Atlas of Pidgin and Creole Language Structures Online

1. Introduction ⇫ ¶

African American English (AAE), which is also referred to by other labels such as African American Language, African American Vernacular English, and more recently Ebonics, is spoken by some African Americans in the United States as a native variety and by still others as a second variety. While different varieties of African American English across the United States share regional patterns with other varieties of American English in the areas in which they are spoken, these varieties of AAE share many core patterns that make it possible to refer to a general African American English. Nevertheless, there is variation among different varieties of AAE spoken throughout the United States. For instance, varieties spoken in southwest Louisiana may share some patterns with Louisiana Creole, which are not produced as part of other varieties of AAE in the United States. Along similar lines, the variety of AAE spoken in Appalachian communities shares some patterns with North Carolina regions that may not occur in other varieties of AAE (Wolfram 2007).

2. Sociohistorical background ⇫ ¶

The sociohistorical conditions under which Africans from Africa and the Caribbean arrived and settled in various United States colonies have been argued to contribute to the language situation of African Americans and the differences between linguistic patterns in mainstream American English varieties and patterns that have come to be associated with varieties of language used by African Americans across the United States. Different scenarios have been offered to characterize the conditions for the development of the progenitor of current African American English: The first groups of slaves who arrived in the United States may have spoken a West African pidgin or creole, or slaves imported from the Caribbean may have spoken some form of Caribbean English Creole, which may have been perpetuated and further developed on plantations. Although slavery in the Unites States colonies was a widespread institution, slave holdings were higher in some areas than in others. For instance, the black population in South Carolina rose to 60% within fifty years after initial British settlement, while blacks in New York made up only about 16% of the population. In addition, from the mid 18th century to the late 19th century, a solid majority of the black population was in the southern colonies/states (Rickford 1986: 225). As such, the ratio of Africans and native English speakers differed from plantation to plantation, so some plantations had large populations of blacks and others had a significantly smaller number of slave inhabitants. These conditions were likely to have played a major role in the variation in development of the linguistic system used by blacks on plantations, especially given the language contact situation with indentured servants and speakers of English and the process of language development.

In the 19th century, some blacks moved from the United States to communities in Nova Scotia, Samaná (Dominican Republic), the Bahamas, Sierra Leone and Liberia, which led to black settlements outside of the United States colonies (for more detailed information on the history of settlement patterns of blacks, see Rickford 1998, Mufwene 2000, Winford 1997, 1998, Poplack 2000, Huber 2004, Hackert & Huber 2007). Given the sociohistorical and sociopolitical background, some of these communities have remained segregated, and early forms of African American English may still be used predominantly in them.

3. Sociolinguistic background ⇫ ¶

African American English is used by African Americans who are members of the working class, and it can be characterized by a list of syntactic, morphological, and phonological features. It is described here according to its systematic properties in the syntactic, semantic (and pragmatic), morphological and phonological components – due to its sociohistorical, sociopolitical, and sociolinguistic background related to African and creole influence and settlement patterns – that are intertwined with patterns from general American English. Some speakers use AAE in varying degrees as a native variety while others may use it only in certain close networks. Speakers of all age groups use the variety; however, in research in the 1960s and 1970s, emphasis was placed on its use by adolescent males, especially those in inner cities. Such emphasis may have led to the general public’s characterization of African American English as slang and a compilation of features commonly associated with males who are part of the popular, or even street, culture.

     As a non-standard variety, African American English is generally not acceptable in educational contexts, but because some children grow up using it as their native language and major means of communication, educators and linguists continue to raise questions about the extent to which use of AAE affects academic performance and whether certain AAE linguistic patterns serve as a barrier to proficient use and comprehension of mainstream American English. Somewhat related to this issue is the question about whether AAE is becoming more or less like mainstream American English, a situation that would have implications for education. Issues such as these were the main focus of the Ann Arbor Black English Case (1979) and the Oakland Unified School District Ebonics controversy (1996).

4. Phonology ⇫ ¶

African American English has twelve monophthongal oral vowels (see Table 1), and the three diphthongs /aɪ/, /aw/, and /ɔj/. There may be variation in the quality of the vowels due to geographical region and other factors. For instance, it is not uncommon to find the diphthongization of /o/, as in [oɪ] (e.g. coach [koɪtʃ]) in certain phonetic environments by AAE speakers in areas in the southern United States. /ɛ/ occurs in AAE but generally not before nasals; the vowel sound [ɪ] occurs in that environment, so ten and tin are homophonous, [tɪn] (see Bailey & Thomas (1998) for a discussion of vowels in AAE).

