Also avoid publisher series, unless the publisher has a true monopoly over the "works" in question. So, the Dummies guides are a series of works. But the Loeb Classical Library is a series of editions, not of works. Home Groups Talk Zeitgeist. The 12 Days of LT scavenger hunt is going on. Can you solve the clues? I Agree This site uses cookies to deliver our services, improve performance, for analytics, and if not signed in for advertising. Your use of the site and services is subject to these policies and terms.

Journal Julien Green Series by cover. Related book awards Image: How do series work? Hatred assumes a subtle character and, in its subtlety, it slowly eats away at the very fabric of the family. The titles of novels by Mauriac reveal how the bitterness, anger and hatred transform the family into a brood of vipers that are unable to peacefully coexist. What makes the Mauriacian literary family even more fragile, and potentially volatile, is that the family members are forced to stay under the same roof.

Moreover, such internally disintegrated families are obliged to present, to the outside world, a picture of unison and togetherness. Hypocrisy thus becomes the only possible bedrock for the survival of the family. The esprit de famille compels family members to stay together, in spite of themselves, and to keep up appearances, although, internally, the family has lost its very soul, being dead, in other words. The esprit de famille ensures that members of the family strike a balance between an attempt to keep displays of temper at a minimal level within the family milieu, whilst at the same time posing as a knit-together and properly functioning unit.

The family environment becomes a uni- verse in which each member is a distant planet that nonetheless gravitates within the orbit of the family: Within the double concentric circles defined by the provincial society and the house where the Mauriacian literary family resides, the family unit becomes a group of prisoners, all fighting for their own survival.

What aggravates this already tragic crisis is that family members do not communicate or do not strive to communicate. However, this willingness comes with a condition: Most who attempt to communicate find that their interlocutor is either not interested or does not understand what is being said. Equally, when he tries to talk to her, she is also indifferent to what he says: This calls to mind J.

All means of communication, verbal or non-verbal, are directed not at expressing oneself but at emotionally injuring others. When a factor like money is added to this already explosive state of affairs, tempers flare to unfathomable levels. Hatred between husband and wife, father and children, as well as between father-in-law and son-in-law blinds all the family members as they all attempt to control the family fortune.

Even though Louis does not eventually disinherit his children, they so blinded by their hatred and love for money, that they do not recognise the transformation that their father goes through towards the end of his life. The deep-cutting and bitter Mauriacian hatred and lack of communication have destructive effects.

Mathilde dies in solitude whilst her husband and mother-in-law stand by. Mauriac paints a rather grim and pessimistic picture of intra- family relationships: Hatred within the family is firmly entrenched on the basis of the fact that individual family members hate themselves or have an extremely low self-esteem. When they hate someone else, it is a reflection of their own deep-seated self-loathing and self-hatred. Pierre de Boisdeffre poses pertinent questions in this respect: Unable to love themselves, these characters find it doubly difficult to love others: Szczepan Babinski concludes on this tragic and pessimistic vision of Mauriacian intra-family relationships characterised by hate: Louis takes this to mean that Isa did not love him but merely accepted him because she has been trying to get over her failed love affair.

Louis does not forgive Isa for this innocent error she makes. The failure of heterosexual relationships in his books, especially in as far as sexuality is concerned, can be read as some sort of hint of the reservations he has of his own sexuality.


  • Esoteric Philosophy-Deeper Teachings in Spiritual Science.
  • Gottsched und Lessing: Zwei Größen der Literaturkritik im 18. Jahrhundert (German Edition).
  • Vincere Fobie e Paure. Come Superare le Tue Paure con la PNL e Vivere la Vita che Desideri. (Ebook Italiano - Anteprima Gratis): Come Superare le Tue Paure ... la Vita che Desideri (Italian Edition).
  • ?
  • .
  • Rescued by the Billionaire (Alpha male, BDSM, male dominant & female submissive);

It is interesting to note that, al- though neither Mauriac nor his literary characters explicitly say anything about their homosexual longings, the texts themselves, through an intricate interplay of semantic buoyancy and silence, deconstruct the sacred domain that homosexuality is. The present account follows the definition of Heine and Kuteva, in that borrowing always includes the transfer of form-meaning units, whereas polysemy copying only includes the trans- fer of meaning-units to a new form.

