objects acquired later in life (e.g. Barry, Hirsh, Johnston, & Williams, 2001; Barry et al., 1997; Bonin, Chalard, Meot, & Fayol, 2002; Bonin, Fayol, & Chalard, 2001; Brysbaert, Lange, & Van Wijnendaele, 2000; Cuetos, Ellis, & Alvarez, 1999; Gerhand & Barry, 1988, 1999; Gilhooly & Gilhooly, 1979; Lachman, 1973; Lachman, Schaffer, & Henrikus, 1974; Morrison, Chappell, & Ellis, 1997; Morrison, Hirsh, Chappell, & Ellis, 2002). The AoA effect also appears to be present throughout a number of tasks that involve; face naming (Moore & Valentine, 1998) and face familiarity classifications (e.g. Moore & Valentine, 1999), word naming (e.g. Brown & Watson, 1987; Ellis & Morrison, 1998; Gerhand & Barry, 1988; Monaghan & Ellis, 2002; Morrison et al., 2002; Yamazaki, Ellis, Morrison, & Lambon Ralph, 1997) auditory word recognition (e.g. Gilhooly & Logie, 1981; Turner, Valentine, & Ellis, 1998) and name generation (Brysbaert, Van Wijnendaele, & De Deyne, 2000; Catling & Johnston, 2005).
The level of lexical phonological representations has often been proposed as the locus for AoA effects. Both Gilhooly and Watson (1981) and Brown and Watson (1987) suggested that the effect of AoA reflects a developmental stage in language acquisition. Brown and Watson proposed that early acquired words are represented in a more 'complete' form in a phonological output lexicon than late acquired words. Consequently, early acquired words are produced more rapidly throughout later life. More recent research findings have called into question the validity of this account. Monaghan and Ellis (2002) tested the phonological completeness account directly. They predicted that if early acquired words are stored in a more complete form, then adult participants should be slower to segment early acquired words compared with later acquired words. However, the results of their segmentation task failed to provide support for this hypothesis, the strong implication being that AoA does not affect the phonological level of processing.
Other research has also found evidence that is not compatible with explanations of AoA that relate solely to the phonological level of representation (e.g. Moore et al., 2004; Moore & Valentine, 1999). Several studies have demonstrated an impact of AoA on tasks that do not seem to require access to names. Brysbaert, Van Wijnendaele and De Deyne (2000) found AoA effects within a semantic categorization task that involved classifying words between those with a 'definable meaning' and 'first names'. Performance in a perceptual task where participants decide whether pictures of stimuli represent real or chimeric unreal objects has also been shown to be influenced by AoA (e.g. Moore et al., 2004; Vitkovitch & Tyrrell, 1995). AoA has also been shown to affect face categorization (Lewis, 1999, but see Moore, Valentine, & Turner, 1999 for a counter argument), face and name familiarity decisions (Moore & Valentine, 1999), and lexical decision tasks (Morrison & Ellis, 1995). These findings suggest that AoA does not act only at the level of phonological representations. This has led to the proposal that the AoA locus of effect is at the level of semantic representation (Van Loon-Vervoon, 1989, cited in Brysbaert, Van Wijnendaele, & De Deyne, 2000; Johnston & Barry, 2005). However, not all the evidence is supportive of AoA working at this level. Izura and Ellis (2002) demonstrated distinct AoA effects for both first and second languages within a translation task. They suggest that if first and second language equivalent words share semantic representations, then the AoA effect of the second language should draw upon the same semantic representations as the first language and therefore show the same characteristics as that of the first language. However, they found that the second language had its own distinct AoA effect that was different from the first language and assert that this demonstrates that the AoA effect cannot be solely located at the level of semantic representation.
Another possibility is that AoA has more than one locus of effect. Moore and Valentine (1998, 1999) and Moore et al. (2004) suggest that the effects of AoA should be found at multiple loci. This would account for the effects of AoA within motor output tasks (as found in verbal and written naming) and perceptual input tasks. Moore and Valentine (1999) established the perceptual effects of AoA by demonstrating AoA effects within a familiarity decision to 50 famous and 50 non-famous people. They suggested that AoA must have a pre-phonological role because of evidence that names are not automatically accessed when familiarity decisions are made to faces.
This multiloci account is supported by the findings of Yamazaki et al. (1997). They used a multiple regression analysis to assess the impact of various variables on the reading time of single-character Kanji words by native Japanese speakers. Interestingly, they found that two measures of AoA exerted significant independent effects upon reading speed; the age at which the words enter a Japanese child's spoken vocabulary and the age at which Japanese children learn the characters that are used to write those words. They suggest that 'the effect of age of spoken acquisition reflects the contribution of the speech output lexicon to Kanji reading, and that the age of written acquisition effect reflects the contribution of the visual input lexicon' (p. 419).
However, AoA effects may not have a localized effect at all; instead, they may be due to an underlying mechanism. Lewis (1999) suggests that the total number of times a word has been encountered can account for the effects of AoA. He claims that the effect of AoA can be demonstrated mathematically by applying the cumulative frequency to curvilinear learning. Using a non-linear equation, he goes on to explain the AoA and frequency effects found in both Gerhand and Barry's (1988) and Carroll and White's (1973) experiments. The cumulative frequency hypothesis would predict that differences between early and late acquired words will be less apparent in older than younger adults. However, Morrison et al. (2002) found that AoA effects were apparent in both old and young adults. This lack of interaction between age and AoA is incompatible with the cumulative frequency theory, which would predict that the impact of AoA would diminish with chronological age.
An adaptation of the cumulative frequency hypothesis comes from Zevin and Seidenberg (2002). They suggest that AoA effects on reading are but manifestations of the weaknesses found within the empirical research. Their main complaint against previous AoA research in reading is that the ratings used to balance important factors such as frequency and familiarity can be inaccurate, suggesting that some ratings of frequency in particular come from corpuses that are relatively small. They state that their preferred ratings are from the Celex database (Baayen, Piepenbrock, & Van Rijn, 1993) or the Educator's word frequency guide (Zeno, Ivens, Millard, & Duvvuri,...
|
