Surprise Negation Sentences (Snegs)

by user








Surprise Negation Sentences (Snegs)
Surprise Negation Sentences (Snegs)
Matteo Greco and Andrea Moro (Iuss/NEtS Pavia)
1. Surprise Negation: a puzzling case study
By “Surprise Negation” we define a particular type of negative marker belonging to the class
of “expletive negations” as discussed in Zanuttini (1997), Zeijlstra (2004) and Horn (2010),
i.e. one which does not reverse the polarity of the sentence, does not license NPI, etc.
Pragmatically, Snegs are limited to a restricted context in which a speaker is struck by an
unexpected fact (hence, the label “Surprise”) and s/he wants to communicate it (hence, the
affirmative polarity and the assertive force). Snegs display a marked intonation blending
acoustic features pertaining to both questions and exclamatives (hence the ?! combined
non è
scesa dal
treno Maria ?!
(Yesterday Sneg be.3rdsing.pres. got off-the train Maria)
“Yesterday, Maria got off the train!”
2. Empirical support for Snegs
The preliminary question is whether Snegs are syntactically encoded or they are just a
semantic/pragmatic phenomenon. We want to argue that the former hypothesis is true.
2.1 Ethical Dative as a diagnostics for Sneg
In Italian, the same negative marker non can express both a propositional negation (PN) and a
Sneg. However the presence of an Ethical Dative (ED) (see Cuervo 2003; Boneh and Nash
2012) can only be associated with the Sneg reading (in the marked intonation):
(2) a. Ieri,
non è
scesa dal
treno Maria?/?!
(Yesterday not be.3 sing.pres. got off-the train Maria)
“Yesterday, Maria did not get off the train / Yesterday, Maria got off the train!”
b. Ieri,
non ti
scesa dal
treno Maria?!
(Yesterday not you.ED be.3rd sing.pres. got off-the train Maria)
“Yesterday, Maria got off the train!”
The co-occurrence of ED and negative marker non qualifies as a diagnostics for Snegs.
2.2 Snegs and the Left-periphery: the incompatibility with Focus Phrases
A further proof that Snegs are syntactically encoded is that a contrastive focalized element
cannot be present (3c), whereas it is usually permitted in affirmative (3a) and negative clauses
(3) a. GIANNI è
sceso dal
treno (non Maria)
(Gianni be.3rd sing.pres. got off-the train (not Maria))
“Gianni got off the train, not Maria”
b. GIANNI non è
sceso dal
treno (non Maria)
(Gianni not be.3 sing.pres. got off-the train (not Maria))
“Gianni did not get off the train but Maria did”
c. * GIANNI non ti
sceso dal treno (non Maria)?!
(Gianni Sneg you.ED be.3rd sing.pres. got off the train (not Maria))
However, Snegs are completely acceptable with Topic Phrase, witness CLLD structures:
(4) Maria non te
vista scendere dal
(Maria Sneg you.ED cl.3rdsing have.3rd sing.pres. seen get
off the train)
“I saw Maria get off the train!”
Notably, a pre-verbal subject-like phrase should be analyzed as a hanging topic, as the
presence of a right dislocation implies:
(5) Maria non ti
(Maria Sneg you.ED arrive.3 sing.pres she)
“Maria arrives!”
2.3 Snegs as root clauses
When it comes to the root vs embedded clausal types, Snegs qualify as root. Consider two
different types of embedded clauses: one introduced by a bridge verb (6), selecting a fullfledge CP layer (Benincà & Poletto 2004), and another introduced by a factive predicate (7),
selecting a reduced CP layer (Haegeman 2005). Snegs cannot co-occur with any of them
regardless the CP-structure.
(6) Luca dice
che non (ti)
scesa dal treno Maria
(Luca say.3 sing.pres. that not you.ED be.3 sing.pres. got off-the train Maria)
“Luca says that Maria didn’t get off the train”
(7) Mi
che non (ti)
scesa dal treno Maria
(To me regret.3rdsing.pres. that not you.ED be.3rdsing.pres.subj. got off the train Maria)
“I regret that Maria didn’t get off the train”
These data support the hypothesis that Snegs are not pragmatic or semantic phenomena; they
are rather encoded syntactically in a specific way.
