GRAMMATIK i FOKUS 33, 2019

Lunds Universitet, LUX-huset (Helgonavägen 3) Torsdagen den 7e februari och fredagen den 8e februari 2019

Torsdagen den 7e februari, C126 13.00–13.05 Symposiet öppnas 13.05–13.35 Dianne Jonas (Goethe-Universität Frankfurt): The so-called Germanic

“blended construction” revisited 13.35–14.05 Gunlög Josefsson (Lunds universitet): Specificity and non-specificity in

Swedish; the case of the two det ‘it, there’ 14.05–14.15 Paus 14.15–14.45 Gerd Carling (Lunds universitet): The cultural component of grammar: an

overview of gender assignment patterns in Indo-European languages 14.45–15.15 Ann-Charlotte Gutsjö: Återbruk av försvunna deklinationsmorfem i

språkhistorien 15.15–15.30 Kaffepaus 15.30–16.00 Mikael Vinka (Umeå universitet & Nord Universitetet, Levanger):

Logophoricity in South Saami 16.00–16.30 Victor Bogren Svensson (Lunds universitet): She wellwrites: Manner Affixes

as Light Adverbs 16.30–16.40 Paus 16.40–17.10 Gunlög Josefsson & Katarina Lundin (Lunds universitetet): Tröskelbegrepp

inom grammatiken 17.10–17.40 Lars-Olof Delsing (Lunds universitet): Kvantorerna mycket och lite som

polaritetselement i modern svenska 18.30 Reception på Språk- och Litteraturcentrum (sal L207, SOL-huset)

Fredagen den 8e februari, C121 9.00–9.30 Shiro Shibata & Takayuki Tohno (Graduate School of Language and Culture,

Osaka University): The “Double Object” Verb-Particle Constructions in Swedish

9.30–10.00 Eirik Tengesdal (Universitetet i Oslo), Björn Lundquist (Universitetet i Tromsø & Universitetet i Oslo) & Ida Larsson (Universitetet i Oslo): Prosodisk variasjon ved partikkelverb i tre norske dialektområde

10.00–10.30 Kaffepaus 10.30–11.00 Tori Larsen & Christer Johansson (Universitetet i Bergen): Anaphor, pronoun,

or other? Investigating PRO using Reactivation Patterns 11.00–11.30 Eva Klingvall (Lund University) & Fredrik Heinat (Linnæus University): Set

focus and reference: an ERP study 11.30–11.35 Paus 11.35–12.05 Nigel Vincent (The University of Manchester): Definiteness in diachrony:

Romance vs Germanic 12.05–12.35 Yvonne van Baal (University of Oslo): Language change in definiteness

marking in American heritage Norwegian Symposiet avslutas

The so-called Germanic “blended construction” revisited

Dianne Jonas Goethe-Universität Frankfurt

This talk is concerned with the diachronic development of the so-called “blended construction” in Faroese, English, and Swedish. The verbs that occur in this construction have dative experiencer subjects and accusative theme objects and are assumed to have changed from the original case-marking pattern of dative subjects and nominative theme objects. The main aspect of the analysis presented here is to take a more fine-grained approach to this construction than has generally been discussed previously.

The construction can actually be divided into 2 sub-types, the first termed here a simple experiencer construction where the subject is dative and the theme object, a co-argument of the experiencer verb, is accusative (1a). In the second subtype, a complex experiencer construction, the verbs involved take an experiencer subject in dative and select for a small clause or non-finite complement whose subject is accusative and is not a co-argument of the higher verb.

I argue here that the two constructions must be kept apart as their diachronic development differs across the 3 languages under consideration.

Grammar in focus, 2018, Gunlög Josefsson, Dept. of Scandinavian languages, SOL, Lund University

Specificity and non-specificity in Swedish; the case of the two det ’it, there’

In my talk I will investigate the distinction between specificity and non-specificity in Swedish. Specificity is a semantic category that appears to have no designated morphology; yet, if we look more closely, we find evidence for it in the Swedish morphosyntactic system.

In my talk I will focus on the pronoun det ‘it, there’. I will argue that there are (at least) two instances of det, one that is specific and one that is non-specific. The two instances of det ’it’ have partly different phonological properties and different syntactic properties in their use as anaphoric pronouns. Furthermore, the specific vs. non-specific distinction is one of the categories that underly the Swedish semantic gender system.

The cultural component of grammar: an overview of gender assignment patterns in Indo-European languages

Gerd Carling, Lund University

Gender assignment in languages with a gender system is a complex issue that is influenced by many different factors. In brief, factors influencing gender assignment can be classified as either semantic, morphological, or phonological (Corbett, 1991, 2013; Corbett & Fraser, 2000). Many Indo-European languages have a three-gender system, distinguishing masculine, feminine, and neuter gender. The system is present in several Indo-European branches, but it has been simplified in other branches or languages, either to a system with a masculine – feminine distinction, or to a system with a masculine/feminine (uter) – neuter distinction. Most historical linguists agree that the three-gender system is an innovation, which is not supposed to be reconstructed to Proto-Indo-European (Luraghi, 2011; Matasović, 2004).