The consonants are shown in Table 2.

       Deletion of r occurs in contexts in which r follows a vowel in the same syllable, so African American English is a non-rhotic variety. It also occurs in some environments in which r is between two vowels, as in barrel [bæә] and Carol [kæә]. In some varieties of AAE, front vowels are centralized before a constricted postvocalic [ɹ], e.g. there [ðʌr] (see Blake, Fix, & Shousterman (2009) for more information on this variable). Vocalization of l also occurs, as indicated in the phonetic representations of barrel and Carol. On the most basilectal level, AAE does not have a voiced interdental fricative [ð], so those would be pronounced as [dos], mother as [mʌvә], and bathe as [bev], in which [d] occurs in initial position and [v] between two vowels and word finally. The voiceless interdental fricative [θ] is invariably produced in initial position (e.g. think [θɪŋk]); however, [f] or [t] occurs in syllable-final position (e.g. birthday [bɚfde], with [wɪt]/[wɪf]).

Some restrictions are placed on the co-occurrence of consonants that can form syllable-final clusters. That is, clusters such as [nt] remain intact syllable finally, but clusters with the same voicing value, such as [st], [pt] ([‑voice]), and [nd] ([+voice]) may not.

(1) accept [æksɛp]

The clusters are more likely to remain intact preceding vowel-initial suffixes.

(2) acceptable [æksɛptәbә]

(3) spending [spɪnɪn], [spɪndɪn]

It should be noted that these clusters may vary depending on the following suffix type. For instance, [nd] may be variably retained preceding -ing, as in [spɪnɪn] and [spɪndɪn].

The description of final consonant restrictions can either be expressed from the angle of cluster reduction in which the cluster is present underlyingly and is reduced by deletion of the final consonant forming the cluster due to syllable or sonority restrictions, or it can be presented from the view that the most basilectal African American English does not have underlying final consonant clusters formed of consonants with the same voicing value. In some contexts, words such as contest [kantɛs] behave as if they end in a single consonant (i.e. s), not a consonant cluster. For instance, the plural of the word is formed by adding the allomorph [ɪz], as in [kantɛsɪz], the same way the plural for dress (dresses) is formed. On the other hand, some consonant clusters remain intact preceding vowel-initial suffixes, as shown in (2) and (3), which can be construed as indicating underlying consonant clusters. Another process that affects word-final consonants is devoicing: big [bɪk] and bed [bɛt].

In addition to segmental properties of AAE, suprasegmental patterns have also been discussed although the research is much more limited. Observations have been made about the wide pitch range and falsetto range, as well as level and falling tones (for more information on prosody in AAE, see Tarone (1973), Foreman (1999), Green (2002), and Thomas (2007)).

5. Noun phrase ⇫ ¶

Both bare and “modified” nouns occur in African American English NPs:

(4) the boys

(5) boys, water

Plural count nouns and mass nouns can occur as bare NPs; however, singular count nouns generally do not occur as bare NPs in many of the varieties of AAE that have been reported in the literature (e.g. Mufwene (1998) and Green (2007)). One exception is the variety reported by Spears (2007, 2008), in which it is stated that singular bare nouns occur and can have generic and definite interpretation, as in the examples (Spears 2008: 528, 529):

(6) Man crazy.

      ‘Man is/human beings are crazy’ or ‘The man is crazy.’

(7) Dog ain’t got no sense.

      ‘Dogs don’t have any sense’ or ‘The dog doesn't have any sense.’

(8) Dog got fleas.

     ‘Dogs got fleas’ or ‘The dog’s got fleas.’

While plural nouns are often morphologically marked, they do not always bear plural morphology, as in (9, 10):

(9) You cooking bean/beans tonight?

      ‘Are you going to cook beans tonight?’

(10) She put fifty cent in the basket.

      ‘She put fifty cents into the basket.’

Number can also be indicated on nouns with human reference in subject and object position by an associative plural marker nem.

(11) a. Sue nem dәn went to the store.

      ‘Sue and the others have gone to the store.’

      b. They already told Sue nem to stay home.

            ‘They already told Sue and the others to stay home.’

      Lexical and pronominal possessives are overtly expressed; and the possessor precedes the possessee. The possessor is generally not inflected, as indicated below:

(12) a. Sue book                    

           ‘Sue’s book’

        b. Sue brother book

            ‘Sue’s brother’s book’

(13) I’m going down there to Faye.

