Much previous research has pointed to the need for a unified framework for language contact phenomena – one that would include social factors and motivations, structural factors and linguistic constraints, and psycholinguistic factors involved in processes of language processing and production. While Contact Linguistics has devoted a great deal of attention to the structural properties of contact phenomena and their sources in the input languages, the field has made much less progress in attending to Weinreich's observation that language contact can best be understood only “in a broad psychological and socio-cultural setting” (Weinreich, Reference Weinreich1953, p. 4). There have been some attempts to establish links between the disciplines that investigate language contact, for example, the psycholinguistic and sociolinguistic (Walters, Reference Walters2005), and the linguistic and psycholinguistic (Myers-Scotton Reference Myers-Scotton2002, Winford Reference Winford2009, among others). Yet, so far, no one has come close to achieving the kind of integrative, multi-disciplinary framework that Weinreich envisaged. Muysken's paper is therefore a welcome reminder of the need for such a framework, and the complexity of the task involved in constructing it, if indeed it can be accomplished. The introduction to the paper outlines a very ambitious objective – “to explore the possibility of unifying these fields, all different approaches to language contact, creating a single framework within which it is possible to link results from different subfields” (Section 1.1).
Muysken's contribution to this endeavor is his “quadrangle model” of contact phenomena. He sets the stage for this model by outlining a number of social, psycholinguistic, and linguistic factors that regulate the outcomes of contact. However, there is no attempt to explore any of these in detail, or to show how they might be integrated into the model he presents. It soon becomes clear that the model is chiefly concerned with the linguistic inputs to language contact and the role they play in shaping the outcomes of that contact. According to Muysken, the key elements of the proposed model are the bilinguals and the bilingual “optimization strategies” at their disposal. These strategies make up the four poles of the quadrangle model, and include the following:
A. Exploit properties of the L1 (abbreviated as “L1”);
B. Exploit shared properties of the L1 and L2 (abbreviated as “L1/L2”);
C. Appeal to universal principles (abbreviated as “UP”);
D. Exploit properties of the L2 (abbreviated as “L2”);
The notion that contact phenomena proceed from the interaction of these four types of contribution is hardly new. Most scholars would agree that some outcomes are the result of heavier reliance on L1 knowledge, while others are the result of reliance on the L2, and still others result from accommodations between L1 and L2, or from innovations that reflect neither L1 nor L2 features (which the author ascribes to “universal principles”). However, there is no real attempt to spell out precisely how these strategies come into play in actual situations of language contact. As a result, the model does not live up to the promise that “it will allow us to see whether the same factors are responsible for specific options in different contact settings” (Section 6). In fact, we are never told what processes or mechanisms are involved in the “speaker optimization strategies” mentioned in the abstract and in Section 1.4, and referred to throughout the paper. Neither are we told exactly how each type of optimization strategy contributes to a particular outcome of contact. For instance, in discussing code switching (CS) patterns, we are told that one pattern, insertion, results from using the grammatical and lexical properties of the first language. A second pattern, congruent lexicalization, results from producing structures and words which share the properties of the L1 and L2. And so on. These are vague descriptions of what speakers do with the two languages, not explanations of how the mixture comes about. Similarly, for creole formation, we are told (in Section 2.3) that the optimization of L1 leads to “relexification or transfer of L1 structures”, while optimization of L2, which is described as “imitation of European vernacular varieties”, leads to creoles which resemble European settler languages. No explanation is given of what linguistic processes or principles are involved in each case. Any model of contact-induced change must explain clearly what mechanisms lead to specific innovations or outcomes. In the absence of such explanation, it is hard to see how the quadrangle model adds anything to our understanding of variation in the outcomes within specific domains such as CS or creole formation. We have just as little basis for understanding how similar kinds of mechanisms produce similar kinds of contact-induced changes across different contact situations.
To compound the problem, Muysken uses the same terms to describe contact phenomena in very different situations, without justifying whether such terms refer to the same processes or results in all cases. For instance, the term “relexification” is used to describe the workings of substrate influence in creole formation as well as the origins of mixed languages like Media Lengua. Is the same process involved in both cases? In what sense do the two phenomena result from the same strategy of “optimizing L1”? Similarly, the term “convergence” is used to describe phenomena such as the changes from OV to VO order in some varieties of Quechua and from VO to OV in the Spanish varieties with which Quechua is in contact. The same term is used to describe the combination of French NP structure with Cree VP structure in Michif. The implication that such different phenomena arise from identical processes is hardly a help.
