Sage Journals: Discover world-class research

Abstract

Aims and objectives:

Code-switching and calquing are two widespread language contact phenomena in bilingual speech. While both phenomena have been discussed extensively in research on language contact in the past decades, only very few studies systematically investigate code-switching and calquing of and within multiword units or constructions. In our contribution, we aim at developing a more differentiated account of code-switching and calquing of and within multiword units and constructions, bringing together recent usage-based psycholinguistic and Construction Grammar theories and modeling with cognitively oriented approaches to language contact and bilingualism.

Design:

We analyze and discuss corpus data from the Corpus of Spanish in Southern Arizona (CESA) from first-, second-, and third-generation bilingual Spanish–English speakers in Arizona.

Data and analysis:

The corpus analysis focuses on code-switching and calquing of and within N Prep N, NN(N), and Adj N/N Adj patterns (N = noun, Prep = preposition, Adj = adjective).

Findings/conclusions:

The analysis highlights the importance of both lexically specific multiword units and partially and fully schematic constructions in code-switching and calquing as well as a continuum between constructions situated closer to the lexical or to the syntactic pole.

Originality:

Whereas multiword sequences have been found to play an important role in (mostly first language [L1]) acquisition, use, and processing, the role of multiword units and constructions in bilingual speech has been much less studied so far.

Significance/implications:

The study shows that both lexically specific constructions or multiword units as well as partially or fully schematic constructions play a central role in code-switching and calquing. The study argues that code-switching and calquing may ultimately be viewed as a continuum, since the speakers rely on similar, highly frequent, and productive constructions and their interlingual correspondences for both phenomena.

Keywords

Language contact code-switching calquing multiword unit Construction Grammar

Introduction: language contact phenomena

In traditional contact linguistics, code-switching and calquing have generally been treated as two distinct language contact phenomena. Code-switching traditionally refers to “[. . .] the juxtaposition within the same speech exchange of passages of speech belonging to two different grammatical systems or subsystems” (Gumperz, 1982, p. 59). By contrast, calquing includes loan translations or loan formations that are coined with material from one language according to a model in the other language (cf. Backus & Dorleijn, 2009).

Moreover, for code-switching and calquing, different classifications have been suggested that generally rely on a distinction between lexicon (or “words”) and syntax, that is, linguistic structure. For example, in his widely known typology of code-switching, Muysken (2000; who himself uses the term “code-mixing”) distinguishes between “insertion” and “alternation”: Whereas “insertion” refers to single-word switches but may also include fully switched noun phrases (NPs) or prepositional phrases (PPs) (cf. Ex. 1) (Muysken, 2000, p. 61), the notion “alternation” is used for switches in grammar and lexicon (cf. Ex. 2) “in which the two languages present in the clause remain relatively separate” (Muysken, 2000, p. 96):

(1) Yo anduve IN A STATE OF SHOCK por dos días. (Pfaff, 1979, p. 296)

“I was in a state of shock for two days”

(2) Sometimes I’ll start a sentence in Spanish y TERMINO EN ESPAÑOL. (Poplack, 1980)

“Sometimes I’ll start a sentence in Spanish and finish in Spanish.”

A third type of code-switching distinguished by Muysken is called “congruent lexicalization,” in which the grammars of both languages are more strongly intertwined due to structural and/or semantic convergence. Congruent lexicalization includes, for example, mixed collocations (e.g., Gibraltar Spanish ponerse on strike “go on strike”) or compounds (e.g., Australian German Beachlandschaft “beach landscape”) (cf. Muysken, 2000, pp. 146-150).

For calquing, Silva-Corvalán (1994, pp. 171-184) distinguishes in her classical study on Spanish in Los Angeles between “single-word calques” (i.e., loan translations such as US–Spanish aplicación “application,” instead of Standard Spanish solicitud), “multiple-word calques” (e.g., US–Spanish máquina de contestar<English answering machine) and “lexico-syntactic calques,” which may also alter grammatical features of the replica language (e.g., US–Spanish esperar por alguien<English wait for someone, instead of Standard Spanish esperar a alguien). Moreover, calquing may also affect “bound collocations, idioms and proverbs” (e.g., US–Spanish Eso está bien conmigo. That’s fine with me.; cf. Silva-Corvalán, 1994, p. 172).¹

As can already be seen from this small number of examples for Spanish–English code-switching and calquing and their suggested classifications, there is in fact no clear-cut distinction between “lexical” and “structural” or “grammatical” contact phenomena, nor between the “single-” and “multiple-words” level. Instead, both code-switching and calquing may involve “mixed” or “calqued” multiword sequences as well as both lexical and functional material.

Going beyond the traditional accounts of code-switching and calquing, the overall goal of our contribution is therefore to explore in a more systematic way the role of “multiword units” (Wray, 1998) or “constructions” (Booij, 2010; Goldberg, 2006) for both types of language contact phenomena. We attempt to show that the conceptual tool of “constructions” as well as their usage-based profile (e.g., in terms of frequency and productivity), can be used in a holistic way to account for different types of single- and multiword switches and calques, including “insertions,” “congruent lexicalizations” as well as “multiple-word” and “lexico-syntactic calques.” To this aim, we bring together usage-based psycholinguistic and Construction Grammar theories (e.g., Arnon & Snider, 2010; Masini, 2009) with recent, more cognitively oriented approaches to language contact and bilingualism (e.g., Backus et al., 2019; Filipović & Hawkins, 2018). These theories highlight the role of multiword units or constructions at the crossroads of lexicon and syntax for linguistic knowledge and processing and ultimately allow for a more unified account of code-switching and calquing within cognitively oriented contact linguistics (cf. also Wiesinger, in press for an overview).

The remainder of this contribution will be structured as follows: In section “Multiword units in language processing and Construction Grammar,” we will outline the importance of multiword units or constructions in both psycholinguistic and constructionist approaches. In section “Corpus analysis”, we will then discuss the role of a number of nominal (e.g., N Prep N, NN or Adj N/N Adj) constructions for analyzing code-switching and, to a more limited extent, also calquing among bilingual Spanish–English speakers in Arizona. The major findings will be summarized in section “Conclusion and outlook on research perspectives.”

Multiword units in language processing and Construction Grammar

Over the past decades, the importance of multiword units for linguistic knowledge, usage, and processing has been highlighted from different theoretical and methodological perspectives, including psycholinguistic as well as Construction Grammar approaches.

In psycholinguistic studies, multiword units are commonly defined as multimorphemic, “prefabricated,” or “chunked” sequences. This means that they are stored and retrieved as a whole in language processing and not generated by syntactic rules (cf. Christiansen & Arnon, 2017, p. 544). More traditional “dual-system” (or “words-and-rules”) theories generally assume that stored forms correspond to units in the mental lexicon, which can be retrieved as such during lexical access, whereas grammatical or syntactic structures are not stored as a unit but have to be computed during language processing. By contrast, “single-system” theories assume similarities between “compositional” and “non-compositional” sequences in language acquisition and processing. A number of recent studies suggest that there is in fact no clear dividing line: Speakers can holistically access frequent, recurrent and thus expected “compositional” structures (e.g., Arnon & Snider, 2010; Cappelle et al., 2010; Snider & Arnon, 2012). Moreover, phrasal frequency or n-gram effects were found for lexicalized compounds as well as for above-word-level sequences in language processing (e.g., Bannard & Matthews, 2008; Arnon & Snider, 2010; Hennecke & Baayen, 2021). These effects are interpreted, for example, by Kuperman et al. (2008), as the speakers’ probabilistic knowledge about co-occurrence frequencies.

