Some verbs are lighter than others
A comparison of Korean ha- ‘do’, sikhi- ‘order’ and mantul- ‘make, produce’
Mogens Mastracchio (he/any)
StuTS 72, 5 November 2022
There are three verbs in Korean that can serve as head to its causative construction, namely ha- 'do', sikhi- 'order' and mantul- 'make/produce'. While ha- is the most frequent among them, their meanings all diminish in this construction in favour of a causative expression, e.g., mantul- 'make/produce' is used where English would use 'make' in the parallel construction 'make sb. (do/...)', but without maintaining its literal meaning of 'producing'. This suggests that these verbs have (at least some) light verb characteristics. In order to determine their supposed lightness, I set a prototype definition for the Korean light verb and corresponding light verb construction along which the three were evaluated. I argue, based on their behaviours in causative contexts (i.e. the periphrastic construction) and non-causative contexts (i.e. light verb constructions and semantically heavy, transitive uses), that these verbs place on three different points of a lightness scale, such that ha- appears as the prototypically lightest and mantul- as the heaviest of the three.
Based on a term paper for the MA seminar Kausalität - Kausation - Kausativkonstruktionen held by Niklas Wiskandt at HHU Düsseldorf (winter semester 2021/22).
Abstract
Topics, roughly
- light verbs (LVs), light verb constructions (LVCs)
and how to define them
- causative constructions
- Korean (a pro-drop agglutinative SOV language)
and a little bit of German and English
Research question
A few verbs in Korean can be used in heavy, light, and causative contexts,
e.g. sikhi- 'to order':
(1)
| cemsim-ul | sikhi- |
| lunch-ACC | order |
'to order lunch'
light (semantically 'primitive') use
(2)
| kyelhon-sikhi- |
| marry-order |
'to marry' (transitive)
causative (periphrastic) use
(3)
'to make (sb.) eat'
| mek-key | sikhi- |
| eat-COMP | order |
heavy ('regular') use
(1)
| cemsim-ul | mantul |
| lunch-ACC | produce |
'to make lunch'
light (semantically 'primitive') use
(2)
| *kyelhon-mantul- |
| marry-produce |
causative (periphrastic) use
(3)
'to make (sb.) eat'
| mek-key | mantul- |
| eat-COMP | produce |
heavy ('regular') use
Not all verbs with causative function seem to work as light verbs, too. But to be causative verbs, they need light characteristics.
So, are there different degrees of verb 'lightness'?
- Defining LVCs for non-European languages
- What characteristics to look at for relative lightness
- Analysis of ha-, sikhi- and mantul-
- Discussion: Are there different degress of lightness?
Overview
To classify a construction as a LVC, you have to separate it from other types of constructions, e.g.
a typical example for a German LVC would be (4).
Criteria for defining LVCs
| zum | Ausdruck | bringen |
| to | expression | bring |
'to express' (lit. 'bring to expression')
(4)
(cf. Kamber 2008: 22)
There are many reasons to classify this construction as LVC, e.g.
- the verb with a diminished meaning ('to bring')
- the type of verb used (movement verb)
- the verb's obligatory complement, and the focus of meaning on it
- the form of the complement (a PP)
- the existence of a simple verb with a parallel meaning (ausdrücken 'to express')
but not all of these criteria are fit for cross-linguistic comparison.
Unless you want to assume that other languages have no LVCs (no you don't).
Kamber's (2008) LVC prototype
(1) light verb
(2) abstract deverbal noun
(3) prepositional phrase
(4) movement verb
Kamber's (2008) LVC prototype
(1) light verb
(2) abstract deverbal noun
(3) prepositional phrase
(4) movement verb
LVC with a light verb, abstract deverbal noun, prepositional phrase, and movement verb
| zum | Ausdruck | bringen |
| to | expression | bring |
'to express' (lit. 'bring to expression')
German,
e.g.
English,
e.g. to take into consideration.
Korean doesn't have these!
But that shouldn't mean Korean doesn't have LVCs.
For discussions of criteria specific to Indo-European languages, cf. Heine 2020, Fleischhauer 2019, Fleischhauer & Hartmann 2021, Knop & Hermann 2020, Bonial 2014
There are words in Korean that consist of a Sino-Korean noun and a native Korean light verb, often ha- 'do'. These are light verb constructions (Bak 2011).
Introduction: Korean verbs
There are also native Korean verbs which are only a stem, no light verb involved.
We're only concerned with the LVCs, and the other uses of light verbs (LVs) like ha-.
| (5) | tochak-ha-yess-e-yo |
| arrive-do-PST-Q-POL 'Have we arrived?' |
| (6) | ka-ss-e-yo? |
| go-PST-Q-POL 'Did you go?' |
N+LV predicate
V predicate
LVCs are best defined by a process of elimination.
