A-not-A question

This is the most atomic form of the A-not-A question, which contains two identical instances of the constituent A separated by negation.

This is a more complex form, containing two instances of the complex constituent AB separated by the negation. AB may represent an embedded clause, a subject joined with a prepositional phrase, or a verb phrase containing a DP.

This form contains two unique constituents, A and AB, separated by the negation. A and AB are similar in that AB contains the entire content of A, but constituents are present in AB that are not present in A.

This form is similar to the A-not-AB form, but the more complex constituent AB occurs before the negation.

This form is only found in instances where A is disyllabic constituent with initial syllable a, and the two constituents are separated by negation.

This form is similar to the a-not-A form with a representing the initial syllable of A and the two separated by negation, but A is joined to another constituent to form the complex constituent AB.

Below are examples of English approximations of the A-not-A question. They are similar to the Sinitic A-not-A in that they present two possibilities and require an echoed response. However, they include an extra segment ("or" in the below examples) in order to read grammatically, which changes these approximations to an alternative question (AltQ) type. This extra segment is not seen in Sinitic A-not-A questions, and in fact the Mandarin segment 還是 haishi 'or' is used to contrast the syntax of the A-not-A form and demonstrate the latter's sensitivity to islands. Nevertheless, for the convenience of understanding this phenomenon from the perspective of an English speaker, the below examples are included to provide context.

(1) Was John at the party or not at the party?

(2) Was John at the party last night or not at the party?

(4) Was John at the party or not at the party last night?

(3) Was John at the party last night or not at the party last night?

It is proposed that the A-not-A sequence is morpheme created by the reduplication of the interrogative morpheme (represented by the A in A-not-A).[4] Though the specific syntactic location of this morpheme is not agreed upon, it is generally accepted that the A-not-A sequence is essentially a word formed by the concatenation of an abstract question morpheme and this duplicated predicate, which likens it to a VP-proclitic. This Morpheme is referred to as NQ in order to represent its character as negative and interrogative.

An extensive cross-dialectic survey conducted in 1985 concluded that the Taiwanese question particle kam appears in the same contexts as the hypothesized Mandarin NQ.[3] From this, it was concluded that kam-type questions and A-not-A questions are in complementary distribution: a language either has kam-type questions or A-not-A questions but not both. It was also interpreted that kam and NQ are "different morphological exponents of the same underlying morpheme".[4]

Unlike the yes/no question type formed using the segment "ma", the A-not-A question can be embedded, and may scope beyond its own clause. This scoping may be blocked if the original location of NQ and its intended final location are separated by an island boundary.[7] These distributional characteristics of NQ are parallel to non-nominal adjunct question particle weishenme 'why'. Due to the uncontroversial nature of the movement-based analysis of weishenme, the similarity of the NQ to weishenme implies that NQ may be subject to the same analysis of its movement.

The dominant view on A-not-A questions is that NQ is similar to a wh-word and related by the movement of NQ.[8][7] This movement is not seen in alternative-type questions using haishi 'or', and therefore delineates A-not-A questions from alternative questions in terms of structure. Due to this syntactic differentiation, A-not-A questions may be contrasted with haishi questions for the purpose of revealing island sensitivity.

In forming A-not-A questions, A must remain the same on both sides. A is essentially a variable which can be replaced with a grammatical particle such as a modal, adverb, adjective, verb, or preposition.

In Mandarin, there are 6 attested patterns of A-not-A: A-not-A, AB-not-AB, A-not-AB, AB-not-A, a-not-A, and a-not-AB of which "A" stands for the full form of the predicate, "B" stands for the complement, and "a" stands for the first syllable of a disyllabic predicate.[9]

Example (3) illustrates that A-not-A pattern, where A is the verb qu 'go', and qu bu qu is 'go not go'.

Example (4) illustrates the AB-not-AB pattern, where AB is the constituent consisting of the verb rende, 'know', as A, and the complement zhe ge ren, 'this CL man', as B, combining to form the AB constituent rende zhe ge ren 'know this CL man'. This produces rende zhe ge bu rende zhe ge ren, 'know this CL man not know this CL man.'