A theory of linguistic knowledge that also highlights the role of multiword units and systematically integrates a gradual cline between lexical and syntactic sequences is usage-based Construction Grammar. According to the latter, linguistic knowledge is fully organized in the form of “constructions,” that is, learned form–meaning pairings or symbolic units. These comprise both “regular” and “idiosyncratic” (or “non-compositional”) structures, which are situated along a lexicon–syntax continuum (cf. Croft, 2001, pp. 14ff.; Goldberg, 2006, pp. 4f.). Moreover, usage-based Construction Grammar assumes that constructions come to be cognitively fixed (or “entrenched”) on the basis of cognitive abilities such as routinization and statistical learning (cf. Bybee, 2010; Langacker, 2008; Tomasello, 2003). According to these principles, linguistic knowledge is, metaphorically speaking, conceived of as an associative network, which is constantly shaped through the speaker’s experience with language in social interaction. Within this network, more schematic constructions are gradually formed out of more concrete, lexically filled ones, to which they relate via hierarchical (or taxonomic) relations or “inheritance links” (cf. Goldberg, 1995, p. 72). As an example, Figure 1 shows the taxonomically related levels for the Italian N Prep N construction, which reach from fully lexically specified constructions at the bottom to fully schematic constructions at the top level (cf. Masini, 2009, p. 262).

Figure 1.

Partial taxonomic network sketch for the [N₁ Prep N₂] construction in Italian (cf. Masini, 2009, p. 262).

As stated above, usage-based Construction Grammar does not assume a strict separation between lexical and grammatical constructions. In this vein, N Prep N is not only a highly frequent schema for Romance word formation but also in syntactic sequences. More specifically, the partially filled mid-level construction N de N represents the most frequent and most productive construction for both more lexical units and more syntactic sequences in Spanish (e.g., libro de bolsillo “pocketbook,” entrevista de trabajo “job interview” vs. libro de Juan “Juan’s book,” entrevista de televisión “televized interview”; cf. Hennecke & Baayen, 2017, 2021)

In total, “single-system” psycholinguistic approaches and usage-based Construction Grammar share the assumption that there is no clear-cut distinction between words as holistically stored units (with phonological content and lexical meaning) and grammar as (potentially more) schematic patterns. Instead, they assign “multiword units” or “constructions” a central role in linguistic representation and processing, which comprise both “compositional” and “non-compositional” as well as lexical(ized) and grammatical(ized) sequences. Both research strands concur that usage-based principles such as the knowledge of statistical (co-)occurrence and, more generally, probabilistic expectations have a strong impact on language learning and processing. However, while psycholinguistic approaches predominantly highlight the ease of activation of (mostly lexically filled) multiword units in language processing, usage-based Construction Grammar assumes that frequency and productivity also determine the routinization (or entrenchment) of constructions as cognitive entities at different levels of schematicity and abstractness.²

Despite these differences, we also observe a slightly growing convergence between the two approaches in recent years. First, the assumption of highly schematic generalizations such as N Prep N has been increasingly questioned in usage-based constructionist approaches, since mid-level constructions were often found to be more predictive with regard to specific types of slot fillers and semantic properties in corpus-based and experimental studies (cf., for example, Pijpops et al., 2021 on English N after N, N by N and similar constructions as well as Perek, 2015 and Dux, 2020 on English argument structure constructions). At the same time, psycho- and neurolinguistic researchers such as Blumenthal-Dramé (2017, pp. 142-145) claim that “it would be interesting to find out […] how these levels interact, and how this interaction can be captured in statistical terms.” Second, and more generally, it is increasingly debated in Construction Grammar whether constructions should be conceived of as cognitively realistic representations or merely as “useful fiction” (Hilpert, 2021). So far, it is an open question whether they account for the fact that frequency in conjunction with other factors speeds up processing and thus essentially strengthens associative links in a more complex, dynamic, and emergent fashion (e.g., Diessel, 2019; Divjak & Caldwell-Harris, 2015; Hilpert & Diessel, 2016; Schmid, 2020).

Whereas a desirable further alignment between psycholinguistic and constructionist approaches with regard to these issues has to await further research, studies on language contact phenomena have also increasingly begun to turn toward cognitive aspects of language usage and processing in recent years. For example, Filipović & Hawkins (2018) and Filipović (2019) argue in favor of bilingualism as a complex adaptive system, which is guided by communicative efficiency and characterized by a constant balancing between maximizing expressiveness and minimizing processing costs. From a constructionist perspective, Höder (2014a, 2014b, 2018) and Höder et al. (2021) assume that bilinguals and second language (L2) learners may form converging, language-unspecific “diaconstructions” that can be flexibly filled with material from either language. Backus (2015, 2019) and Backus et al. (2019) advocate for a usage-based account of different contact phenomena that overcomes a strict dichotomy between lexicon and syntax and also takes into account frequency effects of multiword units (cf. also Backus & Dorleijn, 2009; Hakimov & Backus, 2021). In this regard, it was, for example, shown for different language pairs that various multiword units such as compounds, analytic verbs, idioms, prepositional phrases, or adjective-modified noun phrases are a frequent locus of code-switching and/or calquing (cf. Backus, 2003; Backus & Verschick, 2012; Balam et al., 2014, 2020; Hakimov, 2016, 2021; Treffers-Daller, 2005; Verschik, 2019).³ In line with the approaches outlined in this section, our analysis will, in a first step, primarily explore the conceptual tools of “multiword units” and “constructions” at different levels of schematicity and abstractness to provide a more holistic view of different types of code-switching and calquing phenomena among Spanish–English bilingual speakers in Arizona.

Corpus analysis

Data and sociolinguistic information

The corpus analysis is based on data from the Corpus del Español en el Sur de Arizona (Corpus of Spanish in Southern Arizona [Carvalho 2012-, CESA]). The corpus contains 60 semi-structured sociolinguistic interviews, each approximately 60 minutes, which were recorded between 2012 and 2014 and add up to an overall corpus size of approximately 500,000 words. The interviews were conducted by graduate and undergraduate students (mostly L2 learners) in Spanish, but participants switched to English whenever they wanted to. The interviewees were first-, second-, and third-generation bilingual Spanish–English speakers, living in the Tucson area, in the south of Arizona, close to the US–Mexican border. First-generation speakers were born in Latin America and immigrated into the United States as teenagers or young adults. Second- and third- (or higher-) generation speakers were born in the United States, whereas (one of) their parents or grandparents were born in Latin America. In the interviews, the speakers talked about themselves and their community. The corpus additionally provides ample information on the sociolinguistic profile, the language biography as well as on the linguistic competences of the speakers, which were collected via two self-reporting questionnaires.⁴

The data in CESA are already annotated for some types of language contact phenomena. To establish the dataset for the analysis, all tokens that were annotated as préstamos “loanwords” and cambio de código “code-switching” were extracted from the corpus.⁵ This led to a first dataset of 3,501 occurrences. In a second step, all nominal multiword units were manually extracted (cf. also section “Data analysis” below) and proper names were excluded from the analysis. This second dataset includes 435 occurrences of nominal multiword units. These 435 tokens represent 208 types, which form the basis of the qualitative analysis. In addition, the systematically extracted examples for this type of code-switching are compared to a smaller number of multiword calques that had to be manually extracted from the corpus, since it was not annotated for the latter kind of language contact phenomenon.

