The present study investigated whether in bilingual memory, activation of a node in a lexicon, through presentation of the relevant word stimulus, would spread to a lower node in the other lexicon. Two experiments examined the priming effect of a higher node in one lexicon on the task of translation of a lower node in the other lexicon. In Experiment 1, the primed words representing higher nodes were from subjects' L1 and to-be-translated target words representing lower nodes were from subjects' L2. In Experiment 2, the order of languages was reversed. In both experiments, the primed translation condition was compared with an un-primed translation condition. The results showed a facilitatory effect for the L2 prime-L1 target (L1-to-L2 translation) condition. The result of L1 prime-L2 target (L2-to-L1 translation) condition, however, did not reach significance. The RT difference between L1-to-L2 and L2-to-L1 translation tasks in the un-primed conditions was not significant.

 One of the dominating models of semantic processing in cognitive psychology has been known as spreading activation (Anderson 1983a; Collins & Loftus 1975; Rumelhart & McClelland, 1982; Quillian, 1968; Roelofs, 1992). The model is built on a complex association network in which specific memories are distributed in conceptual space with related concepts that are linked by associations. More specifically, concepts are represented in memory as nodes and relations between the concepts as associative pathways between the nodes. When part of the memory network is activated, activation spreads along the associative pathways to related areas in memory. This spread of activation serves to make these related areas of the memory network more available for further cognitive processing (Balota & Lorch, 1986).

 Spreading activation has in fact been widely used as an explanatory construct in cognitive studies. Several researchers have suggested spreading activation as the underlying search mechanism involved in such tasks as category exemplar production and sentence verification (Loftus, 1974), episodic sentence and word recognition (Anderson, 1983a, 1983b), and perceptual word recognition (McClelland & Rumelhart, 1981).

 Empirical evidence has shown that spread of activation is automatic as opposed to being under strategic control (Balota, 1983; Neely, 1977). It has also been shown that the amount of activation of a concept node is a function of the length of the associative pathway between the node and the source of activation (Lorch, 1982). Moreover, the amount of activation spreading from a given node along a pathway is a function of the strength of that pathway relative to the sum of the strengths of all paths emanating from that node (Reder & Anderson, 1980).

 The spreading activation model also best accounts for an unconscious process called associative priming (Ratcliff & McKoon, 1981). Associative priming refers to the facilitation in access to information when associated items are presented (Anderson, 1995). For example, words are named faster in the context of an associated word (e.g., bread-butter) than in the context of an unassociated word (e.g., table-window) and also a word is recognized more quickly when it is preceded by a word from the same sentence or proposition (e.g., run-fast) than when it is preceded by a word from a different sentence or  proposition (e.g., common-fast).

 Spreading activation model accounts for associative priming by assuming that words which are semantically associated with each other are represented in the form of a network and activation spreads through this network from presented words to their associated words in memory (Anderson, 1995).

 As the present study aims to address the issue of bilingual mental lexicon within the framework of spreading activation, a brief review of bilingual mental lexicon seems to be in order.

Bilingual Mental Lexicon

 The investigation of bilingual mental lexicon has in fact rested on the assumption that generalization from monolingual lexical studies to the more complex domain of bilingual lexicon is not based on any a priori justification. The predominant contribution of priming in this relation has been to determine the possibility of dual ML in bilinguals.

 Most of the current studies have focused on the independency or interdependency of lexical organization of bilinguals. Interdependence model can be described as a common storage system. That is, there is one underlying representation common to each word and its translation equivalent (McCormack, 1977; Potter, So, Von Echardt, & Feldman, 1984; Sharifian, 1996; Samani, 1996). The alternative view (independence model) characterizes bilingual ML in terms of separate, distinct systems for the lexical items of each language. That is, a separate lexical representation is believed to be accessed via each verbal system (Paivio & Desrochers, 1980; Paivio & Lambert, 1981 ). In short, the debate with reference to bilinguals have been raised as whether the two verbal systems available to the bilinguals are distinguishable at the lexical level alone or at a conceptual level as well.

Kolers (1965) conducted a series of experiments regarding the issue of independency or interdependency of bilinguals' ML. Using the recall of a word list, Kolers found that subjects in the unilingual and mixed language conditions recalled nearly an equal number of words from the list. The interpretation of Koler's study supports a common or interdependent language store for the bilingual's two languages.

