Great pizzas, ghost negations:
The emergence and persistence of mixed expressives
Andrea Beltrama and Jackson Lee
University of Chicago
{andremormora,jsllee}@uchicago.edu
Sinn und Bedeutung 19, University of Göttingen
September, 16th 2014
1
The Claims
• Two novel cases of mixed expressives (McCready (2010), Gutzmann (2012))
1. Standard/Emilian Italian gran (adjective → quantifier)
2. Cantonese gwai2 (quantifier → sentential negator)
• Claims:
1. Specific level: In both cases expressive meaning survives through grammaticalization
processes and is preserved on top of newly created truth-conditional meaning.
2. Broad level: Expressivity as a diachronically stable type of (emphatic) meaning.
• Implications:
1. Diachronic point: Emphasis need not disappear throughout trajectories of semantic
change (compare with emphasis in Jespersen’s cycle).
2. Synchronic point: Empirical support to theories assigning to expressive and truthconditional meaning independent semantic representations (Potts 2005, 2007, McCready 2010, Gutzmann 2012).
1.1
Outline of the talk
• Background (Section 2)
• Italian Gran (Section 3)
• Cantonese gwai2 (Section 4)
• Discussion and Conclusions (Section 5)
1
2
Background
2.1
Mixed expressivity in synchrony
• Expressive content: semantic contribution that conveys non truth-conditional information
about the speaker’s attitude/emotive status. Potts (2005, 2007); Potts and Schwarz (2008)
(1)
That bastard Burns is a zombie.
Expressive content: Speaker doesn’t like Burns.
Truth-conditional content: None
(2)
NegAttitude(Burns)EXP
•
BurnsT C
Burns
bastard
NegAttitudeEXP
• Tests for expressivity (Potts 2003, 2005, 2007)
1. Independence - Expressive content can’t be challeged independently from propositional content.
(3)
A: That bastard Burns is a zombie.
# B: No! I like him!
EXPRESSIVE CONTENT CHALLENGED
XB’: No! He’s not a zombie.
PROPOSITIONAL CONTENT CHALLENGED
2. Non displaceability - Expressive content is here and now (Potts 2007)
(4)
That bastard Burns was late for work yesterday.
XI NTENDED: He was on time today.
# I NTENDED: He’s no bastard today, because today he was on time.
3. Non embeddability - Expressive content, even when syntactically embedded, is not
semantically embedded
(5)
Sue believes that that bastard Burns should be fired. #I think he’s a good guy.
4. Ban in predicative position - All predicates that appear in copular position must necessarily fail to be expressive (Potts (2007), Zimmermann (2007)).
2
• Mixed Expressive content: conveys both expressive and truth-conditional meaning (Sawada
(2009), McCready (2010), Gutzmann (2011)).
(6)
Lessing is a Boche.
(Williamson, 2009) discussed in Gutzmann (2011)
Expressive content: Speaker has a negative attitude towards Germans.
Truth-conditional content: Lessing is German
• The two parts do not interact in meaning composition
(7)
NegAttitude(Germans)EXP
•
German(Lessing)T C
Lessing
3
Boche
NegAttitudeEXP
GermanT C
2.2
Mixed expressivity in diachrony
• Intriguing question: how do expressivity and truth-conditional meaning end up co-existing
in a lexical item?
• Two models touch on this issue:
1. Subjectification models:
– Subjectification: Meaning tends to become increasingly based in the speaker’s
beliefs/attitudes (Traugott (1982))
– Mixed expressivity as an intermediate stage along these trajectories (Gutzmann,
2013)
(8)
a. Boor(peasant)T C → Boor(peasant)T C + Neg. att.EXP → Neg. att.EXP
(from Traugott (2003))
b. AT C → AT C + BEXP → BEXP
2. Jespersen’s cycle models: Focus on emphatic meaning.
– Noun: pas → Complex negation: ne...pas → plain negation pas
– Emphasis stems from alternative-based scalar reasoning (Krifka (1995); Eckardt
(2006)).
– Co-existence of truth-conditional and emphatic component is transitory.
Caveat: Emphasis 6= Expressive meaning proper. (See Section 4 - Expressive
meaning grounded in the item’s conventional meaning, emphasis based on scalar
reasoning.)