An LVC should not
- be a complex (e.g. causative) construction
- be a single word
- have a semantically full verb as head
- be phonetically eroded
- be morphologically fossilized/unanalyzable
Criteria for non-European languages
= monoclausal construction
= multi-word expression
= obligatory LV
}
= not grammaticalized
Some criteria can be discarded:
- strong complement restrictions (only PPs, only deverbal nouns, ...)
- obligatory synonyms (take into consideration = consider)
(cf. Bak 2011; Butt 2003, 2010; Butt & Lahiri 2013)
LVC characteristics:
This definition works for Korean!
| (7) | taytap(-ul) | ha-ci | anh-ulkeya |
| answer(-ACC) | do-NEG | not.do-FUT |
'I will not answer.'
(Tiedemann 2012)
What characteristics to look at
For the same verb (e.g. ha-), does the following exist/hold?
- LVC use
[N]-ha- 'to NOUN'
- periphrastic causative use
[V(P)]-key ha- 'to make sb. V'
- heavy use
[NP]-lul ha- 'do NP'
- primitivity ('lightness')
N+LV predicate
V(P)+COMP LV causative
NP-ACC V predicate
Analysis
1. LVC use
- of ha- 'do'
- of sikhi- 'order'
- of mantul- 'make/produce'
'You must come and make Lizzy marry Mr. Collins!'
(Tiedemann 2012)
| (9) | tansin | ili | o-ase | Lizzy-lul | Collins | ssi-wa | kyelhon-sikhi-e-yo |
| you | here | come-CONN | Lizzy-ACC | Collins | HON-with | marry-order-DECL-POL |
| (5) | tochak-ha-yess-e-yo |
|
arrive-do-PST-Q-POL 'Have we arrived?' |
(Tiedemann 2012)
doesn't exist (in Tiedemann 2012)!
N+LV predicate
Analysis
2. periphrastic causative use
- of ha- 'do'
- of sikhi- 'order'
- of mantul- 'make/produce'
| (11) | pi-ta! | halmeni-ka | o-key | ha-yess-e! |
| rain-COP | grandmother-NOM | come-COMP | do-PST-INF |
'Rain! Grandma made it rain!'
(Tiedemann 2012)
| (12) | Edmond-hantey | kul-ul | ssu-key | sikhi-ess-ci-yo |
| Edmond-DAT | writing-ACC | write-COMP | order-PST-SFP-POL |
'I had this guy Edmond write all my papers for me.'
(Tiedemann 2012)
| (13) | John-i | Mary-lul | hakkyo-ey | ka-key | mantul-ess-ta |
| John-NOM | Mary-ACC | school-LOC | go-COMP | produce-PST-DECL |
'John made Mary go to school.'
(mod. from Yoo 1993: 182)
V(P)+COMP LV causative
Analysis
3. heavy use
- of ha- 'do'
- of sikhi- 'order'
- of mantul- 'make/produce'
NP-ACC V predicate
| (14) | ha-ki | silh-un | ke-l | ha-key | ha-yess-ci? |
| do-NOMZ | not.want.ADN | thing.FN-ACC | do-COMP | do-PST-SFP |
'Did she make you do something you didn't want to do?'
(Tiedemann 2012)
| (15) | cemsim | sikhi-si-llay-yo? |
| lunch | order-HON-SFP-POL |
'Do you want to order lunch?'
(Tiedemann 2012)
| (16) | cemsim | mantul-ess-ta |
| lunch | make-PST-DECL |
'I made you lunch.'
(Tiedemann 2012)
is this really heavy? 👀❓
Analysis
4. lightness
- of ha- 'do'
in both LVCs and causatives, possibly outside of that
- of sikhi- 'order'
in both LVCs and causatives
- of mantul- 'make/produce'
only in causatives
Discussion & summary
| LVC use | causative use | heavy use | lightness | |
|---|---|---|---|---|
| ha- | + | + | +/- | + |
| sikhi- | + | + | + | + |
| mantul- | - | + | + | +/- |
Differences:
- mantul- 'make/produce' doesn't exist in LVCs, but has a light meaning in causatives
- ha- 'do' has no proper heavy use - depending on how you define lightness
Open questions:
- Can we (and if so how can we) define lightness categorically -
how do we properly distinguish LVs from verbs (and auxiliaries)? - What does it mean for a verb to not have a heavy use?
Is that an indicator for grammaticalization?
Is there a V > LV > AUX cline (i.e. is the use of ha- between a LV and AUX)?
So, are there different degrees of lightness? Seems so!
ha- 'do' is lighter than sikhi- 'order', which then again is lighter than mantul 'make/produce'.
Discussion & summary
The definition problem for LVCs depends on a lot of things.
