types of topicalization & the role of...

Types of Topicalization & the Role of Reconstruction Elena Callegari

University of Oslo

[email protected]

1 Types of Topicalization

When we think of non-focal movement to the Left Periphery, we generally

think of Clitic-Left Dislocation (CLLD, Cinque 1990):

(1) A casa, i la ga…

The house, they it(cl) have1 (Trevigiano)

In fact, there are several other types of movement operations which front

non-focal constituents:

Simple Preposing (SP):

(2) Dije que terminaría el libro, y el libro he terminado. (Spanish)

I-said that I-would-finish the book, and the book I-have finished2. (Leonetti & Escandel-Vidall 2009: 157)

Quantifier Fronting (QF)(Cinque 1990, Arregi 2003, Giurgea 2015):

(3) “...ragazzi che qualcosa sanno fare, chi più chi meno”

“…kids that something they-can do, who more and who less” (Italian, Paisà corpus)

1 This is part of an actual dialogue which was recorded for an independent

project on subject topics in Northern Italian dialects. 2 This type of movement is referred to as «Verum Focus Fronting» in

Leonetti & Escandel-Vidall (2009). The term “verum focus” however suggests

that the polarity focus is always emphatic, which it need not be.

mailto:[email protected]

Word Order in the Left Periphery, 18/10/2017

2

Bare Neg Fronting (BNF):

(4) Mi scuso per la domanda cretina che poi tanto cretina non è!

I apologize for the question stupid, which very stupid not it-is (Italian, Paisà corpus)

Main Features:

CLLD requires mandatory clitic resumption of any fronted DP. PPs may not

be clitic-resumed (but they must in Sardinian, Spanish and Catalan, see

Cruschina 2010).

SP requires strict identity between the clause where the fronting occurs

(parasitic), and a preceding clause (host). The host always features a

non-finite clausal complement (usually the complement of a volitional/modal

verb), which is rendered finite (perfective) in the parasitic SP structure.

This results in the expression of a contrast in veridicality. Clitic-doubling is

impossible in SP:

(5) Volevo mangiare quella torta, e quella torta ho mangiato3

I-wanted to-eat that cake, and that cake I-have eaten (6) Volevo mangiare quella torta, e quella torta l’ho mangiata

*I-wanted to-eat that cake, and that cake it(cl)-I-have eaten.FEM

BNF fronts DPs or APs to the left periphery of a negated sentence, often

giving rise to a litotes reading (Horn 1989, 1991; van der Wouden 1995)

(see 4 and 7), but not necessarily (see 8):

(7) Il paese fa 13.000 abitanti. Quindi tanto piccolo non è

This town has 13.000 inhabitants. So very small not it-is (8) Trattarlo con un farmaco lo etichetta come malato anche se malato

non è.

Treating him with drugs labels him as ill, even though ill not he-is. (both from Paisà corpus)

3 Unless otherwise indicated, all examples are from Standard Italian.

Elena Callegari University of Oslo

3

Clitic resumption is absent, but may be possible in at least some cases

(more on this in section 4).

QF can be used in litotes environments (9), or take an existential function (10). In the latter case, it only fronts existential quantifiers like

someone/somebody:

(9) A: Luisa ha mangiato poco (litotes)

A: Luisa has eaten little B: Tanto non ha mangiato di certo

B: A-lot not she-has eaten for sure

(10) A: Luisa non ha mangiato nulla (existential)

A: Luisa not has eaten nothing B: Qualcosa ha mangiato, dai

B: Something she-has eaten, come-on = it definitely is the case Luisa ate at least something

In QF environments, clitic resumption is generally barred:

(11) *?Tanto non lo ha mangiato di certo

*?Much not it(cl) she-has eaten for sure

Goals:

- To determine what triggers these types of fronting. - To gain a clearer picture of what determines whether a non-focal

constituent is going to be clitic-resumed or not.

2 Reconstruction & The Locality of Movement

Reconstruction possibilities are going to be crucial in section 4 when

formulating an analysis for the different types of non-focal fronting.

How do the different types of topicalization behave with respect to

reconstruction?


4

SP reconstructs for both binding and scope:

Binding (12) Leoi voleva riscoprire se stessoi, e se stessoi Leoi ha riscoperto.

Leoi wanted to-rediscover him selfi, and him selfi Leoi has rediscovered.