As Table 1 shows, the 435 tokens of nominal multiword units are found in 47 out of the totality of 60 interviews included in CESA and are not evenly distributed among the recorded speakers (the codes refer to the interview numbers in the corpus; E [Spanish entrevistador] stands for the interviewer and P [Spanish participante] for the interviewee).⁶

Table 1.

Speaker information and token distribution in CESA.

Speaker code	Number of tokens of nominal code-switching	Generation and age	Language use with friends (outside family)⁷	Competence in speaking Spanish and English (6 = very good, 1 = very poor)⁸
CESA001 P	3	3rd, 50–59	0% Spanish, 100% English	Spanish 4, English 6
CESA003 E	1	–
CESA003 P	2	2nd, 20–29	50% Spanish,100% English	Spanish 5, English 5
CESA005 P	2	3rd, 20–29	70% Spanish, 90% English	Spanish 5, English 6
CESA006 E	4	–
CESA006 P	11	3rd, 40–49	20% Spanish, 80% English	Spanish 4, English 6
CESA007 E	1	–
CESA007 P	4	2nd, 20–29	50% Spanish, 50% English	Spanish 6, English 6
CESA008 P	8	3rd, 50–59	0% Spanish, 100% English	Spanish 0, English 6
CESA009 E	1	–
CESA009 P	1	2nd, <20	15%–20% Spanish, 75%–80% English	Spanish 4, English 5
CESA010 E	6	–
CESA010 P	18	3rd, 20–29	1% Spanish, 99% English	Spanish 1, English 6
CESA011 E	2	–
CESA011 P	7	3rd, <20	0% Spanish, 100% English	Spanish 4, English 5
CESA013 E	2	–
CESA013 P	3	2nd, 30–39	25% Spanish, 75% English	Spanish 6, English 6
CESA014 E	10	–
CESA014 P	8	2nd, 20–29	75% Spanish, 25% English	Spanish 6, English 3
CESA015 P	1	1st, >60	100% Spanish, 0% English	Spanish 6, English 1
CESA016 P	2	2nd, 20–29	30% Spanish, 70% English	Spanish 4, English 6
CESA017 P	4	1st, 20–29	90% English, 10% Spanish	Spanish 5, English 4
CESA022 P	3	2nd, 20–29	50% Spanish, 50% English	Spanish 5, English 6
CESA023 E	4	–
CESA023 P	6	3rd, 20–29	10% Spanish, 90% English	Spanish 6, English 6
CESA024 P	10	3rd, 20–29	50% Spanish, 50% English	Spanish 3, English 4
CESA025 P	7	3rd, <20	20% Spanish, 80% English	Spanish 4, English 6
CESA027 P	2	3rd, 20–29	10% Spanish, 90% English	Spanish 3, English 6
CESA028 P	1	2nd, <20	0% Spanish, 100% English	Spanish 5, English 6
CESA030 P	2	1st, 20–29	50% Spanish, 50% English	Spanish 6, English 6
CESA031 E	1	–
CESA031 P	8	2nd, 20–29	50% Spanish, 50% English	Spanish 4-5, English 5
CESA032 E	4	–
CESA032 P	26	2nd, 20–29	20% Spanish, 80% English	Spanish 5, English 6
CESA033 E	2	–
CESA033 P	15	2nd, 20–29	30% Spanish, 70% English	Spanish 5, English 6
CESA034 P	2
CESA035 P	11	1st, 20–29	50% Spanish, 50% English	Spanish 5, English 6
CESA036 P	3
CESA037 E	7	–
CESA037 P	12	3rd, 39–39	60% Spanish, 80% English	Spanish 5, English 6
CESA038 E	6	–
CESA038 P	5	2nd, 30–39	40% Spanish, 60% English	Spanish 4, English 6
CESA039 E	2	–
CESA039 P	2	1st, n.i.	100% Spanish, 100% English	Spanish 6, English 6
CESA041 E	6	–
CESA041 P	11	2nd, 20–29	80% Spanish, 20% English	Spanish 6, English 6
CESA042 E	2	–
CESA042 P	5	2nd, 30–39	25% Spanish, 75% English	Spanish 6, English 6
CESA044 P	10	2nd, 20–29	40% Spanish, 60% English	Spanish 4, English 5
CESA045 E	3	–
CESA045 P	29	2nd, 20–29	80% Spanish, 20% English	Spanish 6, English 5
CESA046 E	9	–
CESA046 P	28	2nd, 20–29	20% Spanish, 80% English	Spanish 6, English 5
CESA047 E	1	–
CESA047 P	3	3rd, 20–29	25% Spanish, 75% English	Spanish 4, English 6
CESA048 E	2	–
CESA048 P	8	2nd, 20–29	0% Spanish, 100% English	Spanish 6, English 6
CESA049 P	14	2nd, 20–29	20% Spanish, 80% English	Spanish 5, English 5
CESA050 E	1	–
CESA050 P	5	3rd, 20–29	3% Spanish, 100% English	Spanish 2, English 6
CESA051 E	1	–
CESA051 P	9	3rd, 20–29	2% Spanish, 100% English	Spanish 5, English 6
CESA052 E	5	–
CESA052 P	2	2nd, 20–29	5% Spanish, 95% English	Spanish 4, English 6
CESA053 P	18	3rd, 20–29	100% Spanish, 100% English	Spanish 4, English 6
CESA055 P	7	2nd, 20–29	100% Spanish, 0% English	Spanish 6, English 5
CESA058 P	1	2nd, 20–29	10% Spanish, 90% English	Spanish 6, English 6
CESA059 E	1	–
CESA068 E	1	–
CESA068 P	4	2nd, 20–29	99% Spanish, 5% English	Spanish 6, English 4
CESA069 E	1	–
CESA069 P	6	1st, 20–29	40% Spanish, 80% English	Spanish 6, English 6
Total	435

According to Table 1, most code-switching tokens (349 out of 435, i.e., more than 80%) are found in the interviewees’ (P) speech; however, a certain number of switches also come from the interviewers (E) and will generally not be taken into account in the corpus analysis (when we refer to such examples, this will be explicitly marked below). Most of the interviewees in Table 1 belong to the second- (23 persons) or third- or higher generation (15 persons), that is, they are born in the United States and have parents or grandparents that emigrated from a Spanish-speaking country. Among the 11 speakers that produce more than 10 code-switching tokens (which are also those with the highest number of different types), we find 6 (relatively young) second-generation speakers that indicate fairly balanced competences in speaking Spanish and English (CESA032 P, CESA033 P, CESA041 P, CESA045 P, CESA046 P, and CESA049 P). Four interviewees belong to the third- (or higher) generation (CESA006 P, CESA010 P, CESA037 P, and CESA053 P) and one speaker forms part of the first generation (CESA035 P); with the exception of the participant CESA010 P, these speakers also indicate fairly balanced competences in speaking Spanish and English. However, the sociolinguistic data also show that the speakers differ in other aspects as, for example, the regular use of Spanish and English in their everyday life outside the family, for example, with friends.

The extracted data for the specific multiword units thus seem to confirm the previously found tendency that code-switching generally requires relatively balanced competencies in the respective languages, as typically attributed to second-generation bilinguals (e.g., Poplack, 1980; Silva-Corvalán, 1994). Furthermore, the inter-speaker differences regarding the amount of code-switching does not seem to depend on speaker-related sociolinguistic variables such as language use outside the family. Instead, it may rather be explained in terms of inter- and intra-speaker priming and accommodation effects and/or specific characteristics of the individual interviews (e.g., the topics discussed), the exploration of which is not the aim of the present study (cf. also footnote 3 above).