• In sum: Coexistence of emphatic and descriptive semantic content is transitory
Our proposal:
Expressivity/emphasis can survive semantic change and be preserved atop of newly created truthconditional meanings.
4
3
Italian gran
3.1
Adjectival gran
• Gran derived via truncation from grande (= ‘big’)
• In Standard Italian, the adjective gran is a Mixed Expressive.
(9)
Marco mangio’ una gran pizza lo scorso mese.
Marco ate
a gran pizza the last month
Truth-Conditional:Marco ate an outstanding pizza last month.
Expressive:The speaker is excited about the pizza.
• testing for (mixed) expressivity (Potts (2005, 2007) Gutzmann (2011), McCready and Kauffman (2013)
1. Independence
a. Marco mangio’ una gran pizza lo scorso mese.
Marco ate
a gran pizza the last month
b. No!{#Mangio’ un pizza eccellente, ma non provo nulla/XMangio’ una
pizza normale}.
No!{#He ate an outstanding pizza, but today I don’t care/XHe ate an average pizza}
2. Non displaceability
(10)
(11)
Marco mangio’ una gran pizza lo scorso mese.
Marco ate
a gran pizza the last month
a. X I’m being emotional now (despite past tense)
b. Not: I was being emotional then.
3. Non embeddability
(12)
a. X Marco sostiene di aver mangiato una gran pizza, ma per me era mediocre
Marco says that he ate a gran pizza, but I think it was mediocre.
b. # Marco sostiene di aver mangiato una gran pizza, ma a me non interessa.
Marco says that he ate a gran pizza, but I am not excited.
4. Ban in predicative position:
(13)
a. XMarco mangio’ una gran pizza.
b. *Marco mangio’ una pizza gran.
c. *La pizze mangiata da Marco era gran.
The pizza eaten by Marco was gran.
Note: in Romance languages pre-nominal adjectives are more likely to carry subjectivelyconnotated meanings. (Nespor, 1991), (Demonte, 1999)
5
3.2
Quantificational gran
• In a colloquial variety spoken in the Emilia region (≈ the city of Bologna), gran has developed a quantificational usage with meaning similar to a lot, many.
(14)
Marco mangio’ delle gran pizze lo scorso mese’.
Marco ate
some gran pizzas last month.
Truth-Conditional:Marco ate many pizzas last month.
Expressive:The speaker is excited about such quantity of pizzas.
• Adjectival and Quantificational gran have independent meanings.
(15)
Marco mangio’ delle gran pizze lo scorso mese, X ma non erano nulla di che.
‘Marco ate some gran pizzas last month, X but they were nothing special.’
• Syntactic restriction: Quantificational gran only found with quantificational determiners
dei/delle (≈ some)
• Testing for (mixed) expressivity
1. Independence
a. Marco mangio’ delle gran pizze lo scorso mese.
Marco ate
some gran pizzas the last month
b. No!{#Mangio’ molte pizze, ma non provo nulla/XMangio’ poche pizze }
No!{#He ate many pizzas, but I don’t feel anything/XHe ate few pizzas}
2. Non displaceability
(16)
(17)
Marco mangio’ delle gran pizze lo scorso mese.
Marco ate
many pizzas the last month
a. X I’m being emotional now (despite past tense)
b. Not: I was being emotional then.
3. Non embeddability
(18)
a. X Marco sostiene di aver mangiato delle gran pizze, ma per me erano
poche
Marco says that he ate some gran pizzas, but I think they were few.
b. # Marco sostiene di aver mangiato delle gran pizze, ma a me non interessa.
Marco says that he ate some gran pizzas, but I don’t feel anything about it.
4. Ban in predicative position
(19)
a. XMarco mangio’ delle gran pizze lo scorso mese.
Marco ate many pizzas the last month
b. *Marco mangio’ delle pizze gran lo scorso mese.
c. *Le pizze sono gran.
The pizzas are many.
6
3.3
From mixed-predication to mixed-quantification
• No systematic corpus data available, but plausible evidence that quantificational and adjectival usage are diachronically related
– Quantificational usage only common in a region
– Adjectival usage widespread in all varieties of the language
– Speakers of other regions have a hard time getting the quantificational reading
• Truth-conditional meaning:
OUTSTANDING (x) → |x| > n
CHANGED .
From individuals to cardinalities.
