Mojmir Docekal, Ivona Kucerova docekal@phil.muni.cz, kucerov@mcmaster.ca
SuB 15
SaarbrUcken University
On the scalar representation of total and partial
adjectives: Evidence from Reduplication
The goal:
• Evidence from morphophonological processes in Czech for:
- a structural difference between two classes of antonym adjectives, namely, total and partial adjectives, for example, clean and dirty (Yoon, 1996; Rotstein and Winter, 2004)
- only total adjective have their standard value represented in the derivation
- the relation between partial and total adjectives defined with respect to the standard value of the total adjective represented as the lower bound of its partial counterpart (Rotstein and Winter, 2004)
• Consequences for semantic interpretation of adjectives:
- antonym adjectives must be at least sometimes represented by overlapping scales (contra Kennedy and McNally 2005; Kennedy 2007)
- they cannot be defined as being mirror images of each other with respect to one scale
The bigger question:
• What part of the meaning is obtained from the context (pragmatics) and what must be represented in the grammar (semantics)?
Outline:
1. the data
2. total versus partial adjectives and their semantics
3. the proposal
4. further evidence: context and adverbial modification by almost
5. conclusion
1
Mojmír Dočekal, Ivona Kučerová docekal@phil.muni.čz, kučerov@mčmaster.čá
1 Puzzle
SuB 15
Saarbručken University
• Czech has a productive system of a semantically driven morpho-phonological reduplication (Marantz, 1982; Inkelas and Zoll, 2005)
An example of semantically driven reduplication:
• reduplication of the imperfective verbal morpheme -va- (usually called habitual) encodes iterativity
Imperfective/generic
iterative
(1) a. praco-va-l
WOrk-imperf-pp.m.sg.
'he worked'
b. praco-va-va-l
WOrk-imperf-imperf-pp.m. sg.
'he used to work'
c. praco-va-va-va-l
work-imperf-imperf-imperf-pp.m.sg.
'he used to work' iterative (emphatic)
(2) a.    Petr každe rano     cistí-va-va-l okno.
Petr every morning clean-imperf-imperf-pp.m.sg. window 'Peter used to clean the window every morning.'    /habitual/generic b.  *Petrvcera      rano     cistí-va-va-l okno. Petr yesterday morning clean-imperf-imperf-pp.m.sg. window 'Peter used to clean the window yesterday in the morning.' *episodic
The empirical focus of this talk:
• reduplication in adjectives
• reduplication of a degree morpheme corresponding (roughly) to English
'very'
• the resulting meaning of the adjective may be paraphrased as 'very, very... (clean)', i.e., emphasizing the standard value of the adjective
• native speakers characterize the resulting interpretation as that of reaching the absolute degree of adjectiveness (for example, of cleanness)
2
Mojmír Dočekal, Ivona Kučerová docekal@phil.muni.čz, kučerov@mčmaster.čá
SuB 15
Saarbručken University
The crucial fact:
• only the so-called total adjectives (Yoon 1996, Rotstein and Winter 2004 and Winter 2006) may undergo this process:
(3) čistý 'clean' vs. špinavý 'dirty'
a. cisty — cistounky — cistoulinky — cistoulilinky...  /reduplication
b. špinavý — špinavoulinky — *špinavoulilinký... *reduplication
(4) zavřený 'closed' vs. otevřený 'open'
a. zavreny — zavšeiioulinky — zavšeiioulilinky / reduplication
b. otevšeny — otevšenoulinky — *otevrenoulilinky... *reduplication
(5) Some further examples (source: the Czech National Corpus):
čistý (clean)	cistoulilinky	špinavý (dirty)	*špinavoulilinký
zavřený (closed)	zavřeňoulilinký	otevřený (open)	* otevřeňoulilinký
zdravý (healthy)	zdravoulilinký	nemocný (ill)	*nemocňoulilinký
rovný (straight)	rovňoulilinký	zahnutý (bent)	*zahňutoulilinký
tenký (thin)	tenoulilinký	tlustý (thick)	*tlustoulilinký
jemný (slight)	jemňoulilinký	hrubý (rough)	*hruboulilinký
chabý (faint)	chaboulilinký	pevný (solid)	*pevňoulilinký
křehký (fragile)	křehoulilinký	nerozbitný (unbreakable)	???
