scandinavian verb-particle alternation
DESCRIPTION
Scandinavian Verb-Particle AlternationTRANSCRIPT
The Verb-Particle Alternation in the
Scandinavian Languages
Peter SvenoniusCASTL, University of Tromsø
August 27, 2005Leikanger
1 The Characteristics of the Construction
1.1 Particle Shift
(1) English: Particle shift is optional, sensitive to information structure (Bolinger1971, Svenonius 1996a, Dehe 2002)
a. The dog tore his cap off.b. The dog tore off his cap.
(2) Pronouns precede the particle (Fraser 1976)
a. The dog tore it off.b. *The dog tore off it.
(3) Modifiers prevent particle shift (Emonds 1976)
a. The dog tore the cap right off.b. *The dog tore right off the cap.
(4) Complements prevent particle shift (Svenonius 1996b)
a. The dog tore the cap off his head.b. *The dog tore off the cap his head.c. *The dog tore off his head the cap.
(5) Cross-Scandinavian variation (Taraldsen 1983a;b, Svenonius 1994; 1996c)
a. Danish: No particle shift, ‘discontinuous’ V–DP–Prt order (Herslund 1984)b. Faroese: Particle shift restricted (Sandøy 1976, Svenonius 1996c)c. Icelandic: optional particle shift, as in English (Thrainsson 1979, Collins and
Thrainsson 1996)d. Norwegian 1: optional particle shift, as in English (Afarli 1985)e. Norwegian 2 (e.g. Romsdal): Particle shift obligatory with full DP (but pro-
noun always precedes particle) (Sandøy 1976)f. Swedish 2: Obligatory particle shift with idiomatic particle verbs, even with
pronouns, but optional particle shift with pronouns with spatial particle verbs(Vinka 1999)
g. Swedish 1: Obligatory particle shift, ‘continuous’ V–Prt–DP order (Toivonen2003)
1
1.2 Microparametric variation
(6) Heavy-DP < Prt DP < Prt pronoun < Prt DP < PPDanish yes yes yes yesFaroese no sometimes yes yesIcelandic no optional yes yesEnglish no optional yes yesNorwegian 1 no optional yes yesNorwegian 2 no no yes yesSwedish 2 no no sometimes yesSwedish 1 no no no yes
1.3 Swedish
(7) DP follows bare Prt (Teleman et al. 1999)
a. Hundenthe.dog
slitertears
avoff
mossan.the.cap
‘The dog is tearing off the cap’b. *Hunder
the.dogslitertears
mossanthe.cap
av.off
(8) pn follows bare Prt (Platzack 1998)
a. Hundenthe.dog
slitertears
avoff
den.it
‘The dog tears it off’b. *Hunden
the.dogslitertears
denit
av.off
(9) DP precedes PP: P doesn’t shift away from a complement (Noren 1996, Toivonen2003)
a. Hundenthe.dog
slettore
mossanthe.cap
avoff
husse.owner
‘The dog tore the cap off its owner’b. *Hunden
the.dogslettore
avoff
mossanthe.cap
husse.owner
(10) Modifiers require Prt to remain post-DP (Prt is PP) (Vinka 1999)
a. Hundenthe.dog
slettore
mossanthe.cap
rattright
av.off
‘The dog tore the cap right off’b. *Hunden
the.dogslettore
rattright
avoff
mossan.the.cap
(11) Unlike English: a P complement (Ground) can precede the object (Figure)(Svenonius 2003)
Hundenthe.dog
slettore
avoff
husseowner
mossan.the.cap
‘The dog tore the cap off its owner’
2
2 Identifying the Verb-particle Construction
2.1 Swedish
(12) In Swedish, the word order is almost invariably V–Prt–DP, which gives the samestring as a PP complement to V
(13) a. We ran up a hill.b. *We ran a hill up.
(14) a. We ran up a bill.b. We ran a bill up.
(15) Sometimes considered to be particle verbs, in Swedish
a. JagI
hallerhold
avof
henne.her
‘I like her’b. Vi
wehoppadejumped
iin
vattnet.the.water
‘We jumped in the water’
(16) No such particle verbs in English
a. I yearn for her.b. *I yearn her for.c. We jumped in the water.d. *We jumped the water in.
