Patterns of Code-switching in the Amharic Media

: This study examines patterns of Amharic-English code-switching in the Amharic media language use. It employed a qualitative descriptive approach, which helped break down the recordings and transcripts into intelligible units that helped explain the various language use forms created by code-switching in the Amharic media. Thus, five Amharic media outlets focusing on sport, medical science, and science and technology were purposively selected to study language use in the Amharic media. Hence, 1090 examples which have both Amharic and English elements were selected. A generic step of data analysis was made so as to classify the different types and patterns of code-switching instances. The findings indicated that the three different types of code-switching, namely tag switching, inter-sentential code-switching, and intra-sentential code-switching, are found in the Amharic media language use. Moreover, it was found that there are different patterns of code-switching from English to Amharic. Thus, it is shown that Amharic is the matrix language and is quite resistant against any change of its basic word order. It can also be seen that the mixed constituents found in the Amharic media confirm that the matrix language frame (MLF) model can satisfactorily explain the patterns of Amharic-English code-switching.


Introduction
Code-switching is a widespread phenomenon in bilingual communities.It is the cooccurrence of two languages or varieties.Gardner-Cholors (2009:4) defines codeswitching as bilingual people's use of several languages or dialects in the same conversation or sentence.It affects almost everyone who comes into contact with more than one language or dialect, to a lesser or greater degree.Richards & Schmidt (2010:89) define code-switching as a speaker's (or a writer's) switch from one language or language variety to another.When one speaks in one language and the other responds in a different language, this is known as code-switching.A person may begin speaking one language and then switch to another in the middle of a speech or even in the middle of a sentence.Milroy & Muysken (1995:7) define code-switching as "the alternative use by bilinguals of two or more languages in the same conversation".Similarly, Auer (1998:3) defines code-switching as the alternating use of two or more languages within one conversation.In a similar manner, Clyne (1987:740) states that code switching is "the alternative use of two languages, either within a sentence or between sentences".Bullock & Toribio (2009:2) claim that code-switching is not easy to characterize as it involves a broad range of communication phenomena.This is due to its linguistic manifestation, which may extend from the insertion of a single word to the alternation of languages.Furthermore, patterns of code-switching are not ordered because bilinguals of different proficiency degrees and those who reside in various types of language contact settings produce different patterns.
The contact between two languages will ultimately lead to linguistic interference, where grammar, pronunciation and concepts are transferred from the native language to the foreign language (Galvao, 2009:1).Poplack (1980:614) theorizes three types of CS from a grammatical standpoint: extrasentential, inter-sentential and intra-sentential.Extra-sentential switching is the process of inserting tag elements from one language into a monolingual discourse in another.Inter-sentential switching is defined as a switch from one language variety to another outside the sentence or at the clause level.Intra-sentential switching, on the other hand, is switching from one language variety to another within a single utterance at the clause, phrase, or word level.Muysken (2013:712-713) identifies three types of code-switching.The first type is Alternation, which is a true switch from one language to another with both grammar and lexicon.The second type is Insertion.This is the insertion of lexical items.The third one is congruent lexicalization.This is a situation where two languages share a grammatical structure, which can be filled lexically.Myers-Scotton (1993b) develops the matrix language frame model based on a Swahili/English corpus of recorded Nairobi conversations.It is a non-linear model inspired by psycholinguistic theories.Since its first appearance (Myers-Scotton, 1993a), this model has been in progressive development to become one of the most influential models and further develop established explanations.
The MLF model, which has its basis in bilingual speech, is characterized by an asymmetry regarding the participating languages concerned.The two languages partaking in the process of code-switching do not participate equally.The MLF model is motivated by Joshi's (1985) model, which establishes a formal model for the characterization of intra-sentential code-switching.Joshi uses his Marathi/English code-switching data to apply Garrett's (1975) speech error study and suggests the asymmetry rule, in which the matrix language is established first and then switched to the embedded language.He asserts that closed class items such as quantifiers, determiners, case markers, prepositions, verbal inflections, auxiliary verbs and possessives cannot be switched.Myers-Scotton (1993a) redefines and further develops the concepts of asymmetry and distinction between closed and open class items in the MLF model.The MLF model is used in the analysis of the code-switching data for this study.