Example (5) illustrates the A-not-AB pattern, where A is the verb rende, 'know', AB is the constituent consisting of the A verb rende, 'know', and the complement zhe ge ren, 'this CL man', as B, combining to form the AB constituent rende zhe ge ren 'know this CL man'. This produces rende bu rende zhe ge ren, 'know not know this CL man'.

Example (6) illustrates the AB-not-A pattern, where AB is the constituent rende zhe ge ren, 'know this CL man' consisting of rende, 'know' as A and zhe ge ren, 'this CL man as B; A is likewise rende, 'know', in the second part of the construction. This produces rende zhe ge ren bu rende, 'know this CL man not know'.

Example (7) illustrates the a-not-A pattern, where a is the first syllable, fang, of the disyllabic predicate fangbian, 'convenient', and A is the full predicate fangbian, and fang-bu-fangbian is 'con(venient)-not convenient'.

Example (8) illustrates the a-not-AB pattern, where a is the first syllable, he, of the disyllabic predicate heshi, 'suitable', and AB is the constituent consisting of heshi, 'suitable' as A and jiao quan, 'teaching fist' as B, combining to form the AB constituent heshi jia quan, 'suitable teaching fist'. This produces he-bu heshi jian quan, 'suit(able)-not suitable teaching fist'.

A-not-A can be formed by a verb, an adjective, or an adverb,[11] as well as modals.[1]

In the interrogative clause, A-not-A occurs by repeating the first part in the verbal group (with the option of an auxiliary) and the negative form of the particle is placed in between. However, this clause does not apply when using perfective in aspect. Instead, 沒有; meiyou is used to replace the repeated verb used in A-not-A form.[12]

V-NEG-V type:[1]

Here, the verb qu, 'go', is A, and there is no object.

         (9.a) 你去不去？         A: 去/不去
               ni qu bu qu?         qu/bu qu
               you go not go        go/not go
               DP V-NEG-V           V/NEG V
               Are you going?       Yes/No

V-NEG-V-Object type:[1]

Here the verb kan, 'watch', is again A, and while there is an object, the object is not included in "A", and is therefore not reduplicated.

         (9.b) 你看不看电影？                     A: 看/不看
               ni kan bu kan dianying?              kan/bu kan
               you watch not watch movie            watch/not watch
               N V-NEG-V-N                          V/NEG V
               Will you watch the movie?            Yes/No.

V-Object-NEG type:[1]

Here, the verb kan, 'watch', is likewise A, and while the object is included before NEG, it is not included in A, and is therefore not reduplicated, although it remains an option.

         (9.c) 你看电影不？                     A: 看/不看
               ni kan dianying bu?                kan/bu kan
               you watch movie not                watch/not watch
               N  V-DP-NEG                        V/NEG V
               Will you watch the movie?          Yes/No

V-Object-NEG-V type[1](debatable):

Here, the verb kan, 'watch', is also used for A, and while the object is included before NEG, it is not included in A, and is therefore not reduplicated. A is reduplicated here.

         (9.d) 你看电影不看？                  A: 看/不看
               ni kan dianying bu kan?           kan/bu kan
               you watch movie not watch         watch/not watch
               DP V-DP-NEG-V                     V/NEG V
               Will you watch the movie?         Yes/No

Answers to (9.a), (9.b), (9.c), and (9.d) must be in the form "V" or "not-V"

There is some debate among speakers as to whether or not 3.d. is grammatical, and Gasde argues that it is.

A-NEG-A type:[1]

Here, the adjective hao, 'good', is A, and it is reduplicated. The word ben is a classifier, which means it is a counter word for the noun 'book'.

         (10.a) 这本书好不好？               A: 好/不好
                zhe ben shu hao bu hao?        hao/bu hao
                this CL book good not good     good/not good
                DP          A-NEG-A            A/NEG A
                Is this book good?             Yes/No

A-NEG type:[1]

Here, the adjective hao, 'good', is A, but it is not reduplicated.

         (10.b) 这本书好不？                      A: 好/不好
                zhe ben shu hao bu?                 hao/bu hao
                this CL book good not               good/not good
                DP           A-NEG                  A/NEG A
                Is this book good?                  Yes/No

Answers to (10.a), (10.b) must be in the form A or not-A.