(17) Developing Swedish-specific diagnostics for particle-verbs (Svenonius 2003): HNPS,PP-constituency, P-incorporation
a. Viwe
halldepoured
iin
mjolken.the.milk
‘We poured in the milk’b. Vi
wehoppadejumped
iin
vattnet.the.water
‘We jumped into the water’
(18) HNPS
a. Viwe
mastehad.to
hallapour
iin
medwith
detthe
sammasame
trethree
bagarecartons
medof
mjolk.milk
‘We had to pour in at once three cartons of milk’b. *Vi
wemastehad.to
hoppajump
iin
medwith
detthe
sammasame
30003000
literliter
mjolk.milk
(‘We had to jump in–at once—3000 liters of milk’)
(19) Incorporation
a. Mjolkenthe.milk
blevbecame
ihalld.in.poured
‘The milk was poured in’b. *Vattnet
the.waterblevbecame
ihoppat.in.jumped
3
(20) hoppa i halla iStress on P yes yesHNPS no yesP-DP Constituency yes noP-V in Passive no yes
(21) Locative PP Directional PP Particle
Stress on P no yes yesHNPS no % yesP-DP Constituency yes yes noP-V in Passive no no yes
(22) By these diagnostics, Swedish has more particles than English (or Norwegian?);e.g. ihjal (exx. here from Toivonen 2003)
a. ErikErik
harhas
slagitbeaten
ihjalto.death
ormen.the.snake
‘Erik has beaten the snake to death’b. Erik
Erikharhas
slagitbeaten
ormenthe.snake
blodig.bloody
‘Erik has beaten the snake bloody’
2.2 Danish
(23) In Danish, the order is almost invariably V–DP–Prt
a. BorisBoris
skruedeturned
musikkenthe.music
ned.down
‘Boris turned the music down’b. Boris
Borisskrevwrote
kontraktenthe.contract
under.under
‘Boris signed the contract’
(24) This is the same order as in small clauses
a. I consider the runner out.b. *I consider out the runner.c. They want the cat in.d. *They want in the cat.
(25) Apart from shift, a diagnostic not available in Danish, how can we decide whetherwhat we are looking at is a verb-particle construction?
(26) Verb-particle constructions in English, Icelandic, Norwegian (and Swedish, fol-lowing the diagnostics just presented) are resultative; they denote some sort ofchange of state in the object.
a. I threw the dog out = I caused the dog to go out, by throwingb. They put the party off = They caused the party to go to delayed, by
deliberate actionc. We gave our hobbies up = We caused our hobbies to go out of our lives,
by voluntary action
(27) Non-resultative constructions are not verb-particle constructions
a. I consider the runner out 6= I caused the runner to go out, by consideringb. They want the cat in 6= They cause the cat to go in, by wanting
4
c. We jumped in the water 6= We caused the water to go in, by jumpingd. He yearns for her 6= He causes her to go to his desire-world, by yearning
(28) In addition, verb-particle constructions are frequently idiomatic, whereas verbsdo not readily form idioms with small clause predicates
(29) Furthermore, subjects of small clauses pass certain subjecthood tests, whereasthe post-verbal DP (the object) of the verb-particle construction typically doesnot
3 Incorporation
(30) There are prefixed verbs in North Germanic (and in English), but the patternsare generally unproductive and by and large, particles do not prefix onto activeverbs
a. English: outsource, undergo, overturn, ...b. Norwegian: pasta, undertegne, innga, ...
(31) Deverbal nouns and adjectives formed from particle verbs are always prefixed(though not in English)
a. ga ut ‘go out’ ∼ utgang ‘exit’ *gang utb. laste ned ‘download’ ∼ nedlasting ‘downloading’ *lasting nedc. trekke inn ‘pull in’ ∼ inntrukket ‘pulled in’ *trukket innA
(32) In Swedish, passive participles also incorporate particles, obligatorily
a. Mjolkenthe.milk
blevbecame
ihalld.in.poured
‘The milk was poured in’b. *Mjolken
the.milkblevbecame
halldpoured
i.in
(33) a. Skrapetthe.scrap
mastehad.to
blibecome
utkastat.out.thrown
‘The scrap had to be thrown out’b. *Skrapet
the.scrapmastehad.to
blibecome
kastatthrown
ut.out
(34) This is not true of other varieties, e.g. Danish
a. Hundenthe.dog
blevbecame
smedetthrown
ud.out
‘The dog was thrown out’b. *Hunden
the.dogblevbecame
udsmedet.out.thrown
(35) Since all varieties of Scandinavian have obligatory incorporation in deverbal ad-jectives, adjectival constructions must be distinguished from (eventive) passives
a. Brødetthe.bread
blebecame
skjærtsliced
oppup
avby
Jens.Jens
‘The bread was sliced by Jens’b. Brødet
the.breadvarwas
ferdigready
oppskjærtup.sliced
(*avby
Jens).Jens
‘The bread was already sliced’
5
(36) Controlling for this factor, several varieties of Norwegian, including the Leikangerdialect, have incorporation in the passive
a. Treathe.trees
vartbecame
nedhogne.down.chopped
‘The trees were chopped down’b. ??Trea
the.treesvartbecame
hognechopped
ned.down
(37) Other dialects, for example Tromsø dialect, do not
a. Trærnethe.trees
blebecame
hogdchopped
ned.down
‘The trees were chopped down’b. ??Trærne
the.treesblebecame
nedhogd.down.chopped
(38) Does incorporation pattern with anything?