Aims of the research
Code-switching can occur in all parts of lives, from casual daily communication to formal media language use.According to Leyew (1998:214), the alternate use of the two languages within a single conversation is seen to be a normal part of the communicative behavior of bilinguals.Indeed, among speakers with a good command of both Amharic and English, code-switching has become a commonly used mode of communication.Moreover, because of the large role English plays in Ethiopia, it seems not to be an easy task to keep this situation under control.
In Ethiopia, Amharic-English code-switching is also a repeatedly appearing phenomenon.Leyew (1998:197) states that code-switching is a widespread phenomenon among speakers of different indigenous languages.As a result, it is very common to hear the alternation of codes, especially from one other language to Amharic.Among educated people, code-switching frequently occurs from Amharic to English.Teferra (2008:172) also gives evidence to show how educated people use English mixed with Amharic.He remarks that well-educated people freely intersperse English lexical items in their speech (codeswitching), despite the availability of Amharic equivalents.This might be due to the prestigious status English has in the country.A similar phenomenon concerning code-switching in Ethiopia is explained by Eshetie (2010:9) who states that the majority of society perceives speaking English, or at least mixing English with a local language, as a sign of being educated and/or modernized.This perception would also influence the use of the local language and may show the pressure English exerts on the local languages in Ethiopia.
The purpose of this paper is to discuss the patterns and types of Amharic-English code-switching in the Amharic media.It aims at presenting the instances of code-switching in order to demonstrate the existence of this linguistic phenomenon.Therefore, this article is concerned with answering the following research questions: (a) What are the types of code-switching used in the Amharic media?(b) In what patterns of code-switching are English elements being incorporated into Amharic expressions?

Research design
This is a linguistic analysis of media language use in the Amharic media.The study adopted a qualitative descriptive approach, which helped break down the recordings and transcripts into intelligible units that helped explain the various language use forms created by code-switching in Amharic media.

Samples
In accordance with the focus of the study examining code-switching in Amharic media, it is imperative to clearly define the population.Code-switching in this context specifically refers to one that appears in different Amharic media outlets in Ethiopia.There are different Amharic media outlets in Ethiopia.Thus, it is essential to select those media that are pertinent to the objectives of the current study.
In the current study, the samples selected have a direct relationship with the topic under investigation.In this regard, the selected media fields (sport, science and technology, and medical science) do have a tendency to use numerous English elements in their language use.Therefore, the programs broadcasted and printed for a year in the following Amharic media outlets (four electronics and one print media) in 2007 Ethiopian calendar were purposively selected to study code-switching in the Amharic media.These are EBC Sport, EBS Sport, Tribune Sport, Tech Talk on EBS and Medical Magazine.After a thorough look at the data, 1090 examples which have both Amharic and English elements were selected.

Data collection method
This study investigated the use of Amharic in the media.Thus, the data source was media programs and publications.As a result, document analysis was the data collection method employed.To meet the objectives of the study, various visual and audio mass media records were collected by using the new digital archive, YouTube.Thus, the data-collection method used in this study was document analysis.
Language alteration is typically an involuntary linguistic activity.Thus, the analysis of code-switching started with collecting recordings of the Amharic media, sports, science and technology, and medical science programs, which are the corpus of this study.Thus, this study chose to work on speech data, one-year programs which consist of two television programs and two radio programs of different genres, broadcasted and transmitted on different Amharic channels.In addition, a one-year publication of a weekly Amharic magazine on medical and scientific issues was also chosen.After going through the programs and the magazine, a due focus was made only on the mixed and inserted utterances, consisting of Amharic and English elements.
The analysis model selected was the matrix language frame (MLF) model, which was used to analyze the patterns of code-switching utterances occurring in the Amharic media.This model aims to explain how the ultimate surface configurations, composed of matrix language (ML) and embedded language (EL) morphemes, are structured, and the motivation behind such structures.

Types of code-switching
This section presents the findings of code-switching pertaining to the types of code-switching in the Amharic media language use in sport, science and technology, and medical science outlets.Below are the three types of code-switching that are identified in the data.These are tag switching, inter-sentential code-switching and intra-sentential code-switching.