        ‘I’m going to Faye’s.’ (i.e. Faye’s house/place; or: ‘I’m going to Faye’)

6. Verb phrase and related categories ⇫ ¶

The inventory of transitive and ditransitive verbs in African American English is quite similar to that in other varieties of English; however, verbs such as give and hand can consistently occur without an indirect object or with only one internal argument if the pronominal reference of the indirect object is to first person:

(14) Give my book.

      ‘Give me my book.’

(15) a. Hand my phone.

      ‘Hand me my phone.’

      b. Hand that book (here).

      ‘Hand me that book.’

The subjects in (14) and (15) are second person singular, and the main verbs are in the unmarked forms. In fact, the verb form generally occurs in the unmarked from with 1st, 2nd, and 3rd person singular and plural subjects (e.g., Sue/they eat apples. ‘Sue eats apples’ ‘They eat apples’).

Auxiliary verbs be, do, and have ¹ occur in syntactic environments in which they are stressed (i.e. host a pitch accent) and bear past tense. They can also occur in negative environments preceding not (n’t) and in questions.

(16) Sue runnin/Sue is runnin.

       ‘Sue’s running.’/‘Sue is running.’

Auxiliary be (and copula be) is variable although the likelihood of its occurrence preceding V-ing is low.

      The first person singular I’m and third person singular it’s almost always occur in auxiliary and copula be contexts. While the auxiliary have generally does not occur in neutral declarative sentences, it does occur in perfect contexts to host negation and express emphatic affirmation:

(17) He left./He have left./He haven’t left.

      ‘He has left.’/‘He has left.’/‘He hasn’t left.’

The auxiliary ain’t can also be used to negate verbs in the bare form, progressive, future, and perfect contexts.

(18) He ain’t leave.

      ‘He didn’t leave.’

(19) He not/ain’t leaving.

      ‘He’s not leaving.’

(20) He not/ain’t gon leave.

      ‘He’s not going to leave.’ / ‘He will not leave.’

(21) He ain’t left.

      ‘He hasn’t left.’ / ‘He didn’t leave.’

Another set of auxiliary elements are never inflected for tense, person, and number. They occur closer to the verb and other predicates than the auxiliaries mentioned above. These are the tense-aspect markers be, stressed BIN, and dən. They indicate information about the internal makeup of an event or about the nature of the occurrence of the event or instantiation of the state. These markers can occur with states and events (i.e. eventualities).

Aspectual be indicates that an eventuality recurs: it is a habitual marker. It precedes all types of predicates, including the marker dən. Example sentences are given below:

 (22) a. Sue be singing. (event)

            ‘Sue usually sings.’

      b.   Sue be knowing the answer. (mental state)

            ‘Sue often does something to show that she knows the answer.’

      c.   Sue be at the grocery store. (location)

            ‘Sue is often at the grocery store.’

All of the sentences in (22) have a habitual reading, in which the eventuality expressed by the predicate occurs or holds from time to time. Given the property of be, the marker can apply to activities as well as states that range from temporary (e.g. ready) to more permanent (e.g. know). In the sentences in (22), the habituality is represented by the different occasions on which the eventuality indicated by the predicate occurs or holds with Sue as the agent, theme, or experiencer. In some cases, a plural subject is required for a felicitous reading with aspectual be because it is impossible to get a reading in which some eventualities indicated by predicates can recur. This can be illustrated with the subject baptismal pools and the prepositional phrase behind the choir stands.

(23) a. Baptismal pools be behind the choir stands.

            ‘Baptismal pools are usually located behind the choir stands’ (Green 2007: 410)

      b. *A/The/That baptismal pool be behind the choir stands.

Because once built in a particular place, a baptismal pool is permanently stable and cannot habitually occur from choir stand to choir stand, thus the anomalous reading in (23b). However, given the plural subject in (23a), the reading is felicitous. Baptismal pools can be understood as occurring systematically behind the choir stands. Aspectual be has a special relationship to the verb or predicate phrase following it, but the role of the subject also plays a part in the meaning of the construction (see Green (2000, 2002, 2007) for a discussion of aspectual be).

      The marker BIN situates part or all of an eventuality in the distant past from the viewpoint of the speakers and hearers. It is stressed, at least with respect to the first word in the sentence, and is accordingly written in capital letters in order to distinguish it from unstressed bin. Like aspectual be, BIN can occur with a range of predicates, including the marker dən. In the sentences below, BIN indicates that the eventuality started in the far past. As states, the eventualities indicated by the predicates following BIN continue to the present.