The same problems apply to Muysken's “universal principles” – one of the cornerstones of the model. However, the term is used inconsistently. In a very brief discussion, the author describes universal principles as “general combinatory principles governing improvised language behavior”, listing such examples as reduplication, topic/comment structures, the use of one form for one meaning, and so on (Section 2.2). But in discussing creole formation, he appeals instead to Bickerton's claim that the creole features directly reflect a Language Bioprogram or UG (see Section 2.3). This notion has long been abandoned by most creolists since there is no support for it on either linguistic or sociohistorical grounds. The end result is that we are left unsure what “universal principles” apply in each case of contact, and whether they are indeed the same in all cases. Muysken's formulation of the notion thus misses an opportunity to build on more recent work that relates the universal principles at work in creole genesis to those that operate, for instance, in Second Language Acquisition (SLA), and to compare and contrast them with those that operate in CS or in the formation of bilingual mixed languages. When one thinks of creole genesis, one thinks of universal processes by which learners produce structures that are easier to process (Piennemann, Reference Pienemann1999; Plag, Reference Plag2008), or of internally motivated and contact-induced grammaticalization (Heine & Kuteva, Reference Heine and Kuteva2005; Winford, Reference Winford, Schreier and Hundt2013). One also thinks of the role of “transfer” or imposition, similar to what we find in SLA (Siegel, Reference Siegel2008; Winford, Reference Winford2008). All of these operate according to a number of universal principles which we are beginning to understand much better. None of these applies to classic CS, or to the creation of bilingual mixed languages. Considerations like these render the notion of “optimizing universal principles” as used in the quandrangle model too vague to be useful as a way of finding common ground across contact phenomena.
The task of modeling and interpreting language contact phenomena which Muysken undertakes requires a theoretical framework that addresses, among other things, the nature of the processes underlying contact-induced change. Yet Muysken himself acknowledges that his model fails to make it clear whether “a given strategy reflect[s] a resulting state or an ongoing process” (Section 6). In addition, a model must address both the actuation and implementation of change. This in turn necessitates investigation of both the individual and the community at large. Muysken's model, as he himself acknowledges, makes no distinction between the level of the individual's language behavior and that of the community's language. This touches on a fundamental issue in the study of contact-induced or any other type of linguistic change, namely the complementary roles played by the individual and the community in the origin and spread of change. If, as is generally agreed, the locus of actuation (of a contact-induced innovation) is the individual bilingual, then the mechanisms or processes of change have to be explained in terms of how linguistic systems or inputs interact in the individual mind – i.e., in psycholinguistic terms. At the same time, if the locus of the propagation of change is the set of networks that link individual to individual, then this aspect of change must be explained in sociolinguistic terms (Bachus Reference Bachus2009). By not attending to these distinctions Muysken's model imposes a heavy burden on those who would use it to compare different kinds of contact situations and the processes of change within them.
With regard to the mechanisms involved in the actuation of contact-induced changes, Muysken is right to note that the similarities and differences in the outcomes lie in the way bilinguals implement certain strategies, and that these in turn relate to the roles played by the two languages. But his model does not distinguish clearly among these roles, and consequently, it confounds many different types of change. For instance, the following are all treated as instances of the optimization of L1 strategies:
(a) “insertion” in code switching
(b) “relexification” in creole genesis
(c) “relexification” in the case of Media Lengua
(d) replacement of a native affix by a foreign equivalent
(e) “classical” relexification, the grafting of an L2 phonetic form onto an L1 lemma
(f) the phonological adjustment of loanwords
In a pyscholinguistically-oriented framework such as that of van Coetsem (Reference van Coetsem1988, Reference van Coetsem2000), all of the above phenomena except (b) would be treated as cases of borrowing, accompanied by adaptation to the recipient language (RL) structure. The phenomenon in (b), by contrast, would be treated as a case of imposition, also referred to as ‘transfer’ in SLA research, in which the L1 would be the source language. Similarly, Muysken treats all of the following as manifestations of the optimization of L2 strategies:
(g) European input in creole genesis
(h) the adoption of a grammatical pattern from another language
(i) the generalization of Spanish plural markers in Ecuadorian Quechua
(j) straight lexical borrowing
(k) L1 transfer in L2 pronunciation
(l) transfer of L1 pragmatics into L2 behavior
Of these, it is unclear what process or mechanism is involved in (g)–(h), though it is clear that the source of the input is the L2 in both cases. On the other hand, in the case of (i)–(j), the mechanism that is involved is clearly borrowing, which again presupposes the agency of RL speakers. By contrast, (k)–(l) involve the mechanism of imposition, with L1 acting as source language.
In short, the quadrangle metaphor or model seems to me to confound or group together a number of processes and mechanisms of change that are not in fact the same. It does not show how each of the four optimization strategies operates to produce the relevant outcomes of contact-induced change. It also fails to separate the outcomes from the processes that bring them about, and relate the processes explicitly to the strategies that speakers employ. It therefore falls short of serving as a basis for a unified model of contact-induced change. Instead, it may be considered a template for the empirical investigation of different situations of language contact.