Tag switching
This is an insertion of a tag from one language into the utterances of another.A tag can be inserted in an utterance effortlessly with no violation of syntactic rules because they have negligible syntactic reactions.As a result, they can be used by monolingual speakers.
(1) In (1) "I think" is used in the sentence to show the regular habit of the users.This type of code-switching is classified as tag switching.These switches could be used so repeatedly that audiences may not even notice when speakers have code-switched.Similarly, in (2) the phrase "most probably", in (3) "why not", and "you can imagine" in (4), are tag switches that are used as discourse markers or conversation fillers.These tags are inserted into discourses without disrupting the syntactic structure of the conversational language.
4.2 Inter-sentential switching This is the type of code-switching from one language to another at a sentence boundary, in which one sentence or clause is in one language and the next one is in another language.
In the examples ( 5)-( 8), switching occurs at the end of each language.In inter-sentential switching, a speaker switches from one language to another between different sentences.Similarly, in the examples, the same phenomena took place.In ( 5), ( 6) and ( 8), the utterances begin with complete Amharic sentences followed by English sentences, and in (7), the bilingual clause begins with English and switches to an Amharic clause, then it switches back to an English sentence.As it can be seen, all the switches in the examples in this section take place at the sentence boundary.
( what in the world is material science 4.3 Intra-sentential switching This is a type of code-switching where two languages occur in a single sentence.Speakers switch from one language to another within the same sentence in this type of switching.As a result, a sentence will contain two or more languages.Muysken (1995) says that in an inter-sentential CS, the switch occurs at the clause or sentence boundary where a clause or a sentence is in one language and the latter is in a different language.
(9) sɨlə-DRONE-oʧ l-as-ќaññ-aʧʧhu .The examples from (9) to ( 12) are all examples of intra-sentential code-switching where all the switched elements are within sentence boundaries.So, in the utterances, it is seen that the speakers code-switch between the two languages (Amharic and English) within the same sentence.Therefore, in examples ( 9) and (11), the speakers mixed nouns (N), while in (10) and ( 12), noun phrases (NP) are mixed.The intra-sentential code-switching from Amharic to English will be seen in various syntactical categories in further discussions.
Here in this section, we can also see two categories of words.The first one is science and technology words, and the other one is media-related words.Thus, amongst the intra-sentential code-switched English words in this section, only one word is found to be in the category of science and technology, which is found in (9), "sɨlə-drone-oʧ".However, the rest of the intra-sentential code-switched instances in ( 10), ( 11) and ( 12) fall in to the category of words related to media.

Patterns of the occurrence of English adjectives
English adjectives are found with the Amharic modifiers: demonstrative pronouns and quantifiers.

Amharic demonstrative pronoun + English adjectives
In the data of this study, it is realized that when English adjectives are embedded in Amharic, they are modified by Amharic demonstrative pronouns.Some instances are presented below.
( As it can be seen from the examples ( 13)-( 15), the Amharic demonstrative pronouns јɨhe, which means 'this' and -ja, which means 'that' are used to express the order (distance) of the issue in the context of the discussion.So, the English adjectives critical, fundamental and powerful are modified by the Amharic demonstrative pronouns to show the order in which they are used within the context of the discussion.

Quant bionic eye or artificial eye is-3MS
'The eleventh one is bionic eye or artificial eye.' English adjectives, when embedded in Amharic can also be modified by Amharic quantifiers.In examples ( 16)-( 19), the English adjectives amazing, interesting, professional and bionic occur with the Amharic quantifiers ʔandand meaning 'some', ʔand meaning 'one', sost məto meaning 'three hundred' and ʔasra ʔanddəññaw meaning 'the eleventh'.
5.3 English adjectives + Amharic nouns In the data, it is also found that English adjectives occur alone without the Amharic modifiers within the Amharic grammatical frame.As it can be seen in ( 20)-( 23), the English adjectives senior, normal, legend, artificial and administrative are used to modify the Amharic lexical elements that are found right next to them.