(24) a. Sue BIN singing.

            ‘Sue started singing a long time ago, and she is still singing.’

        b. Sue BIN knowing the answer/Sue BIN knew the answer.

            ‘Sue has known the answer for a long time.’

        c. Sue BIN at the grocery store.

            ‘Sue has been at the grocery store for a long time.’

Note that the verbal predicates occur in the -ing form (e.g. BIN singing, BIN knowing); however, stative verbs can also occur in the past (e.g. BIN knew) and have the ‘for a long time’ reading.

The second property that BIN shares with aspectual be is that it indicates that an event recurs; however, unlike aspectual be, the remote past marker situates the initial point of the habit in the remote past:

(25) Sue BIN singing.

       ‘A long time ago, Sue started the habit of singing and she still sings from time to time.’

Finally, BIN can indicate that the entire event is in the remote past. The verb in the construction is past.

(26)     Sue BIN sung.

           ‘Sue sang a long time ago.’

The state of having sung is situated in the remote past, so focus is on the entire event, and the construction expresses perfective aspect. Only non-stative predicates combine with BIN to give the ‘a long time ago’ reading. Although stative predicates can occur in the past (24b, c), they cannot combine with BIN to give the ‘a long time ago’ reading because states have no end. The predicates in (26) have ended and are in the resultant state (see Green (1998, 2002) for more information on the three uses of BIN: stative, habitual, and resultant state).

     The marker dən indicates that the eventuality indicated by the predicate is in its resultant state (see Terry (2005) for more information about the characterization of dən as a perfect). Dən is unstressed relative to other words in the sentence and is accordingly transcribed with a schwa. It only precedes verbs.

(27) a. Sue dən sung that song.

           ‘Sue has already sung that song.’

        b. Bruce dən left.

            ‘Bruce has already left.’

        c. That horse dən jump the fence.

            ‘That horse has already/just jumped the fence.’

Both aspectual be and remote past BIN can occur with dən to give a compositional meaning. In the case of be dən, the reading is habitual resultant state, and in the case of BIN dən, the reading is remote past resultant state.

(28) a. Sue be dən sung.

            ‘Sue has usually already sung.’

       b. Sue BIN dən sung.

           ‘Sue already sang a long time ago.’

Be followed by dən is the only possible order for the markers in (28a); however, some speakers also allow the order dən BIN in (28b).

      The aspectual markers are quite similar to the participials being, been, and done in English, in form, but it is clear that they are distinct from them and can co-occur with the participial verb forms:

(29) a. Sue be being the teacher.

           ‘Sue plays the role of the teacher from time to time.’

        b. Sue BIN been to Thornwell.

            ‘Sue went to Thornwell a long time ago.’

        c. Sue dən done farm work.

            ‘Sue has done farm work before.’

These tense-aspect markers also occur in yes-no questions and negation but cannot occur in a sentence-initial position in questions and host negation in negative constructions:

(30) a. (Do) Sue be singing?

            ‘Does Sue usually sing?’/’Is Sue usually singing?’

        b. Sue don’t be singing.

            ‘Sue doesn’t usually sing.’/’Sue isn’t usually singing.’

Aspectual had (had_ASP, referred to as “preterite had” in Rickford & Théberge-Rafal (1999)), which is distinguished from past perfect had, also has aspectual functions and does not mark tense. In adolescent and adult AAE, it is often used in narrative contexts and indicates the achievement and accomplishment of an event.

(31) I was writing down the numbers then I had wrote down the wrong thing.

       ‘I was writing down the numbers. Then I wrote down the wrong thing.’

Some main verbs in these aspectual constructions occur in the simple past or participial forms.  For instance, write generally occurs in the simple past form (wrote), but sing occurs in the past participial form (sung).

      Other African American English tense-aspect-modality markers are gon, come, stay, and steady. The habitual marker stay occurs with a range of predicate types, such as activity verb and prepositional phrases, but the remaining markers only occur with verbal predicates. Gon [gɔ̃], which precedes non-finite verb forms and refers to a past event, is an evaluative marker that may express frustration, indignation, or incredulity on the part of the speaker. Come is also an evaluative or disapproval marker; however, it precedes V-ing (see Spears (1990) and Moody (2010) for more information on evaluative and disapproval markers in African American English). Finally, steady, which indicates that some part of the eventuality occurs in an intense and consistent manner, also precedes V-ing.