Regarding the independence model, Tulving and Colotla (1970) conducted an experiment, using a recall task, in which unilingual, bilingual, and trilingual lists were presented for later recall. They found that the memory for items from the lists gradually degenerated from unilingual to trilingual lists. Their results were interpreted as support for separate systems of lexical storage, one for each language. Tulving and Colotla's (1970) interpretation of recall decrements was that semantic information is less easily accessed when encoded in two or more languages.

The results from bilingual lexical studies vary from one experiment to another. Hummel (1993) maintains that the source of conflict relates to the experimental paradigm with regard to the material and demands. Hummel (1993) cites typical cases of word list recall tasks that required the subjects to attend to formal characteristics of the lexical items (e.g., orthographic and phonetic features) and, therefore, concludes that the results of these experiments are not reliable.

Some other studies have addressed the processing aspects of bilingual memory. For example, in a study by Gerard and Scarborough (1989) the effect of cognate/noncognate status of lexical item stimuli was investigated. Their results gathered from their bilingual English/Spanish subjects indicated that, while cross-language facilitation occurred in the case of cognates and homographic noncognates, there was no cross-language facilitation of noncognate translations.

In a similar fashion, de Groot and Nas (1991) provided further evidence that noncognates seemed to be represented as language-specific units in conceptual memory, while cognates appeared to share conceptual representations.

 According to Hummel (1993), the results of these studies support the notion that the employed methodology led to important performance differences which depended on the degree of formal similarities between the lexical items.

It seems likely, therefore, that earlier studies which have failed to control systematically for cognate or noncognate status of words and their translation equivalents could have led to varying results, depending on the relative proportion of each type of word pair, especially when subjects may have been less than balanced bilinguals.

In a series of experiments, Keatly, Spinks, and de Gelder (1994) found, employing lexical decision task, that cross-language priming does occur, but only when the primes are presented in the subject's first language, and the target words are presented in the subject's second language. Keatly et al., (1994) attribute the asymmetry in cross-language priming to stronger connections from L1 to L2, than from L2 to L1 and also suggest that this asymmetric cross-language priming can be accounted for by a language-specific model of bilingual memory.

Regarding the studies that have focused on the degree or the extent of semantic processing, Hummel (1993) believes that the lexical decision tasks do not guarantee that the lexical items are processed at the semantic level. Also, Snodgrass (1984) maintains that the degree of semantic processing in lexical decision tasks do not reveal the organization of semantic memory in bilinguals.

In a similar vein, Smith (1991) suggests that lexical decision tasks are generally data driven and are highly affected by the stimulus characteristics of the words used. The results of the preceding studies generally indicate that lexical decision tasks are not reliable instruments to guarantee the validity of the results.

The key issue regarding the interdependence or independence model of lexical presentation as investigated by lexical decision tasks, according to Frenck and Pynte (1987), lies in the notion of various levels of processing of lexical items in bilinguals. In this relation, Frenck and Pynte (1987) refer to some experimental evidence that at certain levels of analysis, the two linguistic systems are relatively independent. For example, it has been shown that a bilingual can activate one phonological system without experiencing interference from the other system (Altenberg & Cairns, 1983). Kintch (1970) have shown that when a continuous list of words were presented, bilingual subjects could only base their recognition on the physical manifestation or on the meaning of a word. Accordingly, it seems that the argument of Frenck and Pynte (1987) regarding the various levels of processing is temporarily substantiated.

Frenck and Pynte (1987) conducted an experiment to study the semantic organization of the bilingual's two languages. Instead of choosing exact lexical items or translations for the priming in lexical decision tasks, Frenck and Pynte (1987) chose the category names for the bilingual's two languages. Facilitation due to priming was observed both across and within languages. However, it was not observed equally for both languages and not for all the lexical items. The result showed a greater facilitation effect for those words that were identified the most slowly when presented in isolation for less skilled bilinguals.
Frenck and Pynte (1987) suggested that the priming facilitation observed may not have been the result of effortless, automatic processing but was rather due to the conscious, strategic use of primes. They argue that in order to truly substantiate the hypothesis of a common semantic network, it is necessary to show that across-language priming facilitation is the result of non-controlled, automatic processing. Keatly and Gelder (1992) also found that cross-language (but not within-language) priming disappeared when subjects were required to respond at a fixed fast rate.