The current lecture will focus on the semantic and functional factors of gender assignment, and demonstrate patterns of assignment in relation to semantic property as well as cultural function of the artifacts and objects assigned by the genders masculine, feminine, neuter, or commune. The basis is a compilation of around 6,000 nouns of in Indo-European languages, which have been marked for gender and classified according to various semantic properties and cultural parameters. We will test the correlation to these parameters, arguing for a cultural component in gender assigment.

Till symposiet GRAMMATIK I FOKUS vid SOL-institutionen, Lunds universitet, den 7 och 8 februari 2019

Gutsjö, Ann-Charlotte (anncharlotte.gutsjo@gmail.com): Återbruk av försvunna deklinationsmorfem i

språkhistorien

ABSTRACT

Den fornkyrkoslaviska ŭ-deklinationen var på väg att helt försvinna före år 1100. Men i manuskript från 900-,

1000- och 1100-talen finns exempel på återanvändning av de kasusändelser som tillhörde den försvunna ŭ-

deklinationen, tillsammans med ord som tillhörde o-, jo- och i-deklinationerna, men nu med nya grammatiska

roller. I en studie undersöktes återanvändningen speciellt av de kasus och numerus som har särskiljande

böjningsmorfem i ŭ-deklinationen, dvs. D.sg., N.pl. och G.pl., där det finns ett -ov-morfem, som saknas i de tre

andra. Resultatet visar att det framför allt var till o- och jo-deklinationerna som ŭ-deklinationens kasusändelser

spreds, men även till i-deklinationen. Dessa fick följande nya grammatiska funktioner: i D.sg. för att uttrycka

animata substantiv eller adverbial, eller för att personifiera inanimata substantiv; i N. pl. för att uttrycka grupper

av folk, yrken, fåglar och djur; i G.pl. blev ŭ-deklinationens -ov-suffix mycket använt, eftersom formerna i N.sg.

och G.pl. var identiska, och suffixet blev ett sätt att skilja dem åt. Suffixets spridning i genitiv blev så stor att

forskarna inte är överens om ordet tillhörde o-deklinationen men påträffats med ŭ-deklinationens kasusändelser,

eller ett ŭ-deklinationsord som lånat o-deklinationens kasusändelser.

----

Logophoricity in South Saami Mikael Vinka

Umeå universitet och Nord Universitetet, Levanger

Languages with full-fledged logophoric systems accommodate at least two sets of third person pronouns, where each set is characterized by whether it requires co- or contraindexation with another argument that is located in a higher clause (Adesola, 2005; Clements, 1975; Koopman & Sportiche, 1989). South Saami (Finno-Ugric, Sweden and Norway) has three sets of third person pronouns, (1). D-pronouns must be contraindexed with a superordinate subject, whereas S-pronouns require coindexation, (2). One could imagine that the choice between D- and S-pronouns obeys a discourse constraint that requires an argument coindexed with the the topic to be realized as an S-pronoun. Contraindexation results in topic shift, which would rule out the S-pronoun, in favor of the D-pronoun in (2). However, unlike Finnish, South Saami allows third person referential prodrop, which is required in (3), when when the embedded subject refers back to the subject of the main clause. While (3) is predicted by a discourse contraint, such as Grimshaw and Samek-Ludovici's (1996) DropTop, (4), such an account does not carry over to (2). The major difference between (2) and (3), is that fact that the former involves a complement clause, whereas the latter an adjunct clause. It is well-established in the literature, that logophoricity is commonly restricted to contexts involving attitudinal verbs that take a clausal complement (Adesola, 2005; Clements, 1975; Safir, 2004, 2014). Following (Adesola, 2005; Koopman & Sportiche, 1989; Safir, 2004, 2014), I assume that attitudinal or logophoric verbs (verbs of saying, hearing, etc) license a logophoric operator in SpecCP of the complement clause, (5). The operator binds the logophoric pronoun, and the operator itself is bound by the higher subject. (5) straightforwardly accounts for the the occurrence of the S-pronoun in (2), as well as the blocking effect observed in the adjunct clause (3). Since the adjunct clause in (3) is not selected by a logophoric verb, it lacks a logophoric operator, and consequently the S-pronoun fails to be licensed. One classical piece of support for the operator analysis comes from the fact that true logophoric systems are insensitive to islands. That is, the relation between the operator and the logophoric pronoun is not blocked by an intervening island, (6b). Notice that in the absense of a logophoric verb, the S-pronoun is not licit in the relative clause, (6a). Another characteristic is the possibility for interleaving effects, which also show up in South Saami, (7). Thus, either S-pronoun in (7) can refer to either Piere or Læjsa (note that there is a language specific constraint that prohibits an S-pronoun from being A-bound by another S-pronoun). In conclusion, South Saami supports an operator-based account of logophoricity, which successfully explains why the logophoric S-pronoun is licit in the complement of an attitudinal verb, but ruled in adjunct clauses as well as in the complement clause of raising verbs.