Data analysis

The following corpus analysis focuses on the role of different types of constructions, namely, N Prep N, NN(N), and Adj N/N Adj, for phenomena of code-switching and calquing among bilingual speakers of Spanish and English in Arizona. In the analysis, we distinguish between switches of these constructions as a whole (cf. section “Switching of constructions as a whole”) and switches within these constructions (cf. section “Switching within constructions”). This classification cuts across the traditional distinctions between “lexical” and “structural/grammatical” contact phenomena as well as between the “single-” and “multiple-words” level (cf. section “Switching and calquing as a continuum”) and further ensues the view to a more unified understanding of both code-switching and calquing in bilingual speakers. To this aim, our analysis focuses on the respective properties of the constructions or multiword units, namely, their level of schematicity and abstractness (cf. section “Introduction: language contact phenomena”): For the level of schematicity, we distinguish between fully schematic, partially filled, and lexically filled constructions. The degree of abstractness considers the question of whether the construction or multiword unit as a whole can be situated closer to the lexical or the syntactic pole, and, if the switches within the construction concern more lexical or functional constituents.

As outlined in section “Data and sociolinguistic information” above, the analysis is mainly based on the occurrences of nominal multiword units that were systematically extracted from the CESA corpus and manually coded according to the distinction between fully switched constructions and switches within constructions. Table 2 visualizes the underlying coding of the extracted tokens:

Table 2.

Distribution of the tokens in the data.

Code-switching of the whole construction	Construction-internal code-switching	Specification or translation
320	75	40

As shown in Table 2, the majority of the data consist of occurrences where the whole construction or multiword unit is switched. Still, in 75 occurrences, code-switching occurs within the construction. Another 40 occurrences were coded separately, as they fulfill the purpose to specify an aspect in the other language or to translate an already uttered construction.

The same tendency can be seen in the distribution of the underlying types (Table 3).

Table 3.

Distribution of the types in the data.

Code-switching of the whole construction	Construction-internal code-switching	Specification or translation
211	65	34

From a total of 310 types, 211 cases represent code-switching of the whole construction and only 65 cases show construction-internal code-switching. The 34 types that fulfill the purpose of a specification or a translation were coded separately.

All of these language contact phenomena in multiword units will be described by means of examples from the data in the following qualitative analysis, which will be compared to selected examples for calquing in section “Switching and calquing as a continuum.”

Switching of constructions as a whole

As mentioned above, the dataset contains a majority of nominal multiword units that are switched as a whole, which confirms earlier studies on Spanish–English code-switching that have shown that speakers have a strong tendency to switch at the sentence or phrase boundary both in naturalistic and experimental settings (e.g., Beatty-Martínez et al., 2020, p. 7; Poplack, 1980, p. 602). Moreover, it is possible to distinguish different constructions and levels of schematicity and abstractness within our data:

The first group of fully switched constructions are NN and Adj N constructions that form a highly productive pattern in English but are rare in Spanish. Accordingly, NN and Adj N constructions usually occur as fully switched English constructions in the corpus data. As can be seen in the following examples, they include strongly lexicalized (and, possibly, culturally specific) and/or (more or less) collocational and recurrent multiword sequences:⁹

(1) Voy a ser BORDER PARTOL (CESA045 P)

“I’m going to be (a) border patrol”

(2) ella fue la primera que se graduó de la HIGH SCHOOL y también de la universidad. (CESA044 P)

“she was the first to graduate from high school and also from university”

(3) hicimos un CASE STUDY y desde ahí me ha gustado (CESA044 P)

“we did a case study and since then I have liked it”

(4) Eso es mi gran EXISTENTIAL CRISIS (CESA069 P)

“That’s my big existential crisis”

(5) tenían que checarme mi CRIMINAL BACKGROUND (CESA046 P)

“they had to check my criminal background”

(6) se usa un un SPECIAL WIRE (CESA046 P)

“you use a a special wire”

This kind of code-switching is not restricted to NN or Adj N units, but also occurs in more complex nominal compounds that also follow English patterns, such as:

(7) es una una NINE MILIMETER GLOCK y alrededor de cuarenta y cinco balas tienes (CESA046 P)

“it’s a nine millimeter Glock and you have about forty-five shots”

(8) empecé /uh-huh/ en el food bank, en el COMMUNITY FOOD BANK /ah, okay, uh-huh/, bueno fuimos una vez (CESA068 P)

“I started in the food bank, in the community food bank, well, we went there once”

While Example 8 could be considered a more lexicalized complex construction in English (with 73 occurrences in the COCA), the case seems a bit different in Example 7. Here, we see a construction containing the type of the gun as well as the proper name of the arms manufacturer that refers to the gun. In that sense, Example 7 depicts a less frequent and less lexicalized way to express “a nine millimeter gun from the brand Glock” (only two occurrences in the COCA).

In comparison to the large number of Adj N and NN(N) switches to English, syntactically larger switches are much less frequent. These may involve chunk-like multiword constructions such as the moral of the story (306 occurrences in the COCA), but also longer switches that are more closely situated at the syntactic pole, such as in Examples 10 and 11:

(9) ¿Cómo? THE MORAL OF THE STORY era que ell-ella se tenía que cuidar ella sola (CESA069 P)

“What? The moral of the story was that she had to take care of herself”

(10) no sé exactamente mucho OF THE MECHANICS OF THE E-CIGARETTE, pero yo sé que obviamente tienes una batería. (CESA046 P)

“I don’t exactly know much about the mechanics of the e-cigarette, but I do know that you obviously have a battery”

(11) pero hay un SCALE OF SOCIAL DIVISON en Chile (CESA053 P)

“but there is a scale of social division in Chile”

While fully switched English multiword units are very frequent in the Spanish-dominant parts of the corpus data, there are only very few occurrences of fully switched Spanish constructions in the English parts. These generally represent a smaller part of the interviews; in Example 12 below, the fully switched Spanish construction also triggers the continuation of the subsequent discourse in Spanish. The fully switched Spanish constructions involve the highly productive Spanish N Prep N construction and, in particular, the partially filled construction N de N with the functional constituent de (cf. section “Multiword units in language processing and Construction Grammar” above):

(12) And so when they have the like TURISMO DE BICICLETA, NO SÉ LO QUE SEA (CESA050 P)

“And so when they have the like cycling tourism, I don’t know, whatever”

In contrast to the lexicalized and/or highly recurrent English Adj N and NN constructions, turismo de bicicleta has not a single occurrence in the two billion words corpus Corpus del Español: web/dialects (Davies, 2016) with data from 21 different Spanish-speaking countries (including the USA) and thus seems to represent more of an ad hoc formation on the basis of the schematic constructional template.

In the dataset, a specific case of code-switching can be observed in examples in which a fully switched English multiword unit is combined with a complex Spanish construction, as in:

(13) también tienes un una botella de PEPPER SPRAY que es muy potente (CESA046 P)

“you also have a pepper spray bottle which is very powerful”

In Example 13, the English NN construction pepper spray is embedded into a Spanish N Prep N construction, that leads to a mixed N_SP Prep_SP NN_EN construction. The English NN construction specifies and modifies the Spanish head noun in the whole construction. This example may be considered a full switch of a whole construction, when considering the English NN compound, but also a mixed construction, when regarding the complex construction as a whole. Switches within complex constructions will be discussed in more detail in the following section “Switching within constructions”.

