Absolute and Relative Adjectives and their Comparison Classes

Research (PDF Available) · April 2015with 763 Reads
DOI: 10.13140/RG.2.1.3312.0800
Cite this publication
Abstract
This paper investigates core semantic properties that distinguish between different types of gradable adjectives and the effect of context on their interpretation. We contend that all gradable adjectives are interpreted relative to a comparison class (van Rooij to appear), and that it is the nature of the comparison class that constitutes the main semantic difference between the different subclasses of such adjectives: some select a class comprised of counterparts of the individual of which the adjective is predicated, while others - an extensional-category of this individual. The role of the context is to determine the elements that comprise the comparison class. It is proposed, following Kennedy (2007), that the standard of membership is selected according to a principle of economy whereby an interpretation relative to a maximum or a minimum endpoint within a comparison class takes precedence over one relative to an arbitrary point. This proposal captures so-called “standard shift” effects, that is, the influence of context on the interpretation of gradable adjectives from all subclasses, in their positive form and when modified by degree adverbials. Additionally, this proposal captures cases of apparent lack of context sensitivity (e.g. intuitive inference patterns, unacceptability of for-phrases, etc.) Finally, we show that the type of comparison class is aligned with the well known distinction between stage-level and individual-level predicates.
Absolute and Relative Adjectives
and their Comparison Classes
Galit W. Sassoon
ILLC, University of Amsterdam
Assaf Toledo
Utrecht University
Abstract
This paper investigates core semantic properties that distinguish between
different types of gradable adjectives and the effect of context on their interpretation.
We contend that all gradable adjectives are interpreted relative to a comparison
class (van Rooij to appear), and that it is the nature of the comparison class that
constitutes the main semantic difference between the different subclasses of such
adjectives: some select a class comprised of counterparts of the individual of which
the adjective is predicated, while others - an extensional-category of this individual.
The role of the context is to determine the elements that comprise the comparison
class. It is proposed, following Kennedy (2007), that the standard of membership is
selected according to a principle of economy whereby an interpretation relative to a
maximum or a minimum endpoint within a comparison class takes precedence over
one relative to an arbitrary point. This proposal captures so-called “standard shift”
effects, that is, the influence of context on the interpretation of gradable adjectives
from all subclasses, in their positive form and when modified by degree adverbials.
Additionally, this proposal captures cases of apparent lack of context sensitivity
(e.g. intuitive inference patterns, unacceptability of for-phrases, etc.) Finally, we
show that the type of comparison class is aligned with the well known distinction
between stage-level and individual-level predicates.
Keywords:
Comparison class, Counterparts, Gradability, Context sensitivity, Absolute
adjectives, Relative adjectives, Scale structure
The work of Assaf Toledo was supported by a VICI grant number 277-80-002 by the Netherlands
Organization for Scientific Research (NWO). The work of Galit W. Sassoon was carried out in the
project ’On vagueness- and how to be precise enough’, founded by the Netherlands Organization for
Scientific Research (NWO 360-20-201). The theoretical proposal in this paper develops and extends
ideas that were originally presented in Toledo & Sassoon (2011). We thank Chris Kennedy, Louise
McNally, Angelika Kratzer, Robert Van Rooij, Jesse Snedeker, Fred Landman, Susan Rothstein,
Edit Doron, Malka Rappaport Hovav, Avigail Tsirkin-Sadan, Micha Y. Breakstone, Nina Luskin,
Danny Fox, Frank Veltman and Rick Nouwen, as well as the audiences at SALT 21, at the Hebrew
University departmental seminar , at the semantics group at Utrecht University, at Nijmegen semantic
and pragmatic colloquium, and at the Gottingen English Department seminar.
©2011 Sassoon and Toledo
Sassoon and Toledo
1 Introduction
This paper investigates the lexical semantics of gradable adjectives, general princi-
ples of grammar that contribute to their interpretation and the role of context in their
decoding.
In the past several decades, the semantics of gradable adjectives has received a
great deal of attention in the literature representing a number of linguistic traditions.
Various typologies have been proposed to classify gradable adjectives, including
Bierwisch’s (1989) dimensional/evaluative distinction, the partial/total distinction
by Yoon (1996) and Rotstein & Winter (2004), and the relative/absolute distinction
by Unger (1975), Kennedy & McNally (2005) and Kennedy (2007).
