Applicativity tests
In addition to the possession test mentioned above, there are two additional criteria in the literature for determining the height of an applicative projection; that is, the transitivity test and the passivization test (cf. McGinnis 2001; Pylkkänen 2002): The first test has to do with the fact that low applicatives cannot combine with an unergative predicate, as shown by the contrast between (7b) and (8b):
-
(7)
-
a.
John baked a cake. [transitive]
-
b.
John baked Bill a cake.
-
(8)
-
a.
I ran. [unergative]
-
b.
*I ran him.
So one can derive a low applicative such as (7b) from a transitive sentence like (7a), whereas it is impossible for an unergative such as (8a) to derive (8b). On the other hand, high applicatives are quite compatible with an unergative predicate, as evidenced by the Chaga example (9):
-
(9)
N-˝a-˝ı-lyì-í-à__'m-kà. [unergative] (Bresnan and Moshi 1990)
FOC-1s-PR-eat-APPL-FV__1-wife
He is eating for/on (his) wife.
The second test concerns the observation that low applicatives only allow passivization of an indirect object (IO), as shown by the contrast between (10a) and (10b):
-
(10)
-
a.
Bill was baked a cake by John. [ApplPlow, IO passivized]
-
b.
* A cake was baked Bill by John. [ApplPlow, *DO passivized]
High applicatives, on the other hand, do not have such restriction. As pointed out by McGinnis (2001, 2003), both high and low applicatives appear in Kinyarwanda, where benefactive constructions allow both IO and DO to undergo passivization, as illustrated in (11a) and (11b) respectively:
-
(11)
-
a.
umukoôbwak__a-ra-andik-ir-w-a__tk__íbárúwa__n’ûmuhuûngu.
girlk__SP-Pres-write-Appl-Pass-Asp__tk__letter__by.boy
The girl is having the letter written for her by the boy. [ApplPhigh, IO passivized]
-
b.
íbárúwak__i-ra-andik-ir-w-a__umukoôbwa__tk__n’ûmuhuûngu.
letterk__SP-Pres-write-Appl-Pass-Asp__girl__tk__by.boy
The letter is written for the girl by the boy. [ApplPhigh, DO passivized]
So both (11a) and (11b) should be classified as high applicatives. On the other hand, the contrast between (12a) and (12b) shows that only IO but not DO can be passivized in Kinyarwanda locative constructions:
-
(12)
-
a.
ishuûrik__ry-oohere-j-w-é-ho__tk__igitabo__n’úúmwáalímu.
schoolk__SP-send-Asp-Pass-Asp-Loc__tk__book__by.teacher
The school was sent the book by the teacher. [ApplPlow, IO passivized]
-
b.
* igitabok__cy-oohere-j-w-é-ho__ishuûri__tk__n’úúmwáalímu.
bookk__SP-send-Asp-Pass-Asp-Loc__school__tk__by.teacher
The book was sent to school by the teacher. [ApplPlow, *DO passivized]
These locative construals therefore count as instances of low applicatives. As a result, we have secured some reliable tests at hand to map out the topography of applicatives across languages.
Now, we are in a position to apply the tests one by one to outer affectives in Mandarin: First, consider (4a). As revealed by its truth-conditional distinction from (4b), there is no directional possessive relationship between the Affectee wǒ ‘me’ and the Theme 三瓶酒 sān-píng jǐu ‘three bottles of wine’. So it fails the possession test. Secondly, this type of affectives has no trouble with taking an unergative predicate, as in (13), passing the transitivity test with flying colors:
-
(13)
他居然[給我]哭了。
tā__jǖrán__[gěi__wǒ]__kū-le
he__unexpectedly__[AFF__me]__cry-Inc
Unexpectedly, he started crying on me.
Both results indicate that (4a) belongs to the class of high applicatives. The passivization test, on the other hand, is rather inconclusive: Passivization appears to be blocked for both IO and DO, as in (14a) and (14b) respectively:
-
(14)
-
a.
