The Role of Context on Age of Acquisition Effect: Strategic Control in Word Naming in Turkish

Introduction

It has long been acknowledged that readers can attune their reading strategies in response to task demands as determined by context (e.g. Frederiksen & Kroll, 1976; Baluch & Besner, 1991; Paap & Noel, 1991; Rastle & Coltheart, 1999). An example of this flexibility in selecting the most effective strategy is observed when the effect of word frequency is eliminated when subjects name words and nonwords mixed together in a single block. Historically, this phenomenon reported in English and other orthographies had been explained within the dual-route model of reading (e.g., Baluch & Besner, 1991; Coltheart & Rastle, 1994; Monsell, Patterson, Graham, Hughes, & Milroy, 1992; Raman, Baluch & Sneddon, 1996; Rastle & Coltheart, 1999; Reynolds & Besner, 2005; Tabossi & Laghi, 1992; Zevin & Balota, 2000). According to the dual-route model generation of phonology can take place via two qualitatively distinct routes: namely the lexical and the nonlexical route (e.g. Coltheart, 1978; Coltheart and colleagues, 1993; 1999). What characterizes these two routes is that while the nonlexical phonology can be ‘assembled’ via rules, assumptions about generating phonology via the lexical route is twofold: One way to generate lexical phonology is assumed to be via the direct orthography-to-phonology, OP, route where words’ phonology is directly ‘addressed’. A second way of generating lexical phonology is assumed to be via the orthography-to-semantics route where a word’s meaning is activated for the purpose of generating phonology. Some dual-route theorists have argued that the dual-route model is in effect a three-route model, whilst, it is generally assumed that the impact of the semantic route on single-word naming in skilled reading is minimal (e.g. Besner, 1999; Besner & Smith, 1992). This is because the general consensus within the dual-route framework (in terms of RTs) is that attaining phonology from print via the semantic route is the slowest of the two routes. It is further assumed that the involvement of the semantic route in computing words’ phonology is only facilitated when words’ semantic characteristics such as imageability is involved. However, it must be highlighted that systematic investigation of imageability effects in single-word naming is not a widely explored issue in English with the exception of several papers (Strain, Patterson, & Seidenberg, 1995; Hino & Lupker, 1996) and to date just a handful of papers have been reported on other writing systems, e.g. Persian (Baluch & Besner, 2001) and Turkish (Raman & Baluch, 2001). In summary, semantics is assumed to contribute to the computation of phonology from print in orthographies with inconsistent and/or irregular OP representations, such as English and opaque Persian, but not in orthographies with entirely consistent OP representations such as Turkish. It would be naïve to assume that there is no semantic involvement in reading entirely transparent orthographies as semantic information ought to be utilized in order to extract meaning during reading. When OP mappings are entirely transparent, however, the input from semantics in decoding OP mappings that are exception to the rule becomes redundant. Therefore, the claim here is that semantics develops and exerts itself differentially as a function of orthographic transparency across different languages. Indeed, evidence for this claim was presented in Turkish (Raman & Baluch, 2001) and Persian (Baluch & Besner, 2001).