1. Introduction
The phenomenon of code switching (CS) is widespread in many Arab countries. Bentahila & Davies (1983) report that code switching is a style of speaking among Arabic-French bilinguals in Morocco. This is also true in many other countries, such as Tunisia and Algeria. In Lebanon, people code switch between either Arabic and English or Arabic and French. A more limited degree of code switching occurs in Egypt and Jordan, where Arabic and English are used.
Myers-Scotton (1993: p. 3) defines code switching as “the selection by bilinguals or multilinguals of forms from an embedded variety (or varieties) in utterances of a matrix variety during the same conversation”. According to Myers-Scotton’s Matrix Language Frame Model, the matrix language is identified as the language that provides the functional morphemes (such as tense) in a code-switched utterance. Myers-Scotton developed the Matrix Language Frame (MLF) model (Myers-Scotton, 1992, 1993) , which suggests an asymmetrical relationship between the matrix language (ML) and the embedded language (EL) (Boztepe, 2003) .
Some research by Kamhawi (2000) shows that code switching is more dominant in the speech of females than males. This is because English and French are viewed as languages that convey sophistication, elegance, beauty, modernity, and social liberation. Females utilize these languages with their Arabic speech through code switching, thus ascribing to themselves some of the characteristics related to a more prestigious group (Suleiman, 2004) .
Code switches, in an in-group context, according to Suleiman (2004: p. 30) “display…identity negotiation whereby the speaker seeks symbolically to ascribe to the self some of the attributes associated with a more prestigious group.” Henkin-Roitfarb (2011: p. 91) claims that in a spontaneous context, based on her observation of Palestinian Israelis of the Negev, CS determines social borders and the minority feelings of ethnic pride that “[w]e can infiltrate your borders because we speak your language, whereas you cannot infiltrate ours.” However, it is not clear on what she has based her interpretation.
These attitudes ascribed to CS prove that the phenomenon goes beyond a simple linguistic practice; it is deeper in meaning and attribution. Code switching indexes group identity and speaker attitudes.
In this study, the focus is on two localities of Arabs in Israel and how code switching is motivated in their everyday conversation and the syntactic structure used when they code switch from Arabic to Hebrew.
2. Data Collection
The spoken data were collected from 10 subjects, five females and five males, whose ages ranged from 25 to 38 years old. The data was longitudinally collected between the years 2016-2018. The participants were contacted through my close personal network, friends and colleagues, and were recorded in a natural speech setting. Casual conversations were recorded at home and work after obtaining the participants’ consent1. A total of six hours of spoken conversation was recorded. The corpus included 2891 sentences, of which 586 have code switches.
3. Methods of Analysis
The transcribed spoken recordings were entered into an Excel spreadsheet and coded for code switches for every speaker. A Python dictionary script was built to calculate the counts of each unique entry.
To code the topical domains, the Grounded Theory Approach (Strauss & Corbin, 1994) was used. The topic categories emerged from the data. In addition, the switched sentence including the discourse context was considered into this analysis. All examples that did not fit into any specific category were grouped together under the umbrella category of “other”. The data were coded for categories (extra sentential, intersentential, intrasentential, intra word, and mixed), L1/L2-L2/L1 adaptation (i.e., if the word follows L1 phonology or L2 phonology), and switching boundary, e.g., within VP, at the boundary of an adverbial, NP. Moreover, the topic was considered to check which topics attracted most switching.
The research questions are:
RQ1: What topics motivate uses of code switching?
RQ2: How is Hebrew syntactically incorporated into the language use of Palestinian Israelis? In other words, what are the syntactic characteristics of these code switches?
RQ3: Do the two languages keep their own phonologies distinct? If so, what features are involved?
RQ4: Which categories of CS (extra/intersentential, intrasentential, intra word) does CS from Arabic to Hebrew manifest?
My hypothesis is that these switches by bilinguals are largely limited to greeting phrases and are constrained to clause boundaries with one-word switching into Hebrew. I am establishing my hypothesis on the fact that the Arabic language is the main component of Arab identity; as a result, I expect that Palestinian Israelis will try to maintain the “purity” (Suleiman, 2004) of their language (MSA and dialectal varieties) as much as possible. Moreover, according to my personal observations, there is an adaptation of L2 to L1 phonology; however, it depends on the region. The amount of contact can also cause adaptation from L1 to L2 (see Horesh, 2015 ).