Scope (13) Lucia voleva presentare ad ogni ragazzo il suo professore, e ad ogni

ragazzo il suo professore Lucia ha presentato.

Lucy wanted to introduce to each boy the POSS. professor, and to each boy the POSS. professor Lucia has introduced

(a) ✓ “the POSS. professor” is bound by “Lucia”

(b) ✓ “the POSS. professor” is bound by “each boy”

BNF reconstructs for both binding and scope:

Binding (14) Se stessoi, Luigii non ha visto di certo

Himselfi, Luigii not has seen for sure Scope For at least some cases of litotes fronting, it can be shown that the fronted

quantifier must obligatorily reconstruct for scope. One such case is (15):

(15) A: Non ha mangiato nulla

A: Not he-has eaten nothing (=he ate nothing) B: Tutto non ha mangiato di certo

B: Everything not he-has eaten for sure Why reconstruction?

Reconstructed vs. non-reconstructed scope in (15):

(16) ¬ > ∀ = it is not the case that he ate everything (hence: he ate a little, or he ate a lot, ..)

(17) ∀ > ¬ = Everything was not eaten = He ate nothing

- If we follow Giurgea (2015) in assuming that the fronted quantifier

functions like a CT, then the non-reconstructed version would violate

Büring’s (1999) generalization on the usage of CTs (S-Topics): at


5

least one of the questions in the topic value muts remain open after

the utterance.

- The non-reconstructed version is literally a restatement of what stated

by A, so it is a violation of both the maxim of relevance and that of

manner.

Note that BNF is very local:

(18) Vuole essere gentile, ma gentile non è

He-wants to-be kind, but kind not he-is (19) *Vuole essere gentile, ma gentile non credo che sia

*He-wants to-be kind, but kind not I-believe that he-is(subj) (20) Vuole essere gentile, ma gentile non credo che lo sia

He-wants to-be kind, but kind not I-believe that it(cl) he-is(subj)

Does CLLD reconstruct? Two discordant analyses:

Frascarelli (2004), Frascarelli & Hinteholzl (2007). CLLD does not

reconstruct, neither for binding nor for scope.

Cecchetto (2001): CLLD of a DP is always accompanied by reconstruction

for binding, and can or can not be accompanied by reconstruction for scope.

CLLD is indeed compatible with both a surface and an inverse reading, at

least with some types of quantifiers (existentials). It also seems to be the

case that the surface reading is only available (but see section 5) when

clitic resumption applies.

(21) Qualcuno lo amano tutti

Someone him(cl) love everyone = Everyone loves someone ✓ ∀ > ∃ (everybody loves someone different)

✓ ∃ > ∀ (someone specific is loved by everyone)

(22) A: Nessuno ama nessuno

A: Nobody loves nobody B: Qualcuno amano tutti!

B: Someone love everyone!


6

✓ ∀ > ∃ * ∃ > ∀

3 A Prosodic Trigger

Following a tradition started by Vallduví & Engdahl (1996), Zubizarreta

(1998) (see also Hamlaoui & Szendrői (2015), Szendrői (2002, 2003,

2017), Samek-Lodovici (2005, 2006, 2015)), I argue that the (initial)

trigger behind topicalization is prosodic. For the local types of topicalization,

prosody seems to be all there is to it.

The size and location of the material in focus has an effect on whether

topicalization takes place, and what constituent(s) this targets. Consider the

following structures:

(23) Una gonna rossa ce l’ho, sì. (CLLD)

A red skirt, locative(cl) it(cl)-I-have, yes. (24) Volevo andare piano, e piano sono andato (SP)

I-wanted to-go slowly, and slowly I-have gone (25) Brutto non è, quello è sicuro. (BNF)

Ugly he-is not, that is for-sure

At first sight, (23-25) have nothing in common: not the type of constituent

fronted, not the polarity of the sentence, not the presence/absence of clitic

resumption. What they do have in common is the fact that (i) they are all

verb-final, (ii) they all feature a polarity focus:

(26) A red skirt I DO have, yes.

(27) I wanted to drive slowly, and so I DID

(28) Ugly he is definitely NOT, that’s for sure.