(39) DP < Prt pronoun < Prt IncorporationDanish yes yes noFaroese sometimes yes sometimesIcelandic optional yes noEnglish optional yes noNorwegian 1 optional yes noNorwegian 2 no yes yesSwedish 2 no sometimes yesSwedish 1 no no yes
3.1 Agreement
(40) Norwegian dialects with incorporation in the passive tend to have agreement onpassive participles, like Swedish; Norwegian dialects without incorporation tendnot to have agreement
a. Hundenthe.dog
eris
bundentied
∼ Hundanethe.dogs
eris
bundnetied.pl
‘The dog is tied’ — ‘The dogs are tied’ (Leikanger)b. Bikkja
the.dogeis
bundetied
∼ Bikkjenthe.dogs
eis
bundetied
‘The dog is tied’ — ‘The dogs are tied’ (Tromsø)
(41) Incorporation AgreementDanish no noFaroese sometimes yesIcelandic no yesEnglish no noNorwegian 1 no noNorwegian 2 yes yesSwedish 2 yes yesSwedish 1 yes yes
(42) Faroese is a potential counterexample, and Icelandic seems to be a straightfor-ward counterexample. But is it?
6
(43) Agreement in Mainland Scandinavian is triggered by A-movement
a. Trethree
journalisterjournalists
blevbecame
arresterade.arrested.pl
‘Three journalists were arrested’b. Det
itblevbecame
trethree
journalisterjournalists
arresterade.arrested.pl
‘There were three journalists arrested’c. %Det
itblevbecame
arrestertarrested
trethree
journalister.journalists
‘There were three journalists arrested’
(44) Not so Icelandic
a.�rır
threebla�amennjournalists
voruwere
settirset.pl
ıin
var�hald.custody
‘Three journalists were arrested’b.
�a�
itvoruwere
settirset.pl
�rır
threebla�amennjournalists
ıin
var�hald.custody
‘There were three journalists arrested’
(45) Incorporation also seems to be tightly linked to A-movement
(46) Swedish
a. Detit
blevbecame
mangamany
tradtrees
nedhuggna.down.chopped.pl
‘There were many trees chopped down’b. Det
itblevbecame
huggetchopped
neddown
mangamany
trad.trees
‘There were many trees chopped down’
(47) Leikanger
a. Detit
vartbecame
mangemany
tretrees
nedhogne.down.chopped.pl
‘There were many trees chopped down’b. Det
itvartbecame
hoggechopped
neddown
mangemany
tre.trees
‘There were many trees chopped down’
(48) Faroese
a. Ta�it
blivubecame
nogvmany
trøtrees
{?ni�urhøgd/høgddown.chopped/chopped
ni�ur}down
‘There were many trees chopped down’b. Ta�
itblivubecame
høgdchopped
ni�urdown
nogvmany
trø.trees
‘There were many trees chopped down’
(49) No incorporation in the absence of A-movement
a. *Detit
blevbecame
nedhuggetdown.chopped
mangamany
trad.trees
b. ??Detit
vartbecame
nedhoggedown.chopped
mangemany
tre.trees
c. *Ta�it
blivubecame
ni�urhøgddown.chopped
nogvmany
trø.trees
7
(50) Icelandic has agreement without A-movement, and no incorporation in the pas-sive
a. Treinthe.trees
voruwere
hogginchopped.pl
ni�ur.down
‘The trees were chopped down’b.