Examples (1) D-pronouns S-pronouns Null pronouns 3s dihte satne pro 3d dah (guaktah) såtnoe pro 3p dah sijjieh pro (2) Pierei jeehti dihte*i/j/satnei/*j edtja måvhkam bïssedh. Piere.Nom say.Pst.3s D/S.Nom will.Prs.3s pants.Acc wash.Inf 'Piere said that hei/j will wash the pants.' (3) Pierei Maarjamk diervesji P.Nom M.Acc greet.Pst.3s [gosse proi/*k/dihte*i/k/satne*i/*k gaatan byjjelen veedtsi.] when pro/D/S.Nom street.Gen across walk.Pst.3s 'Pierei greeted Maarjak, when s/he crossed the street.' (4) DropTop (Grimshaw & Samek-Ludovici, 1996) Leave arguments coreferent with the topic structurally unrealized (5) … DPi VLog [CP LogOpi … LogProni…]… (6) a *Læjsaj dam gærjamk lohkeminie Laara.Nom the.Acc book.Acc read.Prog [mijk satnei/j tk Lundesne öösti]. wh S.Nom Lund.Iness buy.Pst.3s Intended, but *'Laara is reading the book that he (=Laara) bought in Lund.' b Pierei jeehti [CP LOGi Læjsaj dam gærjamk lohkeminie Piere.Nom say.Pst.3s Laara the.Acc book.Acc read.Prog [mijk satnei/*j tk Lundesne öösti]]. wh S.Nom Lund.Iness buy.Pst.3s 'Piere said that Lisa is reading the book that he (=Piere) bought in Lund.' (7) Pierei veanhta Læjsak jeehti satnei/k satnemi/k lyjhkoe. P.Nom think.Prs.3s L.Nom say.Prs.3s S.Nom S.Acc like.Prs.3s (i) 'Pierei thinks that Læjsak said that hei likes herk.' (ii) 'Pierei thinks that Læjsak said that shek likes himi.' References Adesola, O. (2005). Pronouns and Null Operators. Ph.D. Diss, Rutgers University, New Brunswick, New Jersey. Clements, G. N. (1975). Logophoric Pronouns in Ewe. Journal of West African Languages, 10, 141-177. Grimshaw, J., & Samek-Ludovici, V. (1996). Optimal Subjects. In J. N. Beckman, L. W. Dickey, & S. Urbanczyk

(Eds.), Papers in Optimality Theory. Amherst, Mass: GLSA. Koopman, H., & Sportiche, D. (1989). Pronouns, Logical Variables, and Logophoricity in Abe. Linguistic Inquiry,

20(4), 555-588. Safir, K. (2004). Person, context and perspective. Rivista di Linguistica, 16(1), 107-153. Safir, K. (2014). One True Anaphor. Linguistic Inquiry, 45(1), 91-124.

Grammatik i Fokus 2019 She wellwrites: Manner Affixes as Light Adverbs

Abstract Victor Bogren Svensson

Within current Principles-and-Parameters frameworks there are two main approaches to the syntax of adverbs. One of them is Cartography, which places adverbs in Spec positions, where they are licensed by a semantically appropriated head (which is often phonologically empty). The ordering of these heads is taken to be part of UG (Cinque 1999). This theory also makes predictions regarding the ordering of verbal affixes and auxiliary verbs, since they are analyzed as the overt realizations of the adverb-licensing heads, implying that their ordering ought to mirror that of adverbs. Cartography is often contrasted with scopal theories, which take adverbs to be freely adjoined to the phrase structure, only constrained by semantic restrictions and “weight” (Ernst 2002). Within a cartographic approach, I will discuss my work in progress on verbal affixes encoding manner information in West Greenlandic (Eskimo-Aleut: Greenland), which I will refer to here as manner affixes (e.g. -lluar- ‘well’, -nerlug-, ‘badly’, -rusaar- ‘slowly’ (Fortescue 1984)). In line with a cartographic understanding of verbal affixes, I take them to be the overt realization of heads of functional projections located in the spine of the clause, thus giving them the same status as aspectual affixes and making them a part of the extended IP zone. I further propose that they can be analyzed as light (manner) adverbs, in conformity with light verbs, nouns and adjectives, which similarly have been analyzed as representing realizations of functional categorizing heads v, n and a, respectively (cf. Harley 2014, 233). An implication would be that that these manner affixes only can encode a limited range of semantic content, therefore being restricted to crude distinctions, in contrast to full manner adverbs which have a wider range of semantic content. This analysis would thereby align manner adverbs with cartographic descriptions of attributive adjectives, and make manner affixes structurally parallel to the overt realizations of the different α heads (cf. Cinque 1994 & Julien 2005). During the presentation I will discuss how this model can be applied to West Greenlandic, as well as explore how it can be used to analyze a larger set of languages with manner affixes. I will also attempt relate this model to a larger project where I am also investigating auxiliary verbs encoding manner information (cf. Holmer 2012). References: Cinque, G. (1994). On the evidence for partial N-movement in the Romance DP. In Paths

Towards Universal Grammar: Studies in Honor of Richard S. Kayne, ed. G Cinque, J Koster, J-Y Pollock, L Rizzi, R Zanuttini, pp. 85–110. Washington, DC: Georgetown Univ. Press

Cinque, G. (1999). Adverbs and functional heads: a cross-linguistic perspective. New York: Oxford Univ. Press.