7

Cross-linguistically, the to-go strategy to mark (at least some types) of

polarity focus is to assign main stress to the finite verb4:

(29) A: Ja ne sčitaju Yurija Solomina talantlivym aktiorom (Russian)

A: I not consider Yurija Solomina talented.INSTR actor.INSTR B: Net, on BYL talantlivym aktiorom

B: No, he WAS talented.INSTR actor.INSTR (Pereltsvaig 2007:100-101)

(30) A: Karl har ikke matet hunden (Norwegian)

A: Karl has not fed the-dog B: Han HAR gjort det!

B: He HAS done that!

Languages differ on whether they allow the type of stress shift in (29-30):

- Norwegian, Russian, English: stress shift is possible, resulting in in

situ focalized finite verbs.

- Hungarian (Hamlaoui & Szendrői 2015; Szendrői 2017): stress shift

is available for some constituents which cannot be interpreted

exhaustively, such as universal quantifiers.

- Italian, Spanish: these are stress-rigid languages (Samek-Lodovici

2005, 2006, 2015; Szendrői 2017; Ortega-Santos 2016). The

application of stress shift is limited.

In languages like Italian, the (initial) trigger of topicalization is the

necessity to remove a non-focal constituent from a main stress position,

which explains why all the fronting structures in (23-25) have an identical

focus value ([¬p, p]). 4 Other strategies consist in stressing specific discourse particles (as in

Dutch, see Hogeweg 2009), the complementizer (German and Norwegian, see

Hetland 1992, Lohnstein 2016), clitics (Slovenian, Jasinskaja 2015), or in

adding specific adverbials (alright in English, quidem in Latin, see Danckaert 2014).


8

How does the polarity focus reading arise? Consider a standard SP

construction like (31):

(31) Había que leerse el Quijote, y el Quijote se leyó. (Spanish)

(S)he-had to read the Quijote, and the Quijote she-read (Leonetti & Escandel-Vidall 2009: 171)

The strict SP identity requirement means that every constituent (verb

included) in the parasitic SP sentence is old information, and thus, in a

sense, topical. The only bit of truly new information is the tense specification

on the parasitic verb: [- finite] in the host, and [+ finite] in the parasitic

clause. It is this finite tense specification which triggers the stressing of the

verb, and hence the movement of whatever constituent follows it.

Why polarity focus? The finite tense specification on tha parasitic sentence’s

verb signals how whatever activity was described in the host has taken place.

Polarity focus stresses whether a given proposition is true or false. The tense

specification does something equivalent: it marks the action as finite, thereby

marking it as true.

Note that a stress-shift language like Norwegian can realize structures like

those in (31) without necessarily resorting to fronting:

(32) Atle ville spise brunost, og han SPISTE brunost (Norwegian)

Atle wants to-eat brunost, and he ATE brunost (33) Atle ville spise brunost, og spise/spiste brunost, det gjorde han.

Atle wants to-eat brunost, and eat/ate brunost, that did he.

4 Why Different Types of Topicalization?

Why is only CLLD accompanied by mandatory clitic-resumption?

Some previous accounts which will not work (at least not all the way):


9

(i) Cinque (1990): the fronted bare quantifier in QF structures, being an

operator, exploits the same operator-variable mechanism exploited by

wh-movement and focalization, hence no clitic is necessary.

SP can front constituents which are not operators (see 2, 5, 31) (ii) Arregi (2003): CLLD is a contrastive operation, hence only contrastive

topics can be clitic-resumed.

Not all CLLD topics must be contrastive: (34) A and B have a friend, Paola, who is supposed to come by to

borrow one of A’s cocktail dresses to wear for the inauguration of

her art gallery. Before he leaves to go to work, A tells B she should

lend Paola her blue dress, or perhaps the pink one, as those are

the prettiest she possesses. When A comes home after work, B tells

him:

“A proposito, a Paola alla fine le ho dato il vestito rosa. Quello blu

non le stava”

“By the way, to Paola in-the end to-her(cl) I-have given the dress pink. That blue not to-her fit”

(iii) Constituents which are semantically of type e (individuals) should be

clitic-resumed5-.

SP can front constituents of type e (see again 5, 31)6.

Concerning the property of ranging over individuals: 54 examples of BNF in

the Paisà corpus. Half of them could potentially be turned into CLLD

structures, half of them could not. The latter half mostly consists of mass

nouns/non-atomic entities:

5 This solution is explored in Arregi (2003), but later rejected on the basis

of universal quantifiers, which are of type (et)t but can still be clitic-resumed.

