WiSe19/20: DD@EL

From Lexical Resource Semantics
Jump to navigation Jump to search

Digital Data in English Linguistics: Language in Politics

Course description

The language used in political texts is directed towards a clearer audience than what we find in literary texts and the general attitudes and the current intentions of the author are usually also relatively clear. This makes this text type particularly apt for formal semantic and pragmatic study.

In this course, students will explore central concepts of pragmatics, such as implicatures, presuppositions, and politeness on the basis of real-live texts of political content, including speeches, tweets, blog contributions and others. We will address topics such as gender-inclusive language use, stereotypes, "dog-whistles" and others. The participants will define a research question and pursue it in the form of a small corpus-based project.

This course runs in parallel to a course on "Language and politics" at the University Mainz (lecturer: Ulrike Schneider). The participants of the two courses will present their work in a joint Mainz-Frankfurt Student workshop "Research in English Linguistics" which will take place February 15, 2020. Participation in this workshop is a mandatory course requirement.


Registration: By e-mail to the lecturer: sailer@em.uni-frankfurt.de

Olat course: to be announced

Free corpus tools

We will be working with the free corpus tools available from http://www.laurenceanthony.net/. In particular:

  • AntFileConverter (to convert pdf files to txt files)
  • AntConc (to get concordanced example sentence, word frequencies, etc.)

BYU corpora

BYU corpora: https://corpus.byu.edu

Page to visit for registration: https://corpus.byu.edu/academic_license_password.asp Enter the following information:

Organization: IEAS
Password: (same as the olat password for this course)

COCA tutorials

BNC

BNCWeb: http://bncweb.lancs.ac.uk
Freely usable after registration!

Help with the query language: http://bncweb.lancs.ac.uk/bncwebXML/Simple_query_language.pdf

BNCWeb tutorials

Meeting 2

The examples in the text are based on Shakespeare's play Macbeth. The full text of the play is available on Projekt Gutenberg.

We will use J.K. Rowling's Harry Potter novels for the final exam this term.


Task 2: Model and Interpretation

1. Define a universe that consists of Macbeth and Banquo.

Check your answer

U = { Macbeth, Banquo }

2. Define the interpretation of the names macbeth and banquo in an intuitively plausible way.

Check your answer

I(macbeth) = Macbeth,
I(banquo) = Banquo

3. Define the interpretation of the properties thane1, king1, and witch1 is such a way that Macbeth is a king, both are thanes and neither is a witch.

Check your answer

I(thane1) = {<Macbeth>, <Banquo>},
I(king1) = {<Macbeth>},
I(witch1) = {}

4. Define the interpretation of the 2-place relations mistrust2 and kill2 in such a way that Macbeth and Banquo mistrust each other and Macbeth kills Banquo.

Check your answer

I(mistrust2) = {<Macbeth, Banquo>, <Banquo, Mactbeth>},
I(kill2) = {<Macbeth,Banquo>}

Task 3: Formulae

Write down logical formulae that express the meaning of the following sentences.

1. Banquo is a thane.

Check your answer

thane1(banquo)

2. Macbeth is king and Macbeth mistrusts Banquo.

Check your answer

king1(macbeth) ∧ mistrust2(macbeth,banquo)

3. If Banquo is king then Macbeth does not kill Banquo.

Check your answer

king1(banquo) ⊃ ¬ kill2(macbeth,banquo)

Task 4: Interpreting formulae

Compute the interpretation of the following formulæ step by step.

1. mistrust2(macbeth,macbeth)

Check your answer

[[mistrust2(macbeth,macbeth)]] = 1
iff < [[macbeth]], [[macbeth]] > is in [[mistrust2]]
iff < I(macbeth), I(macbeth) > in I(mistrust2)
iff < Macbeth, Macbeth > in { <x,y> | x mistrusts y } = { <Macbeth, Banquo>, <Banquo, Macbeth> }

Since this is not the case, [[mistrust2(macbeth,macbeth)]] = 0.


2. ¬king(banquo)

Check your answer

[[¬ king1(banquo)]] = 1
iff [[king(banquo)]] = 0
iff < [[banquo]]> is not in [[king1]]
iff < I(banquo> is not in I(king1)
iff < Banquo > is not in { <x> | x is king } = { <Macbeth>}

Since this is the case, [[¬ king1(banquo)]] = 1


3. witch1(banquo) ⊃ king1(macbeth)

Check your answer

[[witch1(banquo) ⊃ king1(macbeth))]] = 1
iff [[witch1(banquo)]] = 0 or [[king1(macbeth) = 1
iff < [[banquo]] > is not in [[witch1]] or < [[macbeth]] > is in [[king1]]
iff < I(banquo) > is not in I(witch1) or < I(macbeth) > is in I(king1)
iff < Banquo > is not in { <x> | x is a witch} = { } or < Macbeth > is in { <x> | x is king} = { <Macbeth>}.

Since both are the case, [[witch1(banquo) ⊃ king1(macbeth))]] = 1.

Task 5: Variables

Provide a g-function that maps the variables x, y, and z to individuals from the universe and compute the interpretation of the following formula with respect to the model and your g.

(i) kill2(z,x)

Check your answer

Example solution (other values for g are equally possible).

g(x) = Macbeth,
g(y) = Banquo,
g(z) = Banquo.

With this variable assignment we can compute the truth value of the formula:

[[kill2(z,x)]]g = 1
iff < [[z]]g, [[x]]g > is in [[kill2]]g
iff < g(z), g(x) > is in I(kill2)
iff < Banquo, Macbeth > is in { <x,y> | x killed y} = { <Macbeth, Banquo> }.

Since this is not the case, [[kill2(z,x)]]g = 0.

Task 6: Quantifiers

Provide logical formulae that expresse the meaning of the following sentences. Are the formulae true in your model (not in the entire play)? Give a short reason (you don’t need to compute the truth value).

1. Banquo was killed by a king.

Check your answer

x (king(x) : kill(x, banquo))

The formula is true in my model, because there is only one king, Macbeth, and Macbeth killed Banquo.
(Note: The English sentence is in passive, but this has no effect on the logical form.)

2. Macbeth mistrusts every witch.

Check your answer

x (witch(x) : mistrust(macbeth, x))

The formula is true in my model, because there are no witches in my model. Therefore, the formula with the universal quantifier is trivially true.


Semantics boot camp