Ernst, T.B. (2002). The syntax of adjuncts. Cambridge, UK: Cambridge University Press. Fortescue, M. (1984). West Greenlandic. London: Croom Helm. Harley, H. (2014). On the identity of roots. Theoretical linguistics, 40(3-4), 225-276. Holmer, A. (2012). Evidence from Formosan for a unified theory of adverb ordering.

Lingua, 122(8), pp. 902-921. Julien, M. (2005). Nominal phrases from a Scandinavian perspective. Amsterdam:

Benjamins.

Tröskelbegrepp inom grammatiken Gunlög Josefsson (Gunlog.Josefsson@nordlund.lu.se) Katarina Lundin (Katarina.Lundin@nordlund.lu.se) Det är många studenter som tycker grammatik är svårt – den upplevelsen delar nog många lärare som har undervisat i momentet. I vårt pedagogiska utvecklingsprojekt Analytiskt tänkande och grammatikundervisningens utmaning har vi identifierat några olika problemområden som är svåra för studenter att tillägna sig och som samtidigt är helt avgörande för förståelse och progression. Analysen pekar ut tre områden som centrala (Josefsson & Lundin 2017, 2018):

• svårigheter med att skilja mellan betydelse, form och funktion, • svårigheter med att skilja mellan en linjär och en hierarkisk dimension, • svårigheter med att korrekt avgränsa satser och fraser.

I föredraget diskuterar vi dessa problemområden med utgångspunkt i tröskelbegreppet, (Thres-hold Concepts, se t.ex. Perkins 1999; 2008, Meyer & Land 2003; 2006, Orsini-Jones 2008, Josefsson & Lundin 2017). Tröskelbegreppet är centralt inom den didaktiska forskningen, och termen sätter namn på erfarenheten av att det finns vissa centrala begrepp, moment och tankesätt som en individ måste förstå på djupet för att kunna gå vidare i sitt lärande. Meyer & Land (2003) liknar tröskelbegrepp vid portaler eftersom förståelsen av ett tröskelbegrepp medför att nya perspektiv och sätt att tänka öppnar sig, som man tidigare inte varit förmögen att se eller förstå. Vi kommer att relatera Perkins tankar om tröskelbegreppet till tankar om olika typer av kunskap: possessive, performative och proactive knowledge (Perkins 2008:5f) samt diskutera hur man ska kunna stödja studenter att ta sig över ”grammatiktröskeln” och skaffa sig djupare grammatikförståelse. Referenser Josefsson, G. & Lundin, K. (2018) Nycklar till grammatik. Lund: Studentlitteratur Josefsson, G. & Lundin, K. (2017) Tröskelbegrepp inom grammatiken. I: Högre utbildning. Vol. 7, Nr 2, 18–34. Meyer, J. H. F., & Land, R. (2003). Threshold concepts and troublesome knowledge:

Linkages to ways of thinking and practising within the disciplines. Improving student learning − Ten years on. OCSLD, (s. 412–424). Oxford.

Meyer, J. H. F., & Land, R. (2006). Threshold concepts and troublesome knowledge: An introduction. In: J. H. F. Meyer, & R. Land (red.), Overcoming barriers to student understanding: Threshold concepts and troublesome knowledge (s. 3–18). London: Routledge.

Orsini-Jones, M. (2008) Troublesome language knowledge: identifying threshold concepts in grammar learning. In: J. H. F. Meyer, & R. Land (red.), Overcoming barriers to student understanding: Threshold concepts and troublesome knowledge (s. 213–226). London: Routledge.

Perkins, D. (1999). The many faces of constructivism. In: Educational Leadership, 57(3), 6– 11.

Perkins, D. (2008). Beyond Understanding. In: R. Land, J.H.F. Meyer & J. Smith (red.), Threshold Concepts within the Disciplines (s. 3–9). Rotterdam: Sense Publishers.

Grammatik i fokus Lars-Olof Delsing 7-8 februari 2019 Språk- och litteraturcentrum Lunds universitet Lunds universitet

Kvantorerna mycket och lite som polaritetselement i modern svenska

I modern svenska har kvantorerna mycket och lite egenskaper som påminner om polaritets-element. Fenomenet är, såvitt jag vet, hittills ouppmärksammat i forskningen om svenska, och föredraget syftar till att beskriva användningen närmare.

Båda kvantorerna kan förekomma i flera olika kontexter, illustrerade med mycket i (1)-(5) nedan.