* 我被他居然給喝了三瓶酒。
wǒ__bèi__tā__jǖrán__gěi__hē-le__sān-píng__jǐu
I__BEI__he__unexpectedly__AFF__drink-Prf__three-bottle__wine
-
b.
* 有三瓶酒被他居然[給我]喝了。
yǒu__sān-píng__jǐu__bèi__tā__jǖrán__[gěi__wǒ]__hē-le
have__three-bottle__wine__BEI__he__unexpectedly__[AFF__me]__drink-Prf
Peripheral construals of Mandarin outer affectives
The next step is to see how Mandarin affectives should be encoded in syntactic structures. In the following discussion, I will attempt to show that the so-called high applicatives are not high enough, and there are even higher applicatives merged to the left periphery (or to the CP phase to the same effect). They display a cluster of properties typically associated with the information structure.
Our first argument is built upon the fact that the outer affective of (15a) ((4a) repeated here) is licensed through a special kind of illocutionary force. As indicated by the following contrast, the presence of the evaluative adverb 居然 jūrán ‘unexpectedly’ is obligatory, without which the sentence becomes declarative, and is ruled out consequently, as in (15b):
-
(15)
-
a.
他居然[給我]喝了三瓶酒!
tā__jǖrán__[gěi__wǒ]__hē-le__sān-píng__jǐu
he__unexpectedly__[AFF__me]__drink-Prf__three-bottle__wine
Unexpectedly, he drank three bottles of wine on me!
-
b.
* 他[給我]喝了三瓶酒。
tā __[gěi__wǒ]__hē-le__sān-píng__jǐu
he__[AFF__me]__drink-Prf__three-bottle__wine
The same restriction does not apply to its inner counterpart in (16a) ((4b) repeated here). As indicated by the well-formedness of (16b), the pseudo double object construal does not require an evaluative adverb and is quite compatible with the declarative usage:
-
(16)
-
a.
他居然喝了我三瓶酒!
tā__jǖrán__hē-le__wǒ__sān-píng__jǐu
he__unexpectedly__drink-Prf__me__three-bottle__wine
Unexpectedly, he drank three bottles of wine on me!
-
b.
他喝了我三瓶酒。
tā__hē-le__wǒ__sān-píng__jǐu
he__drink-Prf__me__three-bottle__wine
He drank three bottles of wine on me.
Interestingly enough, the licenser in question can also be replaced either by a reversal adverb like 卻 qùe ‘however’ in (17a) or by an outer wh-adverb such as 怎麼 zěnme ‘how come’ in (17b):
-
(17)
-
a.
我叫他幫我買酒。他卻給我買了菸!
wǒ__jìao__tā__bāng__wǒ__mǎi__jǐu__tā__qǜe__[gěi__wǒ]__mǎi-le__yān
I__ask__him__help__me__buy__wine__he__however__[AFF__me]__buy-Prf__cigarette
I asked him to help me to buy wine. He, however, bought cigarettes on me!
-
b.
我叫他幫我買酒。他怎麼[給我]買了菸?!
wǒ__jìao__tā__bāng__wǒ__mǎi__jǐu__tā__zěnme__[gěi__wǒ]__mǎi-le__yān
I__ask__him__help__me__buy__wine__he__how.come__[AFF__me]__buy-Prf__cigarette
I asked him to help me to buy wine. How come he bought cigarettes on me?!
In addition, this type of affective construals can also be licensed by imperative and negative moods, as illustrated by (18a) and (18b) respectively:
-
(18)
-
a.
[給我]跪下!
[gěi__wǒ]__gùi-xìa
[AFF__me]__kneel-down
Kneel down for my sake!
-
b.
阿Q從沒[給我]丟過臉!
ākīu__cóng-méi__[gěi__wǒ]__dīu-gùo__lǐan
Akiu__ever-have.not__[AFF__me]__lose-Exp__face
Akiu has never lost face on me!