3. Excluding simple-minded solutions
Starting from the fact that Snegs share some semantic and prosodic features with other
linguistic structures, i.e. Rhetorical questions and Exclamatives, we could considerer them as
belonging to some more general linguistic categories. Unfortunately, this is not the case.
3.1 Snegs are not Rhetorical questions
Snegs share with negative rhetorical questions (NRQ) a few properties: (i) they have an
assertive force, (ii) they involve a negative marker (even though their meaning is affirmative)
and (iii) they display an interrogative intonation. However Snegs can be distinguished from
NRQ for at least two reasons: (i) NRQs are not necessary root clauses (8) unlike Snegs; (ii)
typical yes-no RQ-element such as After all (Han 2002) cannot occur in Snegs (9).
(8) Dopo tutto, dimmi cosa non ha
fatto Laura per Luca?
(After all tell-me what not have.3rd sing.pres. done Laura for Luca)
“After all, tell me what Laura hasn’t done for Luca?”
(9) *Dopo tutto, non ti
scesa dal
treno Maria?!
(After all,
Sneg you.ED be.3rd sing.pres. got off the train Maria)
3.2 Snegs are not Exclamatives
An alternative analysis is to considerer Snegs as negative exclamative clauses since they share
an expletive negation, but this is not the case. According to Zanuttini and Portner (2003),
exclamatives are factive in nature, they generate scalar implicatures, and they cannot be an
answer to a question. Snegs do not match at least two of these features: i) Snegs are not
factives inasmuch they cannot be embedded in factive predicates (7); ii) Snegs can be an
answer to a question:
(10) A: Cosa succede?
What’s happen?
B: a. *Che cosa non ha fatto Gianni!
What did not John do!
b. Non è
scesa dal
“Maria got off the train!”
treno Maria?!
4. Toward a structural analysis
In the previous sections we have seen that Snegs are neither rhetorical questions nor
exclamatives and, at the same time, we showed that they have a specific syntactic behavior: i)
they are affirmative in meaning and do not license NPI; ii) they are root clauses; iii) they are
incompatible with FocP but allows TopP; iv) the pre-verbal subject is in topic position; v) they
are disambiguated by the presence of an ED.
4.1 Proposal
To explain all these facts concerning Snegs, we propose that the entire TP is raised to SpecFoc° and non occupies the position of a higher head in the CP-field (Rizzi 1997) licensing
phonetically null Foco head:
[CP … [X° non ] … TP Foc° [… tTP …] ]
We assume that the Sneg-morpheme activates all foci c-commanded by it, i.e. at least FocP in
CP and FocP in TP (see Belletti 2002 for the latter structure). We also propose that the head
non is externally merged and, due to its propositional nature, it selects the entire TP as an
argument (which is forced to raise in the structure to get a local configuration).
Some consequences
The data seen in the previous sections follow for principled reasons, given the analysis in (11):
Snegs host topicalized elements but not focalized ones: if the [Spec,FocP] is already
occupied by the TP, there is no more space to license another focalized element in
the same clause; on the contrary, the topicalized elements are free to occupy the
Top positions in a higher portion of the CP-structure (Benincà & Poletto2004).
Snegs are root phenomena: usually embedded clauses are endowed with a reduced
CP layer and they lack the FocP (see Haegeman 2004). For this reason Snegs
cannot be embedded. As regards the bridge-verb clauses, selecting a full-fledge CP
layer (Benincà & Poletto 2004), the incompatibility with Snegs is a consequence of
an independent restriction: bridge verbs do not allow the focalization of the
inflected TP (12b) but only of the TP-internal elements (12a). Given that Snegs
focalize the entire TP, they cannot be embedded even in bridge-verb clauses.