Switching within constructions

Most multiword constructions in our corpus that show the insertion of one lexical element from the other language also represent mixed constructions of the type N_SP Prep_SP N_EN, as in:

(14) así, um y hacer dibujos de de FASHION, poner mis dibujos en una bolsa o algo (CESA069 P)

“like, um and make fashion drawings, put my drawings in a bag or something”

(15) no tienen agua del TAP no tienen todo (CESA053 P)

“they don’t have tap water, they don’t have everything”

(16) ella trabajaba como profesora de PRE-SCHOOL (CESA053 P)

“she worked as a pre-school teacher”

(17) Él era él era un veterano del ARMY /aha/ y él creía en mí (CESA045 P)

“He was he was an army veteran and he believed in me”

All of these examples show instances of the highly frequent, partially lexically filled Spanish construction of the type N de N (cf. section “Multiword units in language processing and Construction Grammar” above). The head noun, which in Spanish is typically situated at the left of the construction, and the functional preposition are always uttered in Spanish. Only the modifying noun N2 is filled with an English lexical unit. Semantically, the N Prep N constructions in Examples 14–17 can be situated closer to the lexical pole.

By contrast, other N Prep N structures appear closer to the syntactic pole. In Examples 18 and 19 below, it can be seen that in some of these cases even both nominal elements (as well as possible modifiers) may be taken from English, whereas the prepositional element (en “in” in Example 18 and del “of the” in Example 19) comes from Spanish:

(18) Hago educación de primaria con MINOR en BILINGUAL EDUCATION o algo así (CESA045 P)

“I do primary education with a minor in bilingual education or something like that”

(19) Tenía FORTY PERCENT OFF del SALE (CESA041 P)

“It [The shop] had forty percent off the sale”

While a large number of occurrences of code-switching follow the Spanish N Prep N pattern, other patterns are also possible, albeit much less frequent in the corpus data. In the following examples, an English lexical unit is inserted into the Spanish N Adj construction. In Example 21, the N Adj construction is additionally modified by a prepositional de N structure, in which the noun is again a switched lexical element from English. In Example 22, the speaker first uses the “false friend” cita “appointment” and then corrects to the English word city, which is then modified by a Spanish adjective that is intensified by the Spanish adverbial muy “very”:

(20) voy a … a con mis um … ADVISORS ACADÉMICOS. Ellos me ayudan con mi materia (CESA044 P)

“I go to … to with my um … academic advisors. They help me with my topic”

(21) Un MALL nuevo, aquí en Marana, de OUTLET (CESA042 P)

“A new mall, here in Marana, [an] outlet [mall]”

(22) entonces es una cita, CITY muy histórica donde siempre … siempre escuchas personas hablar sobre ellas (CESA044 P)

“so it is a city [lit. date], [a] very historic city where you always … always hear people speaking of them”

Whereas most intra-constructional switches in the corpus data can be found within Spanish constructions, switches to Spanish within English constructions are only rarely attested in the corpus data and mostly occur within the English NN construction. On the one hand, such code-switching can be observed in the speech of interviewers who are first-language (L1) speakers of English and advanced L2 learners of Spanish. In Example 23, a mixed NN construction with N1 in Spanish and N2 in English (gracias day) is produced when the speaker looks for the equivalent for English Thanksgiving Day (cf. also section “Switching and calquing as a continuum” below). Example 24 represents a co-constructing environment, in which the Spanish N2 organizaciones “organizations” is added by the interviewer:

(23) ¿Tienes planes para (..) Uhm (…) la- LIKE el día de / ¿Gra-gracias?/ gracias DAY? (CESA051 E)

“Do you have plans for … uhm … the like the day of Thanks? Thanks day?”

(24) P: trabajar con … trabajar con … NON-PROFIT (CESA050 P)

“work with … work with … non-profit”

E: ¿Organizaciones? (CESA050 E)

“Organizations?”

P: ¡Sí! ¡Sí! ¡Sí! (CESA050 P)

“Yes! Yes! Yes!”

In the interviewees’ speech, mixed NN(N) constructions seem to occur primarily in contexts where the speaker seeks for a specific expression or for clarification (cf. also section “Switching and calquing as a continuum” below) (Examples 25 and 26). Furthermore, we find some mixed NN constructions with a (partial) doubling of one of the nominal elements (team—equipo in Example 27 and escuela—school in Example 28):

(25) Voy a ir un programa un INTERNSHIP ABROAD programa dentro del colegio (CESA044 P)

“I will go [with] a program an internship abroad program within the school”

(26) salimos como en un DATE, una cita double date y y la muchacha me preguntó (CESA049 P)

“we went out as for a date, a double date [date] and and the girl asked me”

(27) Yo era parte de—del SWIMTEAM equipo. (CESA053 P)

“I was part of—of the swimteam [team]”

(28) yo fui a una escuela PRE-SCHOOL, RIGHT? (CESA053 P)

“I went to a pre-school [school], right?”

Switching and calquing as a continuum

Whereas sections “Switching of constructions as a whole” and “Switching within constructions” have only focused on the role of multiword units in code-switching, this section focuses on the similarities between code-switching and calquing, which may actually be considered as lying on a continuum of bilingual linguistic strategies.

Let us first consider the following fully switched multiword expressions, which represent English repetitions or translations of a Spanish construction with additional information and which are used by the speaker more or less consciously for specifying or exemplification purposes. In Examples 29 and 30 the speaker switches from a fully produced Spanish N Adj construction to an English Adj N or NN construction (and back, as in Example 30):

(29) pues los carros rápidos como los MUSCLE CARS (CESA049 P)

“well, the fast cars like the muscle cars”

(30) como un carro viejito como dicen en inglés OLD SCHOOL o un carro clásico (CESA049 P)

“like an old car as they say in English ‘old school’ or a classic car’

In the following Examples 31–33, the switch to the English Adj N constructions seems to repair an only partially produced Spanish N (Adj) construction:

(31) diciendo con el español /muy así/ BROKEN SPANISH (CESA069 P)

“saying [it] in Spanish (very like) broken Spanish”

(32) pero es ingeniero MECHANICAL ENGINEER SO no es un mecánico (CESA049 P)

“but he is engineer mechanical engineer so he is not a mechanic”

(33) um tener las botas uh SHINY BOOTS (CESA046 P)

“um to have the boots uh shiny boots”

The phenomenon of switched constructional repair or doubling can also be observed with other patterns, for example, starting from English NN(N) constructions, which are then repeated with a Spanish N Prep N (Adj) structure (as in Examples 34 and 35), or vice versa (as in Example 36):

(34) de-de crear en tu salón un LEARNING COMMUNITY una comunidad de aprendizaje (CESA045 P)

“to to create in your living room a learning community a learning community”

(35) me puedo comunicar en AMERICAN SIGN LANGUAGE en lenguaje con signes americano, um, el italiano es muy parecido al español (CESA051 P)

“I can communicate in American Sign Language in American Sign Language, um, Italian is very similar to Spanish”

(36) publicaron mi libro de texto, un TEXT BOOK (CESA041 P)

“they published my text book, a text book”