Patterns of occurrence of English nouns
6.1 English noun + Amharic plural marker English nouns in ML + EL constituents express the notion of plural marking through their mixture with the following Amharic plural suffixes.
(24) bəwɨsť-u dəgmo ј-all-u-t In the examples ( 24)-( 29), the English nouns and noun phrases specialty, quality, library, wireless device, air mass and press conference are realized in the ML of Amharic rather than in English, which is the EL.In Amharic, singular is not morphologically marked, but nouns as we have seen in the above examples, are morphologically marked for plural.The common plural number marking in Amharic is suffixation.The plural marker -oʧʧ is used to express plurality of nouns.Alternation may occur in the process of suffixation based on the type of final consonant or vowel sounds of the nouns.In the code-switched instances ( 27), ( 28) and ( 29), we can see nouns ending in consonant use the Amharic suffix -oʧʧ as a plural marker.On the other hand, in the examples ( 24)-( 26), where the nouns end in vowels, the plural marker suffixes take the form -woʧʧ.
The code-switched instances in this section can also be further categorized into abstract nouns, scientific and technical terms and modern civilization words.As can be seen in the instances the nouns in ( 24)-( 26), which say "speciality-woʧʧ" in ( 24) and "quality-woʧʧ" in (25) can be classified as abstract nouns.Thus, these instances show that the plural marker of English abstract nouns in code-switched instances is realized in ML Amharic.
The second category of code-switched nouns that can be seen in this section is scientific and technical terms.These words are related to science and technology.Therefore, the code-switched instances in ( 27) and ( 28) "wireless device-oʧʧ" and "air mass-oʧʧ" are considered as science-and-technology-related words.In a similar manner, the plural marker of the nouns in this category is also realized in Amharic the matrix language rather than in English.
The third category of code-switched English nouns found in this section is modern civilization words.The instances of this category are found in (26) "library-woʧʧ" and ( 29) "press conference-oʧʧ".Similar to the above two categories in this section, the plural marker of these nouns is also realized in Amharic as it is the matrix language.
Having looked at the English nouns code-switched into Amharic using the Amharic language plural marker, it would be reasonable to see whether singular English nouns code-switching may take place.As it has already been explained, in Amharic singular is not morphologically marked.Thus, singular nouns from English may be code-switched; however, English nouns would be integrated into Amharic as a bare form.
6.2 English noun + Amharic definite article Here, in the following code switched instances, the English noun phrases take the Amharic definite marker suffixes to show definite nouns.Olympic committee-ART preparation-ART-ACC start-3MS-AUX 'The Olympic committee has started its preparation.' Unlike English, Amharic singular indefinite nouns do not need to be specified by a determiner.They remain bare singular nouns, and they can receive indefinite interpretations.Amharic definite article takes gender category.Therefore, it uses the suffixes -u or -w to show definiteness for masculine, and -wa is used to form definite marker in feminine.The suffix -u is used after noun phrases ending in consonants, whereas -w is used after those ending in a vowel.In English, the definite article the occurs before the noun phrase to show definiteness.Here, in the examples (30)-( 32), the noun phrases (current situation, television screen, supervision team) end in consonant sounds.As a result, their Amharic definite article marker is the suffix -u.Whereas, in instances ( 34) and ( 35), in terms of the noun phrases anti-doping agency and "Olympic committee", their ending sounds are vowels, which results in -w becoming the definite article marker.In example (33), the suffix -wa is used to show feminine definite marker after the noun president.In all the examples, the definite article marker is a suffix, which is contrary to English where the definite marker is a prefix.This clearly shows us that Amharic is the ML, which contributes the morpho-syntactic frame for the code-switched utterances.
Furthermore, English nouns and noun phrases in this code-switching section fall into three categories.The first group of words are those which are related to sport as in ( 34) "(wada) anti-doping agency-w" and ( 35) "olympic committee-w".The other group is words that are related to administration.These words include instances in (30) "current situation-u", (32) "supervision team-u" and "president-wa" in (33).The third one is a media-related noun phrase found in (31) "television screen-u".As it can be seen in this section, definiteness in the code-switched English nouns and noun phrases is marked by using the Amharic definite marker suffixes in all the three categories.
6.3 English noun + Amharic possessive marker In Amharic, there are two ways of constructing possessive.The first one is formed by attaching a possessive suffix to the noun stem.The second way of constructing possessive marker in Amharic is by using the genitive prefix jə-.It is attached to an independent pronoun or a noun to express ownership.In the instances below, Amharic/English code-switched utterances are presented to show the English nouns integrated in the Amharic language morpho-syntax by using different Amharic affixes.'This again is the development of their football.' In the examples (36)-( 50), the English nouns tend to be modified by Amharic possessive markers in the bilingual utterances.In ( 36) and (37), the English nouns editor and training use the suffix -e, which means 'my' to form first person singular possession.In (38) and (39), the English nouns audience and role model are modified by the Amharic language possessive marker suffix -h, to show second person masculine possession.In (40), the Amharic possessive marker suffix -ʃ, is used on the English noun phrase academic life, to show possession in second person feminine singular noun.In ( 41)-(42), and ( 43)-( 44), the English nouns performance, squad, mobility and bureau are integrated into the ML morpho-syntax as they are suffixed by the Amharic third person masculine singular markers -u and -w respectively.In ( 45)-( 46), the English nouns program and sponsor, take the Amharic suffix -aʧʧɨn to show the first person plural possession in the bilingual clause.The English nouns audience and comment in the examples (47)-( 48) are modified by the Amharic second person plural possessive marker -aʧʧɨhu.The examples in (49)-( 50) show integration with the ML Amharic as the English nouns management and football are modified by the Amharic third person plural possessive marker -aʧʧəw.In the above examples, it is clearly seen that the English nouns are integrated in the ML Amharic morpho-syntactic frame.
Moreover, the code-switched English nouns and noun phrases in this section are further categorized into three classifications.As a result, the first group of nouns and noun phrases are media-related words, the second group is sport-related words and the third group is words related to administration.Hence, the media-related nouns and noun phrases are those found in (36) "editor-e", (38) "audience-h-n", ( 40) "academic life-ʃ", (45) "program-aʧɨn", ( 46) "sponsor-aʧʧɨn-n", (47) "audience-aʧʧɨhu-n", ( 48) "comment-aʧʧɨhu-n".The second category of English nouns and noun phrases are sport-related words that are found in (37) "training-e-n", (39) "role model-h", ( 41)-( 43) "performance-u", "squad-u", "mobility-w" and ( 53) "football-aʧʧəw".The third group of nouns and noun phrases in this section are words related to administration that are found in (44) "bureau-w" and (49) "management-aʧʧəw".Thus, looking at the code-switched nouns and noun phrases, in all the three categories possession is marked with the matrix language Amharic possessive marker suffixes across all the three categories.
In the following instances ( 51)-( 54), the second way of constructing Amharic possessive is presented.Here again the English nouns and noun phrases are modified by the Amharic possessive marker prefix jə-.So, the English noun phrases in the instances (51) internet access and (53) basketball board are prefixed by the Amharic inflection jə-to express possession.Moreover, the English nouns in ( 52) and ( 54), namely club and humidity are modified by the Amharic possessive marker prefix jə-.In all the above instances, English nouns are well integrated in the Amharic morpho-syntactic frame.
The English nouns and noun phrases that are code-switched into Amharic in this section are classified into two categories.The first group of words fall into science and technology as in (51) "jə-internet access" and (54) "jə-humidity".The second groups of words are nouns related to sport as in (52) "jə-club" and ( 53) "jə-basketbal board".
Unlike the nouns and noun phrases in (36)-( 50) which are marked with Amharic possessive marker suffixes, the instances in ( 51)-( 54) are marked with the Amharic possessive marker prefix jə-.Thus, in all the categories, the groups of English nouns and noun phrases that are code-switched into Amharic possession are marked with the Amharic possessive marker prefix jə.This is because Amharic is the matrix language that supplies the morpho-syntactic frame.