P-NEG-P type:[13]

Here, the preposition zai, 'at', is A, and it is reduplicated.

         (11.a) 张三在不在图书馆？                  A: 在/不在
                Zhangsan zai bu zai tushuguan?        zai/bu zai
                Zhangsan at not at library            at/not at
                DP   P-NEG-P DP                       P/NEG P
                Is Zhangsan at the library?           Yes/No

P-NEG-P type:[13]

Here, the preposition zai, 'at', is A, and it is not reduplicated.

         (11.b) 张三在图书馆不？                A: 在/不在
                Zhangsan zai tushuguan bu?        zai/bu zai
                Zhangsan at library not           at/not at
                DP  P-DP-NEG                      P/NEG P
                Is Zhangsan at the library?       Yes/No

Answers to (11.a) and (11.b) must be in the form P or not-P.

M-NEG-M-V-Object type:[1]

Here, the modal dare is A and it is reduplicated.

         (12.a) 你敢不敢杀鸡?                       A: 敢/不敢
                ni gan bu gan sha ji?                 gan/bu gan
                you dare not dare kill chicken        dare/not dare
                N M-NEG-M -V -DP                      M/NEG M
                Do you dare kill chicken?             Yes/No

The answer to (12.a) must be in the form M or not-M.

Despite having the same negative marker as Mandarin, "不" bat1 is only used in fixed expressions or to give literacy quality,[14] and only "唔" m4 is used as a negative marker in A-not-A questions.[9]

One distinction in Cantonese when compared to Mandarin is that certain forms of A-not-A questions are not attested due to dialectal differences.[2]

Like its Mandarin counterpart, this form is attested in Cantonese as shown by the sentence pair in (13),[9] where in example (13.a), A is the verb lai, 'come', and lai m lai is 'come not come', and in (13.b), A is the verb lai, 'come', and lai bu lai is 'come not come'.

As shown by (14.a), this is not an attested form in Cantonese, unlike the counterpart in Mandarin in (14.b).[2]

Here in (14.a) A is the verb zungji, 'like', and B the noun jamok, 'music', producing the AB form zungji jamok, 'like music'. This would produce the ungrammatical structure zungji jamok m zungji jamok, 'like music not like', which is a poorly-formed sentence in Cantonese.

In the well-formed sentence shown below in (14.b), A is the verb xihuan, 'like', and B is the noun yinyue, 'music', producing the AB form xihuan yinyue, 'like music'. This produces xihuant yinyue bu xihuan yinue, 'like music not like music', a grammatical sentence in Mandarin.

This form is only attested in Cantonese if the predicate is a monosyllabic word as shown by (15.a), where A is the verb faan, 'return', and AB is the constituent faan ukkei, 'return home'. This can be compared to the Mandarin counterpart in (15.b) where A is the verb hui, 'return', and AB is the constituent hui jia, 'return home.[9]

A-not-AB is not attested in Cantonese if the predicate is a bi-syllabic word as shown by (16.a), where A would be the verb zungji, 'like', and AB would be the constituent zungji jamok, 'like music'. This contrasts with its Mandarin counterpart in (16.b), where A is the verb xihuan, 'like', and B is the complement yinyuee, music', combining into the AB form xihuan yinyue, 'like music'.[2] In such cases, Cantonese speakers usually use the form a-not-AB, like (8).[9]

This form is only attested in Cantonese if the predicate is a monosyllabic word A, exemplified in (17.a) with the verb faan, 'return', with an object B, exemplified in (17.a) with the noun ukkei, 'home'. (17.a) is shown below with its Mandarin counterpart in (17.b), where A is the verb hui, 'return', and B is the noun jia, 'home'.[9]

Such forms of AB-not-A in monosyllabic words are used by older generations.[15]

When the predicate is a bi-syllabic word, then AB-not-A form is not attested as shown in (18.a), unlike its Mandarin counterpart in (18.b).[9]

Amoy exhibits A-not-A forms, and differs from Mandarin and Cantonese in its frequent use of modals or auxiliaries in forming these constructions. Amoy forms also differ in that the morphemes for A do not match each other in a given sentence. In these constructions one of the morphemes may also be deleted, as can be seen in Examples (27), (28), and (29), though when this happens it may only be deleted from the negative predicate.