(1) Kvantor vid nominal: mycket pengar (2) Kvantor vid komparativ: mycket klokare/fortare (3) Kvantor vid vebfras: Hon simmar mycket numera (4) Gradadverbial vid positiv: mycket klok/fort

I andra språk (inklusive fornsvenska) görs normalt en skillnad mellan kvantorer (much, viel, beaucoup, mucho) i (1)-(3) och gradadverbial (very, sehr, très, muy = väldigt) i (4). När orden är kvantorer i svenska finns en tryckstark och en trycksvag variant. De tryckstarka (ˈMYke och ˈLIte) förefaller fungera i alla kontexter om de är kontrastiva. Det är de trycksvaga (eller deaccentuerade) som har polaritetsegenskaper, här betecknade ˌmyke och ˌlite.

I föredraget avgränsar jag polaritetskontexter och utesluter vissa tolkningar, men när detta är gjort uppvisar ˌmyke alla tecken på att vara ett starkt polaritetselement, medan ˌlite har helt motsatt distrubution. Se exemplen på affirmativ och negerad sats i (5)-(6).

(5) Hon har FÅTT *ˌmyke pengar. Hon har inte FÅTT ˌmyke pengar.

(6) Hon har FÅTT ˌlite pengar. Hon har inte FÅTT *ˌlite pengar.

I föredraget försöker jag beskriva dirtributionen av ˌmyke och ˌlite i affirmativa, negerade, rogativa, kvesitiva och konditionala satser och ge en översikt över distributionen. Jag diskuterar också particip som tycks utgöra lite av en mellankategori (på svenska och i andra språk).

The “Double Object” Verb-Particle Constructions in Swedish Shiro Shibata & Takayuki Tohno

Graduate School of Language and Culture, Osaka University

In this talk, what we name Double Object Verb-Particle Construction (henceforth DOVPC) will be presented. The

DOVPC is a grammatical construction which consists of a verbal particle and two succeeding argument nominals in

the VP (schematically [VP V PRT NP1 NP2]). The verbal particle here can be realized as several different lexical items

which have prepositional correspondences such as till ‘to’ av ‘off’ and på ‘on’. Examples are given in (1–3):

(1) Jag stack till honom min pasta med röd pesto...

I stuck to him my pasta with red pesto

’I stuck my pasta with red pesto to him…’

(2) …att Pernilla lyckats… dra av honom byxorna och kalsongerna…

that P. succeeded.in pull of him trousers and underpants

’…that P. successfully took off his trousers and underpants…’

(3) att rapa ur sig denna text tar ingen tid…

to belch out.of oneself this text take no time

‘To belch this text out takes no time...’

In previous studies which have handled DOVPC, it has been considered either as a kind of Double Object

Construction (Teleman et al. 1999: 423–424; Toivonen 2001: 91–93; Tohno 2001), or a transitive construction with a

preposed prepositional phrase (Svenonius 2003; see also Åfarli 1985).

The purpose of the study is to further scrutinise the characteristic of DOVPC. In this talk, with Construction

Grammar (Goldberg 1995 inter alia) as a framework and with text data from a blog corpus, we will demonstrate

semantic idiosyncrasies of DOVPC in comparison to other constructions (viz. DOC and prepositional transitive

construction). Specifically, DOVPC bears various specific interpretations according to its particle’s meaning: more

general transfer of possession as in (1), undressing as in (2) and abusive uttering as in (3). Also, DOVPC with a

reflexive pronoun in NP1 slot (reflexive DOVPC; see (3)) is broader than its non-reflexive variant with respect to

both its productivity and its constructional polysemy extension. What is interesting is that the seemingly unrelated

constructional meanings expressed by various reflexive DOVPCs actually fall into indirect middle/reflexive domain

(Kemmer 1993). This indicates that DOVPC is a grammatical construction distinct from both DOC and transitive

construction with a prepositional phrase.

References Åfarli, Tor Anders (1985). “Norwegian Verb Particle Constructions as Causative Constructions”, Nordic Journal of Linguistics 8, 75–98 / Goldberg, Adele E (1995). Constructions: a construction grammar approach to argument structure. Chicago: Univ. of Chicago Press. / Kemmer, S. (1993). The middle voice (Vol. 23). John Benjamins Publishing. / Svenonius, Peter. (2003). “Swedish Particles and Directional Prepositions”. In: Delsing, Lars-Olof, Cecilia Falk, Gunlög Josefsson, & Halldór Ármann Sigurdsson (eds.). Grammar in Focus: Festschrift for Christer Platzack 18 November 2003, II, Dept. of Scandinavian, Lund University, 343–351 / Teleman, Ulf, Hellberg, Staffan & Andersson, Erik. (1999). Svenska akademiens grammatik. 1st ed. Stockholm: Svenska akad. Tohno, Takayuki. (2001). “Suweedengono Nidjuumokutekigo-koobun’no Kakuchoo”. Proceedings of The 123rd Meeting of the Linguistic Society of Japan, 248–258 / Toivonen, Ida (2003). Non-projecting words: a case study of Swedish particles. Dordrecht: Kluwer.