Another argument stems from the fact that outer affectives marked by gěi is strictly speaker-oriented: As mentioned above, the Affectee argument in question can only be first-person singular, as in (5a)–(5c), and there is again no such restriction for their inner counterparts throughout (6a)–(6c). Furthermore, there is a presupposition associated with the affective construal of (15a), i.e., the wine-drinking event should never happen in the first place. As noted by Shu (2008), certain sentential adverbials carry a focus interpretation, and an evaluative adverb like jǖrán ‘unexpectedly’ may well take a focused constituent as its complement. Therefore, it seems reasonable to suggest that the focused portion of (15a) (presumably IP) carries the new information “he drank three bottles of wine”, which is indeed unexpected given the presupposition mentioned above. All these are the hallmark of peripheral construals associated with the information structure and should therefore be encoded syntactically on the complementizer layer.
From a cross-linguistic point of view, there is a class of non-argumental clitics in Modern Greek (and Romance languages in general) which are very much in line with the peripheral construals of Mandarin affectives. They typically express some sort of affectedness between an individual and an event. These so-called ethical datives observe a speaker/hearer-oriented restriction similar to Mandarin high affectives (cf. Cuervo 2003; Jaeggli 1982; Michelioudakis and Sitaridou 2008; Perlmutter 1971, among others), as in (19):
-
(19)
mu/su/?tu arostise i Maria.
me/you/him__fell.ill__on__Mary
Mary fell ill on me/you/him.
Modern Greek ethical datives also require licensing from imperative, optative, subjunctive, or negative moods, as shown by the contrast between (20a) and (20b):
-
(20)
-
a.
na mu prosechis! [imperative]
Subj__eth.dat.1S__take.care
Take care, for my sake!
-
b.
?? mu prosechis. [??declarative]
eth.dat.1S__take.care
You take care, for my sake.
These clitics are often analyzed either as CP adjuncts (cf. Catsimali 1989) or as the realization of the φ-features of a high applicative head which takes the whole event as its argument, merging well beyond vP (cf. Cuervo 2003).
Our observations thus point to the conclusion that the otherwise puzzling restrictions on outer affective construals have a close bearing on the force and/or mood of the entire sentence. In other words, the “extra” argument introduced by gei may well be associated with some functional projection beyond the vP phase, presumably taking its hold in the left periphery.
Mapping the topography of Mandarin high applicatives
One question still remains as to how to separate Mandarin outer affectives from their benefactive/goal counterparts, which are also marked by gěi, as exemplified by (21a) and (21b):
-
(21)
-
a.
阿Q[給我/我們/你/你們/他/他們]烤了一塊蛋糕。 [benefactive]
ākīu__[gěi__wǒ/wǒmen/nǐ/nǐmen/tā/tāmen]__kǎo-le__yí-kùai__dàngāo
Akiu__[for__me/us/you/you(pl.)/him/them]__bake-Prf__one-CL__cake
Akiu baked a cake for me/us/you/you(pl.)/him/them.
-
b.
阿Q[給我/我們/你/你們/他/他們]寫了一封信。 [goal]
ākīu__[gěi__wǒ/wǒmen/nǐ/nǐmen/tā/tāmen]__xǐe-le__yì-fēng__xìn
Akiu__[to__me/us/you/you(pl.)/him/them]__write-Prf__one-CL__letter
Akiu wrote a letter to me/us/you/you(pl.)/him/them.
For one thing, it is impossible to topicalize the outer Affectee along with gěi to the sentence-initial position, as evidenced by the contrast between (22a) and (22b). By contrast, there is no such restriction for benefactive and goal construals, which can be made clear by comparing (21a) and (21b) with (23a) and (23b) respectively:
-
(22)
-
a.
阿Q居然[給我]拿了錢就跑。
ākīu__jǖrán__[gěi__wǒ]__ná-le__qían__jìu__pǎo
Akiu__unexpectedly__[AFF__me]__take-Prf__money__then__run
Unexpectedly, Akiu took the money and ran away on me.