An important factor in borrowing is that borrowed lexemes generally become integrated morphologically and phonologically into the recipient language. Two main questions arise when referring to borrowing. First, it may be questioned if there are general motivations to borrow a certain item from a certain language. Second, it has to be asked if certain items are more suited for borrowing than others. As far as the first question is con- cerned, it is indeed possible to assume certain motivations for borrowing.

The first motivation seems to be very widespread and different examples can be found in differ- ent languages. Cultural loans refer to all kinds of borrowings for cultural reasons, such as agricultural and botanic products like coffee, cocoa or mango in English or new technological innovations such as Internet and Computer in German. These gap-fillers introduce a new concept, which has not been named yet in the recipient language.

The second motiva- tion for borrowing refers to borrowing for prestige reasons of the donor language. These two kinds of borrowings do not necessarily presume a direct geographical language con- tact situation or bilingualism of the respective speakers. For example, words for agricultural products have often followed different paths of borrowing, such as English mango via Portuguese manga from Malay manga, borrowed from Tamil mankay Matras A third motivation for borrowing may be cognitive reasons.

This motivation arises generally in bilingual communities or in situations of direct language contact and is not specific to borrow- ing but may account for different types of language change. In this case, borrowings may first occur as innovations, which are uttered by one speaker. These innovations may arise accidentally or for special purposes.

Works (27)

The first possibility includes that a speaker may use a word from another language in lack of the adequate lexeme in the language currently spoken. According to Lipski , this is a kind of metalinguistic bracketing, while Matras argues in favor of a cognitive trigger that leads to unintentional language choices see chapters 4. The second possibility refers to linguistic innovations for discourse-pragmatic purposes. That is to say that a speaker may employ a word from another language for discourse-strategic reasons, such as emphasis or accentuation or to detach it from the rest of the discourse.

When a word first entered a language through innovation, it may become a borrowing over time through ha- bitualisation Hlavac All these motivations for borrowing will be discussed in detail and with regard to the specific pragmatic marker so in chapter 4. The second main question in borrowing concerns the borrowability of specific lexical items. Are some lexical items more suited for borrowing than other items? It has been subject of discussion that lexical borrowing is far more common than grammatical borrowing, especially of inflected forms.

This discussion suggests that there may indeed be a certain hierarchy in borrowability. Several authors tried to establish borrowability scales. Thomason and Kaufman estab- lished a hierarchy for possible borrowings: Casual contact Category 1: Muysken proposes a different borrowability hierarchy, in which specific word classes are classified accord- ing to their suitability for borrowing Muysken In this hierarchy, Muysken visu- alizes that pragmatic markers are far more suited for borrowing than syntactic elements ibd.

There are indeed broad possibilities to visualize hierarchies of borrowing and it is very difficult to determine one unequivocally valid scale. This is because it is nearly impossible to collect and classify all possible occurrences of borrowing worldwide and to rule out possible borrowings, which may take place in the future. Still, there is consent on some general rules on borrowing. As already mentioned, it is generally recognized that borrowing of content words is more fre- quent than structural borrowing or borrowing of grammatical material, especially of inflectional morphology.

Furthermore, it is assumed that elements that are syntactically peripheral, such as pragmatic markers, are borrowed more frequently than syntactically integrated items. To conclude, the present subchapter aimed to introduce two main notions of contact-induced language change that are replication and borrowing. These two processes are still matters of dis- cussion and it is not always clear how to exactly define and delimitate them.

The present account relies on the differentiation and terminology of Heine and Kuteva , , It has been stated that pragmatic markers show some peculiarities in language contact. It was as- sumed that pragmatic markers are especially suited for contact-induced language transfer, due to their syntactic flexibility and their peripheral positions. This hypothesis will be explained in de- tail in the upcoming subchapter. Furthermore it will be discussed in what way a cross-linguistic analysis of pragmatic markers can be of interest with respect to a general account on pragmatic markers.

In the course of the main debate on pragmatic markers, the focus was clearly on their monolingual characteristics and functions. Still, some more recent studies tried to focus on two cross-linguis- tic issues of pragmatic markers. The first issue concerns the role of pragmatic markers in situa- tions of language contact. Important questions here are for instance what happens to pragmatic markers in language contact and why e. The second important issue concerns the cross-linguistic comparison of pragmatic markers.