Recently, Fox (1996) conducted two experiments to examine whether unattended words in one of a bilingual's languages could influence processing of an attended word in the other language. The results can be summarized as follows: (a) cross-language negative priming from semantic associates was found in the L1-L2 condition (Experiment 1); (b) cross-language negative priming from translation equivalents was found for both L1-L2 and -L2-L1 conditions, but the magnitude of negative priming was greater in the L1-L2 condition (Experiment 2). Fox (1996) concluded that bilinguals access common conceptual representations across languages.

The present study investigates the relationship between the lexical items of a bilingual's two languages within the framework of spreading activation model. In particular, the question is whether activation of a node in a lexicon, through presentation of the relevant word stimulus, would spread to a lower node in the other lexicon. To address this question, the priming effect of a higher node in one lexicon on the task of translation of a target lower node in the other lexicon is investigated. If presentation of a primed word representing the higher node in one lexical store facilitates translation of the target word representing the lower node in the other lexicon, as compared with an un-primed translation of the target lower node, then it may be concluded that activation spreads from a higher node in one of a bilingual's lexicons to the lower nodes in the other lexicon.

 Purpose. The first experiment was an L2-to-L1 translation task. To investigate the facilitatory effect of cross-language priming, superordinate L1 lexical items of the L2 targets were used as prime stimuli for each of the to-be-translated English words. If activation of an L1 superordinate spreads to the lower node in L2, then presentation of the L1 primes should facilitate translation of the L2 targets by reducing the response latencies of the translations.

Method.
 Subjects. The subjects were 30 male and female university students majoring in English at Islamic Azad University, Khorasgan Branch. The subjects ranged in age between 22 and 45 with an average age of (M = 25.5 ). They participated in the experiment in partial fulfillment of a research familiarization requirement.

 Materials. 10 highly frequent English words were chosen as the to-be-translated targets for the first experiment. For each of the L2 targets an L1 superordinate prime was selected. The primes were also high frequent lexical items in L1. The experimental list included 20 targets (i.e., 10 words were primed and 10 were un-primed). The L2 targets and L1 primes were read and digitized by an SB-16 board in a mono environment. For the primed targets, the L1 superordinates were inserted 300 ms before the targets.

 Apparatus. The preparation of the materials and their presentations in the experiment were carried out in a multi-media environment using a 80486-DX2 personal computer. To measure the response latencies of the translation, a sensitive voice switch was attached to the serial port of the computer.

 Procedure. Subjects were instructed verbally in their native language (Persian). The experiment was conducted in the experimenter's presence monitoring the subjects and the computer's program while they were doing the experiment. Subjects were instructed to listen for the English words and translate them as quickly as possible into Persian. For the primed targets, the subjects were instructed to listen to both the Persian and the English words that followed, but only translate the English ones. The order of presentation for primed and un-primed L2 lexical items was randomized. The ITI between translation tasks was varied to minimize the subjects adaptation to the task.

Results
 The subjects' RTs less than 200 ms and greater than 2000 ms were excluded from the sample. The subjects' faulty translations were also excluded from the sample. The subjects' RTs in primed and un-primed conditions were submitted to a t -test for paired samples.
 Although the presentation of L1 primes reduced the response latencies of the translations as compared with the un-primed condition (63 ms), the mean difference between the two conditions, however, did not reach significance (t(29)=1.83, p < .077).

 Purpose. The second experiment was an L1-to-L2 translation task. To  investigate the facilitatory effect of cross language priming, superordinate L2 lexical items of the L1 targets were used as prime stimuli for each of the to-be-translated Persian words. If activation of an L2 supreordinate spreads to the lower node in L1, then presentation of the L2 primes should facilitate translation of the L1 targets by reducing the response latencies of the translations.

Method.
 Subjects. The same subjects participating in Experiment 1 took part in the Experiment 2.

 Materials. 10 highly frequent Persian words were chosen as the to-be-translated targets for the Experiment 2. For each of the L1 targets an L2 superordinate prime was selected. The primes were also high frequent lexical items in L2. Taken together, 20 targets were assigned to the experimental list used in Experiment 2 (i.e., 10 targets were primed and 10 were un-primed). The L1 targets and L2 primes were read and digitized by an SB-16 board in a mono environment. For the primed targets, the L2 superordinates were inserted 300 ms before the targets.