Prosodisk variasjon ved partikkelverb i tre norske dialektområde

Eirik Tengesdal, Universitetet i Oslo Björn Lundquist, Universitetet i Tromsø og Universitetet i Oslo Ida Larsson, Universitetet i Oslo

Ordstillinga i partikkelverbkonstruksjonar kan som kjent variera i norsk (1) (jf. t.d. Aa 2015). Men variasjonen syner seg i tillegg i prosodisk realisering. Her reknar me med minst tre ulike realiseringar: deaksentuert verb og aksentuert partikkel (2a), aksentuert verb og aksentuert partikkel (2b), og samansetjingsaksent på verbet saman med partikkelen og eventuelt andre element (2c). I dette føredraget legg me fram funna frå ein gjennomgang av korpustreff frå Nordisk dialektkorpus (Johannessen et al. 2009) basert på Larsson og Lundquist (2014) med dialektar frå dei tre norske fylka Buskerud, Nord-Trøndelag og Finnmark, med føremålet å granska den prosodiske variasjonen ved partikkelverb.

(1) a. Me kasta ut hunden/%han. b. Me kasta hunden/han ut.

(2) a. kasta ˈut b. ˈkasta ˈut c. ²(kasta ut)

Me vil drøfta tre spørsmål:

i. Kva for syntaktiske restriksjonar finn ein på den prosodiske variasjonen i korpusdataa? ii. Er der kategoriske skilnader mellom dialektane?

iii. Kva for faktorar er det som styrer aksenttildeling?

i) Ein finn samansetjingsaksent berre i kontekstane med verb og partikkel, og i kontekstar med eit trykklett pronomen mellom verb og partikkel. Eit unnatak er Nord-Trøndelag, som i tillegg tillèt og har signifikant fleire lette adverbial som vel og nå mellom verb og partikkel. På den andre sida finn ein aldri samansetjingsaksent når ein full nomenfrase står mellom verb og partikkel. ii) Gjennomgangen syner at alle tre dialektområde har alle tre aksenttypane, sjølv når verb og partikkel står ved sida av kvarandre. Det verkar difor som der finst variasjon i alle dialektane, og ikkje naudsynlegvis kategoriske skilnader mellom dialektane. iii) Kvifor er der ikkje alltid samansetjingsaksent når verb står rett føre partikkel, eller når berre eit lett pronomen står mellom verb og partikkel? Ein kan tenkja seg at t.d. visse hjelpeverb, der hjelpeverbet kanskje ikkje gjerne tek eige trykk, kan medføra aksent berre på partikkelen.

Utvalde referansar:

Aa, Leiv Inge. 2015. The Grammar of Verb-Particle Constructions in Spoken Norwegian. Doktoravhandling. NTNU.

Johannessen, Janne Bondi, Priestley, Joel, Hagen, Kristin, Åfarli, Tor Anders & Øystein Alexander Vangsnes. 2009. The Nordic Dialect Corpus – an Advanced Research Tool. I: Jokinen & Bick (red.), Proceedings of NODALIDA 2009.

Larsson, Ida & Björn Lundquist. 2014. Objektsplacering vid partikelverb i norska dialekter och äldre svenska. I: Johannessen, Janne Bondi & Kristin Hagen (red.), Språk i Norge og nabolanda. Oslo: Novus. 99–131.

Anaphor, pronoun, or other? Investigating PRO using Reactivation PatternsTori Larsen and Christer Johansson

University of Bergen

The intricacies of anaphora resolution continue to puzzle theoreticians. A combination ofsyntactic, semantic, and pragmatic factors make the concept di�cult to analyze. To complicatematters, the exact definition of an anaphor varies across linguistic frameworks. If we considera broad definition, an anaphor is an NP that refers to something that has been previously beenmentioned using a di�erent referring NP (antecedent) (Löbner, 2013). Under this definition, ananaphor can be a pronoun, as in example (1).

(1) My cousin [antecedent] does not eat her vegetables like she [anaphor] should.

If we consider a narrower definition of an anaphor, more specific to theories of GenerativeGrammar, then an anaphor and a pronoun are defined as di�erent types of NPs. This di�erencestems from the Government and Binding Theory and whether or not an NP obligatorily receivesits meaning from another intersentential NP. The binding restrictions for pronouns di�er fromthose of anaphors. The former must be free within its binding domain, while the latter mustbe bound within its binding domain (c.f. Chomsky, 1981; 1982).

The concept of an anaphor becomes even fuzzier when we include empty categories. PRO, acovert NP that holds the subject position of the non-finite clause in a Control sentence, displayscharacteristics of both anaphors and pronouns (see example (2)). As a result, we cannot relyon the Binding Theory when investigating the coreferential relationship between PRO and itsantecedent. What rules and regulations determine PRO antecedent assignment?

(2) My cousin [antecedent] promised me to PRO [anaphor] eat her vegetables.