The present subchapter aims to introduce these two main issues on bilingual pragmatic markers and their cross-linguistic comparison. As pointed out in the previous subchapter, pragmatic markers are considered very susceptible to borrowing and other cross-linguistic influences because of their syntactic detachability. Still, this is not the only reason as to the interest of researchers in bilingual pragmatic markers.

It has been considered extensively that pragmatic markers are generally difficult to translate and that they may have more than one translation equivalent. Furthermore, their semantic value is often difficult to determine and it is not even clear if they encode conceptual or procedural meaning see.

Another interesting factor comes from diachronic analysis of pragmatic markers, in that the moment of the emergence of their discourse-pragmatic functions is generally difficult to determine. This is because most pragmatic markers emerged through processes of grammaticali- zation or pragmaticalization from already existing lexical items. But it is also due to the fact that pragmatic markers generally occur in spoken language and often only in very informal speech.

Therefore it is hard to retrace their diachronic evolution by means of written corpus data see chapter 4. All these peculiarities make of pragmatic markers a very interesting field of research, especially from a cross-linguistic perspective. While it is clear that pragmatic markers have specific characteristics in comparison to other lexical items, it is still not clear what happens to them in situations of strong language contact.

Clyne spotted the transfer of German pragmatic markers into English discourse of Ger- man-speaking immigrants in Australia. Mougeon and Beniak examined the core borrow- ing of pragmatic markers in Canadian French discourse. Torres and Potowski tried to classify the possible outcomes of pragmatic markers in contact, which they describe as follows: In this case, the whole German set of markers was replaced by English markers ibd.

Still, it is unclear if these three outcomes are mutually exclusive or if long-term and intense lan- guage contact always results in outcome number three, that is the complete replacement of one set of pragmatic markers. The opposite option would be that sets of pragmatic markers from two languages might co-occur over a long period of time without too strong of an influence on each other. This question will be discussed in more detail and with regard to a special language contact situation in chapter 4.

Another related problem concerns the question why pragmatic markers show these important peculiarities in language contact situations. Mougeon and Beniak assume that all mark- ers first occur in the respective other language as code-switches and then gradually become bor- rowings. It can be stated here that most researchers regard the transfer of form-meaning units of pragmatic markers as a case of borrowing e. The borrowing of pragmatic markers will be regarded in detail in chapter 4.

To conclude, it can be stated that pragmatic markers are very susceptible to language contact and that sets of pragmatic markers may behave differently in different language contact situa- tions. The peculiarities of pragmatic markers in language contact are mostly due to their general characteristics. But the use of pragmatic markers from one language in the other may have ad- ditional reasons see 4. The analysis of pragmatic markers in language contact can indeed offer important new insights into the functioning and structure of pragmatic markers.

But also a cross-linguistic comparison of monolingual pragmatic markers may offer important research opportunities.

That is to say that cross-lin- guistic comparison can not only help to identify common paths in the emergence and evolution of pragmatic markers; it can also help to clarify the motivations for change as well as functioning of pragmatic markers. Consequently, cross-linguistic comparison can help to gain insights from a synchronic and from a diachronic perspective. To underline the usefulness of cross-linguistic comparison, two methods will be presented.

The first one concerns the comparison of functions of pragmatic markers by means of separate corpus data. The second method concerns the trans- lation of pragmatic markers and the creation of semantic fields. The first method appears to be the more common one, in that it resembles very much the mono- lingual procedure to analyze the functioning of pragmatic markers. Fleischman and Yaguello compared English like and French genre, to identify if languages may develop inde- pendently pragmatic markers with similar functions.

In their positive conclusion, they try to establish pragmatic universals: Mihatsch compared these pragmatic markers not for the aim of estab- lishing pragmatic universals, but to explore their grammaticalization paths and possible bridging contexts from a diachronic perspective. Still, this diachronic comparison also helps to identi- fy current possible grammaticalization processes, such as the emergence of new meanings and functions in Canadian French comme see 4.

The comparison of possibly equivalent markers in corpus material from different languages is not only an important method to retrace the diachronic and synchronic evolution of pragmat- ic markers. It also provides evidence for the polyfunctionality and the polysemy of pragmatic markers. Several authors proved extensively that pragmatic markers in different languages may share certain meanings and functions but may vary in others e.