• Expressive meaning: PRESERVED. The excitement component survives.
excitedspeaker (x) → excitedspeaker (|x|)
(20)
Diachronic trajectory of gran
‘big’:
‘outstanding’ + expressive:
⇒
Jgrande he, tiK
Jgran he, ti•he, uiK
⇒
‘many’ + expressive:
Jgran het, ti•het, uiK
For the expressive type u, see Gutzmann (2011, 2012), cf. Potts (2005, 2007).
7
4
Cantonese gwai2
(21)
Literal meaning: Cantonese gwai2 = ‘ghost’
(22)
Non-literal and productive:
Gwai2 as an infix – a pure expressive (not discussed in this paper)
Gwai2 not as an infix – a mixed expressive (the focus of this paper)
(Yip and Matthews (2001: 157-160), Lee and Chin (2007), Yu (2007: 134-135), and
Matthews and Yip (2011: 52-54, 184))
4.1
Gwai2 = ‘nobody’ + expressive
(23) Gwai2 as a mixed expressive #1 —
Truth-conditional meaning: negative quantifier (= nobody)
Expressive: The speaker is in a ‘heigthened emotional state’ (Potts (2007); = goddamn)
(24) Gwai2 ho2ji5 loeng5 lin4 duk6 jyun4 bok3si6.
GHOST can
two year study finish PhD
‘No goddamn person can get a PhD in two years.’
(25) Gwai2 sik1
GHOST know
‘Nobody knows shit.’ (cf. “vulgar minimizers” like shit, Postal (2002))
(26)
Diagnostic tests for mixed expressivity
a. Independence: Third-party objection to (25) can deny the truth-conditional content
but not the expressive meaning.
i. X M4hai6 – kei4sat6 ngo5 sik1
No – actually I know.
ii. # Lei5 m4 lau1
You’re not mad.
b. Non-displaceability:
i. Gwai2 wui5 ting1jat6 heoi3
GHOST will tomorrow go
‘No goddamn person will go tomorrow.’
3 I’m being emotional now (despite future reference).
7 I will be emotional tomorrow.
c. Interaction with Plain Negation: The default negation m4 ‘not’ interacts with and
flips the polarity of gwai2 (with expressivity taking the widest scope, see (Potts and
Kawahara, 2004))
i. Gwai2 m4 sik1
GHOST not know
‘Every goddamn person knows.’
8
4.2
Gwai2 = ‘not’ + expressive
(27)
Synchronically, Gwai2 as a mixed expressive #2 —
Truth-conditional meaning: sentential negator (= not)
Expressive: goddamn
(28)
a. Keoi5 gwai2 ho2ji5 loeng5 lin4 duk6 jyun4 bok3si6.
s/he GHOST can two year study finish PhD
‘He can’t goddamn get a PhD in two years.’
b. Compare (without gwai2):
Keoi5 ho2ji5 loeng5 lin4 duk6 jyun4 bok3si6.
s/he can two year study finish PhD
‘He can get a PhD in two years.’
(29)
Keoi5 gwai2 sik1
s/he GHOST know
‘He doesn’t goddamn know.’
(30)
Diagnostic tests for mixed expressivity
a. Independence: Third-party objection to (29) can deny the truth-conditional content
but not the expressive meaning.
i. X M4hai6 – kei4sat6 keoi5 sik1
No – actually he knows.
ii. # Lei5 m4 lau1
You’re not mad.
b. Non-displaceability:
i. Keoi5 gwai2 wui5 ting1jat6 heoi3
s/he GHOST will tomorrow go
‘He won’t goddamn go tomorrow.’
3 I’m being emotional now (despite future reference).
7 I will be emotional tomorrow.
c. Interaction with plain negation: The default negation m4 ‘not’ interacts with and
flips the polarity of gwai2 (with expressivity taking the widest scope).
i. Keoi5 gwai2 m4 sik1
s/he GHOST not know
‘He goddamn knows.’
4.3
The emergence and persistence of mixed expressivity of gwai2
2 steps:
• Emergence of gwai2 as a mixed expressive —:
Gwai2 ‘ghost’ ⇒ ‘nobody’ + Expressivity:
Implicature: “Ghosts” ⇒ “No human beings”. (World knowledge: ghosts do not exist).