 Apparatus. The same instrumentation used in Experiment 1 was also used in Experiment 2.

 Procedure. Experiment 2 was procedurally the same as Experiment 1.

Results and Discussion
 The subjects' RTs less than 200 ms and greater than 2000 ms were considered as misses and were excluded from the sample. The subjects' faulty translations were also excluded from the sample. The subjects' RTs in primed and un-primed conditions were submitted to a t test for paired samples.
 

TABLE 1. Mean RTs for Primed and Un-primed conditions in Experiments 1 and 2.
    _______________________________________________

                      L2-to-L1 Translation     L1-to-L2 Translation
                                   Exp. 1                        Exp. 2
    _______________________________________________
    Primed                   599.7                         513.3
    Un-primed              663                            648
    _______________________________________________
 

 The result of RT analysis showed a high facilitatory effect for L2 priming  in reducing the overall translation response latencies as compared to the un-primed cases (t(29) = 4.92, p < .0001). This indicates that activation of L2 superordinate nodes have spread to their L1 lower nodes and, therefore, significantly reduced the translation response latencies in L2-L1 priming condition. Moreover, the RT difference between L1-to-L2 and L2-to-L1 translation tasks in the un-primed conditions was not significant (t(29) = 0.51, p < . 612).

 The main question addressed in the present study was whether in bilingual memory, activation of a node in a lexicon, through presentation of the relevant word stimulus, would spread to a lower node in the other lexicon. To answer this question, the cross-language priming technique was employed. The major assumption was that if activation of a supreordinate in one language spreads to the lower node in the other language, then presentation of the primes should facilitate translation of the targets by reducing the response latencies of the translations.
 The main results can be summarized as follows: (1) there was no significant facilitation from a prime in the subject's first language in L2-to-L1 translation; (2) significant facilitation occurred from a prime in the subject's second language in L1-to-L2 translation; (3) there was no significant difference between L1-to-L2 and L2-to-L1 translation tasks in the un-primed conditions.
 The results of Experiment 2 clearly support the view that activation of a node in one of a bilingual's two lexicons would spread to its semantically associated nodes in the other lexicon.
 As for the results of Experiment 1, due to the 63 ms facilitation obtained, it is hard to maintain that activation of the L1 superordinates has not spread to L2 lower nodes. It seems that we should interpret the results in terms of the strength of connections between the lexical items of the two languages. That is, lexical connections from L2 to L1 seem to be stronger than those from L1 to L2. If this is the case, then we can expect a higher facilitatory effect for the L2 primes than the L1 primes.
 The results of this study also provide support for the hypothesis which assumes a common representational system in bilinguals (Fox, 1996; Sharifian, 1996, 1997, Samani, 1996). The type of asymmetry in cross-language priming apparent in the results of the present study, however, is not in consonance with the asymmetry observed in the findings of Keatly et al. (1994). Contrary to the findings of the present study, Keatly et al., (1994) found a cross-language priming in the L1-L2 condition and not in the L2-L1 condition.
 Finally, a point which requires due attention is that subjects' level of L2 proficiency may be an important factor in studies of this kind. Further research seems to be needed to address the current issue by manipulating the variable of L2 proficiency.