Arregi (2003) thus concludes that it is the clitic itself which is interpreted as

ranging over individuals, which cannot be correct either given the grammaticality of

the reconstructed reading in (21). 6 Note that the grammaticality of (5) also excludes an explanation in terms

of the referentiality of the fronted DP. Referentiality seems however to be a factor

in Greek, where only referential topics can undergo CLLD (Alexopoulou and

Kolliakou 2002; Alexopoulou and Folli 2010)


10

(35) “(...) l'abissale ignoranza (...) di chi evoca premesse "filosofiche"

e scientifiche per una disciplina come la medicina che SCIENZA NON

E'!!”

“(…) The abysmal ignorance (…) of those who resort to "philosophical" and scientific premises to describe a discipline like medicine that SCIENCE NOT IT-IS !!

… *che scienza non lo è!!!

... *that science not it(cl) it-is!!!

This talk: BNF and SP exhibit total reconstruction (Saito 1989). The trigger behind the displacement is arguably prosodic (which explains cross-linguistic

differences and identical focus values contexts).

Both types of movement can be captured in terms of PF movement

(Sauerland & Elbourn 2002). Movement of the fronted element takes place

at PF, which explains the lack of semantic/syntactic effects.

Supporting evidence 1: locality (as seen in BNF). Unlike CLLD (Cinque

1990, Abels 2012), the fronted constituent in BNF environments will only go

as far as it takes to be outside of a main stress sposition.

Supporting evidence 2: the extreme variation characterizing the target of SP

fronting. SP can front nominal constituents, APs, adverbials, and in fact it

may target pretty large constituents:

(36) Volevo mettermi a scrivere in soggiorno, e a scrivere in soggiorno mi

sono messa.

I-wanted to-start to write in-the living room, and to write in-the living room I have started


11

5 Non-Reconstructed QF: a Fourth Type of Topicalization

From section 2: clitic-less movement reconstructs. Consider however existential

QF:

(37) A: Mario ha mangiato tutto

A: Mario has eaten everything B: Qualcosa non ha mangiato

B: Something not he-has eaten * ¬ > ∃

✓∃ > ¬

Why no reconstruction?

“Qualcosa” is a PPI, hence it cannot appear in downward entailing

environments.

This appears to be a fourth type of topicalization:

i. CLLD with no reconstruction

ii. CLLD with reconstruction (possible with fronted existentials)

iii. PF movement (SP and BNF)

iv. Existential QF (no clitic resumption but no reconstruction)

6 Conclusions

- The (initial) trigger of topicalization is prosodically motivated.

- A prosodic analysis of topicalization explains why structures like SP

come with a polarity focus.

- SP and BNF are examples of total reconstruction. BNF can also be

shown to be very local. The properties of both structures follow from

a PF movement analysis.

- If we analyze topicalization in terms of reconstruction, we find there

are four types of topicalization: CLLD without reconstruction, CLLD


12

with reconstruction (restricted), cliticless fronting with reconstruction,

cliticless fronting without reconstruction (restricted).

7 References

Abels, K. (2012). The Italian left periphery: A view from

locality. Linguistic Inquiry, 43(1), 229-254.

Alexopoulou, D., & Folli, R. (2010). Indefinite topics and the syntax of

nominals in Italian and Greek. In The 28th West Coast Conference on Formal Linguistics (WCCFL 28) online proceedings.

Alexopoulou, T., & Kolliakou, D. (2002). On linkhood, topicalization and

clitic left dislocation. Journal of Linguistics, 38(2), 193-245.

Arregi, K. (2003). CLLD is contrastive topicalization. In Proceedings of the 26th annual Penn Linguistics Colloquium (Vol. 9, p. 1).

Bocci, G. (2013). The Syntax Prosody Interface: A cartographic perspective with evidence from Italian (Vol. 204). John Benjamins Publishing.

Cinque, G. (1990). Types of Ā-dependencies. MIT press.

Cecchetto, C. (2001). Syntactic or semantic reconstruction? Evidence

from pseudoclefts and clitic left dislocation. Semantic interfaces, 90-144.

Cruschina, S. (2010). Syntactic extraposition and clitic resumption in

Italian. Lingua, 120(1), 50-73.

Danckaert, L. (2014). Quidem as a marker of emphatic

polarity. Transactions of the Philological Society, 112(1), 97-138.