Another possibility to explore cross-linguistic differences of pragmatic markers is the translation method by Aijmer and Simon-Vandenbergen , , The translation method contributes to specifying how markers function intra- linguistically, how they relate to other, semantically and pragmatically similar items in the same language, and how semantic fields in different languages relate to one another. This method does not only aim to determine different translation equiva- lents of a specific pragmatic marker: Furthermore, this method helps to provide even more evidence for the polysemous hypothesis of pragmatic markers.

A semantic map tries to capture not only the diachronic evolution of a respective pair of pragmatic markers; it also rep- resents the synchronic comparison of cross-linguistic or partial equivalent marker pairs. Here, a semantic map is particularly suited to highlight differing degrees of polysemy between differ- ent markers and to point out their degree of equivalency see Beeching Furthermore, a semantic map can visualize the relatedness of functions of pragmatic markers in a very suitable way.


  1. .
  2. Across the Great Sparkling Water (The Peacemaker Series Book 2)!
  3. Journal Julien Green | Awards | LibraryThing?
  4. .
  5. .
  6. Sogni Oscuri (Italian Edition);
  7. An example of a semantic map of the markers bon, enfin, well and I mean, taken from Beeching As a conclusion, it can be stated that cross-linguistic comparison and the analysis of pragmatic markers in language contact constitute an important tool for linguistic analysis. The present corpus analysis chapter 4 follows Fleischman and Yaguello as well as Mihatsch in its methodology.

    It only differs in that almost all examples from both languages will be taken from the same set of bilingual corpus data, presented in chapter 3. Before introducing and analyzing the underlying corpus data in detail, the upcoming chapter aims to introduce current problems, models and approaches on bilingual language processing. Therefore, the next chapter puts the focus on a psycholinguistic perspective, which will be cru- cial for the analysis in chapter 5 and 6.

    The research on bilingual lan- guage processing is extremely complex and therefore it is even more important to provide a clear theoretical basis for the upcoming chapters. On the one hand, the modeling of the bilingual mental lexicon includes the same problems as of the monolingual lexicon, for instance questions about the overall structure of language stor- age and lexical retrieval.

    Here, the focus of research lies on the questions of how phonological, orthographical, morphological, semantic, syntactic and pragmatic information of a given lexical item is stored in the mental lexicon, how this information is linked and how it can be accessed and integrated into speech comprehension and production. On the other hand, the modeling of the bilingual mental lexicon raises new issues. The most controversial discussions concern the is- sue of independence or interdependence of both languages and of selectivity or nonselectivity in lexical access.

    These questions, treated in this introductory part, are crucial for an understanding of the different models of bilingual language representation in chapter. The second subchapter, 2. Thus, chapters and 2. On this basis, it is possible to formulate the present understanding and perspective of bilingual lexical access. This is crucial to clearly define the role of pragmatic, semantic and conceptual representations in bilingual lexical access. Building on this, part 2. All current models of bilingual language representation agree on the fact that the mental lexicon is divided into different levels that are, at least, a conceptual and a lexical level.

    The lexical level contains all language specific word information; the conceptual level represents non-language specific abstract information. This ambiguity in the terminology may be confusing, especially in crosslinguistic research. Therefore it is crucial for the present work to strictly differentiate between conceptual and semantic representations. The present study relies on the differentiation of Pavlenko , who distinguishes the following three components of words: The models, discussed in the present part and in part 2.

    They put the emphasis on a specific level of language processing and neglect other levels. This issue will be discussed in detail in the upcoming parts of this chapter. If it is assumed that bilingual language representation may be visualized by means of different levels, the question arises how two or more languages are stored in the multilingual mental lexi- con. At first, it seems plausible to suggest that both languages are stored and accessed separately in different memory stores. This hypothesis is known as the separate or independent hypothesis e.

    In contrast, the shared or interdependent hypothesis claims that both languages are stored in one memory store and that words are tagged as belonging to one language in word retrieval e. At first sight, it seems unclear why there is experimental evidence for both hypotheses, but there may be an explanation for these results. It may be the case that the results depend largely on the respective experimental tasks and designs. While tasks, relying on semantic and conceptual pro- cesses often give evidence for the shared memory hypothesis, tasks relying on lexical processes point more often towards the separate memory hypothesis Heredia These findings lead to the assumption that both languages may share a conceptual and semantic level but have a separate lexical level.