Altenberg, E. P., & Cairns, H. S. (1983). The effects of phonetactic constraints  on lexical processing in bilingual and
        monolingual subjects. Journal of Verbal Learning and Verbal Behavior, 22, 174-188.
Anderson, J. R. (1983a). The architecture of cognition. Cambridge, MA:  Harvard University Press.
Anderson, J. R. (1983b). A spreading activation theory of memory, Journal of  Verbal Learning and Verbal Behavior, 22,
        261-295.
Anderson, J. R. (1995). Cognitive psychology and its implications (4th ed.).  New York: W. H. Freeman.
Balota, D. A. (1983). Automatic semantic activation and episodic memory  encoding. Journal of Verbal Learning and
        Verbal Behavior, 22, 88-104.
Balota, D. A., & Lorch, R. F. (1986). Depth of automatic spreading activation:  Mediated priming effects in pronunciation but
        not in lexical decision. Journal of Experimental Psychology: Learning, Memory, Cognition, 12, 336-345.
Collins, A. M., & Loftus, E. F. (1975). A spreading activation theory of  semantic priming. Psychological Review, 82,
        407-428.
De Groot, A. M. B. & Nas, G. (1991). Lexical representation of cognates and  noncognates in compound bilinguals. Journal
        of Memory and Language,  30, 90-123.
Fox, E. (1996). Cross-language priming from ignored words: Evidence for a  common representational system in bilinguals.
        Journal of Memory and  Language, 35, 353-370
Frenck, C., & Pynte, J. (1987). Semantic representation and surface forms: A  look at across-language priming in bilinguals
        Journal of Psycholinguistic  Research, 16, 383-395.
Gerard, L. D., & Scarborough, D. L. (1971). Language-specific lexical access  of homographs by bilinguals. Journal of
        Experimental Psychology:  Learning, memory, and cognition, 15, 305-315.
Hummel, K. M. (1993). Bilingual memory research: From storage to processing  issues. Applied Psycholinguistics, 14,
        267-284.
Keatly, C. W., & De Gelder, B. (1992). The bilingual primed lexical decision  task: Cross-language priming disappears with
        speeded responses. European Journal of Cognitive Psychology, 4, 273-292.
Keatly, C. W., Spinks, J., & De Gelder, B. (1994). Asymmetrical cross- language priming effects. Memory and Cognition,
        22, 70-84
Kintch, W. (1970). Learning, memory, and conceptual processes. New York:  Wiley.
Kolers, P. (1965). Bilingualism and bicodalism. Language and Speech, 8, 122- 128.
Loftus, E. F. (1974). Activation of semantic memory. American Journal of  Psychology, 86, 331-337.
Lorch, R. F. (1982). Priming and search processes in semantic memory: A test  of three models of spreading activation.
        Journal of Verbal Learning and  Verbal Behavior, 21, 468-492.
McClelland, J. I., & Rumelhart, D. E. (1981). An interactive activation model  of context effects in letter perception: Part 1. an
        account of basic  findings. Psychological Review, 88, 375-407.
Neely, J. H. (1977). Semantic priming and retrieval from lexical memory: Roles  of inhibitionless spreading activation and
        limited capacity attention. Journal of Experimental Psychology: General, 106, 226-254.
Paivio, A., & Desrochers, A. (1980). A dual-coding approach to bilingual  memory. Canadian Journal of Psychology, 34,
        390-401.
Paivio, A., & Lambert, W. (1981). Dual coding and bilingual memory. Journal  of Verbal Learning and Verbal Behavior,
        20, 532-239.
Potter, M. C., So, K. F., Von Echardt, B., & Feldman, L. B. (1984). Lexical and  conceptual representation in beginning and
        proficient bilinguals. Journal  of Verbal Learning and Verbal Behavior, 23, 23-28.
Quillian, M. R. (1968). Semantic memory. In M. L. Minsky (Ed.), Semantic  information processing. Cambridge, MA: MIT
        Press
Ratcliff, R., & McKoon, G. (1981). Does activation really spread? Psychological Review, 88, 454-462.
Reder, L. M., & Anderson, J. R. (1980). A partial solution of the paradox of  interference: The role of integrating knowledge.
        Cognitive Psychology,  12, 447-472.
Roelofs, A. (1992). A spreading-activation theory of lemma retrieval in  speaking. Cognition, 42, 107-142.
Rumelhart, D. E., & McClelland, J. L. (1982). An interactive activation model  of context effects  in letter perception: Part 2.
        The contextual  enhancement effect  and some tests and extensions of the model.  Psychological Review, 89, 60-94.
Samani, R. (1996). Lexical representation and bilingual memory: A  methodological approach. Unpublished Master's
        thesis. Islamic Azad  University, Khorasgan Branch.
Sharifian, F. (1996). The system of mental representation of lexical meanings.  In C. Lucas (Ed.), Proceedings of the
           International Conference on  Intelligent and Cognitive Systems. Tehran: IPM
Sharifian, F. (1997). A network model of mental compiling mechanisms.  Manuscript submitted for publication.
Smith, M. C. (1991). On the recruitment of semantic information for word  fragment completion: Evidence from bilingual
        priming. Journal of  Experimental Psychology: Learning, Memory, and Cognition, 17, 234- 244.
Snodgrass, G. (1984). Concepts and their surface representation. Journal of  Verbal Learning and Verbal Behavior, 23,
        3-22.
Tulving, E., & Colotla, V. (1983). Strategies of discourse comprehension, New  York: Academic Press.