THE LEXICON OF SUBJECTIVITY
A one-day workshop at the
University of the Basque Country
Department of Linguistics and Basque Studies
Micaela Portilla Research Centre, room 02
(Justo Vélez de Elorriaga, 1)
Vitoria-Gasteiz
JUNE 13, 2017
In recent years, one of the most vivid debates in philosophy of language has been that between contextualism, relativism, and expressivism about a series of subjective, or perspectival, expressions (such as predicates of personal taste, aesthetic and moral terms, epistemic vocabulary, gradable adjectives, epistemic modals etc.). The main arguments discussed so far can be categorized into two types: arguments based on intuitions about the truth-value of sentences containing the expressions in question in various situations (disagreement, eavesdropping, retraction etc.) and arguments based on syntactic considerations (binding, licensing, control, embeddings under various attitude verbs etc.).
While such arguments are no doubt important, not much attention has been devoted to investigating the expressions in question from the point of view of lexical semantics. There are many lexical theories on the market, but they can be divided into two broad types. Thus, according to one type of theories, lexical entries of natural language expressions are rich, comprising a significant amount of information, possibly structured along several dimensions, which get selected in context. According to the other type, lexical entries of natural language expressions are rather skeletal, comprising only minimal information that is further enriched in context. Intermediary positions are possible, too: according to those, lexical entries of some expressions are rich, while those of other expressions are minimal.
This workshop aims to explore the possible significance for the debate mentioned above of a lexical approach to subjective, or perspectival, expressions. The general purpose of such an endeavor is to gain insight into the lexical configuration of subjective expressions and bring to the fore possible consequences for their syntax, as well as for their semantics. Thus, among the questions raised and tackled in the workshop are the following:
- What is the lexical configuration of subjective, or perspectival, expressions?
- What lexical theory is best suited to adequately account for such expressions?
- How does the lexical configuration of such expressions affect their syntax and their semantics?
- What are the consequences of such findings for the contemporary debate between contextualism, relativism and expressivism?
The workshop is organized by Dan Zeman, Marina Ortega and Agustin Vicente, with the financial support of the Lexical Meaning and Concepts (FFI2014-52196-P) project and the Basque Research Group for Theoretical Linguistics (HiTT). Attendance is free, but please write an email to danczeman[at]gmail.com if you want to participate. Poster by Marina Ortega.
A one-day workshop at the
University of the Basque Country
Department of Linguistics and Basque Studies
Micaela Portilla Research Centre, room 02
(Justo Vélez de Elorriaga, 1)
Vitoria-Gasteiz
JUNE 13, 2017
In recent years, one of the most vivid debates in philosophy of language has been that between contextualism, relativism, and expressivism about a series of subjective, or perspectival, expressions (such as predicates of personal taste, aesthetic and moral terms, epistemic vocabulary, gradable adjectives, epistemic modals etc.). The main arguments discussed so far can be categorized into two types: arguments based on intuitions about the truth-value of sentences containing the expressions in question in various situations (disagreement, eavesdropping, retraction etc.) and arguments based on syntactic considerations (binding, licensing, control, embeddings under various attitude verbs etc.).
While such arguments are no doubt important, not much attention has been devoted to investigating the expressions in question from the point of view of lexical semantics. There are many lexical theories on the market, but they can be divided into two broad types. Thus, according to one type of theories, lexical entries of natural language expressions are rich, comprising a significant amount of information, possibly structured along several dimensions, which get selected in context. According to the other type, lexical entries of natural language expressions are rather skeletal, comprising only minimal information that is further enriched in context. Intermediary positions are possible, too: according to those, lexical entries of some expressions are rich, while those of other expressions are minimal.
This workshop aims to explore the possible significance for the debate mentioned above of a lexical approach to subjective, or perspectival, expressions. The general purpose of such an endeavor is to gain insight into the lexical configuration of subjective expressions and bring to the fore possible consequences for their syntax, as well as for their semantics. Thus, among the questions raised and tackled in the workshop are the following:
- What is the lexical configuration of subjective, or perspectival, expressions?
- What lexical theory is best suited to adequately account for such expressions?
- How does the lexical configuration of such expressions affect their syntax and their semantics?
- What are the consequences of such findings for the contemporary debate between contextualism, relativism and expressivism?
The workshop is organized by Dan Zeman, Marina Ortega and Agustin Vicente, with the financial support of the Lexical Meaning and Concepts (FFI2014-52196-P) project and the Basque Research Group for Theoretical Linguistics (HiTT). Attendance is free, but please write an email to danczeman[at]gmail.com if you want to participate. Poster by Marina Ortega.