A central question addressed in this paper is whether the standard of membership
employed in the interpretation of gradable adjectives can be determined without
reference to context. In this connection, the next two sections review recent accounts
anchored in scale structure theory, which pertain to the interpretation of the positive
form of adjectives, as in (1).
(1) a. John is tall
b. This shirt is dirty
c. This knife is clean
d. This glass is full
1.1 Absolute Adjectives: Arguments against context sensitivity
Kennedy & McNally (2005) and Kennedy (2007) argue that the scales of gradable
adjectives come in four different forms, as in (2):
(2) Typology of Scale Structure
a. Open (e.g. tall,short)h h
b. Lower closed (e.g. dirty,wet)x h
c. Upper closed (e.g. clean,dry)h x
d. Totally closed (e.g. full,empty)x x
Evidence for this typology comes from the distribution of degree modifiers such as
slightly and perfectly, which, according to this theory, pick out the minimum and the
maximum degree on the scale, respectively:
(3) The Distribution of Degree Modifiers
a. {#slightly, #perfectly} tall
b. {slightly, #perfectly} dirty
2
Absolute and Relative Adjectives and their Comparison Classes
c. {#slightly, perfectly} clean
d. {slightly, perfectly} full
In discussing context sensitivity of adjectival membership standards in relation
to, e.g., the positive forms in (1), Kennedy & McNally (2005) posit the following
generalization:
(4)
An adjective Ais interpreted relative to a context-dependent standard if and
only if A’s scale is linguistically open. If Ais associated with a closed scale,
a (non-contextual) endpoint standard is employed in As interpretation.
To support this proposal, Kennedy (2007) stipulates a general grammatical principle
of economy, quoted in (5):
(5)
Economy Principle: Maximize the contribution of the conventional meanings
of the elements of a sentence to the computation of its truth conditions.
(Kennedy 2007: 35)
The above principle dictates that, in selecting a standard of membership, an adjec-
tive’s scale structure – which is part of its conventional meaning – takes precedence
over contextual properties. It follows that a context-dependent standard is resorted
to only in the event that the lexically encoded (’fixed’) scale of an adjective lacks
an endpoint. Thus, for each adjective in (1), a different standard of membership
surfaces:
(6) Standards of Membership
a. tall: open scale a context-dependent midpoint on the scale
b. dirty: lower closed scale the scale’s minimum endpoint
c. clean: upper closed scale the scale’s maximum endpoint
d. full: totally closed scale the scale’s minimum or maximum endpoint
Thus, according to Kennedy & McNally (2005) and Kennedy (2007), the typology
of relative/absolute adjectives pivots on the type of standard employed in their
interpretation. The standard of relative adjectives ( e.g. tall/short,expensive/cheap,
etc.) lies at some midpoint on the scale and is context-dependent. The standard
of absolute adjectives (e.g. dirty/clean,full/empty, etc.) is located at the scale’s
absolute minimum or maximum point, in which case the context plays no role in
interpretation.
This analysis seems to be supported by the distribution of for-phrases, which
restrict the contextual domain that determines the standard (Siegel 1979):
(7) a. John is tall for a ten-year old
b. #This shirt is dirty for a T-Shirt
3
Sassoon and Toledo
c. #This knife is clean for a kitchen knife
d. #This glass is full for a wine glass
According to Kennedy (2007: 37), "Interpretive Economy dictates that the absolute
truth conditions are the ones that should surface" – which explains the infelicity
of (7b), (7c) and (7d), where an unwarranted move is made to shift the absolute
standard selected by dirty,clean and full.
Another piece of evidence in favor of this analysis comes from inference patterns
such as those illustrated in (8):
(8) a.
Relative (Contextual-std) Adj: X is taller than Y
;
X is tall / Y is not tall
b. Absolute (Minimum-std) Adj: X is dirtier than Y X is dirty
c. Absolute (Maximum-std) Adj: X is emptier than Y Y is not empty
The above disparity in inference patterns between relative adjectives, such as
tall, and absolute adjectives, such as dirty and clean, can be attributed to the different
kind of standards evoked in their interpretation. A relative adjective like tall selects
a contextual midpoint standard, and therefore a comparative such as "X is taller
than Y" entails that the height of X exceeds the height of Y, but does not entail that
either of them is above or below the selected contextual standard. By contrast, the
comparative of a minimum-standard adjective like dirty entails that one individual
necessarily exceeds the minimum standard and therefore counts as dirty. Similarly,
the comparative of a maximum-standard adjective like clean entails that one in-
dividual necessarily fails to exceed the maximum standard and therefore counts
as not-clean. This account of the given inference patterns rests on the assumption
that the standard of absolute adjectives is fixed semantically to be an absolute-scale
endpoint. It is not affected by contextual considerations.