Patterns of bilingual compound verb formation
The data in this research display that in this kind of verbs English verbal stems appear as the first constituent.Speakers use the English content morphemes in a definite way when English verb forms avail themselves as a better form to meet their purpose.In such instances, the Amharic verbal inflections do not attach themselves to the stem, but the Amharic lexical element, which is the root for "to do" verb, emerges.In addition, all the necessary verbal inflections from Amharic language are added to it.This is how the bilingual verbs are formed.Below different patterns of Amharic-English verbal expressions are discussed.
7 In the instances presented in section 7.1, the English (embedded language) verb stems appear in mixed constituents as bare forms.This is to mean no English verbal inflections are attached to them.They function as content lexical items.Amharic supplies the 'do verb', occurring after the stems to attach the inflections to them.This do verb is a transitive auxiliary verb, which is marked for tense and aspects.
These simple verbs are not the only verbs that are found in the bilingual utterances, but English phrasal verbs are also introduced through the process of 'do construction' in the same way.In the Amharic-English bilingual verb formation, English phrasal verbs occur as a unit.Then, the Amharic 'do' verb emerges and all the necessary inflections are added to it.
A few examples of English phrasal verbs in the bilingual utterances from the data are presented below in ( 59) and ( 60 Up to now, we have seen English verb stems occurring as bare forms, but in the data of this research, there are also some instances of non-finite English verb forms.These are English verb stems appearing together with the past participle marker attached to them.Here as well, the bilingual verbs appear in a bare form with helping verbs from Amharic. 7.3 Past participle form of English verb + 'do' verb+ Amharic inflections The non-finite English verb stems appearing in Amharic bilingual clauses.In ( 61)-( 64), the past participle verb forms of English occur as a bare form in Amharic bilingual clause.Here again, the Amharic helping verb ('do' verb) carries all the required inflections.