PROGRAM
9.30-10.45
John Collins, "Lexical Items as Atomic Syntactic Constituents: Active and Inactive Aspects of Words"
10.45-12.00
Elena Castroviejo, "From Evaluativity to Intensification. The Cases of Romance 'Good' and 'Well'"
12.00-12.15
Coffee break
12.15-13.30
Lisa Bylinina: "Evidence for Subjective Statements"
13.30-15.00
Lunch
15.00-16.15
Sara Packalen, "A Two-Dimensional Defense of Subjectivism about Normative Predicates"
16.15-16.30
Coffee break
16.30-17.45
Hazel Pearson, "Predicates of Personal Taste and the De Se: Similarities and Differences"
17.45-19.00
Carla Umbach, "On the Meaning of Dimensional and Aesthetic Predicates"
20.30
Dinner
ABSTRACTS
John Collins, "Lexical Items as Atomic Syntactic Constituents: Active and Inactive Aspects of Words"
Minimally, lexical items are atomic syntactic constituents in the sense that they submit to combinatorial syntactic operations independently of varied aspects of the items' contents. In particular, the relevant syntactic operations appear to be mostly, if not entirely, semantically blind. Concomitantly, as constituents of sentences in use, the items appear to express rich aspects of the subjectivity of the speaker and wider contextual factors that affect truth conditions. What is the source of such apparent enrichment? Starting from the idea of lexical items as minimal syntactic atoms, the paper argues that there are no good reasons to enrich lexical items in a way that bears upon their syntactic profile. Instead, lexical items are best conceived as containing structurally active and inactive aspects, with context-relevant features invariably falling into the latter category. The extent of the active features remains open to dispute. The paper will pay special attention to aesthetic predicates as an examplar.
Elena Castroviejo, "From Evaluativity to Intensification. The Cases of Romance Good and Well"
The semantics of the adjective good is interesting for various reasons. It is an evaluative predicate with a commending function (Hare 1952) or at most a quasi-denotational (and highly context-dependent) meaning (Umbach 2016). It is also a subsective modifier that has been claimed to predicate over events (Larson 1998) or contextual variables standing for certain roles (Szabo 2001), or to encode a functional relation over individuals and degrees with respect to a property of eventualities or dispositions (Asher 2011). Additionally, from the morpho-syntactic perspective, in Romance languages such as Spanish and Catalan, an interesting correlation has been established between the pre- vs. postnominal position of certain adjectives (including good) and their subsective vs. intersective interpretation (Demonte 1982, 1999).
Building on this literature and our own previous work, we aim to account for a different phenomenon, namely that Spanish buen and Catalan bon ‘good’ seem to behave as degree intensifiers when they subsectively modify certain nouns. This is exemplified in (1) for Catalan.
(1) una bona dosi ‘a good dose, ~a big dose’, un bon problema ‘a good problem, ~a big problem’, un bon idiota ‘a good idiot, ~a big idiot’
These examples are to be compared to examples where bon has a plain evaluative interpretation, (2), or where it is ambiguous between the intensifying and the plain evaluative interpretation, (3).
(2) un bon amic ‘a good friend’, un bon llibre ‘a good book’, una bona notícia ‘good news’
(3) un bon esmorzar ‘a good breakfast’, un bon massatge ‘a good massage’, un bon pernil ‘a good ham’
The key piece of data that we aim to account for is the apparent positive polaritive behavior of intensifying bon, as illustrated in (4).
(4) (*No) tens un bon problema.
NEG have you a good problem
‘You (*don’t) have a good problem.
Intended: ‘You (*don’t) have a big problem.’
To do so, we reject previous ideas that relate positive polarity with expressive meaning and concentrate on the lexical meaning of good (its subjective nature as well as its quasi-denotation that relies on criteria, as shown in Umbach 2016) and on the internal structure of the noun. Specifically, we establish a correlation between unidimensionality and the emergence of intensification.
Time permitting, we will extend our analysis to the adverbial domain, so as to cover for the following Catalan data, whereby well can modify gradable adjectives to yield high degree:
(5) ben alt ‘well tall, ~really tall’, ben bonic ‘well pretty, ~really pretty’, ben ple ‘well full, ~really full’.