(32) a. Sue stay at the mall.

           ‘Sue goes to the mall often.’

        b. Sue stay travelling.

            ‘Sue travels frequently.’

        c. I went to her house. She gon go to sleep on me.

            ‘I went to her house, and she had the nerve to fall asleep while I was there.’

        d. Sue come telling me that I be giving out too much information.

            ‘Sue had the audacity to tell me that I generally give out too much information.’

        e. Sue steady singing.

            ‘Sue is singing intensely and consistently.’

(33) Sue gon come telling me that I be giving out too much information. (cf. 32d)

       ‘Sue had the audacity to tell me that I generally give out too much information.’

7. Simple sentences ⇫ ¶

The word order in matrix declarative sentences in African American English is Subject-Verb-Object. The direct object of the ditransitive verb in AAE can immediately follow the verb, and the indirect object can be omitted or implicit if the pronominal reference of the direct object is to first person, as in (14) and (15) above.

Reflexives occur in the same clauses as their antecedents and often take the infix own to indicate emphasis or independence.

(34) a. Let him do it by hisownself.

           ‘Let him do it by himself.’

       b. All those people need to find a seat theyownself.

           ‘All those people need to find a seat themselves.’

The call -self reflexive construction is used in evaluative contexts on the part of the speaker who is assessing the state or event carried out by a participant. In general, the evaluation is one in which the speaker does not perceive the participant to be meeting the (personal) standards for the event or state. The following sentence is from Green (2002: 21):

(35) He call hisself cooking.

        Intended: ‘He thinks he’s cooking but is merely playing around in the kitchen. That

        is, he isn’t doing anything remarkable.’

The sentence in (35) would be used in a context in which the speaker does not find the participant to be doing an impressive job of cooking. These constructions have the distributions of reflexives, and are governed by the same type of restrictions. For instance, hisself can only refer to Bruce, not Tyrone in the following sentence: [Tyrone say [Bruce call hisself cooking]]. That is the call -self construction must find its antecedent in the same clause in which it occurs.

      Existential constructions in African American English, which refer to the existence of an entity, have the following components:

Existential pronoun  –  linker   –  NP associate  –  (PP)

The existential pronouns are it and dey, and the linkers that occur in existential constructions are -’s, be, have, and got. The pronoun can be null in existential constructions, as in (36d-e):

(36) a. It’s a fly on my back.

           ‘There’s a fly on my back.’

       b. Dey have a fly on my back.

           ‘There’s a fly on my back.’

       c. It/Dey got a fly on my back.

           ‘There’s a fly on my back.’

       d. __ got two books on the table.

           ‘There are two books on the table.’

       e. __ be two books on the table.

          ‘Generally, there are two books on the table.’

       f. Sometimes it didn’t have no chalk, no book, no teacher.

           ‘Sometimes there weren’t any chalk, any book or any teacher.’ (Green 2002: 77)

      Negation can be marked in multiple positions in simple sentences, resulting in agreement of the negative elements or a single negative meaning.

(37) a. I don’t never have no problems.

           ‘I don’t ever have any problems.’

       b. No game don’t last all night.

           ‘No game lasts all night.’ (Green 2002: 80)

Inversion of the negated auxiliary to the sentence-initial position can also occur in simple sentences.

(38) Don’t no game last all night. (cf. 37b)

        ‘Not a single game lasts all night.’ (Green 2002: 78)

Yes-no questions in African American English exhibit variation in the occurrence of non-modal auxiliaries. Non-modal auxiliaries are optional; modals are obligatory. In questions, past tense is indicated on the verb in instances in which auxiliaries are not produced. Modals carry information other than tense, so they are generally overt in questions, although there may be some exceptions. The following types of questions occur in African American English:

(39) a. He in your group?

           ‘Is he in your group?’

        b. You traded your other one in?

            ‘Did you trade your other one in?’ (Green 2002: 129)

        c. I can draw you a picture?/Can I draw you a picture?

            ‘May I draw you a picture?’

The questions in (39) are all true yes-no questions; they are not rhetorical questions or simply rising declaratives. Past tense in (39b) is expressed on the verb traded, so the event is interpreted as occurring before the utterance time. The modal can (39c) can invert or it can stay in its base position. As questions, the sentences in (39) are produced with final intonational contours that have been associated with yes-no questions in African American English. A number of studies (e.g. Foreman 1999, Thomas 2007) have addressed the final level and falling tones on contours associated with true yes-no questions in African American English.