Patterns of English adverbs in mixed constituents
In this section, grammatical patterns of adverbs, i.e., singly occurring adverbs are integrated into Amharic expressions.The data show that English adverbs are inserted without any change in their form.In the data of this research, English adverbs occur in the code-switched instances.The English adverbs in the bilingual clauses below from ( 65)-( 69) appear in a bare form, but they are well integrated into the Amharic morpho-syntactic frame. (

PDEM already to-finish stage on that-found is-3MS
'This already is found on the finishing stage.' In the above instances in section 8, the embedded language (English) adverbs appear as a bare form.Thus, the embedded language content morphemes (English adverbs) are occurring freely in the matrix language (Amharic) morpho-syntactic frame.
Under the MLF model, content morphemes are those which either assign or receive thematic roles.These morphemes are semantic in the sense that they contribute to the semantics of the expression.In the case of adverbs mentioned so far, it has been seen that the lexical category of adverbs comes from the EL, i.e.English in this case.In the construction of mixed constituents, EL can provide content morphemes.The surface morpheme order of these examples is the empirical proof that Amharic is the source of the ML.In the grammatical configuration set up by Amharic, English content morphemes are occurring freely.

Embedded language (EL) islands
In the previous sections, we have been dealing with singly occurring EL (English) elements integrated into the morpho-syntactic frame of the ML (Amharic).In these instances, English nouns, adjectives and adverbs show full integration in the Amharic morpho-syntax.When we look at the English verbs, however, having made no changes in their forms, code-switching in the verbs shows a different process of morphological integration.Here, the Amharic verbalizer ('do' verb) plays a significant role by taking all the necessary ML inflections and making it possible for the EL verb to be well integrated.Now, in this section a different kind of EL incorporation, which is a sequence of more than one EL elements occurring together as a unit in Amharic morpho-syntactic frame, will be focused.
In a situation where the problem of congruence between the grammars of the languages involved in the code-switching appears, this might be the reason behind a speaker's using an EL island.The differences between the languages at an abstract level often may block accessing an embedded language element in any other way than as a full phrase (Myers-Scotton, 2006:266).There might be different intentions for the speaker to use EL content morphemes.Having those different intentions when speakers encounter insufficient congruence between the two languages grammars, there are different compromise strategies in which speakers can access EL elements.One of these compromise strategies are EL islands.
Though they are framed by the matrix language frame, embedded language islands must show structural dependency to the embedded language.This means that words in an EL island show a hierarchical structure.Just two EL words occurring next to each other in the ML are not necessarily islands (Ibid).
9.1 Classification of the types of EL islands This section explains how instances of EL islands found in the data are classified.First, they are classified into two major sections named non-formulaic and formulaic expressions.Among the most formulaic expressions may be idioms and proverbs.An idiom is a phrase or clause whose meaning is not the meaning of the sum of its parts.

Formulaic and non-formulaic expressions
In the data, elements which are described under this classification include adjuncts.They are "add ons" to the main structural/semantic elements of the clause.These may be adverbial phrases or prepositional phrases.
In contrast, formulaic expressions can be classified into speech formulae, idioms and proverbs.Those expressions, which are used in conversations such as sayings, clichés, slangs and some collocations and the like, are speech formulae.Below, different types of phrase instances in the data are presented to show non-formulaic structures.of-sport infrastructure-PL NEG-PASS-know-ART-NEG for that matter 'The sport infrastructures were not known for that matter.' In the instances above (prepositional phrases in (a) and adverbial phrases in (b)), we have observed that adjuncts, which are prepositional phrases and adverbial phrases, are located outside the predicate-argument structure formed by the main verb of the clause.Thus, they only give additional meanings to the structural elements of the clause.However, these are not the only islands found in the data under this category.Noun phrases are also spotted as EL islands.
(c) NP islands In examples ( 76)-( 79), the EL islands (English noun phrases), which are lack of determination, man of the match, guard of honor and speed of light are well formed in English.Though these EL islands are well formed in the EL (English) structure as a unit, all of them are well integrated in the structure of the ML (Amharic) in all the instances they appear.

EL formulaic expressions
Having looked at the non-formulaic expressions, let us turn to formulaic structures.In this category, EL islands appear in the form of noun phrases, which are collocations.These are combinations of words which often occur side by side as a single phrase.In this category, in all the instances only EL morpheme is used, not even a single word from the ML appears in the islands.Since the occurrence of one word activates that of the other, therefore, these cases seem to be more formulaic than the instances in the previous category.In addition to this, different kinds of speech formulae are also found.These are various expressions used in conversation to make our speech organized.They can also facilitate free and easy flow of ideas in conversation.
(a) NPs (collocations) The instances given above ( 80)-( 83) are collocations, displaying situations in which a common grouping of words is used.The elements of EL islands regularly appear together as a unit to transmit some meaning.The meaning of such combinations is realized by the meaning of its components.Besides, the expressions above show structural dependency allowing no inclusion of ML element in between them.As a result, they appear as a single unit.
( 1PL-invite-AUX-1PL 'By the way, concerning the today's game we will invite a guest.' The EL islands shown in the instances above are referred to as speech formulae.These are expressions used in organizing discourse to convey the speaker's attitude to other participants and also to facilitate the flow of interaction (Cowie, 2001:1).As it has been shown in the instances, speech formulae is used to depict such functions.Expressions in the instances are used as a single unit there by displaying structural dependency.
Idioms and proverbs were not seen in the data of this research.Thus, they were not described in the analysis.
So far, we have been dealing with EL islands that are displayed in the Amharic/English code-switched utterances.All the instances support the theories made by the MLF model.

