Pernilla Hallonsten Halling

be-hofshiut.
with-freedom
manner

‘He acts freely.’

2. 
   Hu
   

3sg
cava
painted
et
acc
ha-kir
det-wall
be-adom
in-red
resultative

‘He painted the wall red.’
Manner adverbs, depictives, and resultatives have thus repeatedly been treated as be- longing to one and the same domain conceptually (Schultze-Berndt & Himmelmann 2004; Himmelmann & Schultze-Berndt 2005b; van der Auwera & Malchukov 2005; Loeb-Diehl 2005; Verkerk 2009). Apart from motivations such as cross-linguistic evidence of encoding overlaps and conceptual affinity in general, the explanations for this differ. Loeb-Diehl (2005) and Verkerk (2009) treat the phenomena as instances of secondary predication (see further discussion in section 3.3.2). They attribute similarities and differences within this domain to participant- versus event-orientation (cf. discussion in section 2.3.2 above) and simultaneity with the main event. Depictives and resultatives are then argued to share the feature of being participant-oriented, whereas manner adverbs are event-oriented. De- pictives and manner adverbs, however, are argued to share the feature that what they describe is simultaneous with the main event (e.g. Verkerk 2009: 117), whereas resultatives are consecutive to it (e.g. Loeb-Diehl 2005: 14). Loeb-Diehl (2005) argues that resulta- tives and manner adverbs are the most distant from each other. This can be compared to the comprehensive comparison of English and Japanese resultatives by Washio (1997), who shows that there are different subtypes of resultatives, some of which are closer to manner adverbs.⁹ In addition to the features that unite depictives and resultatives on the one hand, and depictives and manner adverbs on the other, Verkerk (2009: 118) proposes a third feature that unites manner adverbs and resultatives: “manner predications and resultatives do not refer to the subject participant introduced in the primary predicate”. This factor is argued to motivate examples such as that from Hebrew in (2.20). Although

⁸ The difference in glossing in (2.20) is based on the original sources (as indicated in the example), whereas Verkerk glosses both instances of be- as ‘in’ (2009: 123).

⁹ For another comprehensive study of resultatives in European languages, see Riaubiene˙ (2015).

subject-oriented resultatives are attested, these are argued to be marginal in this con- text (2009: 117-118). In Verkerk’s sample, the most common strategy is that manner adverbs, resultatives, and depictives have the same encoding. Verkerk (2009) illustrates this with examples from Lao, which has a serial verb construction in all three functions. These examples can perhaps be questioned, since they are zero-marked, which is not very distinctive – other examples would be useful here, but these are the only ones at hand.

(2.21) Lao (Tai-Kadai) (Enfield 2007: 398, 401, 410)

1. 
   man2
   

3sg
kin3
eat
paa3
fish
nii4
dem
vaj2.
fast
manner

‘He ate this fish fast.’

2. 
   man2
   

3sg
kin3
eat
siin4
meat
dip2
raw
depictive

‘He eats meat raw.’ (also: ‘He eats raw meat.’)

3. 
   laaw2
   

3sg.fam
ñing2
shoot
nok1
bird
taaj3
die
resultative

‘She shot a bird dead.’
Both Verkerk (2009) and Loeb-Diehl (2005) motivate identical encoding of manner expressions, depictives, and resultatives with the argument that all three functions are instances of secondary predication. Himmelmann & Schultze-Berndt (2005b) instead sep- arate secondary predicates, including depictives and resultatives, from adverbials, arguing that this distinction is often drawn based on participant orientation. Whereas secondary predicates are participant-oriented, adverbials are event-oriented (again, cf. section 2.3.2). Still, Himmelmann & Schultze-Berndt (2005b) show that this traditional distinction does not align with the distribution of depictives and adverbials. In conclusion, depictives, re- sultatives and adverbials are naturally treated as belonging to one and the same domain. Importantly though, it does not appear to be necessary to make secondary predication the basis for such a domain for Himmelmann & Schultze-Berndt (2005b).
The second most common encoding for manner in Verkerk’s (2009) sample appears to be for a language to have a separate strategy here (though it should be noted that more than one strategy is often found in a specific language). This is the case in Icelandic, where depictives and resultatives are encoded in the same way, with adjectives agreeing with their NP head in terms of gender, case, and number. In contrast, manner adverbs always take the form of a neuter accusative singular adjective. Very similar patterns are attested in Swedish and Russian (Indo-European).
(2.22) Icelandic (Indo-European) (Whelpton 2006: 7, 10)

1. 
   þeir
   

they.m.nom.pl
voru aD
were to
keyra
drive
allt
all
of
too
hratt.
fast.neut.acc.sg
manner

‘They were driving way too fast.’

2. 
   ViD
   

we.nom
kláruDum
finished
kjötbollurnar
meatball.def.f.acc.pl
kaldar.
cold.f.acc.pl
depictive

‘We finished the meatballs cold.’

3. 
   JárnsmiDurinn
   

blacskmith.def
barDi
pounded
málminn
metal.defm.acc.sg
flatan.
flat.m.acc.sg
resultative

‘The blacksmith pounded the metal flat.’

Thus, although it is more common for depictives, resultatives, and manner adverbials to take the same encoding strategy, the second most common pattern in Verkerk’s sample is for manner to have its own strategy. This is an interesting tendency against the back- ground of treating this area as one single domain, as advocated by all works on the topic. Whether the shared domain is due to the nature of secondary predication or whether only depictives and resultatives are secondary predicates, while manner adverbials have a different function, remains unclear. This matter will be further discussed in chapter 3.

3.