Matrix wh-questions follow patterns of inversion that are similar to that of yes-no questions. The following are from Green (2002: 85–86).

(40) a.             What did you eat?

            ‘What did you eat?’

        b. What they was doing?

            ‘What were they doing?’

        c. How you knew I was here?

           ‘How did you know I was there?’

        d. Who you be talking to like that?

            ‘Who are you usually talking to like that?’

As all of the wh-questions show, while auxiliary inversion is optional, the wh-word obligatorily occurs in the initial position in the sentence.

8. Complex sentences ⇫ ¶

A number of different types of structures in African American English occur in complex sentences in embedded contexts. Clauses without overt relative pronouns in subject position can be used to modify or make a comment about nouns. These constructions have been referred to as bare relative clauses and topic-comment structures in the literature.

(41) a. It was a nurse and a nurse’s aid __ used to stand up at the door.

            ‘There was a nurse and a nurse’s aid who used to stand up at the door.’

            (Green 2002: 90)

       b. He the man __ got all the old records.

           ‘He’s the man who has got all the old records.’

            (Martin & Wolfram 1998: 32; gloss mine)

       c. I think Aunt M. had a daughter __ lived off.

           ‘I think Aunt M. had a daughter who lived far away.’ (Green 2002: 90)

In line with the topic-comment structure analysis, the claim would be that the phrase a nurse and a nurse’s aid is the topic and used to stand up at the door (41a) is the comment about the topic. Under this view, there is no missing relative pronoun. These constructions are especially compatible with existential (41a) and copula (41b) structures, as also noted in Martin & Wolfram (1998). However, not all such constructions follow those patterns (e.g. 41c).

       Embedded declaratives can be introduced by a sequence of tell… say, as in the following:

(42) a. Sue told him say, “You trying to buy too many cell phones.”

            ‘Sue said to him, “You are trying to buy too many cell phones”.’

       b. Bruce tell me say she ask him if she could borrow $10.

           ‘Bruce told me that she asked him if she could borrow $10.’

Tell…say can introduce a direct quote but also an embedded statement that is indirect speech (42b). Tell can be inflected for tense, as in (42a) told, and it can occur in its -ing form with aspectual be (e.g. She be telling ’em say… ‘She generally tells them that…’).

      Embedded questions with the structure of direct questions also occur in African American English. That is, subject-auxiliary inversion (e.g. do it, will he, did she) can also occur in embedded clauses in indirect question contexts (Green 2002: 88):

(43) a. I wonder [do it be like the water we drink].

           ‘I wonder if it’s like the water we drink.’

       b. Then after the scripture, I’m gonna ask Brother Wall [will he come and play

           “Bless Ye the Lord” for us].

            ‘Then after the scripture, I’m going to ask Brother Wall if he will come and play

            “Bless Ye the Lord” for us.’

        c. I meant to ask her [did she want it].

            ‘I meant to ask her if she wanted it’

        d. I wanted to see [was it the one we bought].

            ‘I wanted to see if that was the one we bought.’

Overwhelmingly, the embedded questions are introduced by verbs such as wonder, ask, and see, which take interrogative complements, but the questions are not direct quotes.

Negative inversion constructions, which occur as matrix sentences, are also licensed in embedded contexts.

(44) a. They told me [(that) didn’t none of the children see anything, but you never

             know].

            ‘They told me that not a single child saw anything, but you never know.’

         b. Let me know [*(if) don’t nobody wanna ride the bus].

            ‘Let me know if not a single person wants to ride the bus.’ (Green 2010)

In the sentence in (44a), the negative inversion construction can be introduced by the overt complementizer that, or it can occur without an overt complementizer (the example was produced in spontaneous speech without a complementizer but it can occur in the structure. On the other hand, the embedded negative inversion construction in (44b) is obligatorily introduced by the complementizer if. The embedded constructions in (44) are quite similar to the embedded indirect questions in (43). There are some important differences, however. One difference is that the embedded constructions in (43) are question complements, and those in (44) are not. Another difference between embedded questions and negative inversion clauses is that complementizers cannot occur with embedded auxiliary inversion in question contexts, but they are optional with negative inversion in some cases and obligatory in others. That the complementizer is required with the embedded clause in (44b) is evidence that the structure is not a question because complementizers cannot co-occur with embedded auxiliary inversion.