9
(31)
Gwai2 sik1
GHOST know
Literal: ‘Ghosts know.’
⇒ Truth-conditional meaning: ‘Nobody knows.’ + Expressive: goddamn
• Persistence of gwai2 as a mixed expressive —
‘nobody’+Expressive ⇒ ‘not’+Expressive
The dynamics:
1. Utterances like (24) and (32) are re-analyzed as having a null subject (Cantonese is a
topic-comment language allowing null subjects, cf. pro drop in Romance languages).
(32) Gwai2 sik1
GHOST know
‘Nobody knows shit.’ (cf. “vulgar minimizers” like shit, Postal (2002))
2. An overt subject prepends them as in (33),
(33)
Keoi5 gwai2 sik1
s/he GHOST know
‘He doesn’t goddamn know.’
3. Gwai2 is interpreted as a sentential negator.
• Evidence for this diachronic story —
Corpus data of early/mid 20th century Cantonese: A Linguistic Corpus of Mid-20th Century
Hong Kong Cantonese (Chin, 2013):
3 Wide attestation of gwai2 as ‘nobody’+Expressive
7 No attestation of gwai2 as ‘not’+Expressive
⇒ gwai2 as ‘not’+Expressive did not emerge until quite recently.
(34)
Diachronic trajectory of gwai2:
‘ghost’:
‘nobody’ + expressive:
⇒
Jgwai2 he, tiK
Jgwai2 het, ti•ht, uiK
⇒
‘not’ + expressive:
Jgwai2 ht, ti•ht, uiK
For the expressive type u, see Gutzmann (2011, 2012), cf. Potts (2005, 2007).
4.4
‘Ghost’ versus ‘steps’
• From ghosts to negation:
– world knowledge that ghosts don’t actually exists
⇒ source of grammatical negation with gwai2 ‘ghost’
10
– By the Cooperative Principle and the Maxim of Quality,
when someone says “ghosts know” in Cantonese, it generates the implicature that...
⇒ Only non existing things know.
⇒ Every existing thing doesn’t know.
⇒ Nobody knows (the actual interpretation of “ghosts know” in Cantonese).
• From pas to negation:
– Pas’ cycle (Eckardt 2006):
1. Literal meaning (e.g. pas = “step” )
2. Emphatic focus in negative polarity context (“puzzling usages”, neither literal
meaning nor part of a negation. Example from Eckardt 2006, p. 139)
(35)
N’i
ot rei, prince ne baron, Qui pas m’i
pust contrester.
not-there has king prince nor baron who pas me-there could contradict.
There is no king, prince or baron [[who could even in the least contradict me there]]P .
Scalar reasoning: The focused proposition containing pas, (i.e. the “smallest
amount” of contradiction) leads to the exclusion of all alternative propositions
which contain a “larger amount” of contradiction, reinforcing the negation.
⇒ Emphasis is the result of focus of the NPI in the scope of a higher negation (or
in other downward entailing contexts for other nouns, see Eckardt 2006, p. 169)
3. The loss of emphasis: As soon as scalar reasoning ceases to be recognized, emphasis vanishes and is replaced by a pure functional meaning (via a brief stylistic
emphasis stage).
Differences:
– Contrary to French pas, gwai2 can be used in positive polarity utterances
– French negators like pas came into being by quantity-based reasoning and the subsequent loss thereof. But gwai2-negation is more of the quality type.
– For pas, emphasis and other levels of semantic content (truth-conditions, scalar reasoning, etc.) are strictly connected; For Gwai2 they are more independent.
(In addition: gwai2 is an expressive in the more general grammar of Cantonese, outside
the domain of negation; full details in Lee (2014))
• gwai2 : An incomplete Jespersen’s cycle:
XLiteral NP meaning → Functional meaning
7 Loss of emphasis in the process
11
5
Conclusions and implications
• Two new cases of mixed expressives: Italian gran and Cantonese gwai2
• Both mixed expressives have recently undergone a shift in meaning, while maintaining expressivity.
• Implications:
1. Mixed expressivity need not represent just a transitional stage of semantic change, but
can be a stable category, capable of persisting through semantic shifts.
2. Evidence that the expressive and the truth-conditional meaning diachronically proceed
in a parallel fashion, interacting very little in the process.
⇒ Diachronic independence mirrors synchronic independence.