We have chosen to investigate PRO by analyzing Norwegian PRO antecedent reactivatione�ects during sentence processing. Previous research on overt NP trace processing has shownsupport for a Trace Reactivation Hypothesis (TRH) (Nicol and Swinney, 1989). This hypothesisstates that reference-dependent items cause reactivation of their antecedent NPs. NP traceshave been found to display immediate antecedent reactivation at the trace position (Nicol andSwinney, 1989; Hornstein, 1999). This implies coreference processing and referent assignmentoccurs immediately during real-time sentence parsing. Do covert NPs, such as PRO, presentsimilar reactivation patterns? Do they even reactivate their antecedents? If no reactivation ispresent during sentence processing then it is possible that a simple search-and-find (memory)mechanism is being used for antecedent assignment, instead of a coreference relationship.

Our research on PRO antecendent reactivation patterns in Control sentences has shownthat only the subject NP of the matrix clause displays a significant reactivation (facilitationwhen primed) e�ect, whether or not it is the correct antecedent of PRO. The data suggeststhat 1) the subject NP remains activated throughout sentence processing and 2) there is arule/mechanism that prevents the activated subject NP from being assigned as the antecedentof PRO in instances of Object Control.

We will discuss the possible theoretical implications of these reactivation patterns and specu-late about the involvement of semantics and pragmatics in this primarily syntactic phenomenon.We will also examine how the TRH allows for the investigation of theory-based items using ex-perimental methods.ReferencesChomsky, N. (1981). Lectures on Government and Binding. Foris.Chomsky, N. (1982). Some concepts and consequences of the theory of government and binding. MIT

Press.Hornstein, N. (1999). Movement and control. Linguistic Inquiry, 30(1):69–96.Löbner, S. (2013). Understanding semantics. Routledge.Nicol, J. and Swinney, D. (1989). The role of structure in coreference assignment during sentence com-

prehension. Journal of Psycholinguistic Research, 18(1):5–19, https://doi.org/10.1007/BF01069043.

Set focus and reference: an ERP study

Eva Klingvall (Lund University) & Fredrik Heinat (Linnæus University)

In this talk, we present the results from an Event Related Potentials (ERP) studyon the processing of anaphoric reference to quantified expressions (QEs) in Swedish.QEs pick out proportions of possible members of some set for which a propertyholds. In (1a) and (1b), for example, some or few members of the set of studentsattended the lecture.

(1) a. Some students attended the lecture.

b. Few students attended the lecture.

(2) a. They found it very interesting.

b. They stayed at home instead.

Some and few di↵er in polarity: some is positive (upward entailing) while few isnegative (downward entailing) (Peters and Westerstahl, 2006) and this is of import-ance when referring back to the QE using anaphoric expressions. The sentence in(1a) is naturally followed by (2a), which is about the students attending the lecture(the reference set, refset). The sentence in (1b), in contrast, is naturally followedby (2b), which is about the students not attending the lecture (the complement set,compset) (e.g. Moxey and Sanford, 1987). While (1b) can in fact be followed eitherby (2a) or (2b), (1a), cannot be followed by (2b).

160 experimental items (from Heinat and Klingvall 2018) of four sentences eachwere manipulated along two dimensions: polarity (positive vs negative quantifier,nagra vs fa in (3)), and set (refset vs compset targeting disambiguating adjective,duktiga vs daliga in (3)).

(3) Nagra/Fasome/few

studenterstudents

skrevwrote

brawell

paon

tentanthe-exam

igaryesterday

ochand

attthat

deCW

theyvarwere

saso

duktiga/daligaCW

good/badforbrylladeconfused

professorn.the-professor

There were four lists with 40 sentences from each condition. Each participant (31)only saw one sentence from each item, but saw all types of manipulation. In total,each participant read 400 sentences (160 test items, 240 fillers).

Unlike Filik et al. (2011) we found that positive QEs showed a pronouncedpositivity over the central region in the compset condition relative to negativeQEs, in the P600 time span after the onset of the critical word (the disambiguatingadjective,‘bad’). We interpret this to mean that for positive QEs, a new discoursereferent needs to be introduced following compset reference, while for negative QEsthis discourse referent is already available (Burkhardt, 2007). In the talk we discusswhat consequences this e↵ect has on our understanding of anaphoric reference toquantified expressions in Swedish.

1

Definiteness in diachrony: Romance vs Germanic

Nigel Vincent

The University of Manchester nigel.vincent@manchester.ac.uk

Definiteness and determiners lie at the intersection of three, all too often

independent, research traditions: semantic, syntactic and diachronic. For the first two, the issues turn on the nature of the conceptual primitives and the kinds of structures they map onto, whereas for the last the key question is the way links between form and meaning shift over time and what historical trajectories are discernible (König 2018). The challenge then is to find ways of integrating the results which emerge from these different lines of inquiry. In addressing this challenge, I will build on the approach set out in Börjars et al (2016) in which neither cartographic structure (as for example in Giusti 2015) nor a universal spine (Wiltschko 2014, especially ch. 6) is assumed to pre-exist the development of articles. Rather, syntax and semantics are modelled independently of each other with the changes in the formal mapping between the two being explained in diachronic and functional terms.