• this restriction is puzzling because it does not apply to its semantically closest variant, i.e., adverbial modification by velmi 'very':
(6) No restriction on adverbial modification:
a. velmi cisty very clean
b. velmi špinavy very dirty
• similarly, the closest English paraphrase (the repetition of 'very') is compatible with both total and partial adjectives as well:
(7) No restriction for English adverbial modification:
a. very very very clean
b. very very very dirty
• crucially, the restriction cannot be explained in phonological or morphological terms because neither semantic class of the adjectives forms a phonological or morphological natural class
3
Mojmír Dočekal, Ivona Kučerová SuB 15
docekal@phil.muni.cz, kučerov@mčmaster.čá Saarbrücken University
2   The semantics of total and partial adjectives
• sčalar semantičs for adječtives: the positive form of an adječtive denotes a subinterval of the sčale SA
• the subinterval depends on a standard value dA in the sčale
• the sčale is ordered by a relation RA defined with respečt to the standard value
dA G Sa
• the standard value variable dA is čontext dependent (e.g., a big house x a big mouse)
• the denotation of the positive form of an adječtive čan be formalized as in (8) (after Rotstein and Winter 2004, ex. (18))
(8)     [A] =f {x g Sa : Ra(da,x)}
• (Note: the denotation of an adječtive in (8) must be mapped on the set of entities for the degree of A-ness to be inčluded in [A] otherwise the intersečtion interpretation of the AP within an NP yields a type-mismatčh. For the sake of simpličity, we set this issue aside)
Relevant types of adjectives:
• partial: indičate some amount of the relevant property (moisture, dirt, sičk-ness etč.)
• total: indičate no amount of this property (e.g., a dirty obječt has some degree of dirtiness, but it is not nečessarily free of čleanliness; in čontrast, a člean obječt is free of dirtiness)
• (relative: no member of an antonym pair has it standard value set independently of the čontext; Kennedy and MčNally 2005; Kennedy 2007)
... and their denotation:
• we follow Rotstein and Winter (2004) in formalizing total v. partial adječtives as overlapping sčales:
4
Mojmir Docekal, Ivona Kucerova docekal@phil.muni.cz, kucerov@mcmaster.ca
SuB 15
SaarbrUcken University
P   =d d
min      t p
Q-\-
T
ma
p
max
-►o
-o
Figure 1: Total and Partial adjective scales
• SP: a partial scale, without a fixed standard value (context-dependent)
• ST: a total scale; its standard value is fixed as the lower bound of its partial counterpart (a healthy man is a man that lacks any illness)
• ST and SP are ordered inversely
• Sp may partially overlap with the St', some amount of the relevant partial property doesn't exclude some amount of the complementary total property
• e.g., if a coat is dirty it can mean that it is stained on sleeves but the rest of the coat is clean
3 Proposal
• we argue that total and partial predicates must be semantically represented by a scale and a standard value
• we define the relation between partial and total adjectives with respect to the standard value of the total adjective represented as the lower bound of its partial counterpart (following Rotstein and Winter 2004 contra Kennedy and McNally 2005)
• crucially, the scales are partially overlapping and the impression of their antonymous interpretation (not clean lh dirty) comes from an interaction of their interval boundaries and the standard values
The denotation of partial adjectives:
• crucially, the standard value of partial adjectives is determined contextually
• dp E Sp; Sp ... closure of the partial scale
5
Mojmir Docekal, Ivona Kucerova docekal@phil.muni.cz, kucerov@mcmaster.ca
SuB 15
SaarbrUcken University
• Consequence: the standard value of partial adjectives has no structural representation
The denotation of total adjectives:
• the standard value of the total member of the adjectival pair is defined as the lower bound of its partial counterpart:
• dT = Pmin £ St', ST ■ ■ ■ closure of the total scale The denotation of reduplication:
• the intuition is that reduplication corresponds to semantic modification: adjectives with reduplicated morphemes denote some interval close to the standard value
• since this is semantic modification, it depends on the type of scale in the denotation of the adjective with which it combines
• more formally, the denotation of the reduplication corresponds to a limit function where the limit is defined as the standard value of the total adjective:
• the resulting denotation corresponds to approaching the standard value of the adjective
Consequences:
• since reduplication is a morphophonological process, it may apply only to the material present in the derivation
• in other words, reduplication is possible only if the standard value is structurally represented
• consequently, reduplication applies only to total adjectives since only total adjectives have their standard value structurally represented as some value in the closure of the total scale
• in contrast, the standard value of a partial adjective is determined contextually and may fall anywhere within the interval
• thus, there is no structural representation of the standard value
• consequently, there is no material that could be used for reduplication
6
Mojmir Docekal, Ivona Kucerova docekal@phil.muni.cz, kucerov@mcmaster.ca
SuB 15
SaarbrUcken University
4   Further evidence: context and the standard value
• English adverbial modification by almost shows similar properties to the Czech adjectival reduplication
• almost usually combines with total adjectives but not with partial adjectives:
(9) (from Rotstein and Winter 2004, ex. (9))
a. The work is almost complete/*incomplete.
b. The patient is almost dead/*alive.
c. The explanation is almost clear/*unclear.