• Further work: Two types of emphatic content?
1. Expressive meaning: Not linked to scalar reasoning. More stable.
2. Focus-based emphasis: Based on scalar reasoning. Transitory.
References
Chin, Andy Chi-On. 2013. New resources for Cantonese language studies: A Linguistic Corpus of
Mid-20th Century Hong Kong Cantonese. Newsletter of Chinese Language 92(1): 7–16.
Demonte, Violeta. 1999. A minimal account of spanish adjective position and interpretation. In
Grammatical analyses in basque and spanish linguistics. Jo.
Eckardt, Regine. 2006. Meaning change in grammaticalization. New York: Oxford University
Press.
Gutzmann, Daniel. 2011. Expressive modifiers & mixed expressives. In Empirical issues in syntax
and semantics, vol. 8, 123–141.
Gutzmann, Daniel. 2012. Use-conditional meaning. studies in multi-dimensional semantics. Ph.D.
thesis, University of Frankfurt, Frankfurt/Main.
Gutzmann, Daniel. 2013. Pragmaticalization and multidimensional semantics. Paper presented at
the Workshop on Systematic Semantic Change, University of Texas, Austin, 5-6 April.
Krifka, Manfred. 1995. The semantics and pragmatics of polarity items. Linguistic Analysis 25(34): 209–257.
Lee, Jackson L. 2014. Demystifying the Cantonese ghost. University of Chicago, ms.
Lee, Peppina Po-Lun and Andy Chi-On Chin. 2007. A preliminary study on Cantonese gwai
‘ghost’. In Joanna Ut-Seong Sio and Sze-Wing Tang (eds.), Studies in Cantonese Linguistics 2.
Hong Kong: Linguistic Society of Hong Kong.
Matthews, Stephen and Virginia Yip. 2011. Cantonese: A Comprehensive Grammar. London:
Routledge, 2nd ed.
McCready, Eric. 2010. Varieties of conventional implicature. Semantics and Pragmatics 3: 1–58.
McCready, Eric and Magdalena Kauffman. 2013. Maximum intensity. Paper presented at the
Semantics Workshop, Keio University, 29 November.
12
Nespor, Marina. 1991. Il sintagma aggettivale. In Grande grammatica di consultazione. Il Mulino.
Postal, Paul. 2002. The structure of one type of American English vulgar minimizers. In Skeptical
Linguistic Essays. New York: Oxford University Press.
Potts, Christopher. 2005. The logic of conventional implicature. Oxford: Oxford University Press.
Potts, Christopher. 2007. The expressive dimension. Theretical Linguistics 33: 165–198.
Potts, Christopher and Shigeto Kawahara. 2004. Japanese honorifics as emotive definite descriptions. In Kazuha Watanabe and Robert B. Young (eds.), Proceedings of Semantics and Linguistic
Theory 14, 235–254. Ithaca, NY: CLC Publications.
Potts, Christopher and Florian Schwarz. 2008. Exclamatives and heightened emotion: Extracting
pragmatic generalizations from large corpora. Unpublished maniscript.
Sawada, Osamu. 2009. Pragmatic aspects of scalar modifiers. Ph.D. thesis, University of Chicago.
Traugott, Elizabeth. 1982. From propositional to textual and expressive meanings; some semanticpragmatic aspects of grammaticalization. In Winfred P. Lehmann and Yakov Malkiel (eds.),
Perspectives on historical linguistics, 245–271. Amsterdam: Benjamins.
Traugott, Elizabeth. 2003. Constructions in grammaticalization. In Brian D. Joseph and Richard D.
Janda (eds.), A handbook of historical linguistics, 624–647. Oxford: Blackwell.
Williamson, Timothy. 2009. Reference, inference and the semantics of pejoratives. In Joseph
Almog and Paolo Leonardi (eds.), The philosophy of david kaplan, 137158. Oxford: OUP.
Yip, Virginia and Stephen Matthews. 2001. Intermediate Cantonese: A Grammar and Workbook.
London: Routledge.
Yu, Alan C. L. 2007. A Natural History of Infixation. Oxford: Oxford University Press.
Zimmermann, Malte. 2007. I like that damn paper -three comments on christopher potts’ the
expressive dimension. Theoretical Linguistics 33(2): 247–254.
13
© Copyright 2024 Paperzz