- the amount of internal fixedness allowed between LV and complement
( = their distinction from proper arguments; cf. Lee 2011)
- the LV definition
- whether a LV is seen as step between light verbs and auxiliaries, or not
(cf. Bak 2011; Butt 2003, 2010; Kamber 2008; Rhee & Koo 2014) - whether you want to assume multiple lexical entries for the same verb
(e.g. as verb, LV, AUX, or even clitic; cf. Chae 1995, 1996)
- whether a LV is seen as step between light verbs and auxiliaries, or not
- whether you want to stay within the same language family or go cross-linguistic
Bak (2011: 75) introduces the idea of "pseudo-noun incorporation" to solve the problem of LVCs most often looking like noun-verb compounds.
If LVC one word, then why 'construction'?
| cekkwun-i | tali-lul | [phakoy-(lul) | ha-yess-ta] |
| enemy-NOM | bridge-ACC | destroy-(ACC) | do-PST-DECL |
'The enemy destroyed the bridge.' (lit. destruction-do the bridge)
(x)
(cf. Bak 2011: 79)
In (x), the LVC which usually has the basic form phakoy-ha- ('destroy') is separated by ACC marking. But since the LVC functions as transitive verb, it also takes an ACC object.
Noun incorporation usually involves an argument, and phakoy isn't one. This no proper noun incorporation, but a slightly different operation. This LVC is not a compound, but a construction that can be broken up by case marking to aid the information structure in the sentence.
"The nominal complement of the LV is marked with the accusative case particle when it is focused” (Bak 2011: 75), i.e. not as it would be as argument of a transitive verb. LVCs function as constructions, just not overtly so every context.
References
| 📖 | Bak, Jaehee. 2011. The light verb construction in Korean. University of Toronto dissertation. |
| 📖 | Bonial, Claire Nicole. 2014. Take a look at this! Form, function and productivity of English light verb constructions. University of Colorado at Boulder dissertation. |
| 📖 | Butt, Miriam. 2003. The Light Verb Jungle. Harvard Working Papers in Linguistics 9. 1–28. |
| 📖 | Butt, Miriam. 2010. The Light Verb Jungle: Still Hacking Away. Complex predicates in cross-linguistic perspective. 48–78. |
| 📖 | Butt, Miriam & Aditi Lahiri. 2013. Diachronic pertinacity of light verbs. Lingua 135. 7–29. |
| 📖 | Chae, Hee-Rahk. 1995. Clitic Analyses of Korean ‘Little Words’. In Proceedings of the Korean Society for Language and Information Conference, 97–102. |
| 📖 | Chae, Hee-Rahk. 1996. Light verb constructions and structural ambiguity. In Proceedings of the Korean Society for Language and Information Conference, 99–108. |
| 📖 | Fleischhauer, Jens. 2019. The distribution of meaning components – The composition of affectedness and light verb constructions. Habilitation dissertation. |
| 📖 | Fleischhauer, Jens & Stefan Hartmann. 2021. The emergence of light verb constructions. Yearbook of the German Cognitive Linguistics Association 9(1). 135–156. |
| 📖 | Kamber, Alain. 2008. Funktionsverbgefüge – empirisch. Max Niemeyer Verlag. |
| 📖 | Knop, Sabine & Manon Hermann. 2020. Funktionsverbgefüge im Fokus: Theoretische, didaktische und kontrastive Perspektiven. Vol. 89. Walter de Gruyter GmbH & Co KG. |
| 📖 | Lee, Juwon. 2011. Two Types of Korean Light Verb Constructions in a Typed Feature Structure Grammar. In Proceedings of the Workshop on Multiword Expressions: from Parsing and Generation to the Real World, 40–48. |
| 📖 | Rhee, Seongha & Hyung Jung Koo. 2014. Grammaticalization of causatives and passives and their recent development into stance markers in Korean. Poznan Studies in Contemporary Linguistics 50(3). 309–337. |
| 📖 | Yoo, Eun Jung. 1993. Subcategorization and case marking in Korean. Papers in Syntax. 178–198. |
| 📺 | Tiedemann, Jörg. 2012. Parallel Data, Tools and Interfaces in OPUS. In Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC 2012). Istanbul, Turkey: European Language Resources Association (ELRA). |
Glosses
| ACC | accusative | FUT | future tense | N | noun |
| ADN | adnominal | HON | honorative | NP | noun phrase |
| COMP | complementizer | INF | informal | POL | polite |
| CONN | connective | LOC | locative | PST | past |
| COP | copula | LV | light verb | Q | question particle |
| DAT | dative | NEG | negation | SFP | sentence-final particle |
| DECL | declarative | NOM | nominative | V | verb |
| FN | formal noun | NOMZ | nominalizer | VP | verb phrase |
I'm also upset about this topic online! :-)
Twitter: @mostracchio
Mastodon: @mostracchio@mastodon.social
Email: brmas100@hhu.de
no categories,
only gradients