Placing Romance and Germanic developments within a broader typological picture (Dryer 2014), I will focus on three case studies: (a) the different diachronic trajectories and hence synchronic distributions of Romance articles deriving from Latin ipse and ille (see data and references in Vincent 2017); (b) the contrasting patterns of development and eventual status of end articles in North Germanic and Romanian (Börjars 1994, Faarlund 2009, Dahl 2010); (c) the evolution of the indefinite article in Romance and Germanic and the place of plural indefinites such as Spanish unos/-as.

References: • Börjars, K. (1994): Swedish double determination in a European typologicalperspective.Nordic Journal of Linguistics17: 219-252. •Börjars, K., P. Harries & N. Vincent (2016): Growing syntax: the development of a DP in North Germanic. Language 92, e1–37. • Dahl,Ö. (2010)Grammaticalization in theNorth:NounPhrase Syntax in ScandinavianVernaculars.University of Stockholm Dept of Linguistics. • Dryer, M. (2014): Competing methods for uncovering linguistic diversity: the case of definite and indefinite articles. Language 90, e232-249. • Faarlund,J-T.(2009):OnthehistoryofdefinitenessmarkinginScandinavian.JL45:617-639.•Giusti, G. (2015): Nominal Syntax at the Interfaces: A Comparative Study of Languages with Articles. Newcastle: Cambridge Scholars Press. • König, E. (2018): Definite articles and their uses: diversity and patterns of variation. In D. Olmen et al (eds) Aspects of Linguistic Variation, Berlin: De Gruyter, 165–184. • Vincent, N. (2017): Determination and quantification. In: A. Dufter & E. Stark (eds) Manual of Romance Morphosyntax and Syntax, Berlin: De Gruyter, 725–770. • Wiltschko, M. (2014): The Universal Structure of Categories. Cambridge University Press.

Language change in definiteness marking in American heritage Norwegian

Yvonne van Baal, University of Oslo In this talk, I will discuss compositional definiteness (CD) in American Norwegian (AmNo), a heritage variety of Norwegian spoken in the US by 3rd and 4th generation immigrants. I will show how AmNo differs from homeland Norwegian, and provide an analysis of the language change in AmNo.

Results from two elicitation tasks reveal that although homeland-like CD is found in AmNo (1), modified definite phrases typically only contain the suffixed article (2a-b). This makes AmNo different from homeland Norwegian. Phrases with only the prenominal determiner (2c) are much less frequent and only used by a few speakers. This suggests that the suffix is stable in AmNo whereas the determiner is not.

(1) den hvite hest-en “the white horse” (2) a. brone hest-en “the brown horse” b. grønne bok-a “the green book” c. den grønne bil “the green car”

Within Scandinavian varieties, modified definite phrases without a determiner are used in Icelandic and Northern Swedish, and in homeland Norwegian with a limited set of adjectives. Julien (2002, 2005) suggests that these are the result of movement of αP (which includes the adjective, the noun and the suffix) to Spec-DP. This movement has two restrictions: (i) αP cannot move across cardinal numbers and (ii) αP cannot move when there is ellipsis of the noun. In both cases, D is obligatory spelled out.

If αP-movement accounts for AmNo phrases without the determiner (2a-b), the same restrictions are predicted to be active. However, these predictions are not borne out: phrases with a cardinal number or ellipsis where the determiner is absent are found in AmNo (3) and accepted by speakers in a judgment task. These findings suggest that there is no αP-movement in AmNo. Since there is enough evidence to conclude that AmNo still has a DP layer, I propose that the determiner has become optional in AmNo modified definite phrases.

(3) a. to brone hund-ene “the two brown dogs” b. hvite “the white one”

The patterns found in AmNo are strikingly similar to those found in monolingual children, who acquire the suffix very early, and the determiner much later. Phrases without the determiner are both frequent and persistent during acquisition (Anderssen 2007, 2012). In their bilingual environment, heritage speakers do not only receive less input than monolingual children, they also only receive input from spoken language and not from written language or formal registers (cf. Montrul 2016). This leads to a quantitative and qualitative different input.

I propose that this different input leads to a new grammar. Whereas monolingual children go through a stage of optional determiners, this optionality is maintained in the heritage language. As a result, modified definite phrases without the determiner are found in all contexts in AmNo. The bilingual acquisition has, I suggest, caused change in AmNo. This change has not made AmNo more like English, but rather led to a grammar that is unlike the other Scandinavian varieties. References: Anderssen, M. (2007) The acquisition of compositional definiteness in Norwegian. Nordlyd 34(3), 252-276. Anderssen, M. (2012) A spanning approach to the acquisition of double definiteness in Norwegian. IBERIA: An International

Journal of Theoretical Linguistics 4(1), 1-34. Julien, M. (2002) Determiners and word order in Scandinavian DPs. Studia Linguistica 56(3), 246-314. Julien, M. (2005) Nominal Phrases from a Scandinavian Perspective. Amsterdam/Philadelphia: John Benjamins. Montrul, S. (2016) The Acquisition of Heritage Languages. Cambridge: Cambridge University Press.