13

Dimroth, C./Andorno, C./Benazzo, S./Verhagen, J. (2010). Given

claims about new topics. How Romance and Germanic speakers link changed

and maintained information in narrative discourse. Journal of Pragmatics 42, 3328-3344.

Frascarelli, M. (2004). DISLOCATION, CLITIC RESUMPTION AND

MINIMALITY A COMPARATIVE ANALYSIS OF LEFT AND RIGHT TOPIC

CONSTRUCTIONS. Romance Languages and Linguistic Theory 2002: Selected Papers from" Going Romance," Groningen, 28-30 November 2002, 256, 99.

Frascarelli, M., & Hinterhölzl, R. (2007). Types of topics in German

and Italian. On information structure, meaning and form, 87-116.

Giurgea, I. BARE QUANTIFIER FRONTING AS CONTRASTIVE

TOPICALIZATION1. Bucharest Working Papers in Linguistics.

Hamlaoui, F., & Szendrői, K. (2015). A flexible approach to the

mapping of intonational phrases. Phonology, 32(1), 79-110.

Hetland, J. (1992). Polaritätsfokus, VERUM-Fokus, Kopffokus. STUF-Language Typology and Universals, 45(1-4), 3-16.

Hogeweg, L. (2009). The meaning and interpretation of the Dutch

particle wel. Journal of Pragmatics, 41(3), 519-539.

Horn, L. (1989). A natural history of negation.

Jasinskaja, K. (2016). Information structure in Slavic. In The Oxford Handbook of Information Structure.

Lohnstein, H. (2015). Verum focus. In The Oxford Handbook of Information Structure.


14

Leonetti, M., & Escandell-Vidal, V. (2009). Fronting and verum focus

in Spanish. Focus and background in Romance languages, 155-204.

Lyding, V., Stemle, E., Borghetti, C., Brunello, M., Castagnoli, S.,

Dell'Orletta, F., & Pirrelli, V. (2014). The PAISA corpus of italian web

texts. In Proceedings of the 9th Web as Corpus Workshop (WaC-9) (pp. 36-43).

Neeleman, A., Titov, E., Van De Koot, H., & Vermeulen, R. (2009).

A syntactic typology of topic, focus and contrast. Alternatives to cartography, 1551.

Neeleman, A., & Vermeulen, R. (Eds.). (2013). The syntax of topic, focus, and contrast: An interface-based approach (Vol. 113). Walter de

Gruyter.

Ortega-Santos, I. (2016). Focus-related operations at the right edge in Spanish: Subjects and ellipsis (Vol. 7). John Benjamins Publishing Company.

Pereltsvaig, A. (2007). The universality of DP: A view from

Russian. Studia linguistica, 61(1), 59-94.

Rizzi, L. (1997). The fine structure of the left periphery.

Saito, M. (1989). Scrambling as semantically vacuous A'-

movement. Alternative conceptions of phrase structure.

Samek-Lodovici, V. (2005). Prosody–syntax interaction in the expression

of focus. Natural Language & Linguistic Theory, 23(3), 687-755.


15

Samek-Lodovici, V. (2006). When right dislocation meets the left-

periphery.: A unified analysis of italian non-final focus. Lingua, 116(6), 836-873.

Samek-Lodovici, V. (2015). The interaction of focus, givenness, and prosody: A study of Italian clause structure. Oxford University Press.

Sauerland, U., & Elbourne, P. (2002). Total reconstruction, PF

movement, and derivational order. Linguistic Inquiry, 33(2), 283-319.

Szendrői, K. (2002). Stress-focus correspondence in Italian. Romance Languages and Linguistic Theory 2000: Selected Papers from" Going Romance" 2000, Utrecht, 30 November-2 December, 232, 287.

Szendrői, K. (2003). A stress-based approach to the syntax of

Hungarian focus. The Linguistic Review 20: 37–78.

Szendrői, K. (2017). The syntax of information structure and the PF

interface. Glossa: a journal of general linguistics, 2(1).

Vallduví, E., & Engdahl, E. (1996). The linguistic realization of

information packaging. Linguistics, 34(3), 459-520.

Van der Wouden, T. (1996). Litotes and downward

monotonicity. Negation: a notion in focus, 7, 145.

Zubizarreta, M. L. (1998). Prosody, Focus, and Word Order (Linguistic inquiry monographs; 33). MIT Press.

types of topicalization & the role of...

Documents