4. Findings and Discussion
In this section, I present the results for code switching in the spoken Corpus, which comprises 2891 sentences, of which 586 have code switches, for a frequency of 20.3%.
The CS was classified into four categories depending on where in the sentence the code switch was made: extra sentential, intersentential, intrasentential, and intra word. Some entries mixed two of these categories. The results are shown in Table 1 below.
The most frequent category is intrasentential, followed by intra word CS. Extra sentential is the least frequent category that appeared in the data.
Intrasentential entries were examined for part of speech (POS), and it was found that nouns followed by adjectives, NP, ADV, DM, verbs and verb phrases are the most common POS where the code switch in made, while PP and plural markers are the least frequent, as shown in Table 2 below.
![]()
Table 1. CS categories in the spoken corpus.
![]()
Table 2. POS of intrasentential entries in the spoken corpus.
The corpus includes several examples of switching that occur at syntactic boundaries. CS occurs at the boundary of both maximal and minimal syntactic categories. Switches were found within clause and word boundaries, as well. Some instances of Arabic-Hebrew extra sentential switching can be seen in the following examples:
(1)
[ɣajar-it batˁ:arij:i-t i-s:j:a:ra mar:a, naxon?]2
changed-you battery-poss def.-car once, right?
“You’ve changed the car’s battery once, right/have not you?”
In example (1), the tag question “right?” in the sentence is in Hebrew. All the examples from (1 to 20) are taken from the spoken data (recordings).
The following are some examples of intersentential switches.
Two main clauses:
(2)
[fiʃ ʔimkanij:i θanji. b-muʕed bet ʔen ma laʕsut]
Negation possibility second prep.-appointed time second negation what to do
There is no other possibility. In the second appointed time for the exam there is nothing to do!
[According to the context: “There is nothing that I can do now! I will succeed in the second appointed time for the exam.”]
Example (2) shows two separate sentences that are connected in content to each other, but each one is in a different language.
Interrogatives:
(3)
[tatsiaʕ lo, ʃu ilmoʃkili]
offer for him, what Def. problem
“offer (it) to him, what’s the problem?!”
In this example (3), the first part of the sentence is in Hebrew, but the second part that has the question is in Arabic.
After a conditional:
(4)
[baʕrif-iʃ ʔiða ʔifʃar lakaħat ʔoto bi-ħinam]
I know-negation if possible to take it (object) prep-free
“I do not know if it is possible to take it for free.”
In example (4), the first part of the sentence including the conditional “if” appears in Arabic; however, the “clause” that comes after is in Hebrew.
In intrasentential CS, switches were found within various types of phrases, such as NP, VP, PP, AdvP, as in the following:
Verb and Object:
(5)
[ha:d il-kurs biʕtˁi-k kili-m kef titʕamal maʕ il-ʔomu:r]
this def. course gives-you tool-plural how to deal with def. matters
“This course provides tools to be able to deal with different matters.”
In example (5), the noun “course” is in Hebrew (loanword from English), while the determiner “this” and the definite article is in Arabic. Moreover, the Verb “gives you” is in Arabic, and the object “tools” is in Hebrew.
Verb and PP:
(6)
[iħdari-ha bi sbort ħameʃ]
watch-it on sport five
“watch it on Sport Five Channel.”
In example (6), the verb and the object “watch it” are in Arabic, while the prepositional phrase is in Hebrew “on Sport Five Channel.”
NP and Adj:
(7)
[fi-ʃ iʃi mijuħa:d]
neg. thing special
“There is nothing special.”
This example includes an NP “nothing” in Arabic followed by an adjective “special” in Hebrew. The switch happened after the NP.
VP and AdvP:
(8)
[ʕimil iʃi bkavanah]
He did something deliberately (with intention)
“He did something deliberately.”
The VP “made something” is in Arabic, and then after the VP there is a CS to Hebrew for the adverb “deliberately.”
Within an equational sentence with a Predicative Adjective:
(9)
[ah hu3 lo baʃel]
yes he not mature
“Yes, he is immature.”