The fact that speakers seem to be aware of systematic correspondences between English (e.g., NN) and Spanish (e.g., N Prep N) constructions also plays a central role for language contact phenomena that have been described as multiword or lexico-syntactic calques. As outlined in section “Introduction: language contact phenomena,” calquing involves the coinage of multiword units with material from one language according to a model in the other language. In Example 37 below, the speaker reproduces the English NN construction Thanksgiving Day, highly emblematic for US–American culture, with the Spanish N Prep N construction día de dar gracias (instead of Día de Acción de Gracias in Standard Spanish). Similarly, the colloquial US–American NN constructions Turkey Day (also for Thanksgiving Day) and Bunny Day (for Easter) are systematically rendered with Spanish N Prep N constructions (día del pavo and día del conejo):

(37) No sé si ella decoraba para navidad, el día de dar gracias , desde mantel, flores. Cambiaba diferentes cosas en la casa. El día del pavo se hacía la comida el pavo y todo. […] Por ejemplo para el día de San Valentín, ella un corazoncito, un dulcecito algo nos ponía a un lado de la de la almohada de la cabecera con una tarjetita. […] El día de las madres y este, cada estación pues. Para Halloween ir a pedir dulces, disfrazarse, decorar y así. Entonces yo Halloween no lo, no lo celebro mucho. […] Y luego para el día de dar gracias ella se ha dado cuenta pues desde que nació que, hemos hecho el pavo. Bueno, lo hacemos en la casa de mi madre lo celebramos, ¿no? Pero ella sabe el significado. P’al día del conejo sus dulces. Pa’ el día de San Valentín igual. (CESA026 P)

“I don’t know if she decorated for Christmas, Thanksgiving, from tablecloths to flowers. She changed different things in the house. On Turkey Day, the food was made, the turkey and everything. [… ] For example, for Valentine’s Day, she put us a little heart, a little candy, something on one side of the pillow on the headboard with a little card. [… ] For Mother’s Day etc., for every season then. For Halloween going to trick-or-treat, dress up, decorate and so forth. So I don’t, I don’t celebrate Halloween very much. […] And then for Thanksgiving she has realized that since she was born, we have made the turkey. Well, we make it at my mother’s house, we celebrate it, don’t we? But she knows the meaning. For Bunny Day her sweets. For Valentine’s Day the same.”

Example 38 outlined below does not represent a case of calquing in the strict sense, but also clearly shows that the speaker is aware of the constructional correspondences between NN and N Prep N constructions. The speaker P is looking for the Spanish expression for South America, relying on the Spanish N Prep N construction (América de Sur). However, he is corrected by speaker E, who suggests the alternative expression Sur América.¹⁰ In response, speaker P explicitly comments on the fact that Sur América (the structural equivalent to South America in English) corresponds to the reverse (“backwards”) version of the expected N Prep N construction América de Sur:

(38) P: Uum pues sí me gustaría visitar varios lugares […] um, Centro América, um Am- América de Sur ? IS THAT HOW YOU SAY IT? (CESA049 P)

“Uum well yes I would like to visit several places […] um, Central America, um South America? Is that how you say it?”

E: Sur América (CESA049 E)

“South America”

P: Sur América . I WASN’T SURE IF IT WAS BACKWARDS. (CESA049 P)

“South America. I wasn’t sure if it was backwards.”

Both multiword calquing and inter- and intra-constructional code-switching, as discussed in sections “Switching of constructions as a whole” and “Switching within constructions”, thus have in common a reliance on similar constructions and their interlingual correspondences, such as between the highly frequent and productive English NN and the Spanish N Prep N constructions. Moreover, (intra-constructional) code-switching and calquing can be ultimately considered as more or less “creative” or playful slot filling within more schematic patterns, whose instantiations may be situated closer to the lexical or to the syntactic pole. From this perspective, the difference between code-switching and calquing ultimately boils down to the fact that those cases in which the model construction is explicitly attested in one language and then repaired or doubled by the corresponding construction in the other language are usually considered as code-switching. By contrast, those cases in which the model construction in one language is only implicit but reproduced with a construction and lexical material from the other language, are considered as calquing (as in Example 37 above). Our data suggest that similar cognitive mechanisms such as the recourse to high-frequency patterns and interlingual correspondences (or potential “diaconstructions” in Höder’s sense, cf. section “Multiword units in language processing and Construction Grammar”) are at play in both phenomena, in which the speakers make use of the full range of constructions and potential fillers available to them in their multilingual repertoire.¹¹

Conclusion and outlook on research perspectives

The central aim of the present contribution was to investigate the role of multiword units and constructions for code-switching and calquing in a corpus of bilingual speech of first-, second-, and third- (or higher) generation Spanish–English speakers in Arizona. As outlined from a theoretical perspective in sections “Introduction: language contact phenomena” and “Multiword units in language processing and Construction Grammar,” the focus on multiword units and constructions, as established in usage-based psycholinguistics and Construction Grammar, cuts across the traditional distinctions between “lexical” and “structural/grammatical” contact phenomena as well as between the “single-” and “multiwords” level, which is usually adopted in more traditional approaches to code-switching and calquing. From an empirical perspective, our corpus analysis has shown that both lexically specific constructions or multiword units as well as partially or fully schematic constructions play a central role in these language contact phenomena. It reveals mostly switching of and within highly frequent and productive constructions, such as English NN(N) and Adj N and Spanish N Prep N (in particular N de N) and N Adj, which are also involved in phenomena that have been traditionally classified as calquing.

In general, the corpus analysis shows that the speakers in our data tend to switch fully produced constructions, whereas switches within constructions are less frequent. Moreover, these switches are mostly to English since most interviews are predominantly held in Spanish (cf. section “Data and sociolinguistic information”). Fully switched English constructions usually follow the English NN(N) and Adj N patterns and correspond to more or less recurrent multiword units that are usually situated closer to the lexical pole. By contrast, fully switched Spanish constructions following the Spanish N de N template and more complex switches situated closer at the syntactic pole are less frequently attested.

Switching within constructions concerns mostly Spanish N de N constructions situated closer to the lexical pole, in which the first noun and the functional preposition are uttered in Spanish, whereas the second noun is typically switched from English. The switching of nouns in other Spanish templates such as N Adj is found less frequently. Longer switches following the Spanish N Prep N template and situated more at the syntactic pole, in which both nouns may come from English, are also less systematic. Construction-internal switches of the first or second noun from Spanish within the English NN construction are infrequent and mostly found in L2 learners of Spanish and in contexts of hesitation, repetition, or clarification.

In section “Switching and calquing as a continuum”, we have argued that code-switching and calquing may ultimately be viewed as a continuum, since the speakers rely on similar, highly frequent, and productive constructions and their interlingual correspondences for both phenomena. In our corpus data, the bilingual speakers frequently “translate,” duplicate or “repair” English and Spanish constructions, such as in switching from Spanish N Adj to English Adj N/NN or from English NN(N) to Spanish N Prep N constructions for specifying or clarification purposes. The same awareness for systematic correspondences between Spanish and English constructions are also at play in calquing of English NN constructions in the form of Spanish N Prep N constructions, mostly situated at the lexical pole in the lexicon–syntax continuum.