�a�
itvoruwere
hogginchopped.pl
ni�urdown
morgmany
tre.trees
‘There were many trees chopped down’
(51) Hence, a possible generalization: There is incorporation in the passive if andonly if there is A-sensitive participle agreement
(52) Incorporation A-sensitive AgreementDanish no noFaroese sometimes ?Icelandic no noEnglish no noNorwegian 1 no noNorwegian 2 yes yesSwedish 2 yes yesSwedish 1 yes yes
4 Conclusion
(53) Regarding the Verb-particle construction: There is microparametric variationalong a range of parameters
a. Particle shiftb. Shift with pronounsc. Shift with full PPsd. Incorporation under passivee. Incorporation in deverbal formsf. The inventory of particles
(54) Some of these parameters seem to be interrelated, or related to other factors (e.g.participle agreement)
(55) In order to compare “constructions” cross-linguistically, some theoretically so-phisticated understanding of the contruction is necessary
(56) Conceivably, Swedish has developed a distinct particle system, in which casesome of what I have taken as diagnostics (e.g. resultativity) are really additionalpoints of microparametric variation.
(57) a. JagI
tyckerthink
omabout
henne.her
‘I like her’b. Vi
wegickwent
paon
bussen.the.bus
‘We got on the bus’
(58) Ultimately, cross-linguistic comparison will be necessary to determine wherethere are accidental gaps
8
References
Afarli, Tor A. 1985. Norwegian verb particle constructions as causative constructions.Nordic Journal of Linguistics 8 1: 75–98.
Bolinger, Dwight. 1971. The Phrasal Verb in English. Harvard University Press, Cam-bridge, Ma.
Collins, Chris and Hoskuldur Thrainsson. 1996. VP-internal structure and object shiftin Icelandic. Linguistic Inquiry 27 3: 391–444.
Dehe, Nicole. 2002. Particle Verbs in English: Syntax, Information Structure, and Into-nation. John Benjamins, Amsterdam.
Emonds, Joseph E. 1976. A Transformational Approach to English Syntax: Root,Structure-Preserving, and Local Transformations. Academic, New York.
Fraser, Bruce. 1976. The Verb-Particle Combination in English. Academic Press, NewYork.
Herslund, Michael. 1984. Particles, prefixes, and preposition stranding. Topics in DanishSyntax (Nydanske Studier & Almen Kommunikationsteori 14) .
Noren, Kerstin. 1996. Svenska partikelverbs semantik . Acta Universitatis Gothobergensis,Gothenberg.
Platzack, Christer. 1998. Svenskans inre grammatik: Det minimalistiske programmet .Studentlitteratur, Lund.
Sandøy, Helge. 1976. Laust samansette verb i vestnordisk: ein samanliknande leddstill-ingsanalyse for islandsk, færøysk og romsdalsmal . Universitetet i Oslo, Oslo.
Svenonius, Peter. 1994. Dependent Nexus: Subordinate Predication Structures in Englishand the Scandinavian Languages . Ph.D. thesis, University of California at Santa Cruz.
Svenonius, Peter. 1996a. The optionality of particle shift. Working Papers in Scandina-vian Syntax 57: 47–75.
Svenonius, Peter. 1996b. Review of den Dikken 1995, Particles. Language 74: 816–820.
Svenonius, Peter. 1996c. The verb-particle alternation in the Scandinavian languages.Ms. University of Tromsø; available at www.ling.auf.net/lingBuzz/000046.
Svenonius, Peter. 2003. Swedish particles and directional prepositions. In Grammarin Focus: Festschrift for Christer Platzack 18 November 2003 , edited by Lars-OlofDelsing, Cecilia Falk, Gunlog Josefsson, and Halldor A. Sigur�sson, vol. II, pp. 343–351. Department of Scandinavian Languages, Lund University, Lund.
Taraldsen, Knut Tarald. 1983a. Parametric Variation in Phrase Stucture: A Case Study .Ph.D. thesis, University of Tromsø.
Taraldsen, Knut Tarald. 1983b. Some phrase structure dependent differences betweenSwedish and Norwegian. Working Papers in Scandinavian Syntax 9: 1–45.
Teleman, Ulf, Steffan Hellberg, and Erik Andersson. 1999. Svenske Akademiens Gram-matik . Norstedts, Stockholm.
Thrainsson, Hoskuldur. 1979. On Complementation in Icelandic. Garland, New York.
Toivonen, Ida. 2003. Non-projecting Words: A Case Study of Swedish Particles. Kluwer,Dordrecht.
Vinka, Mikael. 1999. Predicative and non-predicative verb particle constructions. InWCCFL 18 Proceedings , edited by S. Bird, A. Carnie, J. Haugen, and P. Norquest,pp. 570–585. Cascadilla Press, Somerville, Ma.
9