Lisa Bylinina, "Evidence for Subjective Statements"
The class of subjective, or judge-dependent, expressions in natural language is heterogeneous. The differences concern whether they can be embedded under subjective attitude verbs like ‘find’ in their bare and/or comparative form, whether they induce subjective or objective order on their domain, whether they allow for an overt expression of the judge argument, and more. One property of a subclass of subjective predicates — the core group of predicates of personal taste (‘tasty’, ‘fun’ etc.) is that they give rise to so-called ‘direct experience requirement’ (Pearson 2013 a.o.). In earlier work (Bylinina 2016), I implemented this requirement as a lexical presupposition of predicates of personal taste that indirectly references an experience event. This presupposition, I suggested, has consequences for the behaviour of such predicates: ability to take overt ‘judge arguments’, requirement for evidential marking in languages like Japanese, possibly more. This picture suggests that there are two clear-cut classes of subjective predicates: ones that have internal, experience-based meaning, and ones for which this is not the case. In this talk, I will discuss problems with such view. I will focus on both (cross-)linguistic and conceptual considerations that point in the direction of more fine-grained balance between experiential and normative in the grounds for subjective statements of different types — moral, aesthetic, statements with vague predicates and predicates of taste.
Sara Packalén, "A Two-Dimensional Defense of Subjectivism about Normative Predicates"
By subjectivism I intend the following semantic thesis concerning normative language. Let c be the context of utterance and let Sc be the speaker of the context of utterance. Furthermore, let F be a predicate of action. Semantic clauses for normative sentences then have the following general form:
(1) [[F-ing is wrong]] (c) = 1 iff Sc disapproves of F-ing
(2) [[F-ing is right]] (c) = 1 iff Sc approves of F-ing
‘Disapproval’ and ‘approval’ are supposed to be placeholder terms for some suitable ‘pro-’ and ‘con-’ attitudes, characterizable in non-normative terms.
The subjectivist semantic clauses have three characteristics. First, assuming that approval and disapproval in themselves are non-normative properties, the right hand sides of the equivalences contain only non-normative terms. Hence, subjectivism is a naturalist position in that the truth-conditions for normative sentences are given in non-normative terms. This sets subjectivism apart from non-naturalist semantic frameworks, according to which the truth-conditions of normative sentences contain irreducibly normative terms. Second, according to subjectivism, normative sentences have truth-conditions in the same way as descriptive, non-normative sentences. This sets the position apart from expressivist positions, according to which normative sentences do not state facts about the world or the speaker’s views but rather express an attitude or a state of mind of the speaker. Third and finally, normative sentences are indexical according to subjectivism in that the truth value of normative sentences vary with the context of utterance and in particular with the speaker. This final feature is the source of the semantic problems that are the focus of this paper: the modal problem and the problem of genuine disagreement (Schroeder, 2008: 16-17).
Hazel Pearson, "Predicates of Personal Taste and the De Se: Similarities and Differences"
A number of authors on both the relativist and contextualist side of the debate have pursued the view that there is an intimate connection between predicates of personal taste and de se construal (see for example Stephenson 2007, Moltmann 2009, Pearson 2013). This is generally implemented by means of a variable occupying the internal argument slot of the predicate, which in attitude reports is bound by the attitude verb. I discuss novel data that present a challenge for this family of views. The observations I discuss present a fundamental challenge for the relativist stance that whether or not something is tasty (say) depends on who is speaking.
Carla Umbach, "On the Meaning of Dimensional and Aesthetic Predicates"
Dimensional and aesthetic predicates (like tall and beautiful) exhibit a number of parallels. They are, for example, gradable and relative. But whereas the truth of the proposition in (1a) hinges on a property inherent to the object it is predicated of (i.e. the size of the statue), the property in (1b) is merely ascribed to the object – there is no matter of fact of whether something counts as beautiful.
(1a) This statue is tall.
(1b) This statue is beautiful.
Still, when denied in discourse both (1a) and (1b) may be subject to so-called faultless disagreement, (that is, the intuition that, although they assert contradictory propositions, neither discourse participant can be blamed to be wrong), thereby raising the question of what it is in the lexical meaning of these predicates that makes them behave in the same way.
It will be proposed in this paper that dimensional predicates as well as aesthetic predicates (and, more general, evaluative predicates) share a core characteristics: Their primary meaning consists in a rule-like valuation component expressing that something is outstanding or noteworthy, be it approving or disapproving. Their denotational meaning is only secondary, determined by criteria of approval or disapproval agreed on by conversational participants.