In conclusion, highly frequent and productive constructions such as English NN(N) and Adj N and Spanish N de N and N Adj, which are characteristic for both more lexical and more grammatical constructions in (“monolingual”) English and Spanish, also strongly determine both code-switching and calquing in the data from the bilingual speakers. These constructions are mostly filled with sequences that represent strongly lexicalized or (more or less) collocational and recurrent combinations (“fully switched constructions”), which can be attributed to statistical learning of syntagmatic and paradigmatic associations (i.e., the cognitive entrenchment of constructions). However, there is no clear dividing line between language contact phenomena situated closer to the lexical and syntactic pole, since similar (partially) schematic patterns can be detected at the full range of the lexicon–syntax continuum. The bilingual speakers generally show more systematicity in fully switched or mixed constructions situated closer to the lexical pole. In the case of the former, switched English NN constructions dominate and may also be integrated as a whole into Spanish N Prep N constructions. As for switches within constructions, there is a clear tendency in the corpus data to switch the second noun within Spanish N Prep N constructions to English.

Overall, our analysis thus confirms an important overlap in the domain of fully, partially, and non-schematic constructions (or multiword units) in mono- and bilingual speakers. Although the bilingual speakers’ linguistic knowledge, especially in the domain of lexically specific multiword units, may not exactly correspond to that of monolingual speakers, the former productively draw on their multilingual and interrelated constructional knowledge at different levels of schematicity and abstractness, as attested in code-switching and calquing. These bilingual strategies could thus both be considered as a continuum between creative and routine slot filling within more schematic constructions, for which the bilingual speakers make use of the full repertoire of constructions and potential fillers available to them. Although the resulting language contact phenomena may appear “creative” in comparison to monolingual speech, they ultimately rely on the same cognitive processes as those found in monolingual speakers, namely, the production of recurrent constructions and the more or less flexible and playful filling of their slots with elements that are available in the speaker’s linguistic repertoire.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) received no financial support for the research,authorship,and/or publication of this article.

ORCID iD

Evelyn Wiesinger

Author biographies

Inga Hennecke is an associate professor for Romance linguistics at the University of Tübingen. Her general research interests fall in the areas of language contact,multilingualism,pragmatics,morphology,experimental phonetics,experimental psycholinguistics as well as corpus linguistics.

Evelyn Wiesinger is junior professor for Romance linguistics at the University of Tübingen. Her research focuses on variation and language change,multilingualism,language contact and creole studies and is situated in functional-cognitive as well as usage-based and constructionist approaches.

References

Arnon

Snider

(2010). More than words: Frequency effects for multi-word phrases. Journal of Memory and Language, 62, 67–82.

Backus

(2003). Units in codeswitching: Evidence for multimorphemic elements in the lexicon. Linguistics, 41, 83–132.

Backus

(2014a). Towards a usage-based account of language change: Implications of contact linguistics for linguistic theory. In Nicolaï

(Ed.), Questioning language contact: Limits of contact, contact at its limits (pp. 91–118). Brill.

Backus

(2014b). A usage-based approach to borrowability. In Zenner

Kristiansen

(Eds.), New perspectives on lexical borrowing: Onomasiological, methodological and phraseological innovations (pp. 19–39). Mouton de Gruyter.

Backus

(2015). A usage-based approach to code-switching: The need for reconciling structure and function. In Stell

(Ed.), Code-switching between structural and sociolinguistic perspectives (pp. 19–37). De Gruyter.

Backus

(2019). Code-switching as a reflection of contact-induced change. In Grant

A. P.

(Ed.), The Oxford handbook of language contact (pp. 192–214). Oxford University Press.

Backus

Dorleijn

(2009). Loan translations versus code-switching. In Bullock

B. E.

Toribio

A. J.

(Eds.), The Cambridge handbook of linguistic code-switching (pp. 75–93). Cambridge University Press.

Backus

Verschick

(2012). Copiability of (bound) morphology. In Johanson

Robbeets

(Eds.), Copies versus cognates in bound morphology (pp. 123–149). Brill.

Backus

Winter-Froemel

Zenner

(2019). Introduction: Placing usage, meaning and mind at the core of contact-induced variation and change. In Zenner

Backus

Winter-Froemel

(Eds.), Cognitive Contact Linguistics: Placing usage, meaning and mind at the core of contact-induced variation and change (pp. 1–20). De Gruyter Mouton.

10.

Balam

Parafita Couto

Stadthagen-González

(2020). Bilingual verbs in three Spanish/English code-switching communities. International Journal of Bilingualism, 24(5–6), 952–967.

11.

Balam

Prada Pérez

Mayans

(2014). A congruence approach to the study of bilingual compound verbs in Northern Belize contact Spanish. Spanish in Context, 11, 243–265.

12.

Bannard

Matthews

(2008). Stored word sequences in language learning: The effect of familiarity on children’s repetition of four-word combinations. Psychological Science, 19(3), 241–248.

13.

Beatty-Martínez

A. L.

Navarro-Torres

C. A.

Dussias

P. E.

(2020). Codeswitching: A bilingual toolkit for opportunistic speech planning. Frontiers in Psychology, 11, 1699.

14.

Bencini

Goldberg

(2000). The contribution of argument structure constructions to sentence meaning. Journal of Memory and Language, 43(4), 640–651.

15.

Betz

(1949). Deutsch und Lateinisch: Die Lehnbildungen der althochdeutschen Benediktinerregel. H. Bouvier u. Co.

16.

Blumenthal-Dramé

(2017). Entrenchment from a psycholinguistic and neurolinguistic perspective. In Schmid

H.-J.

(Ed.), Entrenchment and the psychology of language learning. How we reorganize and adapt linguistic knowledge (pp. 129–152). De Gruyter Mouton.

17.

Booij

(2010). Construction morphology. Oxford University Press.

18.

Bybee

J. L.

(2010). Language, usage, and cognition. Cambridge University Press.

19.

Cappelle

Shtyrov

Pulvermüller

(2010). Heating up or cooling up the brain? MEG evidence that phrasal verbs are lexical units. Brain and Language, 115(3), 189–201.

20.

Carvalho

A. M.

(2012-). Corpus del Español en el Sur de Arizona (CESA). http://cesa.arizona.edu

21.

Christiansen

M. H.

Arnon

(2017). More than words: The role of multi-word sequences in language learning and use. Topics in Cognitive Science, 9, 542–551.

22.

Croft

(2001). Radical Construction Grammar: Syntactic theory in typological perspective. Oxford University Press.

23.

Davies

(1990-). Corpus of Contemporary American English (COCA). https://www.english-corpora.org/coca/

24.

Davies

(2016). Corpus del Español: Web/dialects. http://www.corpusdelespanol.org/web-dial/

25.

Diessel

(2019). The grammar network. How linguistic structure is shaped by language use. Cambridge University Press.

26.

Divjak

Caldwell-Harris

C. L.

(2015). Frequency and entrenchment. In Dąbrowska

Divjak

(Eds.), Handbook of Cognitive Linguistics (pp. 53–75). De Gruyter.

27.

Dux

(2020). Frame-constructional verb classes. Change and theft verbs in English and German. Benjamins.

28.

Filipović

(2019). Bilingualism in action. Theory and practice. Cambridge University Press.

29.

Filipović

Hawkins

J. A.

(2018). The complex adaptive system principles model for bilingualism: Language interactions within and across bilingual minds. International Journal of Bilingualism, 23(6), 1223–1248.

30.

Fricke

Kootstra

G. J.

(2016). Primed codeswitching in spontaneous bilingual dialogue. Journal of Memory and Language, 91, 181–201.

31.

Goldberg

A. E.

(1995). Constructions: A Construction Grammar approach to argument structure. University of Chicago Press.

32.