    This idea will be explained in more detail on the basis of the respective lexicon models in the upcoming part of this chapter 2. This chapter focuses on models of bilingual language representation, which do not differentiate particularly between speech comprehension and production. If both languages are indeed interconnected to some degree, another question arises: Is lexical access and retrieval selective or non-selective? This is to say, do speakers only access one language at a time or are both languages always activated in the lexical retrieval process?

    And if language processing is non-selective, how is the right language se- lected and the other one suppressed? These questions will be addressed in part 2. The revised hierarchical model by Kroll and Stewart Figure 7 is based on the hierarchi- cal models by Potter et al. According to Potter et al. According to the word association mod- el, translating from the L1 into the L2 should take slower than picture naming in the L2.

    The concept mediation model predicts that there should be no difference between picture naming in the L2 and translating. They found out that both groups of bilinguals responded faster to pictures than to words, regardless of language proficiency. This result is consistent with the con- cept mediation model, but not with the word association model Potter et al.

    They conclude that languages may not have associations between lexical pairs from both languages Potter et al.

    Pourquoi suis-je moi? Journal - Free Online Library

    In contrast, Kroll and Stewart chose another approach to the two hierarchical models and, in consequence, their own evolution of these models. To provide evidence for this model, they performed three picture naming and bilingual translation tasks with fluent Dutch-English bilinguals.

    JE SUIS CONTRE L'IMMIGRATION

    In these tasks, they ma- nipulated the semantic context of the translations randomized or categorized. In Experiment 1 and 3 response latencies were slower in the categorized than in the randomized conditions. Furthermore, category interference in Experiment 3 only occurred when translating from the L1 into L2 and not vice versa.

    Kroll and Stewart concluded from these findings that a conceptual representation is used to retrieve a lexical entry. When multiple lexical entries are activated by a conceptual representation, this creates interference. Furthermore, they conclude that translating from L1 to L2 passes via the conceptual level whereas translating from L2 to L1 can be com- pleted on the lexical level. That is to say, L1-L2 translation is conceptually mediated whereas L2-L1 translation is lexically mediated. To visualize these findings, Kroll and Steward proposed the revised hierarchical model RHM , which still is one of the most influential of the bilingual lexicon models.

    The RHM proposes two independent but interrelated language stores. The storage for the L1 is larger, because bilinguals generally know more words in their L1 than in their L2. These links are less strong from the L1 to the L2, because bilinguals associate L2 words stronger with the L1 translation than vice versa. Furthermore, links from the L1 to the conceptual level are stronger than the respective links from the L2 to the conceptual level. This can be explained by the fact that bilinguals link L1 words naturally to the respective concept, whereas L2 words are first associated to the L1 translation equivalent.

    Still, the RHM is not a static model, since it assumes a possible shift in the dependence of the L2 words on the L1 translation equivalent with increasing language proficiency. Nevertheless, the RHM has been challenged and questioned regularly in previous years.

    Series by cover

    Brys- baert and Duyck suggest that the RHM is antiquated; because a large amount of research has been conducted since its development and new insights have been made. The criticism is primarily directed at the independence of the two lexica and the selective lexical access of the RHM.

    Furthermore, the RHM does not consider the different levels of lexical access and does not give any information about semantic, phonological or orthographic representation. It is indeed incontestable that the RHM does not provide any detailed evidence about lexical access in language production or comprehension. It is rather a more general sketch of the bilingual lexicon that primarily considers the representation of both languages and not the detailed lexical access. It is true that a large amount of evidence in current research indicates that language selec- tion is non-selective see chapter 2.

    Still, the exact interdependence between both languages and the location and mechanism of selection rests controversial. Despite some inaccuracies on the different levels and some outdated assumptions, the RHM is still relevant. Especially the idea of weaker and stronger links has received particular attention in research on bilingual lexical access see the weaker links model in 2.

    The RHM predicts that L1 and L2 translations are sensitive to semantic context, and this idea is also relevant for other models of the bilingual mental lexicon.

    Pourquoi suis-je moi? Journal 1993-1996.