Discussion on the patterns of code-switching in the Amharic Media
The objective of this research is to find out the types of code-switching which are found in the Amharic media and to explore the patterns of code-switching by using the matrix language frame model.Hence, from the analysis of the data three types of code-switching are found in the Amharic sport, science and technology, and medical science media.These are tag switching, inter-sentential code-switching, and intra-sentential code-switching.This finding seems to echo Muysken's (2003) which deals with the types of code-switching.

Tag switches
As it has clearly been elaborated in the analysis of this study, the Amharic media language use comprises this type of code-switching.In the analysis, the journalists in the Amharic media are found to use this type of code-switching called tag switching for different purposes.

Inter-sentential switching
This type of code-switching does not frequently occur in the Amharic media.While using inter-sentential code-switching type, a speaker switches from one language to another between different sentences.Therefore, when such type of code-switching occurs, the utterance of the speaker is divided into two different languages.As we have seen in the analysis of the data, in this case, one of the languages is Amharic while the other is English.Hence, the utterance may begin in a complete Amharic sentence, and then followed by a complete English sentence.This seems to back up the assertion by Kebeya (2013:228) who implies when intra-sentential code-switching occurs the speech of an individual is divided into two sentences, one sentence will be in one language while the other sentence will be in a totally different language.
10.3 Intra-sentential switching This type of code-switching occurs within the same clause or sentence boundary.Here, the code-switching segment must confirm to the structural system of the recipient language.This type of code-switching may occur in the beginning, in the middle and at the end of a sentence.Hence, the analysis of the data confirmed that there is the occurrence of intra-sentential code-switching in the language use of the Amharic media.Therefore, the analysis of the examples in this regard was focused on the bilingual clause.This is to mean that its focus is on a clause comprising one or more morphemes from more than one language.
10.4 Patterns of occurrence of English adjectives The analysis of the data also explored the different patterns of code-switching in the Amharic media.By using the matrix language frame model the study identified the two particular principles named (1) the morpheme order principle, and (2) the system morpheme principle.In the analysis, we have seen different patterns of English adjectives occurring in Amharic.The morpho-syntactic patterns of the adjectives in the bilingual clauses in the above examples show the acceptability of the MLF model, and demonstrate that it is Amharic that supplies the morpho-syntactic frame for the ML + EL constituents.As is clearly shown in the examples, all the English adjectives embedded in Amharic follow the surface morpheme order of Amharic.As a result, it is a confirmation that supports the morpheme order principle (MOP).Both the modifiers, i.e., demonstrative pronouns and quantifiers, are Amharic system morphemes, which are used to modify the English content morpheme adjective.Thus, the occurrence of the system morphemes from Amharic proves that Amharic is the ML.

Code-switching patterns of nominal constituents
In the data analysis, we have also seen different examples to show grammatical patterns of nominal constituents.The morpho-syntactic patterns of the examples presented correspond with the principles of the MLF hypothesis.The mixed constituents analyzed in the instances from the data follow the syntactic order of Amharic.Consequently, it is logical to say it is congruent with the MOP.There are noticeable instances that can lead to this.First, in all the examples the clauses follow the Amharic word order.The Amharic word order is SOV, which is different from SVO of English word order.Second, the Amharic definite marker, which is a bound morpheme, is a suffix on the noun (N + ART).Contrary to this, the pattern in English is ART + N where the article is a content morpheme.In addition, the Amharic definite marker takes gender category, which we cannot see in English, can be seen as typical of Amharic.Furthermore, all the system morphemes are from Amharic.There is nowhere in the examples that we can see English system morphemes used.As a result, it can be evidence that the instances correspond with the SMP.
It is seen in the above instances that Amharic is the one that provides the system morphemes: plural markers, definite articles and possessive markers.These three are system morphemes used with the embedded language content morphemes (English nouns in this case).This pattern can be a proof that Amharic the ML, which further proves that it is the one that provides the morpho-syntactic frame for the ML + EL constituents.
10.6 Code-switching patterns of adverbs In the analysis of patterns of code-switching of the adverbs, it is found that the English adverbs are code-switched in the Amharic bilingual clauses.As it has clearly been mentioned in the MLF model, that content morphemes come from the EL because these morphemes are those either assign or receive thematic roles.In the bilingual constituents, all the adverbs are from English, which is the EL.The surface morpheme order of the above instances can confirm that Amharic is the source of the ML.In the grammatical patterns of the instances, the English content morphemes (adverbs) are well integrated into the Amharic morpho-syntactic frame.Thus, this confirms that the above instances support MOP.In addition, all the system morphemes in the bilingual clauses are from Amharic, which also maintains the SMP.When considering the roles of the two languages, Amharic provides the morpho-syntactic frame and all the system morphemes required, and English, on the other hand, contributes content morphemes only.Hence, it is possible to say that Amharic is the source of the ML whereas English is the EL.

