Documentation

Linglib.Phenomena.Presupposition.ProjectiveContent

Strong Contextual Felicity (SCF).

A trigger has SCF if it requires its projective content to be established in the utterance context prior to its use.

Examples with SCF:

  • Pronouns: "She left" requires antecedent in context
  • Demonstratives: "That cat" requires indication established
  • too (salience): "John came too" requires salient alternative

Examples without SCF:

  • Expressives: "That damn cat" doesn't require prior annoyance
  • Appositives: "Lance, an Ohioan,..." informative content allowed
  • Factives: "John knows it's raining" can be informative
Instances For
    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      Obligatory Local Effect (OLE).

      A trigger has OLE if, when embedded under a belief predicate, its projective content must be part of the attitude holder's beliefs.

      Examples with OLE:

      • "John believes Mary stopped smoking" → John believes Mary used to smoke (obligatory local reading)
      • "John believes it's raining" (from "John knows...") → The embedded proposition is part of John's beliefs

      Examples without OLE:

      • "John believes Lance, an Ohioan, will win" → Speaker (not John) commits to Lance being Ohioan
      • "John believes that damn cat is outside" → Speaker (not John) is annoyed at the cat
      Instances For
        Equations
        • One or more equations did not get rendered due to their size.
        Instances For

          The four classes of projective content from @cite{tonhauser-beaver-roberts-simons-2013}.

          Each class is defined by a combination of SCF and OLE values.

          • classA : ProjectiveClass

            Class A: SCF=yes, OLE=yes Examples: pronouns (existence), too (existence)

          • classB : ProjectiveClass

            Class B: SCF=no, OLE=no Examples: expressives, appositives, NRRCs, possessive NPs

          • classC : ProjectiveClass

            Class C: SCF=no, OLE=yes Examples: stop, know, only, almost

          • classD : ProjectiveClass

            Class D: SCF=yes, OLE=no Examples: too (salience), demonstratives, focus (salience)

          Instances For
            Equations
            • One or more equations did not get rendered due to their size.
            Instances For

              Reconstruct the class from SCF and OLE values.

              Equations
              • One or more equations did not get rendered due to their size.
              Instances For

                Types of projective content triggers, following @cite{tonhauser-beaver-roberts-simons-2013}.

                Each trigger type is associated with a projective class and a description of the content it triggers.

                Instances For
                  Equations
                  • One or more equations did not get rendered due to their size.
                  Instances For

                    The projective class for each trigger type.

                    Equations
                    Instances For

                      A projective content item, combining a trigger with its content.

                      This extends the basic PrProp to track what kind of projective content is involved.

                      • The trigger type

                      • content : WBool

                        The projective content as a proposition

                      • atIssue : WBool

                        The at-issue content (if any)

                      • projectivityDegree : Option

                        Gradient projectivity degree (TBD2018), if empirically measured

                      • atIssuenessDegree : Option

                        Gradient at-issueness degree (TBD2018), if empirically measured

                      Instances For

                        Convert a ProjectiveItem to a PrProp.

                        The projective content becomes the presupposition, and the at-issue content becomes the assertion.

                        Equations
                        Instances For

                          Get the projective class for this item.

                          Equations
                          Instances For

                            Projection behavior describes how content behaves under embedding.

                            All projective contents share the core property that they can project past certain operators, but they differ in their default behavior.

                            • projectsPastNegation : Bool

                              Projects past negation by default

                            • projectsPastQuestions : Bool

                              Projects past questions by default

                            • projectsPastModals : Bool

                              Projects past modals by default

                            • projectsPastConditionals : Bool

                              Projects past conditionals (from antecedent)

                            Instances For
                              Equations
                              • One or more equations did not get rendered due to their size.
                              Instances For

                                Default projection behavior: all projective contents project by default.

                                Tonhauser et al. argue that projection is the default for all projective contents. The differences are in SCF and OLE, not in whether they project.

                                Equations
                                Instances For

                                  At-issueness status of content (binary classification).

                                  Following @cite{roberts-2012}, at-issue content addresses the Question Under Discussion (QUD), while not-at-issue content is backgrounded.

                                  Note: @cite{tonhauser-beaver-degen-2018} show that at-issueness is gradient, not binary. For the gradient version, see Core.Discourse.AtIssueness.AtIssuenessDegree. This binary enum is recoverable from a degree + threshold via AtIssueness.ofDegree.

                                  Instances For
                                    Equations
                                    • One or more equations did not get rendered due to their size.
                                    Instances For

                                      Recover binary at-issueness from a gradient degree and threshold. With the default threshold of 0.5, content with degree > 0.5 maps to .atIssue, ≤ 0.5 to .notAtIssue.

                                      Equations
                                      • One or more equations did not get rendered due to their size.
                                      Instances For

                                        Challengeability with "Hey wait a minute!"

                                        The HWAM test distinguishes at-issue from not-at-issue content:

                                        • At-issue content can be challenged with "No, that's not true"
                                        • Not-at-issue content can be challenged with "Hey wait a minute!"
                                        Instances For
                                          Equations
                                          • One or more equations did not get rendered due to their size.
                                          Instances For

                                            Projective content is typically not-at-issue and HWAM-challengeable.

                                            This diagnostic identifies projective content.

                                            Equations
                                            • One or more equations did not get rendered due to their size.
                                            Instances For

                                              Attribution of projective content under belief predicates.

                                              When "x believes that S" is uttered, who is committed to the projective content of S?

                                              Instances For
                                                Equations
                                                • One or more equations did not get rendered due to their size.
                                                Instances For

                                                  Traditional classification of projective phenomena.

                                                  This maps the traditional terminology onto the Tonhauser et al. taxonomy. Note that this is a simplification — the paper argues that traditional categories don't carve at the joints.

                                                  Instances For
                                                    Equations
                                                    • One or more equations did not get rendered due to their size.
                                                    Instances For

                                                      Map triggers to traditional categories (rough approximation).

                                                      The paper argues that this traditional classification is inadequate — the four-class taxonomy based on SCF and OLE is more explanatory.

                                                      Equations
                                                      Instances For

                                                        The paper uses data from English and Paraguayan Guaraní to establish that the SCF/OLE distinctions are cross-linguistically valid.

                                                        The four-class taxonomy is supported by data from both languages.

                                                        • The trigger tested

                                                        • englishSupport : Bool

                                                          English data supports class assignment

                                                        • guaraniSupport : Bool

                                                          Guaraní data supports class assignment

                                                        Instances For

                                                          The taxonomy is cross-linguistically supported.

                                                          Both English and Guaraní provide evidence for the four-class distinction based on SCF and OLE.

                                                          Equations
                                                          Instances For