    Another very early model of the bilingual brain is the distributed conceptual feature model DCFM from de Groot see Figure 8. This model implies lexical nodes that are asso- ciated with a distributed set of conceptual features. The most important aspect of this model is that the author assumes a different degree of overlap of conceptual features for different L1-L2 word pairs. This overlap of conceptual features then determines the translation performance of participants in translation tasks.

    As a first assumption, the authors of the DCFM assume that concrete words in two languages have more semantic overlap than abstract words. De Groot a performed three word-trans- lation tasks, two normal translation tasks and one translation-recognition task, with Dutch-Eng- lish bilinguals. In the normal translation tasks, participants were confronted with words from their first language Dutch , which they had to translate into their second language English. In the translation-recognition task, participants were confronted with English-Dutch word pairs and they had to decide if these words were translation equivalents.

    In all three experiments, de Groot found a concreteness effect. That is to say, in the normal translation tasks, concrete words were translated faster than abstract words and, in the translation-recognition task, responses on concrete word pairs were faster and with lower error rates than responses on abstract word pairs.

    In an additional normal translation task, de Groot obtained faster response latencies for cognate words than for noncognate words. These results lead de Groot to establish the DCFM, in which concrete words, abstract words and cognates vary in the overlap of their conceptual representations. However, this model does not differentiate adequately between conceptual and semantic representations and both terms are often used interchangeably. Additional support for these results comes from van Hell and de Groot Still, the DCFM has been subject to criticism for different reasons.

    One point of critique has been highlighted by Heredia and Brown While some concrete translation equivalents overlap completely in their conceptual representations others do not. For instance, taking the concrete English word ball e. In contrast, some abstract words from two languages may actually overlap or not in all of their features. Considering this aspect, the solution of the DCFM may indeed be too simple. A further point of criticism comes from Finkbeiner et al.

    They point out that the DCFM cannot provide a satisfactory explanation for the translation asym- metry. Translation asymmetry describes the phenomenon that masked priming occurs from the L1 to the L2 but not vice versa e. In an unmasked priming paradigm, the exposure of a visual stimulus influences the response latencies of a later stimulus. Response latencies vary depending on the semantic, phonological or orthographic overlap of the prime stimulus and the target stimulus. In the masked priming para- digm, the prime stimulus is only exposed very shortly, generally for 50 or ms.

    Furthermore, the prime stimulus is preceded and in many cases also followed by a pattern mask of several hash keys, which is exposed for ms. This technique relies on a subliminal processing of the prime stimulus and therefore an unconscious effect on the target stimulus see chapter 6. In a consequence, Finkbeiner et al. The Sense Model relies on the assumption that most words are polysemous and that each word has different semantic senses. Accordingly, ostensible translation equivalents may vary in the overlap of their semantic senses, or meaning patterns as they are called in the present work.

    Finkbeiner and colleagues refer to this partial equivalence as representational asymmetry. The Sense Model adapted from Finkbeiner et al. This task differ- ence describes the phenomenon that L2-L1 priming effects are significantly stronger in semantic categorization than in lexical decision for a detailed discussion see 6. To provide evidence for the Sense Model, Finkbeiner et al. In Experiment 1—3, Japanese-English bilinguals accomplished masked priming tasks, either with lexical decision or semantic categorization. As a result, they replicated the task difference in the translation asymmetry, in that L2—L1 priming occurred in semantic categorization and not in lexical decision.

    In Experiment 4—6, the underlying assumptions of the Sense Model were tested in a within-language setting with English speakers. The aim of these experiments was to analyze the priming effect from many-to-one senses e. Their results indicate that priming was found in a semantic cate- gorization task for both directions.

    But for lexical decision, priming was only obtained in the many-to-one direction. These results suggest an analogous priming asymmetry similar to the bilingual Experiments 1—3. In contrast, in lexical decision, all senses of the words are activated. In the L1—L2 direction, all senses of the L2 word are primed and consequently a priming occurs. In the L2—L1 direction, the L2 word cannot prime all senses of the L1 word and no or few priming effects occur. Further support for the Sense Model comes from Wang and Forster , who performed four differ- ent experiments to test hypotheses of the Sense Model.