The semantics of almost in a nutshell:
• almost cross-categorically denotes negation of the denotation of the constituent it modifies:
(10) a.    John almost passed the exam ^ John didn't pass the exam
b. Almost every student passed the exam ^ Not every student passed the exam
c. John is almost healthy ^ John isn't healthy
• Rotstein and Winter (2004): the interval associated with the phrase almost A denotes degrees that are adjacent to the standard value of A and are in the opposite direction from the ordering ofthe scale associated with the adjective
A
• if the standard value of a partial adjective equals the standard value of a total adjective (just on the opposite scale), then the adjectives are complementary
• consequently, the partial adjective cannot be modified by almost because there is no complement interval between dP and 0
The predictions:
• since the standard value of a partial adjective is not fixed in the structure but instead it is contextually dependent, we should be able to modify the standard value in a way so it would no longer be complementary to the total adjective
• more precisely, we should be able to modify the standard value so there would be an interval between dP and 0 that could feed into the denotation of almost
• consequently, if such modification is possible, a partial adjective should become modifiable by almost
7
Mojmir Docekal, Ivona Kucerova docekal@phil.muni.cz, kucerov@mcmaster.ca
SuB 15
SaarbrUcken University
this prediction is borne out:
(11)
(from Kennedy 2007)
a. We need a rod that is bent in an angle of 90 degrees. Let's pick up that rod over there and bend it a little: it should be easy, as it's almost bent already.
b. We consider a glass dirty and wash it as soon as there are five spots on it. This glass is now almost dirty - it has four spots on it.
(12) a.    We need a TALL basketball player - one whose height is at least 1.95
meters. But we cannot take John, who is 1.90 meters - he's just almost tall.
b.    The publisher considers a book long if it's 300 pages or more. This book is almost long - it's 298 pages.
• the crucial question is whether Czech adjectival reduplication of partial adjectives could be fixed by the context as well
• the prediction is that if the standard value of a partial adjective gets contex-tually fixed, reduplication should still fail to apply, in contrast to the English almost-modification facts
• the reason is that the contextual fixation happens only later in the derivation (after Spell-out), thus, it cannot affect the PF side of the derivation
• this prediction is indeed borne out:
(13) [modeled after Rotstein and Winter (2004)]:
a. This glass is certainly not clean, since it has several big spots on it and I am not willing to drink from it even if you insist. The glass is simply. . .
*spinavoulilinka 'very very dirty'
b. This glass is certainly not dirty, since it has absolutely no dirty spots on it. The glass is simply. . .
/cistoulilinka 'very vert clean'
8
Mojmír Dočekal, Ivona Kučerová docekal@phil.muni.čz, kučerov@mčmaster.čá
5 Conclusion
SuB 15
Saarbručken University
• we have argued the Czech data show a structural difference between partial and total adjectives (Yoon, 1996; Rotstein and Winter, 2004)
• only total adjective have their standard value represented in the derivation, the standard value of partial adjectives is derived from the context
• consequently, antonym adjectives cannot be represented by adjacent scales but instead they must be allowed to partially overlap (in agreement with Rotstein and Winter 2004 and contra Kennedy and McNally 2005; Kennedy
2007)
• crucially, the semantics of antonym adjectives must be formalized as a combination of grammatically encoded (semantics) and contextually-determined (pragmatics) meanings
• the different representations are empirically testable
Inkelas, Sharon, and Cheryl Zoll. 2005. Reduplication: Doubling in morphology. New York: Cambridge Univ Press.
Kennedy, Christopher. 2007. Vagueness and grammar: the semantics of relative and absolute gradable adjectives. Linguistics and Philosophy 30:1-45.
Kennedy, Christopher, and Louise McNally. 2005. Scale structure, degree modification, and the semantics of gradable predicates. Language 81:345-381. Marantz, Alec. 1982. Re reduplication. Linguistic inquiry 13:435-482.
Rotstein, Carmen, and Yoad Winter. 2004. Total adjectives vs. partial adjectives: Scale structure and higher-order modifiers. Natural Language Semantics 12:259288.
Winter, Yoad. 2006. Closure and telicity across categories. In Proceedings of SALT XVI, ed. M. Gibson and J. Howell. Ithaca, NY: CLC Publications, Cornell University.
Yoon, Youngeun. 1996. Total and partial predicates and the weak and strong interpretations. Natural Language Semantics 4:217-236.
References
9