Goldberg

A. E.

(2006). Constructions at work: The nature of generalization in language. Oxford University Press.

33.

Gries

S. T.

Wulff

(2005). Do foreign language learners also have constructions? Evidence from priming, sorting, and corpora. Annual Review of Cognitive Linguistics, 3, 182–200.

34.

Gumperz

(1982). Discourse strategies. Cambridge University Press.

35.

Guzzardo Tamargo

R. E.

Mazak

C. M.

Parafita Couto

. (Eds.). (2016). Spanish-English codeswitching in the Caribbean and the US. Benjamins.

36.

Hakimov

(2016). Effects of frequency and word repetition on switch-placement: Evidence from Russian-German code-mixing. In Robinson

J. A.

Reif

(Eds.), Culture and cognition in bilingualism (pp. 91–125). De Gruyter.

37.

Hakimov

(2021). Lexical frequency and frequency of co-occurrence predict the use of embedded-language islands in bilingual speech: Adjective-modified nominal constituents in Russian-German code-mixing. Journal of Language Contact, 13(3), 501–539.

38.

Hakimov

Backus

(2021). Usage-based contact linguistics: Effects of frequency and similarity in language contact. Journal of Language Contact, 13(3), 459–481.

39.

Hennecke

Baayen

(2017). A quantitative survey of N Prep N constructions in Romance languages and prepositional variability. Quaderns De Filologia: Estudis Lingüístics, 22, 129–146.

40.

Hennecke

Baayen

(2021). Internal constituent variation in visual word recognition: An eye-tracking study of N Prep N constructions in French, Spanish and Portuguese. The Mental Lexicon, 16(1), 98–132.

41.

Hilpert

(2021). Networks all the way down? Reflections on constructions and the links between them. Presentation at the University of Erlangen-Nürnberg.

42.

Hilpert

Diessel

(2016). Entrenchment in construction grammar. In Schmid

H.-J.

(Ed.), Entrenchment and the psychology of language learning. How we reorganize and adapt linguistic knowledge (pp. 57–74). De Gruyter.

43.

Höder

(2014a). Constructing diasystems. Grammatical organisation in bilingual groups. In Åfarli

Mæhlum

(Eds.), The Sociolinguistics of grammar (pp. 137–152). Benjamins.

44.

Höder

(2014b). Convergence vs. divergence from a diasystematic perspective. In Braunmüller

Höder

Kühl

(Eds.), Stability and divergence in language contact. Factors and mechanisms (pp. 39–60). Benjamins.

45.

Höder

(2018). Grammar is community-specific. Background and basic concepts of Diasystematic Construction Grammar. In Boas

H. C.

Höder

(Eds.), Constructions in contact. Constructional perspectives on contact phenomena in Germanic languages (pp. 37–70). Benjamins.

46.

Höder

Prentice

Tingsell

(2021). Additional language acquisition as emerging multilingualism. A Construction Grammar approach. In Boas

H. C.

Höder

(Eds.), Constructions in contact 2. Language change, multilingual practices, and additional language acquisition (pp. 309–337). Benjamins.

47.

Kootstra

G. J.

van Hell

J. G.

Dijkstra

(2010). Syntactic alignment and shared word order in code-switched sentence production: Evidence from bilingual monologue and dialogue. Journal of Memory and Language, 63, 210–231.

48.

Kuperman

Bertram

Baayen

(2008). Morphological dynamics in compound processing. Language and Cognitive Processes, 23(7–8), 1089–1132.

49.

Langacker

R. W.

(2008). Cognitive Grammar. A basic introduction. Oxford University Press.

50.

Masini

(2009). Phrasal lexemes, compounds and phrases: A constructionist perspective. Word Structure, 2(2), 254–271.

51.

Muysken

(2000). Bilingual speech: A typology of codemixing. Cambridge University Press.

52.

Onysko

(2019). Reconceptualizing language contact phenomena as cognitive processes. In Zenner

Backus

Winter-Froemel

(Eds.), Cognitive Contact Linguistics. Placing usage, meaning and mind at the core of contact-induced variation and change (pp. 23–50). De Gruyter Mouton.

53.

Perek

(2015). Argument structure in usage-based Construction Grammar. Benjamins.

54.

Pfaff

C. W.

(1979). Constraints on language mixing: Intrasentential code-switching and borrowing in Spanish/English. Linguistic Society of America, 55(2), 291–318.

55.

Pijpops

Speelman

Van de Velde

Grondelaers

(2021). Incorporating the multi-level nature of the constructicon into hypothesis testing. Cognitive Linguistics, 32(3), 487–528.

56.

Poplack

(1980). Sometimes I’ll start a sentence in Spanish Y TERMINO EN ESPAÑOL: Toward a typology of code-switching. Linguistics, 18(7–8), 581–618.

57.

Poplack

(1983). Bilingual competence: Linguistic interference or grammatical integrity? In Elías-Olivares

(Ed.), Spanish in the US setting: Beyond the southwest (pp. 110–129). National Clearinghouse for Bilingual Education.

58.

Schmid

H.-J.

(2020). The dynamics of the linguistic system. Usage, conventionalization and entrenchment. Oxford University Press.

59.

Silva-Corvalán

(1994). Language contact and change: Spanish in Los Angeles. Clarendon Press.

60.

Smead

R. N.

(2000). Phrasal calques in Chicano Spanish: Linguistic or cultural innovation? In Roca

(Ed.), Research on Spanish in the U.S (pp. 162–172). Cascadilla.

61.

Snider

Arnon

(2012). A unified lexicon and grammar? Compositional and non-compositional phrases in the lexicon. In Divjak

Gries

S. Th.

(Eds.), Frequency effects in language representation (pp. 127–164). De Gruyter Mouton.

62.

Stefanowitsch

Flach

(2016). The corpus-based perspective on entrenchment. In Schmid

H.-J.

(Ed.), Entrenchment and the psychology of language learning. How we reorganize and adapt linguistic knowledge (pp. 101–127). De Gruyter.

63.

Tomasello

(2003). Constructing a language. A usage-based theory of language acquisition. Harvard University Press.

64.

Toribio

A. J.

(2011). Code-switching among US Latinos. In Díaz-Campos

(Ed.), The Handbook of Hispanic sociolinguistics (pp. 530–552). Wiley-Blackwell.

65.

Treffers-Daller

(2005). Evidence for insertional codemixing: Mixed compounds and French nominal groups in Brussels Dutch. International Journal of Bilingualism, 9(3–4), 477–506.

66.

Valenzuela

Rojo

(2008). What can language learners tell us about constructions? In De Knop

De Rycker

(Eds.), Cognitive approaches to pedagogical grammar. Volume in honour of René Dirven (pp. 197–229). Mouton de Gruyter.

67.

Verschik

(2019). English-Estonian code-copying in blogs: Combining a contact linguistic and cognitive approach. In Zenner

Backus

Winter-Froemel

(Eds.), Cognitive contact linguistics. Placing usage, meaning and mind at the core of contact-induced variation and change (pp. 51–80). De Gruyter.

68.

Wiesinger

(in press). Language contact and construction grammar. In Cerno

Döhla

H.-J.

Gutiérrez Maté

Hesselbach

Steffen

(Eds.), Handbook of contact varieties of Spanish and Spanish-lexified contact varieties. De Gruyter Mouton.

69.

Wray

(1998). Protolanguage as a holistic system for social interaction. Language and Communication, 18, 47–67.