User:Magnap/Completeness Proposal

From The Toaq Wiki
Revision as of 17:57, 26 September 2024 by Magnap (talk | contribs) (Add the "to tell" exhaustivity case study)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Depends on: Inquisitive Semantics Proposal

What does it mean to be an answer to a question? This question aims to clear that up.

Note that we cannot rely solely on our intuitions to lead us to unambiguous agreement! Take the following scenario, in which German-speakers and English-speakers have different intuitions for the exhaustivity of "to correctly predict"/"korrekt vorhersagen" (adapted from this paper):

We have a string of letters, "XXXX", and I predict that the two middle letters will change. Then, the string becomes "XOOO".

  • Was my prediction correct? No.
  • War meine Vorhersage korrekt? Ja. ("Was my prediction correct? Yes.")
  • Ma due shíudua jıbo móq? ("Was my prediction correct?")


Case studies

Who passed the test?

Imagine 3 students A/B/C and 3 teachers X/Y/Z. A test is held, and while A and B pass, C doesn't.

  • X says: "A passed the test, but I don't know if B or C did"
  • Y says: "A and B both passed, but I don't know about C"
  • Z says: "A and B both passed, and C failed"
Who told you what?
Who told you: X? Y? Z?
Someone who passed Y Y Y
Everyone who passed N Y Y
Who passed and who didn't N N Y

How can we unambiguously ask all three questions in Toaq?

Prime knowledge

  • Ma shê, ꝡä bu dua jí, mä reutoaı gú rá sáq rá jó, nä dua jí, ꝡä reutoaı hí mea gú róı sáq róı jó móq? (if I don't know whether 2 nor 3 nor 4 are primes, do I know which among 2, 3, and 4 is/are prime?)
  • Ma shê, ꝡä dua jí, ꝡä reutoaı jó, rú bu dua jí, mä reutoaı gú rá sáq ,nä dua jí, ꝡä reutoaı hí mea gú róı sáq róı jó móq? (if I know that 4 is a prime and I don't know whether 2 nor 3 are primes, do I know which among 2, 3, and 4 is a prime / are prime?)
  • Ma shê, ꝡä dua jí, ꝡä reutoaı gú rú sáq, rú bu dua jí, mä reutoaı jó, nä dua jí, ꝡä reutoaı hí mea gú róı sáq róı jó móq? (if I know that 2 and 3 are primes and I don't know whether 4 is a prime, do I know which among 2, 3, and 4 is a prime / are prime?)
  • Ma shê, ꝡä dua jí, ꝡä reutoaı gú rú sáq, rú dua jí, ꝡä bu reutoaı jó, nä dua jí, ꝡä reutoaı hí mea gú róı sáq róı jó móq? (if I know that 2 and 3 are primes and I know that 4 is not a prime, do I know which among 2, 3, and 4 is a prime / are prime?)
  • Ma shê, ꝡä dua jí, ꝡä reutoaı gú, rú dua jí, ꝡä bu reutoaı jó, rú bu dua jí, mä reutoaı sáq, nä dua jí, ꝡä reutoaı hí mea gú róı sáq róı jó móq? (if I know that 2 is a prime, and that 4 is not a prime, and I don't know whether 3 is a prime, do I know which among 2, 3, and 4 is a prime / are prime?)


TODO Background

  • Roelofsen, F., Theiler, N., & Aloni, M. (2014). Embedded interrogatives : the role of false answers. Presented at the 7th questions in discourse workshop, Göttingen, Sept 2014.
  • Theiler, N. (2014). A multitude of answers: Embedded questions in typed inquisitive semantics. MSc thesis, University of Amsterdam, supervised by M. Aloni and F. Roelofsen.

TODO paint the rest of the owl

We will need to add a new phrase to the grammar, the completeness phrase, which will have a completeness operator as its head, and we'll have to pick exactly one of the following as its complement: a , a , or a . Completeness operators map a proposition to the (not necessarily downwards closed! thus, not necessarily a proposition) set of suitable answers to it in some world (each of which is a truth-conditional proposition). TODO describe the completeness operators. However, it will usually have a null head. Alternatively, we could have completeness be lexical for clause-embedding words, which would avoid building a concept of a "suitable answer" into the semantics (as long as we don't attempt lexicosemantics on predicates).