The following negative markers are used.[16] Alternate transliterations are shown in bold.

         (19)   a.  m        negative of volition (m-1)                                
                b.  m        negative simplex (m-2)
                c.  bo       negative possessive/existential/affirmative aspect 
                    bou         
                d.  bue      negative potential/possibility
                e.  be       negative perfective aspect

While m-1 occurs as a free morpheme with its own semantic feature indicating volition, m-2 cannot function by itself as a verb and works only to express negation. It is attested only with a limited amount of verbs.

Shown below are A-not-A constructions in Amoy.[17]

The following is a list of A-not-A constructions in Amoy with auxiliary verbs which may function as the main verb of a sentence.

The auxiliary verb u here functions as an aspectual marker indicating that an action has been completed. In u — bou A-not-A constructions, u functions as the first A, corresponding with the auxiliary 'have', while bou functions as the second A of the A-not-A construction, corresponding with the negative counterpart 'not have'. Example (20) illustrates the use of this construction.

The auxiliary verb bat functions as an aspectual marker indicating experience. In bat — m bat A-not-A constructions, bat functions as the first A, corresponding with an auxiliary expressing the sense of 'to have experienced', while m bat functions as the second A of the A-not-A constructions, corresponding with the negative counterpart 'not to have experienced'. Example (21) illustrates the use of this construction.

The auxiliary verb si works to express emphasis. In si — m si constructions A-not-A constructions, si functions as the first A, roughly corresponding with 'to be', and m si as the second A, indicating the negative counterpart 'not to be'. Example (22) illustrates the use of this construction.

The following is a list of A-not-A constructions in Amoy with auxiliary verbs which may never be used as the main verb of a sentence.

The use of a beq — m construction is used to express an intention or an expectation. In these constructions, beq functions as the first A, indicating 'to want to', and m as the second A, here working with beq to express its negative counterpart 'not want to.' Example (23) illustrates the use of this construction.

The use of a tiouq — m bian construction expresses a sense of obligation. In these constructions, tiouq functions as the first A, indicating 'must', and m bian as the second A, here indicating the negative counterpart 'must not'. Example (24) illustrates the use of this construction.

The use of a t'ang — m t'ang construction expresses a sense of permission. In these constructions, t'ang functions as the first A, indicating 'may', and m t'ang as the second A, here indicating the negative counterpart 'may not'. Example (25) illustrates the use of this construction.

The use of an e — bue construction expresses a sense of possibility or probability. In these constructions, e functions as the first A, indicating 'could', and bue as the second A, here indicating the negative counterpart 'could not'. Example (26) illustrates the use of this construction.

The use of an e tang — bue tang construction expresses a sense of the ability to do something. In these constructions, e tang functions as the first A, indicating 'can', and bue tang as the second A, here indicating the negative counterpart 'can't'. Example (27) illustrates the use of this construction as well as an instance of deletion from the negative predicate.

The use of an e sai — bue sai construction expresses a sense of a potential ability to do something. In these constructions, e sai functions as the first A, indicating 'could', and bue sai as the second A, here indicating the counterpart 'couldn't'. Example (28) illustrates the use of this construction as well as an instance of deletion from the negative predicate.

The use of an e hiau — bue hiau construction expresses a sense of one's knowledge. In these constructions, e hiau functions as the first A, indicating 'to know how', and bue hiau as the second A, here indicating the negative counterpart 'not to know how'. Example (29) illustrates the use of this construction as well as an instance of deletion from the negative predicate.

Both an and mos can precede the predicate in A-not-A questions.

Example (26) illustrates the use of an, short form for ani-, which expresses simple negation. Here A is ca-ni, 'sleep-COMP', and ca-ni an ca-ni is 'sleep-COMP not sleep-COMP'.

Example (27) illustrates the use of mos, which expresses impossibility or inability. Here A is ka-ss-ni, 'go-PAST-COMP' and ka-ss-ni mos ka-ss-ni is 'go-PAST-COMP cannot go-PAST-COMP'.

Korean has three negative predicates that can form A-not-A question, molu-, eps-, and ani-.

