Ch.+24+by+Judith+Kroll+&+Ton+Dijkstra


 * CHAPTER 24 **

 **THE BILINGUAL LEXICON**

 Judith F. Kroll

 Ton Dijkstra

 This chapter reviews the way in which models of the bilingual lexicon reflect different assumptions about the formation and processing of words in two languages and considers three main questions about the lexical access:

 1. What codes are activated?

 2. When are those codes activated?

 3. What are the critical factors that affect lexical selection?

 **//MODELS OF THE BILINGUAL LEXICON//**

 **The Revised Hierarchical Model (RHM)**

 The initial attempts to model the bilingual lexicon proposed a hierarchical arrangement to represent word form and meaning (e.g. Potter, Von Eckhardt and Feldman 1984). At the level of word form they proposed an independent lexical representation for each language, but at the level of meaning they assumed a single conceptual system. The connection between the lexical form in L1 and L2 might be active early in L2 acquisition, but by the time the bilingual becomes proficient in L2, words in each language can access concepts directly.

 Kroll and Stewart (1994) proposed the //revised hierarchical model// (RHM). They argued that in early acquisition the lexical level connections between words in the two languages provided a means for transfer, L1 could provide meaning for an L2 word if L2 had an equivalent translation.

 The RHM assumed independent lexical representation for words in each language.

 **The Bilingual Interactive Model (BIA)**

 The BIA model accounts for asymmetries observed in unbalanced bilinguals (stronger effects from L1 on L2 than vice versa) by assuming that relative to L1 words the frequency of L2 words is lower for participants with lower L2 proficiency.

 **//COMPREHENSION//**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> What Codes Are Activated?

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> According to Kroll and Dijkstra (1999), lexical decision is one of the most commonly used tasks in monolingual and bilingual word recognition. Presentation of a word to monolinguals induces activation not only of orthographic codes but of phonological and semantic codes likewise (Frost, 1998; James, 1975). Besides, monolingual studies that involve vague words suggest that different meanings of these words are initially activated during recognition. Thus, during monolingual word recognition there is a so-called //intralingual coactivation// of lexical candidates with overlap in meaning or form.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> The authors assume that the nonselective access view suggest that it should not matter whether the coactivated lexical candidates belong to the same language or not. Thus, it can be inferred that there will be //interlingual activation// during bilingual word recognition as well. In contrast, the language-selective access view states that a presented word will activate the form and meaning representations only from the language that is selected at that time.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **When Are These Codes Activated?**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> Under particular experimental circumstances, form and meaning representations of lexical candidates belonging to different languages become activated and may have an influence on the patterns of results.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> A recent study by Dijkstra, Timmermans and Schriefers (2000) proposes that coactivation of lexical candidates from dissimilar languages occurs until comparatively late in the word recognition process. The study assumes that low proficiency bilinguals might show relatively weak effects from their L2 on their L1 lexicon (though the vice versa effect is quite strong). The aforementioned points lead to considering and pondering over the critical factors that may affect the selection of lexical candidates during the bilingual word recognition process.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **Critical Factors That Affect Lexical Selection**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> A number of influential factors have been distinguished by Grosjean (1998), such as //L2-proficiency, language intermixing, task demands,// and //instruction//.

<span style="font-family: 'Arial','sans-serif'; font-size: 18.6667px;">- //<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">Language intermixing //<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> refers to whether an experiment contains items that belong only to one language (mixed representation). <span style="font-family: 'Arial','sans-serif'; font-size: 18.6667px;">- //<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">Effects of instruction //<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> assumed that bottom-up factors, such as language intermixing and stimulus characteristics (frequency, code similarity), are the more significant ones.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> All in all, bilingual comprehension studies suggest that:

