Adjectives like happy or unhappy introduce, like events, variables denoting states. Typical roles for adjectives are Attribute, Value, and Degree. Adjectives that say something about mental or physical states of person have the roles Experiencer and Stimulus. There are various modifiers of adjectives that say something about the degree an adjective is applied to an indivudual. The following scale exemplifies these modifications in the meaning of adjectives:

happy not happy
very happy rather happy unhappy
extremely happy slightly happy
happiest least happy
From this diagram you can read the entailment predictions: "extremely happy" entails "very happy", "rather happy" entails "happy", and "unhappy" entails "not happy" (but not vice versa).

Aspectual Verbs
Aspectual verbs like start or stop are analysed as sub-events from the event they say something about. This relation is modelled by the PartOf role. Although aspectual verbs are known to be presupposition triggers, they aren't analysed that way. Some of the entailments supported by aspectual verbs can be derived via first-order axioms.

  • Kadmon (2001): Formal Pragmatics, p. 208

A conditional sentence is something iffy. Conditionals introduce universal quantification over objects. This is encoded in the semantic representation of a sentence by introducing two new boxes connected by an implication symbol.

  • Kadmon (2001): Formal Pragmatics, p. 34
  • Kamp and Reyle (1993): From Discourse to Logic, p. 141

Discourse Relations
Following recent theories of discourse we assume that a text is hierarchically structured. The relations between units of discourse can be coordinating or subordinating. Examples of discourse relations are: NARRATION, BACKGROUND, RESULT, CONTINUATION, PARALLEL, CONTRAST, ELABORATION, INSTANCE, TOPIC, EXPLANATION, PRECONDITION, COMMENTARY. These relations can be lexically triggered or be the results of updating the discourse by including new sentences in the interpretation.

  • Asher (1993): Reference to Abstract Objects in Discourse.
  • Asher and Lascarides (2003). Logics of Conversation.
  • Asher and Vieu (2005): Subordinating and coordinating discourse relations. Lingua 115: 591–610

Indexical pronouns introduce special variables in the meaning representation: for the first person pronoun I this is the free variable speaker, and for the second person pronoun you it's hearer. These variables can be bound by contexts presenting direct speech.

  • Bos (2017): Indexicals and Compositionality: Inside-Out or Outside-In?

Negation introduces explicit scope in the meaning representation with a negation operator. Expessions that introduce negation are not, never, and instead of.

  • Kadmon (2001): Formal Pragmatics, p. 29
  • Kamp and Reyle (1993): From Discourse to Logic, p. 99

Numeral Expressions
Numeral expressions such as twenty, introduce the thematic role Quantity linking a concept with a numeral value. For vague expressions, for instance about 6, we use a set of comparison operators to indicate approximative, lower, or higher values.

  • Kennedy (2015): A "de-Fregean" semantics (and neo-Gricean pragmatics) for modified and unmodified numerals. Semantics and Pragmatics 8: 1-44

Plurals and Groups
The model behind the DRSs used in the PMB deals with plural phrases by treating them as collections (groups). Every group associates with a cardinality. Singular entities are seen as groups with exactly one member.

  • Link (1998): Ten Years of Research on Plurals - Where do we Stand?
  • Schwertel (2005): Plural Semantics for Natural Language Understanding. A Computational Proof-Theoretic Approach. PhD thesis, University of Zurich

Proper Names
Proper names single out a specific individual. They are used to name persons, animals, organisations, locations, vehicles, buildings, events, planets, and natural phenomena (such as hurricanes). They can appear as single words (for instance, Tom) or as multi-word expressions, such as Marilyn Monroe). Semantically they are analysed as "the X named Y", where X is a predicate specifying the concept to which the named individual belongs, and Y the name given to it.

  • Geurts (1997): Good news about the description theory of names.

Reflexive Pronouns
Some pronouns like himself and herself behave in a reflexive manner: they refer (usually) to the subject of the clause. We deal with this in the syntax-semantics interface by assigning them a category that changes a transitive verb into an intransitive verb, and specify in the lexicon that the same variable abstracts over both the subject and object.

  • Van Benthem (1991): Language in Action, p. 127-128
  • Szabolsci (1992): Combinatory Grammar and Projection from the Lexicon.

Repetition Triggers
Scopal modifiers that presuppose the existence of a similar event are not thoroughly analysed in the Parallel Meaning Bank. Examples are also, again, too, and yet. Repetition triggers are hard to analyse semantically, because they interact with focus and require complex representational operations. So they are simply ignored! This is fine for the purpose of giving a sentential semantic representation, as repetition triggers usually do not add any meaning to the sentence, but rather put constraints on the context.

  • von Stechow (1996): The Different Readings of Wieder 'Again': A Structural Account. Journal of Semantics 13: 87-138

Temporal Deixis
The temporal adverb ago is related to the time of utterance. In this example, Monroe died at a time point that meets the beginning of a 33-year-long period, which end meets the time of utterance (now). We represent this with the abut operator.

  • Kamp and Reyle (1993): From Discourse to Logic, p. 573

Universal Quantification
Words like everyone and each introduce universal quantification over objects. This is represented in the semantic representation of a sentence by introducing two new boxes connected by an implication symbol.

  • Kamp and Reyle (1993): From Discourse to Logic, Chapter 2