Additional evidence consistent with the typology proposed by Kennedy & Mc-
Nally (2005) and Kennedy (2007) comes from experimental work by Syrett, Bradley,
Kennedy & Lidz (2006). Investigating the use of the definite article with nouns
modified by gradable adjectives, these researchers discovered a striking contrast
between tall and full, namely:
(9) The tall one vs. the full one:
When subjects are presented with two glasses, both neither tall nor full, they
respond to requests for the tall one, identifying the latter as the taller of the
two glasses, but reject requests for the full one. They consider these requests
to be inappropriate in the given context.
This observation aligns with the proposal that the standard of relative adjectives is
context sensitive: Speakers can accommodate a standard of tallness for the situation
at hand, which only one of the glasses exceeds – thereby rendering the use of the
4
Absolute and Relative Adjectives and their Comparison Classes
definite article licit. With an absolute adjective, such accommodation is impossible,
since the standard is fixed semantically.
The disparity in the sensitivity of relative and absolute adjectives to context is
developed by Kennedy (2007) to a comprehensive theory of vagueness. Kennedy
argues that vagueness occurs only in the interpretation of relative adjectives and
accounts for this phenomenon by appealing to their contextual midpoint standard.
1
Absolute adjectives, on Kennedy’s account, are interpreted relative to a semantically
fixed endpoint, and therefore do not give rise to vagueness.
Despite the advantages of the typology proposed by Kennedy & McNally (2005)
and Kennedy (2007), independent evidence, to which we now turn, suggests that the
interpretation of absolute adjectives is not as impervious to contextual influences as
these theories have it.
1.2 Absolute Adjectives: Arguments for Context Sensitivity
Consider (10), which is taken from Cruse (1980), and (11):
(10) a. This kitchen knife is clean
b. This surgical instrument is clean
(11) a. This child’s shirt is dirty
b. This tuxedo is dirty
Admittedly, the standard of membership of the adjectives clean and dirty depends
on the object these properties are predicated of. Thus, the standard for clean must
necessarily be lower in reference to a kitchen knife than to a surgical instrument,
and the standard for dirty depends on whether at issue is a child’s shirt or a tuxedo.
A similar rationale applies to the adjective full: for a gasoline tank to count as
full it has to be filled to the top or almost to the top, but – as observed by McNally
(2011) – a wine glass is usually considered to be full when filled up to about half of
its capacity. In Italy, a completely full Espresso cup is less full than a half-full tea
cup, and so on and so forth. It seems that contextual effects in the form of standard
shifts triggered by linguistic and extra linguistic features are far from rare. In fact,
it is hard to come up with an example where full is used in relation to a standard
exactly corresponding to the scale maximum.
To Rotstein & Winter (2004) the data in (10) and (11) indicates that absolute
adjectives may exhibit vagueness and context sensitivity just like relative ones.
On their account, the standard of partial adjectives (Kennedy and McNally’s 2005
1
Kennedy’s argument is based on a set of observations concerning characteristics of vague predicates:
(1) vague predicates display context variability; (2) vague predicates give rise to the Sorites paradox;
and (3) vague predicates have borderline cases.
5
  • Conference Paper
    Full-text available
    • Galit W. Sassoon
      Galit W. Sassoon
  • Article
    • Berit Gehrke
    • Elena Castroviejo
      Elena Castroviejo
    This chapter provides an overview of semantic issues concerning manner and degree. This includes going over the theoretical questions that arise in the literature on VP modification and event semantics, as well as the research that has been carried out on gradable adjectives, scale structure and vagueness. This introduction further situates the papers and commentaries in this volume within this discussion. We end by outlining commonalities between manner and degree.
  • Article
    • Elena Castroviejo
      Elena Castroviejo
    • Berit Gehrke
    While we can name examples of modifiers (e.g. adjectives, adverbs, PPs, relative clauses), it is not uncontroversial to what extent “modifier” is a syntactic term and how we should represent modification as part of a semantic model. This being true, modification is not only interesting because it challenges a simple composition system that proceeds through application of functions to arguments. In this introduction we present four papers that show that research on modification proves to be relevant for current investigations on the syntax–semantics interface as well as the language–cognition interface.