Example (28) illustrates the use of molu-, which means 'don't know'.

Example (29) illustrates the use of esp-, which means 'do not have; do not exist'.

Example (30) illustrates the use of ani-, which means 'is not'.

Meaning 'desist from', mal follows an affirmative polar question, and will occur instead of a reduplicated full verb that has a post predicate negation, meaning that there is only one full verb in this type of A-not-A question.

However, the modal auxiliary verb mal is restricted in that it does not co-occur in predicative adjectives or the factual complementizer ni. Moreover, with mal being a bound form, it cannot be the echo negative answer. Instead, the full negative verb will be provided as the answer, taking an negation, as illustrated in (32).

The A-not-A operator is defined as an MWd. The A-not-A operator can only lower to a MWd which is immediately dominated by the maximal projection that is also immediately dominated by the maximal projection of the A-not-A operator. An SWd cannot be the target for the A-not-A operator. In addition, if there is an intervening MWd or SWd between the A-not-A operator and its target, the A-not-A operation fails.[13]

A-not-A operator lowering must satisfy four conditions:

1. The A-not-A operator targets the closest MWd that is the X′-theoretic head that it c-commands.
2. Closeness of the head is qualified by: (i) The closest head is a X′-theoretic head of the maximal which is immediately dominated by the maximal projection of the A-not-A operator. (ii) The target must have overt phonological realization.
3. There is not any non-X′-theoretic head or SWd intervening between the A-not-A operator and its target.
4. Intervention is defined by c-command relation.

Grammatical A-not-A operator lowering to adjacent MWd verb xihuan. Xihuan is the closest X′-theoretical head that the A-not-A operator c-commands, thus satisfies lowering conditions.

Ungrammatical A-not-A operator lowering to adjacent MWd hen. Even though hen is the closest adjacent MWd, it is not an X’-theoretical head, thus lowering to it violates A-not-A lowering conditions

After lowering, the A-not-A operator triggers reduplication on the target node. The reduplication domain can be the first syllable of the targeted element, the targeted element itself, and the maximal projection that contains the targeted element. Reduplication is linear and the A-not-A operator cannot skip the adjacent constituent to copy the next constituent.[13]

In first syllable reduplication, the A-not-A operator copies the first syllable of the adjacent MWd and moves the reduplicant, i.e. copied syllable, to the left of the base MWd. Then the negation is inserted between the reduplicant and base to form a grammatical sentence. In (33.a), the A-not-A operator copies the first syllable tao of the MWd taoyan. The reduplicant tao is put at the left of the base taoyan and then the negative constituent bu is inserted in between. In figure (33.b) *Zhangsan taoyan Lisi-bu-tao is ungrammatical because tao cannot be put to the right of the maximal projection VP, taoyan Lisi.

In MWd reduplication, the A-not-A operator copies the adjacent MWd and moves the reduplicant MWd overtly to the left of the base MWd or to right of the base maximal projection containing the MWd. Otherwise, the reduplicant can move covertly, i.e. in such a way that there is no overt surface evidence, to the right of the base maximal projection containing the MWd. The negation is then inserted between the reduplicant and base to form a grammatical sentence. In (34.a) the A-not-A operator copies the MWd taoyan. The reduplicant taoyan is overtly put at the left of the base taoyan and then the negative constituent bu is inserted in between. In (34.b) the A-not-A operator copies the MWd taoyan. The reduplicant taoyan is overtly put at the right of the base taoyan Lisi and then the negative constituent bu is inserted in between. In (34.c) the A-not-A operator copies the MWd taoyan. The reduplicant taoyan is covertly put at the right of the base taoyan Lisi after which the negative constituent bu is inserted.

In maximal projection reduplication, the A-not-A operator copies the maximal projection that contains the adjacent MWd and moves the reduplicant either to the left or to the right of the base. The base may be just the MWd or the maximal projection containing the MWd. The maximal projection may be any XP (VP, AP, PP etc.). The negation is then inserted between the reduplicant and base to form a grammatical sentence. In (35) the A-not-A operator copies the maximal projection VP taoyan Lisi. The reduplicant taoyan Lisi is put at the left of the base taoyan Lisi and then the negative constituent bu is inserted in between.