(12) a. Dico
che, ROMA, pensano
sporca (non NY)
(Say.3rd sing.pres. that Rome think.Istplur.pres. be.3rdsubj.pres. dirty (not NY))
“I just say that ROME they consider dirty (not NY)”
b. * Dico
SPORCA pensano
(Say.3rdsing.pres. that, that Rome be.3rdsubj.pres. dirty
The unavailability of the pre-verbal subject: since the Sneg morpheme activates all
foci in its c-command domain, the subject can only occupy the lower focus position
right above the VP (Belletti 2002). If the subject is in a pre-verbal position, it is
topicalized (since the focus is already occupied by the TP).
If non in Snegs is not a real instance of a propositional negation, as follows by
assuming that the Sneg-morpheme is generated in the CP layer rather than in a TP
internal position (Zanuttini 1997), then it is easy to explain why Snegs cannot host
NPI, since the adequate negative context is lacking.
Along with (i)-(iv), our analysis also predicts that WH-elements cannot co-occur with Snegs
because they both compete for the same Spec-FocP position (distinguish Snegs from NRQs,
which allow WH-elements):
Da quale treno non (ti)
scesa Maria?/?!
(From which train not you.ED be.3 sing.pres. got Maria)
“From which train didn’t Maria get off?”
Toward a comparative analysis of Snegs
Our analysis also suggests a comparative program of research, especially since Snegs
obligatorily involve a post-verbal subject position. The immediate prediction is that Snegs
should not exist in non-pro-drop languages. As for independent evidence on the existence of a
Foc° head licensed by the negative morpheme, interesting data come from Latin. In this
language, in nonne – used in rhetorical questions - the clitic head –ne shows up that is
arguably the head of a high CP-phrasal head combined with the negative morpheme non in
Snegs (see Ernout and Thomas 1953 for a detailed analysis of Latin -ne). Vidisti-ne Romam?
(Did not you see Rome?) vs. Nonne Romam vidisti? (S/he not saw-ne; “Didn’t s/he see
Rome?!”); and notably ne is also a subordinate conjunction in “negative” final clauses in Latin.
Belletti (2002), Aspect of the low IP area, in Rizzi (eds.) The structure of CP and IP, Oxford University Press,
Oxford: 16-51. Benincà - Poletto (2004), Topic, Focus and V2: Defining the CP sublayers, in Rizzi (eds.) The
structure of CP and IP, Oxford University Press, Oxford: 52-75. Boneh - Nash (2012), Core and non-core datives
in French, in Fernández & Exteparre (eds.) Variation in Datives, Oxford, Oxford University Press, pp. 22-49.
Cuervo (2003), Datives at Large, Ph.D. dissertation MIT. Ernout - Thomas (1953), Syntaxe Latine, Paris,
Éditions Klicksieck. Haegeman L. (2004). Topicalization, CLLD and the left periphery, in Maienborn, Frey &
Shaer (eds.) Proceedings of the Dislocated Elements Workshop, ZAS Papers in Linguistics (35): 157-192. Han
(2002), Interpreting interrogatives as rhetorical questions, Lingua, (112): 201–229. Horn (2010), The expression
of negation, Berlin-New York, De Gruyter Mouton. Moro (2003), Notes on vocative case: a case study in clause
structure, in Schroten, Sleeman & Verheugd, Romance languages and linguistic theory, John Benjamins,
Amsterdam and Philadelphia: 251– 265. Rizzi (1997), The fine structure of the left periphery, in Haegeman (eds.)
Elements of grammar: A handbook of generative syntax, Kluwer, Dordrecht: 281-337. Rowlett (1998), Sentential
Negation in French, Oxford University Press, Oxford. Zanuttini (1997), Negation and Clausal Structure, Oxford
University Press, Oxford. Zanuttini – Portner (2003), Exclamative clauses: At the syntax-semantics interface,
Language 79(3): 39–81. Zeijlstra (2004), Sentential Negation and Negative Concord, Ph.D. Dissertation,
University of Amsterdam, LOT Publications, Utrecht.
Fly UP