  • Conference Paper
    Full-text available
    • Helena Aparicio
    • Ming Xiang
      Ming Xiang
    • Christopher Kennedy
      Christopher Kennedy
  • Chapter
    • Elena Castroviejo
      Elena Castroviejo
    • Louise McNally
      Louise McNally
    • Galit W. Sassoon
      Galit W. Sassoon
    In this chapter we present an overview of three main issues that have surrounded the study of gradable properties—vagueness, measurement, and dimensionality—and how they have been pursued from the perspectives of philosophy, linguistics, and psychology. We then provide a brief summary of each chapter in the volume, together with a guide to how the chapters relate to each other thematically.
  • Conference Paper
    • Elena Castroviejo
      Elena Castroviejo
    • Berit Gehrke
    We provide a semantic account of the Catalan ad-adjectival modifier ben ‘well’, which yields intensification by, we argue, positively evaluating a property ascription. Formally, this translates as applying the predicate good to the saying event available to any utterance. We treat the output of this modification as a Conventional Implicature (rather than at-issue content), which is responsible for its positive polarity behavior. Additionally, we exploit the semantic similarity between this intensifier use of well with other readings of well, including manner (well written) and degree (well acquainted), which we analyze as ‘manner-in-disguise’. In our proposal, they all predicate goodness of an event.
  • Article
    Full-text available
    • Silvia Gumiel
      Silvia Gumiel
    • Isabel Pérez-jiménez
      Isabel Pérez-jiménez
    The distribution of the Spanish copulas ser and estar (‘be’) (specifically in the syntactic context ) has been generally accounted for in the literature in aspectual terms, more explicitly, in terms of the distinction between individual-level and stage-level predicates. Our claim is that the distributional properties of adjectives in the structure can be better described if the scalar properties of the adjectives are taken into account, crucially those properties related to the relative vs. absolute distinction (in the sense of Kennedy & McNally 2005): relative adjectives combine with ser, absolute adjectives combine with estar. From this hypothesis, a better theory arises about aspectual composition in the domain of stative predications (in the line of Husband 2010, 2012).
  • Conference Paper
    Full-text available
    • Yael Greenberg
      Yael Greenberg
    • Moria Ronen
      Moria Ronen
    In this paper we look at three approximators: Almost (and the Hebrew kim'at), more or less (and the Hebrew paxot o yoter) and be-gadol (literally in-big, which can be translated as "by-and-large" / "loosely speaking" / " basically "). We examine a range of similarities and differences between the three approximators and propose to account for them by assuming that all of them are scalar operators, whose semantics combines a polar (negative) component, rejecting the prejacent under an actual value to a relevant parameter, is false, and a proximity (positive) component, accepting an alternative which is lower and close to the prejacent, but that the three particles differ in the choice of the 'relevant parameter' in the negative component (a world, a precision standard, a standard of answerhood to the Question Under Discussion, etc.), and hence also in the alternatives to the prejacent in the positive component.
  • Article
    • Heather Burnett
      Heather Burnett
    The paper presents both new data and a new analysis of the semantic and pragmatic properties of the class of absolute scalar adjectives (ex. dry, wet, straight, bent, flat, empty, full…) within an extension of a well-known logical framework for the analysis of gradable predicates: the delineation semantics framework (DelS) (see Klein, Linguist Philos 4:1–45, 1980; van Benthem, Pac Philos Q 63:193–203, 1982; van Rooij, J Semant 28:335–358, 2011b, among many others). It has been long observed that the context-sensitivity, vagueness and gradability features of absolute scalar predicates give rise to certain puzzles for their analysis within most, if not all, modern formal semantic frameworks. While there exist proposals for solving these puzzles within other major frameworks (such as the degree semantics framework), it has been argued that some of their aspects are particularly challenging for the analysis of absolute scalar predicates within the delineation approach. By combining insights into the relationship between context-sensitivity and scalarity from the DelS framework with insights into the relationship between tolerance/similarity relations and the Sorites paradox from Cobreros et al.’s (2012) Tolerant, Classical, Strict (TCS) framework, I propose a new logical system, called Delineation TCS (DelTCS), in which to set analyses of four classes of adjectival predicates. I argue that this new framework allows for an analysis of absolute scalar adjectives that answers these challenges for delineation-based frameworks, while still preserving the heart of the Klein-ian approach.