The sentence in example (9) is an equational sentence in Arabic. Arabic and Hebrew do not have the verb “to be,” unlike English. This sentence would have the verb “to be” conjugated in the present tense if it were in English. Example (9) starts in Arabic (“yes he”) followed by the Hebrew adjective in the negation form (“not mature”). The adjective “not mature” is considered a predicative adjective because it fills the role of the predicate in this Arabic\Hebrew equational sentence. In this case “not mature” completes the idea about the subject “he.”
Object of a preposition (NP within a PP):
(10)
[ruħit ʕa be:t miʃpat]
I went to court.
“I went to court.”
The sentence in example (10) starts in Arabic (“I went to”), and immediately after the preposition there is a switch to the Hebrew NP “court.” The switch here occurs within the PP where the preposition is in Arabic and the NP is in Hebrew. In other words, the switch happens inside the Prepositional Phrase. This example, can also be considered a borrowing, depending on the area and how strong is the language contact.
Determiner and Noun:
(11)
[kul:-ajat-ha tsefe]
all-(fem) plural-obj. pron expectation
“These are all expectations.”
The example above starts with the quantifier determiner “all” in Arabic, but the noun that follows is in Hebrew. The switch in this sentence happens between the determiner and the noun that it modifies.
Verb switching in the middle of a sentence:
(12)
[hi4 hitsiʕa ʔili]
she suggested to me
“She suggested (this) to me.”
Example (12) above includes a switch to a Hebrew verb in the middle of the sentence. The subject and the object are in Arabic, while the verb in between is in Hebrew.
Noun switching in middle of a sentence:
(13)
[ʔistaɣil:-i ha:d mivtsaʕ bidʒan:in]
take advantage-obj. (it) this sale great
“Take advantage of it, it is a great sale!”
The sentence above starts and ends in Arabic, but in the middle, there is a noun switch. The Hebrew noun is followed by an Arabic adjective. In Hebrew and Arabic, the noun precedes the adjective.
(14)
[il-dira5 il:i ilha xamis daqajiq min mustaʃfa il-karmil bʕidi]
def.-apartment that for her five minutes from hospital def.-Karmil far away
“Her apartment is five minutes far away from the Karmil Hospital.”
This sentence also has a one-word noun switch to Hebrew in the middle of the sentence. The definite article is in Arabic. This is also considered an intra word switch since the definite article is part of the noun in both languages and not separate as is the case in English.
Complementizer and the clause it introduces:
(15)
[ʔawal iʃi qal-u inu dkira ʕal rekaʕ liʔumani]
first thing said-they that stabbing on background nationalistic
“At first they said that it is a stabbing on a nationalistic background.”
The first part of the sentence in (15) is in Arabic up until the complementizer “that,” while the full clause after the complementizer is in Hebrew.
As regards intra word CS, there are some examples that show switches inside a word, as illustrated below.
Arabic conjugation/form of Hebrew verbs:
(16)
[ʔana batak:in min wara:h-a]
I making corrections from after-Fem.
“I am making corrections after her.” (i.e., “I correct her mistakes”)
The sentence in this example is in Arabic and the verb is Hebrew; however, although it is a Hebrew verb root, it gets an Arabic verb form. This is evident in the dialectal prefix [b] and the geminated [k:] (in bold).
(17)
[ʔiða ba- tfos ʔaj ʃuɣul hon, b- atrik ʃuɣl-i hna:k]
if I. present- catch any work here I. present leave work-poss. there.
If I get a job here, I will leave my job there.
In this sentence the verb is in Hebrew, and it is also conjugated into an Arabic form. This is a conditional sentence; the whole sentence is in Arabic with a one-word switch into a Hebrew verb conjugated into Arabic that appears right after the conditional “if” in Arabic.
Sentential complement with a null complementizer:
(18)
[qulti-l:u ʔen ma laʕsut]
told-him negation what to do.
“I told him (that) there is nothing to do (in this case).”
The sentence above is a sentential complement, i.e., the Arabic verb “told him” needs a sentence as its complement. The whole complement sentence is in Hebrew. The sentence has a null complementizer, because the complementizer “that” does not appear after the Arabic verb “I told him.”
Arabic negation of Hebrew verbs:
(19)
[ma-titħajv-iʃ inu tru:ħ-i]
neg.-commit-neg that go-fem.