<span style="font-family: 'Arial','sans-serif'; font-size: 18.6667px;">- <span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">Lexical codes from dissimilar languages are activated together with the input. <span style="font-family: 'Arial','sans-serif'; font-size: 18.6667px;">- <span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">Selection of lexical candidates takes place rather late in the recognition process. <span style="font-family: 'Arial','sans-serif'; font-size: 18.6667px;">- <span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">Several factors influence the final result patterns, such as L2 proficiency level, the requirements of the task, the mixed/blocked presentation of items from different languages.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **PRODUCTION**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **What Codes Are Activated?**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">When a picture is showed or a single word is needed to be expressed a chain of processes is needed for speaking the word. This is a complex process in itself, even when the speaker speaks a single language. So for bilinguals it becomes more complex process, because they must chose the language in which they want to speak the word. A surprising fact, during the process of expressing the meaning, the word in two languages of bilingual are activated.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">Number of studies used the picture-word interference to investigate the form of the codes active during lexical access. Studies show that lexical alternatives are available in both languages, at least during the early stages of production, when abstract lexical representations corresponding to the intended meaning of utterance are activated.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **When Are These Codes Activated?**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">There was a study held by Kroll and Peck (1998), which investigated the time course of English-Spanish bilinguals to speak the name of the picture. As the L1 is the dominant and first language, the time of picture naming was faster for L1 than L2. Sometimes it is also possible the activation of both languages simultaneously. The author assumes that if L2 picture naming normally involves the activation and subsequent inhibition of L2, then mixing of L1 and L2 will have little effect on L2 performance. So when the picture is named in L1, there is no influence of L2. But when the word is named in L2, there is certainly a L1 influence.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **Inhibitory control**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> According to Green inhibitory control is a mechanism which allows a bilingual to complete the activity effectively in the nontarget language and avoid speaking in that language.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> There have been a lot of experiments on inhibitory control mechanism. Most of the experiments were conducted with the bilinguals switching the languages. The results were that the larger switches were noted when the bilinguals switches to their dominant language. one more worth-mentioning point is that inhibitory mechanism is more global than local.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **Cues to language selection**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> It is said in the chapter that in order to achieve early language selection is to rely on cues that signal one language. Two tasks are compared: picture naming and translation. And after several experiments it was found out that language cues available in the translation task makes language selection occur earlier than in picture naming.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> The experiments include Kroll’s, Dietz’s and Green’s experiments and Stroop translation study conducted by Miller in 1997.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> **Similarities and differences between comprehension and production**

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> Comprehension and production are similar in many ways. They both have nonselective access to words in both languages. They both share the consequences of the lower L1 than L2 proficiency in unbalanced bilinguals. In both cases there is much L1 influence on L2 than viceversa.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> The differences between the two can be notices in the orthographic, phonological and semantic fields.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> Also there are number of factors (variables) connected with the cognitive resources which may differ in the two cases.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;">A number of steps involved in the course of lexical access require making choices. A lot of scholars (De Bot and Schreuder (1993), Paulisse and Bongaerts (1994)) claim that “language” is among the features employed for the choice of lemmas. It follows that in the preverbal stage the bilingual speaker has to decide which language should be used to convey the intended meaning. The solution imposing that either //language A// or //language B// should be chosen is often criticized. As an integral part of the conceptual specification language is used in the process of retrieval during which semantic cues of the lexical element are matched with the semantic information. There might be observed to this or that degree random CS and thus not all the switches are necessarily to be initiated and regulated by a certain higher level process.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> Research on CS has revealed that words from distinct languages may be randomly mixed in specific settings (Swigart, 1992). In certain instances language appears to be a less significant cue than other semantic cues. Sometimes a word from another language is chosen as it meets the conceptual conditions better than a word from the cued language.

<span style="font-family: 'Times New Roman','serif'; font-size: 18.6667px;"> Due to Levelt (1989) one can orally utter from 2 up to 5 words per second. Bilinguals need more time to find the necessary words to express the intended meaning than the monolinguals do. Nonnative speakers often neglect less significant cues when faced with time constraints.