-
b.
* [給我], 阿Q居然拿了錢就跑。
[gěi__wǒ]__ākīu__ jǖrán__ná-le__qían__jìu__pǎo
[AFF__me]__Akiu__unexpectedly__take-Prf__money__then__run
Unexpectedly, Akiu took the money and ran away on me.
-
(23)
-
a.
? [給他們], 阿Q烤了一塊蛋糕。
[gěi__tāmen]__ākīu__kǎo-le__yí-kùai__dàngāo
[for__them]__Akiu__bake-Prf__one-CL__cake
Akiu baked a cake for them.
-
b.
[給他們], 阿Q寫了一封信。
[gěi__tāmen]__ākīu__xǐe-le__yì-fēng__xìn
to__them__Akiu__write-Prf__one-CL__letter
Akiu wrote a letter to them.
This indicates that gěi may well serve as a true applicative head in outer affective construals, while it is merely a preposition forming a constituent with the Beneficiary/Goal argument.
For another thing, it is possible to distinguish the outer Affectee in question from a Beneficiary argument by their word ordering relative to a manner adverb such as 慢慢地 mànmàndi ‘slowly’. As illustrated in (24a), the reading is affective when the gěi phrase precedes the manner adverb, and we may change it into benefactive by placing the gěi phrase after the manner adverb, as in (24b):
-
(24)
-
a.
阿Q居然[給我]慢慢地擦地板。 [affective]
ākīu__jǖrán__[gěi__wǒ]__mànmàndi__cā__dìbǎn
Akiu__unexpectedly__[AFF__me]__slowly__wipe__floor
Unexpectedly, Akiu slowly wiped the floor on me.
-
b.
阿Q居然慢慢地[給人家]擦地板。 [benefactive]
ākīu__jǖrán__mànmàndi__[gěi__rénjīa]__cā__dìbǎn
Akiu__unexpectedly__slowly__[for__people]__wipe__floor
Unexpectedly, Akiu slowly wiped the floor for others.
In fact, the two construals can marginally co-occur in the same sentence, separated by the manner adverb, as evidenced by (25):
-
(25)
? 阿Q居然[給我]慢慢地[給人家]擦地板。
ākīu__jǖrán__[gěi__wǒ]__ mànmàndi __[gěi__rénjīa]__cā__dìbǎn
Akiu__unexpectedly__[AFF__me]__slowly__[for__people]__wipe__floor
Unexpectedly, Akiu slowly wiped the floor for others on me.
The intuition explored above can then be implemented under the split CP approach pioneered by Rizzi (1997), as visualized in the following diagram (irrelevant details omitted):
-
(26)
The topography of Mandarin outer affectives and benefactives
Here, we envision the affective gěi to be a high applicative head, which hosts the outer Affectee. It raises to the evaluative head subsequently to license the peripheral construals of outer affectives. As for the benefactive (and goal) PP, we place it at the peripheral area of vP as a working hypothesis (also cf. Cinque 1999), separated from the inflectional layer only by the manner adverb, presumably a delimitator for the vP phase.
For one thing, it may seem a bit overstretched to place an applicative projection as high as the complementizer layer. Nevertheless, given the definition that a high applicative dependency holds between an individual and an event, we should expect a natural extension to the left periphery if the applied argument is involved in a speaker/addressee-oriented construal. In fact, it has been suggested by Buell (2003) that DP locative applicatives in Zulu must merge above the Agent (hence outside vP). A similar proposal is also made by McGinnis and Gerdts (2003) for Kinyarwanda benefactives, based on the locality effects observed in various types of multiple applicatives. It is thus not unimaginable that, through the process of grammaticalization along the line of Roberts and Roussou (1999), an applicative can evolve a step further into the domain of peripheral construals.
Our treatment also accounts for the fact that the passivization test turns out to be inconclusive in Mandarin. This is because the affectivity involved is far beyond the core argument structure associated with vP, and therefore does not observe the usual morpho-syntax of passives.