“Do not commit to go” (i.e., “do not promise and make the commitment to go (to an event)”).
This sentence starts with an Arabized Hebrew verb, which is in the colloquial Palestinian Arabic negation form. This negation form is a circumfix—a prefix and a suffix to indicate negation. While the circumfix affix is Arabic, the verb that it circumfixes is in Hebrew. The rest of the sentence is in Arabic.
Determiner and Noun (frequent in the data):
(20)
[ʔaxad-it kul in-nikud-ot]
received-I all def.-point-plural
“I have all the points (that I need for my degree).”
This intra word switch is common in the data, where the noun is in Hebrew and preceded by an Arabic definite article. The definite article is part of the word itself in Arabic and not separated from it like in English.
It seems that CS is used in many domains in the spoken corpus. The subjects tend to code switch when talking about topics such as education, employment, and social relationships, followed by technology and health. In contrast, greetings, construction, and swear words are the domains that have the least CS in this corpus. See Figure 1.
Moreover, the data show that almost all utterances are affected by L1 (Arabic) phonology: 99.1% are affected by L1, and only 0.9% preserve L2 (Hebrew) phonology.
Some examples:
(21)
fi hasaʕot6
“There is transportation.”
In this example the bold letter /ʕ/ is pronounced [ʕ] followed by a long [o] in Arabic and as a glottal stop [ʔ]followed by a short [o] in standard Hebrew. The word “transportation” in the example was pronounced with the Arabic sound [ʕ]. This means that the Hebrew word for “transportation” is adapted to the Arabic phonological system.
(22)
latan:u jiʃaħreru otanu
“until they allow us to go”
The Hebrew word “let go” is pronounced as jiʃaħreru, yet in standard Hebrew it is pronounced jɛʃəxɣɛɣu. The phonemes [r] and [ħ] are not pronounced as in Hebrew [ɣ] and [x].
In contrast, see the following example:
(23)
xamuts ha:d noʕo za:ki
“Sour, it is delicious.”
The Hebrew word for “sour” is pronounced with [x]. In Hebrew orthography it is written with < ח > which is parallel to the phoneme [ħ], but pronounced in standard Hebrew as [x]. In this example it is pronounced as [x] in Arabic as well, despite the fact that [ħ] exists in Arabic, and in most cases where there is < ח > [ħ] in Hebrew it is pronounced as [ħ] in Arabic. This example shows that there is no phonological adaptation of the word meaning “sour” into Arabic and Hebrew pronunciation is kept.
From my personal observation, the letter < r > in Hebrew, which is pronounced as [ɣ] in Ashkenazi dialect (standard Hebrew), is almost always pronounced as [r] in Arabic. Yet, [ħ] sometimes is pronounced as [x], and [ʕ] is sometimes softened a little to become similar to [ʔ]. This indicates that not all phonemes follow Hebrew pronunciation. Refraining from incorporating all Hebrew phonology might suggest that the speaker is trying to keep some indicators as a hint of an Arabic identity. Moreover, in Israel there are basically two dialects of Hebrew: Ashkenazit and Mizrachit (western vs. eastern dialects). Ashkinazit is the standard current dialect; while Mizrachit was a traditional older Israeli dialect, especially used by Sephardim. Today Mizrachit is disappearing and only used by older Sephardim or Yeminites. Arabs in Israel usually follow the Mizrachi dialect which preserves some of the historical pronunciations of Hebrew that are Arabic-like. Consequently, when a code switch has Arabic phonology, it may be reflective of the Arabic dialect of Hebrew, which needs systematic study.
To summarize, in the case of Palestinian Israelis, it is not necessarily expected that speakers in code-switching will preserve the phonology of the second language (L2), contrary to what one might expect. If the pronunciation of Hebrew is kept, then this is most likely a code switch, if it is not, it can still be a code switch unless the Hebrew word or phrase is used very frequently in Arabic, then it becomes a loanword instead. The added complication is that Israeli Arabs’ dialect of Hebrew, which seems to preserve the historical Arabic-like pronunciations, needs further study given that a code switch with Arabic phonology may also reflect Israeli Arabs’ dialect of Hebrew.
5. Discussion and Conclusion
The percentage of CS in the corpus is 20.3% of all sentences. The domains with high occurrences of CS were employment, education, social relationships, and technology.