Much previous research has pointed to the need for a unified framework for language contact phenomena – one that would include social factors and motivations, structural factors and linguistic constraints, and psycholinguistic factors involved in processes of language processing and production. While Contact Linguistics has devoted a great deal of attention to the structural properties of contact phenomena and their sources in the input languages, the field has made much less progress in attending to Weinreich's observation that language contact can best be understood only “in a broad psychological and socio-cultural setting” (Weinreich, Reference Weinreich1953, p. 4). There have been some attempts to establish links between the disciplines that investigate language contact, for example, the psycholinguistic and sociolinguistic (Walters, Reference Walters2005), and the linguistic and psycholinguistic (Myers-Scotton Reference Myers-Scotton2002, Winford Reference Winford2009, among others). Yet, so far, no one has come close to achieving the kind of integrative, multi-disciplinary framework that Weinreich envisaged. Muysken's paper is therefore a welcome reminder of the need for such a framework, and the complexity of the task involved in constructing it, if indeed it can be accomplished. The introduction to the paper outlines a very ambitious objective – “to explore the possibility of unifying these fields, all different approaches to language contact, creating a single framework within which it is possible to link results from different subfields” (Section 1.1).
Muysken's contribution to this endeavor is his “quadrangle model” of contact phenomena. He sets the stage for this model by outlining a number of social, psycholinguistic, and linguistic factors that regulate the outcomes of contact. However, there is no attempt to explore any of these in detail, or to show how they might be integrated into the model he presents. It soon becomes clear that the model is chiefly concerned with the linguistic inputs to language contact and the role they play in shaping the outcomes of that contact. According to Muysken, the key elements of the proposed model are the bilinguals and the bilingual “optimization strategies” at their disposal. These strategies make up the four poles of the quadrangle model, and include the following:
A. Exploit properties of the L1 (abbreviated as “L1”);
B. Exploit shared properties of the L1 and L2 (abbreviated as “L1/L2”);
C. Appeal to universal principles (abbreviated as “UP”);
D. Exploit properties of the L2 (abbreviated as “L2”);
The notion that contact phenomena proceed from the interaction of these four types of contribution is hardly new. Most scholars would agree that some outcomes are the result of heavier reliance on L1 knowledge, while others are the result of reliance on the L2, and still others result from accommodations between L1 and L2, or from innovations that reflect neither L1 nor L2 features (which the author ascribes to “universal principles”). However, there is no real attempt to spell out precisely how these strategies come into play in actual situations of language contact. As a result, the model does not live up to the promise that “it will allow us to see whether the same factors are responsible for specific options in different contact settings” (Section 6). In fact, we are never told what processes or mechanisms are involved in the “speaker optimization strategies” mentioned in the abstract and in Section 1.4, and referred to throughout the paper. Neither are we told exactly how each type of optimization strategy contributes to a particular outcome of contact. For instance, in discussing code switching (CS) patterns, we are told that one pattern, insertion, results from using the grammatical and lexical properties of the first language. A second pattern, congruent lexicalization, results from producing structures and words which share the properties of the L1 and L2. And so on. These are vague descriptions of what speakers do with the two languages, not explanations of how the mixture comes about. Similarly, for creole formation, we are told (in Section 2.3) that the optimization of L1 leads to “relexification or transfer of L1 structures”, while optimization of L2, which is described as “imitation of European vernacular varieties”, leads to creoles which resemble European settler languages. No explanation is given of what linguistic processes or principles are involved in each case. Any model of contact-induced change must explain clearly what mechanisms lead to specific innovations or outcomes. In the absence of such explanation, it is hard to see how the quadrangle model adds anything to our understanding of variation in the outcomes within specific domains such as CS or creole formation. We have just as little basis for understanding how similar kinds of mechanisms produce similar kinds of contact-induced changes across different contact situations.
To compound the problem, Muysken uses the same terms to describe contact phenomena in very different situations, without justifying whether such terms refer to the same processes or results in all cases. For instance, the term “relexification” is used to describe the workings of substrate influence in creole formation as well as the origins of mixed languages like Media Lengua. Is the same process involved in both cases? In what sense do the two phenomena result from the same strategy of “optimizing L1”? Similarly, the term “convergence” is used to describe phenomena such as the changes from OV to VO order in some varieties of Quechua and from VO to OV in the Spanish varieties with which Quechua is in contact. The same term is used to describe the combination of French NP structure with Cree VP structure in Michif. The implication that such different phenomena arise from identical processes is hardly a help.