Embedded language islands
In the analysis of the data, the occurrence of embedded language islands of prepositional phrases, adverbial phrases and noun phrases is found.In all the instances, the phrase structures are well formed in the embedded language (English) grammar; however, it doesn't fit into the morpheme order of the matrix language (Amharic) grammar.This finding seems to be confirmed by Myers-Scotton (2006:261) who states that they are constituents comprising only EL morphemes which occur in a bilingual clause as phrases not full clauses that are framed by the ML.Thus, this finding may serve as evidence of embedded language islands in the Amharic/English code-switching data.This finding is in harmony with that of Myers-Scotton (2002:141) which says adverbial phrases are major type of embedded language island.In addition to this, in the analysis of this study the occurrence of collocation and speech formulae was also manifested.These combinations of words often occur side by side as a single phrase.They are used to make one's speech more organized and to facilitate the free flow of ideas.These expressions, which occur as a single unit, show structural dependency and forbid inclusion of matrix language elements in between them.Hence, evidences of embedded language islands in the Amharic/English code-switching seem to be confirmed in the analysis.
One of the key findings in the analysis of this research is the appearance of the three different types of code-switching, namely tag switching, inter-sentential code-switching, and intra-sentential type of code-switching which are found in the Amharic media language use.Moreover, in the analysis it is found that in the language use in the Amharic media, there are different patterns of code-switching from English to Amharic.
In the analysis of Amharic-English code-switching, English elements of different parts of speech are discussed.As a result of the analysis, it is found that the morpho-syntactic patterns of mixed nominal expressions conform with the matrix language hypothesis.The two basic principles of the matrix language hypothesis are validated in the collected data.The mixed constituents, which are of Amharic and English mixture, seem to follow the syntactic order of Amharic.As a result, they confirm the morpheme order principle.A common instance is the occurrence of the Amharic definite marker after English nouns while it occurs in a reverse order in English.In addition, Amharic plural markers, quantifiers and possessive markers are used when English nouns are embedded in Amharic.Hence, this confirms that Amharic is the matrix language, which is the source of the morpho-syntactic frame for the mixed constituents in which the Amharic system morphemes appear without any restriction.
When English verbs are embedded in mixed constituents, they are likely to appear as single lexical items with no English language inflections attached to them.Thus, while English verbal forms occur in the mixed constituents, Amharic supplies the 'do verb' occurring after the stems to attach the inflections to them.
Regarding embedded language islands, they are constituents consisting only of embedded language morphemes well formed in the embedded language and occurring in the matrix language frame.Both formulaic and non-formulaic expressions have been found to come from English.
Although there is a growing influence of English on Amharic, and we have seen certain consequences of code-switching in the Amharic media language use, there is no evidence that indicates contact-induced changes in the syntactic domain of Amharic.Furthermore, it is shown in the data that Amharic is the matrix language and is quite resistant against any changes of its basic word order.Thus, it is seen that the influence of English has only been noticed on lexical level.Generally, it can be concluded that the matrix language model is able to describe the Amharic-English code-switching patterns studied.This study has made an effort to understand the language contact situation through the analysis of English-Amharic code-switching with a structural approach.The data of the research is focused on language use of the Amharic media on sport, science and technology, and medical science.A study with more semantic fields included could have generated an additional understanding of the matter under investigation.However, it does not mean that the instances presented in the data of this research are not reliable but indicates that a different sequence discernible in the findings of this study could be found.
viewed by paying for the type of service called pay per view.'