The most inserted elements code switches are nouns and NPs, for instance:
(24) جبت كرتيس موريم?
dʒibit kartis murim?
“did you bring a teachers’ card?”
In this example “a teacher’s card” are NP switch.
The higher occurrences of nouns compared to other elements is supported in the CS literature ( Alhazmi, 2016 ; Alrowais, 2012, cited in Alhazmi, 2016 ; Bowers, 2006 ). Myers-Scotton (2002: p. 76) claims that verbs are harder to fit into the recipient language, since they carry more “syntactic baggage.” Forslund (2009) also believes that nouns are freer than other word classes when it comes to grammatical restrictions. Even though Arabic and Hebrew are typologically similar and share many grammatical aspects as well as some lexical roots, nouns and NPs are still the most frequent items in CS. It seems that nouns and NPs in the two languages share more common ground than verbs, particularly given that verbs must be conjugated using certain templates, adding stress and gemination depending on the verb class.
It is worth mentioning that code switching does not make the switched item part of the matrix language, and it generally keeps the grammars of both languages separate. However, there are exceptions in the speaking of Arabs in Israel.
The most frequently used type of switch was intrasentential. Using this type of switching requires a high proficiency in both languages. Speakers need to develop a high level of syntactic knowledge in both languages to be able to use this type of switch properly (Alhazmi, 2016) . Poplack (1983) found that balanced bilinguals use more intrasentential switches in their speech, compared to bilinguals who are dominant in Spanish who use a lot of intersentential switches. Intersentential switching requires less syntactic knowledge, hence it is found more in the speech of less fluent bilinguals (Kanakri & Ionescu, 2010) .
The code switch boundary occurs in slightly more different places in the spoken corpus than in the CMC corpus. This might lead to a prediction for more switching variation in different spoken corpora in the future, particularly in the intrasentential category, which allows wider options for switching. The data includes examples at various minor boundaries, such as switching within NP, PP, and VP. These findings are consistent with the results of Alhazmi (2016) , Redouane (2005) , Bentahila and Davies (1983) , and Alrowais (2012, cited in Alhazmi, 2016 ), who found similar examples of switching in the data they examined. Furthermore, switches found between a complementizer and the clause it introduces countered Gumperz (1977, cited in Alhazmi, 2016 ), who claimed that switching is impossible in this position. However, this finding supports that of Alhazmi (2016) , Bentahila and Davies (1983) , and Redouane (2005) .
CS is still mostly influenced by L1 phonology. Moreover, code switches can still be code switches (not loanwords) even if they have been adapted to Arabic phonology. This has to do with the fact that both languages are typologically similar and share many phonological aspects. There might be some suprasegmental adaptation, like stress.
6. Challenges and Limitations
The sample size is relatively small, primarily due to the challenges of convincing people to participate and be recorded. Extensive explanations were provided face to face and detailed pages were sent via email to ensure participants understood the purpose of the study and the handling of the collected data, assuring them of anonymization and subsequent destruction of the raw data.
Moreover, the data in this research is taken from educated subjects, and this affected the topics discussed, which in turn may have affected the frequency and type of CS s. For example, it explains why the topic of education appeared the most, since some of the participants were teachers.
NOTES
1This research was approved by the Indiana University Institutional Review Board as it is part of my Ph.D. dissertation in 2019.
2The Arabic part is underlined, and the transcriptions are phonetic for the most part. Note that the code switched Hebrew of the Arabs in Israel may contain Arabic features that are part of the Israeli Arabs’ dialect of Hebrew.
3The 3rd person singular masculine pronoun هو hu is the same in Arabic and Hebrew, this makes the example slightly ambiguous. However, my sense is as a native speaker of Arabic with fluent proficiency in Hebrew is that the pronoun is most likely Arabic.
4The 3rd person singular feminine pronoun hɪ is ambiguous here as well, because it is the same in Arabic and Hebrew. However, since the object pronoun is in Arabic, I assumed that the subject pronoun is in Arabic as well.
5This is a one-word code switching. In the data the Arabic word for apartment or house appeared as well, so I consider it a code switch rather than a borrowing.
6I use underlined words to indicate Arabic, and bold letters are those that help me decide regarding phonological adaptation.