    Further evidence for the fact that response times may be sensitive to semantic senses comes from Rodd et al. They performed three monolingual lexical decision tasks with polysemous words, homonymous words and unambiguous words. Their results indicate that homonymous words are processed slower than unambiguous words and that polysemous words are processed faster than unambiguous words.


    1. .
    2. .
    3. !
    4. They reason from these findings that the character of semantic representations has an impact on response latencies in lexical decision. In conclusion, it may be noted here that the models of bilingual language representation pre- sented in this subchapter seem to be very different at the first sight. The RHM and the DCFM do not include the different levels of language processing and only differentiate between a lexical and a conceptual level.

      However, the Sense Model does not include a conceptual level but differentiates between the lexical form representation and the lexical semantic representations. Still, all models have some aspects in common. The authors of all three models do not make any statement about the mechanism and the location of language selection. In this context, all three models fail to make adequate assumptions about language selectivity.

      The main problem still concerns the question to which extent the different levels in the bilingual brain are independent or interdependent. As shown in the preceding chapter, most researchers today assume that the two languages are, at least to some ex- tent, interdependent and influence each other in lexical retrieval and selection. This assumption raises a range of new questions when trying to model the exact stages of lexical access. Lexical access is known as the process when speakers try to retrieve the appropriate lexical representation for a certain concept.

      The term lexical access may be used for language production and percep- tion, despite the difference in the underlying processes. Therefore, in the present approach the term lexical retrieval will be used referring to language production processes, whereas the term lexical access will be employed for processes of language reception.

      The current chapter aims to discuss different models of bilingual lexical access. All these models in only regard the process of speech production and cannot be adopted in the same manner for speech perception. The brief over- view of the models is limited to information that is relevant to the present study. As it is common- ly assumed in models of the bilingual lexicon, the two languages share a common conceptual level and differ only at the lexical and feature level. This should lead to more problems in lexical retrieval in bilingual than in monolingual speakers because representations from both languages can be activated in the lexical retrieval process.

      To account for this disadvantage, researchers have proposed several models of bilingual lexical access that try to explain the retrieval and selection process. According to Finkbeiner et al. Furthermore, it is commonly accepted that bilingual speakers cannot switch one language completely off; certain interference is always given. Two general approaches of bilingual lexical access can be differentiated, a competitive approach and a weaker links approach. Most recent models assume that bilingual lexical retrieval is a competitive process, that is that lexical nodes compete for selection.

      Here, only two current models are presented for a more detailed overview and discussion see Finkbeiner et al. The first model to be presented is the model of language-specific access by Costa et al. Language-specific lexical access adapted from Costa et al. The lexical level is language specific, and 3. Hence, this model does not consider lexical retrieval as a competitive process at the lexical level but only at the level of semantic representations. Experimental evidence comes from the bilingual picture word inter- ference paradigm, and in particular the semantic interference effect between languages Costa et al.

      In both studies, they found a facilitatory translation equivalent effect when picture and distractor word were the same, independent of the respective language. In contrast, they found slower naming latencies when the distractor was a semantically related word.

      These results led them to the assumption that there is no competition of languages at the lexical level. This explanation accounts for the fact that translation equivalent may function just as identi- ty primes whereas semantically related words evoke semantic interference effects Costa et al. Still, Costa and colleagues found a language effect in their picture-naming task. Participants showed slower response latencies when unrelated distractor words were in the target language than in the non-target language.

      Series: Journal Julien Green

      This effect challenges the view of language-specific selection because parallel activation of translation equivalents should lead to equal interference in picture naming Finkbeiner et al. Many competitive models of lexical access therefore propose a shared lexical level between lan- guages. But how do these models account for the retrieval and selection of the respective lan- guage? The Inhibitory Control Model ICM by Green , acts on the assumption of a shared lexical level in which all lexical representations are tagged with a language node.

      In the speech production process, inhibitory control mechanisms suppress the non-relevant languages; trans- lation equivalents are not activated and cannot be selected Green The following figure shows a simplified version of the ICM by Finkbeiner et al. It is regrettable that Green does not differentiate sufficient- ly between the conceptual and the semantic level so that no clear presumptions concerning this differentiation can be made on the basis of the ICM.

      Empirical evidence for this model can be found in studies on asymmetrical language costs in code switching. They detected that bilinguals showed more switching costs when switching into their more dominant language L1 than into their less dominant language L2.