The same problems apply to Muysken's “universal principles” – one of the cornerstones of the model. However, the term is used inconsistently. In a very brief discussion, the author describes universal principles as “general combinatory principles governing improvised language behavior”, listing such examples as reduplication, topic/comment structures, the use of one form for one meaning, and so on (Section 2.2). But in discussing creole formation, he appeals instead to Bickerton's claim that the creole features directly reflect a Language Bioprogram or UG (see Section 2.3). This notion has long been abandoned by most creolists since there is no support for it on either linguistic or sociohistorical grounds. The end result is that we are left unsure what “universal principles” apply in each case of contact, and whether they are indeed the same in all cases. Muysken's formulation of the notion thus misses an opportunity to build on more recent work that relates the universal principles at work in creole genesis to those that operate, for instance, in Second Language Acquisition (SLA), and to compare and contrast them with those that operate in CS or in the formation of bilingual mixed languages. When one thinks of creole genesis, one thinks of universal processes by which learners produce structures that are easier to process (Piennemann, Reference Pienemann1999; Plag, Reference Plag2008), or of internally motivated and contact-induced grammaticalization (Heine & Kuteva, Reference Heine and Kuteva2005; Winford, Reference Winford, Schreier and Hundt2013). One also thinks of the role of “transfer” or imposition, similar to what we find in SLA (Siegel, Reference Siegel2008; Winford, Reference Winford2008). All of these operate according to a number of universal principles which we are beginning to understand much better. None of these applies to classic CS, or to the creation of bilingual mixed languages. Considerations like these render the notion of “optimizing universal principles” as used in the quandrangle model too vague to be useful as a way of finding common ground across contact phenomena.
The task of modeling and interpreting language contact phenomena which Muysken undertakes requires a theoretical framework that addresses, among other things, the nature of the processes underlying contact-induced change. Yet Muysken himself acknowledges that his model fails to make it clear whether “a given strategy reflect[s] a resulting state or an ongoing process” (Section 6). In addition, a model must address both the actuation and implementation of change. This in turn necessitates investigation of both the individual and the community at large. Muysken's model, as he himself acknowledges, makes no distinction between the level of the individual's language behavior and that of the community's language. This touches on a fundamental issue in the study of contact-induced or any other type of linguistic change, namely the complementary roles played by the individual and the community in the origin and spread of change. If, as is generally agreed, the locus of actuation (of a contact-induced innovation) is the individual bilingual, then the mechanisms or processes of change have to be explained in terms of how linguistic systems or inputs interact in the individual mind – i.e., in psycholinguistic terms. At the same time, if the locus of the propagation of change is the set of networks that link individual to individual, then this aspect of change must be explained in sociolinguistic terms (Bachus Reference Bachus2009). By not attending to these distinctions Muysken's model imposes a heavy burden on those who would use it to compare different kinds of contact situations and the processes of change within them.
With regard to the mechanisms involved in the actuation of contact-induced changes, Muysken is right to note that the similarities and differences in the outcomes lie in the way bilinguals implement certain strategies, and that these in turn relate to the roles played by the two languages. But his model does not distinguish clearly among these roles, and consequently, it confounds many different types of change. For instance, the following are all treated as instances of the optimization of L1 strategies:
(a) “insertion” in code switching
(b) “relexification” in creole genesis
(c) “relexification” in the case of Media Lengua
(d) replacement of a native affix by a foreign equivalent
(e) “classical” relexification, the grafting of an L2 phonetic form onto an L1 lemma
(f) the phonological adjustment of loanwords
In a pyscholinguistically-oriented framework such as that of van Coetsem (Reference van Coetsem1988, Reference van Coetsem2000), all of the above phenomena except (b) would be treated as cases of borrowing, accompanied by adaptation to the recipient language (RL) structure. The phenomenon in (b), by contrast, would be treated as a case of imposition, also referred to as ‘transfer’ in SLA research, in which the L1 would be the source language. Similarly, Muysken treats all of the following as manifestations of the optimization of L2 strategies:
(g) European input in creole genesis
(h) the adoption of a grammatical pattern from another language
(i) the generalization of Spanish plural markers in Ecuadorian Quechua
(j) straight lexical borrowing
(k) L1 transfer in L2 pronunciation
(l) transfer of L1 pragmatics into L2 behavior
Of these, it is unclear what process or mechanism is involved in (g)–(h), though it is clear that the source of the input is the L2 in both cases. On the other hand, in the case of (i)–(j), the mechanism that is involved is clearly borrowing, which again presupposes the agency of RL speakers. By contrast, (k)–(l) involve the mechanism of imposition, with L1 acting as source language.
In short, the quadrangle metaphor or model seems to me to confound or group together a number of processes and mechanisms of change that are not in fact the same. It does not show how each of the four optimization strategies operates to produce the relevant outcomes of contact-induced change. It also fails to separate the outcomes from the processes that bring them about, and relate the processes explicitly to the strategies that speakers employ. It therefore falls short of serving as a basis for a unified model of contact-induced change. Instead, it may be considered a template for the empirical investigation of different situations of language contact.